commit 4a0b95bec5f46c2300a512b94ea5f03342c1193a Author: ModelHub XC Date: Mon Jun 15 14:41:51 2026 +0800 初始化项目,由ModelHub XC社区提供模型 Model: Neelectric/Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.04 Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..52373fe --- /dev/null +++ b/.gitattributes @@ -0,0 +1,36 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..0cd29b0 --- /dev/null +++ b/README.md @@ -0,0 +1,60 @@ +--- +base_model: meta-llama/Llama-3.1-8B-Instruct +datasets: Neelectric/MoT_science_Llama3_4096toks +library_name: transformers +model_name: Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.04 +tags: +- generated_from_trainer +- trl +- open-r1 +- sft +licence: license +--- + +# Model Card for Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.04 + +This model is a fine-tuned version of [meta-llama/Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) on the [Neelectric/MoT_science_Llama3_4096toks](https://huggingface.co/datasets/Neelectric/MoT_science_Llama3_4096toks) dataset. +It has been trained using [TRL](https://github.com/huggingface/trl). + +## Quick start + +```python +from transformers import pipeline + +question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?" +generator = pipeline("text-generation", model="Neelectric/Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.04", device="cuda") +output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] +print(output["generated_text"]) +``` + +## Training procedure + +[Visualize in Weights & Biases](https://wandb.ai/neelectric/open-r1_science/runs/jfkt36wm) + + + +This model was trained with SFT. + +### Framework versions + +- TRL: 1.0.0.dev0 +- Transformers: 4.57.6 +- Pytorch: 2.9.0 +- Datasets: 4.8.2 +- Tokenizers: 0.22.2 + +## Citations + + + +Cite TRL as: + +```bibtex +@software{vonwerra2020trl, + title = {{TRL: Transformers Reinforcement Learning}}, + author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin}, + license = {Apache-2.0}, + url = {https://github.com/huggingface/trl}, + year = {2020} +} +``` \ No newline at end of file diff --git a/all_results.json b/all_results.json new file mode 100644 index 0000000..7d4b5d8 --- /dev/null +++ b/all_results.json @@ -0,0 +1,9 @@ +{ + "ewc_loss": 1.71661376953125e-05, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.923321149223798, + "train_runtime": 37069.0608, + "train_samples": 145693, + "train_samples_per_second": 11.791, + "train_steps_per_second": 0.737 +} \ No newline at end of file diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000..0ab931a --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,121 @@ +{{- bos_token }} +{%- if custom_tools is defined %} + {%- set tools = custom_tools %} +{%- endif %} +{%- if not tools_in_user_message is defined %} + {%- set tools_in_user_message = true %} +{%- endif %} +{%- if not date_string is defined %} + {%- set date_string = "26 Jul 2024" %} +{%- endif %} +{%- if not tools is defined %} + {%- set tools = none %} +{%- endif %} + +{#- This block extracts the system message, so we can slot it into the right place. #} +{%- if messages[0]['role'] == 'system' %} + {%- set system_message = messages[0]['content']|trim %} + {%- set messages = messages[1:] %} +{%- else %} + {%- set system_message = "You are a helpful AI Assistant that provides well-reasoned and detailed responses. You first think about the reasoning process as an internal monologue and then provide the user with the answer. Respond in the following format: +... + + +... +" %} +{%- endif %} + +{#- System message + builtin tools #} +{{- "<|start_header_id|>system<|end_header_id|>\n\n" }} +{%- if builtin_tools is defined or tools is not none %} + {{- "Environment: ipython\n" }} +{%- endif %} +{%- if builtin_tools is defined %} + {{- "Tools: " + builtin_tools | reject('equalto', 'code_interpreter') | join(", ") + "\n\n"}} +{%- endif %} +{{- "Cutting Knowledge Date: December 2023\n" }} +{{- "Today Date: " + date_string + "\n\n" }} +{%- if tools is not none and not tools_in_user_message %} + {{- "You have access to the following functions. To call a function, please respond with JSON for a function call." }} + {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }} + {{- "Do not use variables.\n\n" }} + {%- for t in tools %} + {{- t | tojson(indent=4) }} + {{- "\n\n" }} + {%- endfor %} +{%- endif %} +{{- system_message }} +{{- "<|eot_id|>" }} + +{#- Custom tools are passed in a user message with some extra guidance #} +{%- if tools_in_user_message and not tools is none %} + {#- Extract the first user message so we can plug it in here #} + {%- if messages | length != 0 %} + {%- set first_user_message = messages[0]['content']|trim %} + {%- set messages = messages[1:] %} + {%- else %} + {{- raise_exception("Cannot put tools in the first user message when there's no first user message!") }} + {%- endif %} + {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}} + {{- "Given the following functions, please respond with a JSON for a function call " }} + {{- "with its proper arguments that best answers the given prompt.\n\n" }} + {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }} + {{- "Do not use variables.\n\n" }} + {%- for t in tools %} + {{- t | tojson(indent=4) }} + {{- "\n\n" }} + {%- endfor %} + {{- first_user_message + "<|eot_id|>"}} +{%- endif %} + +{%- for message in messages %} + {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %} + {%- if message['role'] == 'assistant' %} + {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }} + {% generation %} + {{- message['content'] | trim + '<|eot_id|>' }} + {% endgeneration %} + {%- else %} + {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }} + {%- endif %} + {%- elif 'tool_calls' in message %} + {%- if not message.tool_calls|length == 1 %} + {{- raise_exception("This model only supports single tool-calls at once!") }} + {%- endif %} + {%- set tool_call = message.tool_calls[0].function %} + {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}} + {% generation %} + {%- if builtin_tools is defined and tool_call.name in builtin_tools %} + {{- "<|python_tag|>" + tool_call.name + ".call(" }} + {%- for arg_name, arg_val in tool_call.arguments | items %} + {{- arg_name + '="' + arg_val + '"' }} + {%- if not loop.last %} + {{- ", " }} + {%- endif %} + {%- endfor %} + {{- ")" }} + {%- else %} + {{- '{"name": "' + tool_call.name + '", ' }} + {{- '"parameters": ' }} + {{- tool_call.arguments | tojson }} + {{- "}" }} + {%- endif %} + {%- if builtin_tools is defined %} + {{- "<|eom_id|>" }} + {%- else %} + {{- "<|eot_id|>" }} + {%- endif %} + {% endgeneration %} + {%- elif message.role == "tool" or message.role == "ipython" %} + {{- "<|start_header_id|>ipython<|end_header_id|>\n\n" }} + {%- if message.content is mapping or message.content is iterable %} + {{- message.content | tojson }} + {%- else %} + {{- message.content }} + {%- endif %} + {{- "<|eot_id|>" }} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }} +{%- endif %} \ No newline at end of file diff --git a/config.json b/config.json new file mode 100644 index 0000000..06df27b --- /dev/null +++ b/config.json @@ -0,0 +1,36 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "dtype": "bfloat16", + "eos_token_id": 128009, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "pad_token_id": 128009, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 8.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": false, + "transformers_version": "4.57.6", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..50f6077 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,9 @@ +{ + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128009, + "pad_token_id": 128009, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.57.6" +} diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors new file mode 100644 index 0000000..9baec50 --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:571724eca45b23d6954d9aa7f1c0ea09e8d9c140cf175559814efb2f2878cb57 +size 4976698672 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..6871050 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90886ac8e6ab1a219aabe65f5fe5321748b7a1a32058bd6ffba8872841f2ab7c +size 4999802720 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..008eeeb --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4fc9d98ef36d8108cd827fa0f35fa29d4d19789ac18fd32a74a5f59a429adeac +size 4915916176 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..b246507 --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a11a734180c854f36987988afb881cd84a7289027c529370e2b7a6bcafef266 +size 1168138808 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000..5c64f1e --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,299 @@ +{ + "metadata": { + "total_parameters": 8030261248, + "total_size": 16060522496 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..9d4773c --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,11 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": "<|eot_id|>", + "pad_token": "<|eot_id|>" +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..1c1d8d5 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..3beeacc --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,2063 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|eot_id|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "<|eot_id|>", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/train_results.json b/train_results.json new file mode 100644 index 0000000..7d4b5d8 --- /dev/null +++ b/train_results.json @@ -0,0 +1,9 @@ +{ + "ewc_loss": 1.71661376953125e-05, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.923321149223798, + "train_runtime": 37069.0608, + "train_samples": 145693, + "train_samples_per_second": 11.791, + "train_steps_per_second": 0.737 +} \ No newline at end of file diff --git a/trainer_state.json b/trainer_state.json new file mode 100644 index 0000000..29a2c7d --- /dev/null +++ b/trainer_state.json @@ -0,0 +1,273224 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 3.0, + "eval_steps": 500, + "global_step": 27318, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.00010981770261366133, + "ewc_loss": 0.0, + "grad_norm": 6.513947010040283, + "learning_rate": 0.0, + "loss": 1.5235, + "mean_token_accuracy": 0.6157128810882568, + "num_tokens": 30018.0, + "step": 1 + }, + { + "epoch": 0.00021963540522732265, + "ewc_loss": 0.0, + "grad_norm": 6.408624649047852, + "learning_rate": 3.6603221083455343e-10, + "loss": 1.4062, + "mean_token_accuracy": 0.628243088722229, + "num_tokens": 58744.0, + "step": 2 + }, + { + "epoch": 0.000329453107840984, + "ewc_loss": 1.249373597200093e-20, + "grad_norm": 7.2841010093688965, + "learning_rate": 7.320644216691069e-10, + "loss": 1.4743, + "mean_token_accuracy": 0.6152434945106506, + "num_tokens": 84293.0, + "step": 3 + }, + { + "epoch": 0.0004392708104546453, + "ewc_loss": 1.0570971181733668e-18, + "grad_norm": 7.0338873863220215, + "learning_rate": 1.0980966325036603e-09, + "loss": 1.5503, + "mean_token_accuracy": 0.5984405279159546, + "num_tokens": 113639.0, + "step": 4 + }, + { + "epoch": 0.0005490885130683066, + "ewc_loss": 5.7462715141731735e-18, + "grad_norm": 7.096231937408447, + "learning_rate": 1.4641288433382137e-09, + "loss": 1.5147, + "mean_token_accuracy": 0.6076140403747559, + "num_tokens": 139615.0, + "step": 5 + }, + { + "epoch": 0.000658906215681968, + "ewc_loss": 1.0354130747236567e-17, + "grad_norm": 6.862222194671631, + "learning_rate": 1.8301610541727671e-09, + "loss": 1.6123, + "mean_token_accuracy": 0.5816936492919922, + "num_tokens": 168855.0, + "step": 6 + }, + { + "epoch": 0.0007687239182956292, + "ewc_loss": 7.676151381197371e-17, + "grad_norm": 7.638943195343018, + "learning_rate": 2.1961932650073206e-09, + "loss": 1.5225, + "mean_token_accuracy": 0.602878212928772, + "num_tokens": 192650.0, + "step": 7 + }, + { + "epoch": 0.0008785416209092906, + "ewc_loss": 2.1163626406917047e-16, + "grad_norm": 6.320235729217529, + "learning_rate": 2.562225475841874e-09, + "loss": 1.5188, + "mean_token_accuracy": 0.6040030717849731, + "num_tokens": 225812.0, + "step": 8 + }, + { + "epoch": 0.000988359323522952, + "ewc_loss": 4.180683577104105e-16, + "grad_norm": 8.854935646057129, + "learning_rate": 2.9282576866764274e-09, + "loss": 1.5177, + "mean_token_accuracy": 0.6042245626449585, + "num_tokens": 245223.0, + "step": 9 + }, + { + "epoch": 0.0010981770261366132, + "ewc_loss": 5.793976409762536e-16, + "grad_norm": 6.878784656524658, + "learning_rate": 3.294289897510981e-09, + "loss": 1.5663, + "mean_token_accuracy": 0.595893919467926, + "num_tokens": 273619.0, + "step": 10 + }, + { + "epoch": 0.0012079947287502745, + "ewc_loss": 2.3037127760972e-15, + "grad_norm": 7.196942329406738, + "learning_rate": 3.6603221083455343e-09, + "loss": 1.3787, + "mean_token_accuracy": 0.6412514448165894, + "num_tokens": 297418.0, + "step": 11 + }, + { + "epoch": 0.001317812431363936, + "ewc_loss": 2.8033131371785203e-15, + "grad_norm": 6.064698219299316, + "learning_rate": 4.026354319180088e-09, + "loss": 1.393, + "mean_token_accuracy": 0.6357642412185669, + "num_tokens": 328966.0, + "step": 12 + }, + { + "epoch": 0.0014276301339775973, + "ewc_loss": 3.6637359812630166e-15, + "grad_norm": 7.5634846687316895, + "learning_rate": 4.392386530014641e-09, + "loss": 1.3892, + "mean_token_accuracy": 0.638329267501831, + "num_tokens": 353003.0, + "step": 13 + }, + { + "epoch": 0.0015374478365912585, + "ewc_loss": 4.3576253716537394e-15, + "grad_norm": 7.200562000274658, + "learning_rate": 4.758418740849194e-09, + "loss": 1.4653, + "mean_token_accuracy": 0.6210613250732422, + "num_tokens": 377957.0, + "step": 14 + }, + { + "epoch": 0.0016472655392049198, + "ewc_loss": 2.0539125955565396e-14, + "grad_norm": 8.110922813415527, + "learning_rate": 5.124450951683748e-09, + "loss": 1.4743, + "mean_token_accuracy": 0.6144688129425049, + "num_tokens": 398957.0, + "step": 15 + }, + { + "epoch": 0.0017570832418185812, + "ewc_loss": 2.5424107263916085e-14, + "grad_norm": 7.211446285247803, + "learning_rate": 5.490483162518302e-09, + "loss": 1.5626, + "mean_token_accuracy": 0.5936265587806702, + "num_tokens": 424867.0, + "step": 16 + }, + { + "epoch": 0.0018669009444322424, + "ewc_loss": 2.7977620220553945e-14, + "grad_norm": 6.961252212524414, + "learning_rate": 5.856515373352855e-09, + "loss": 1.4469, + "mean_token_accuracy": 0.6198854446411133, + "num_tokens": 450753.0, + "step": 17 + }, + { + "epoch": 0.001976718647045904, + "ewc_loss": 3.375077994860476e-14, + "grad_norm": 6.822975158691406, + "learning_rate": 6.222547584187408e-09, + "loss": 1.4527, + "mean_token_accuracy": 0.6170147657394409, + "num_tokens": 478721.0, + "step": 18 + }, + { + "epoch": 0.002086536349659565, + "ewc_loss": 3.6415315207705135e-14, + "grad_norm": 8.597040176391602, + "learning_rate": 6.588579795021962e-09, + "loss": 1.5914, + "mean_token_accuracy": 0.5925307273864746, + "num_tokens": 499326.0, + "step": 19 + }, + { + "epoch": 0.0021963540522732263, + "ewc_loss": 1.305622276959184e-13, + "grad_norm": 7.617938995361328, + "learning_rate": 6.954612005856515e-09, + "loss": 1.4421, + "mean_token_accuracy": 0.6253232955932617, + "num_tokens": 522720.0, + "step": 20 + }, + { + "epoch": 0.002306171754886888, + "ewc_loss": 1.8207657603852567e-13, + "grad_norm": 6.330581188201904, + "learning_rate": 7.3206442166910685e-09, + "loss": 1.4216, + "mean_token_accuracy": 0.6322718262672424, + "num_tokens": 551422.0, + "step": 21 + }, + { + "epoch": 0.002415989457500549, + "ewc_loss": 1.9984014443252818e-13, + "grad_norm": 8.043166160583496, + "learning_rate": 7.686676427525622e-09, + "loss": 1.5617, + "mean_token_accuracy": 0.6012513041496277, + "num_tokens": 572912.0, + "step": 22 + }, + { + "epoch": 0.0025258071601142102, + "ewc_loss": 2.2115642650533118e-13, + "grad_norm": 5.998161792755127, + "learning_rate": 8.052708638360176e-09, + "loss": 1.411, + "mean_token_accuracy": 0.6289289593696594, + "num_tokens": 604172.0, + "step": 23 + }, + { + "epoch": 0.002635624862727872, + "ewc_loss": 2.611244553918368e-13, + "grad_norm": 7.537014007568359, + "learning_rate": 8.418740849194728e-09, + "loss": 1.4468, + "mean_token_accuracy": 0.6209278106689453, + "num_tokens": 627461.0, + "step": 24 + }, + { + "epoch": 0.002745442565341533, + "ewc_loss": 2.824407374646398e-13, + "grad_norm": 6.369226455688477, + "learning_rate": 8.784773060029282e-09, + "loss": 1.4483, + "mean_token_accuracy": 0.6270922422409058, + "num_tokens": 657516.0, + "step": 25 + }, + { + "epoch": 0.0028552602679551946, + "ewc_loss": 3.055333763768431e-13, + "grad_norm": 6.298149585723877, + "learning_rate": 9.150805270863836e-09, + "loss": 1.5438, + "mean_token_accuracy": 0.5959436297416687, + "num_tokens": 691937.0, + "step": 26 + }, + { + "epoch": 0.0029650779705688557, + "ewc_loss": 3.9612757518625585e-13, + "grad_norm": 8.514826774597168, + "learning_rate": 9.516837481698389e-09, + "loss": 1.4647, + "mean_token_accuracy": 0.6267452239990234, + "num_tokens": 711900.0, + "step": 27 + }, + { + "epoch": 0.003074895673182517, + "ewc_loss": 1.1226575225009583e-12, + "grad_norm": 7.253059387207031, + "learning_rate": 9.882869692532943e-09, + "loss": 1.4916, + "mean_token_accuracy": 0.6121397018432617, + "num_tokens": 737085.0, + "step": 28 + }, + { + "epoch": 0.0031847133757961785, + "ewc_loss": 1.3216094885137863e-12, + "grad_norm": 8.068613052368164, + "learning_rate": 1.0248901903367497e-08, + "loss": 1.5753, + "mean_token_accuracy": 0.6034032106399536, + "num_tokens": 759817.0, + "step": 29 + }, + { + "epoch": 0.0032945310784098397, + "ewc_loss": 1.4637180356658064e-12, + "grad_norm": 7.557466506958008, + "learning_rate": 1.0614934114202049e-08, + "loss": 1.4181, + "mean_token_accuracy": 0.6269631385803223, + "num_tokens": 783353.0, + "step": 30 + }, + { + "epoch": 0.003404348781023501, + "ewc_loss": 1.5845103007450234e-12, + "grad_norm": 8.3798246383667, + "learning_rate": 1.0980966325036603e-08, + "loss": 1.4704, + "mean_token_accuracy": 0.6186267137527466, + "num_tokens": 804197.0, + "step": 31 + }, + { + "epoch": 0.0035141664836371624, + "ewc_loss": 1.6484591469634324e-12, + "grad_norm": 6.9016642570495605, + "learning_rate": 1.1346998535871157e-08, + "loss": 1.5128, + "mean_token_accuracy": 0.6059359908103943, + "num_tokens": 830796.0, + "step": 32 + }, + { + "epoch": 0.0036239841862508236, + "ewc_loss": 1.8189894035458565e-12, + "grad_norm": 6.58912467956543, + "learning_rate": 1.171303074670571e-08, + "loss": 1.4162, + "mean_token_accuracy": 0.640106201171875, + "num_tokens": 858092.0, + "step": 33 + }, + { + "epoch": 0.0037338018888644848, + "ewc_loss": 2.1032064978498966e-12, + "grad_norm": 7.311476230621338, + "learning_rate": 1.2079062957540264e-08, + "loss": 1.4557, + "mean_token_accuracy": 0.6129831075668335, + "num_tokens": 882474.0, + "step": 34 + }, + { + "epoch": 0.0038436195914781464, + "ewc_loss": 2.2311041902867146e-12, + "grad_norm": 8.547355651855469, + "learning_rate": 1.2445095168374816e-08, + "loss": 1.4568, + "mean_token_accuracy": 0.6254299879074097, + "num_tokens": 902361.0, + "step": 35 + }, + { + "epoch": 0.003953437294091808, + "ewc_loss": 2.3874235921539366e-12, + "grad_norm": 7.353146553039551, + "learning_rate": 1.281112737920937e-08, + "loss": 1.5078, + "mean_token_accuracy": 0.6058182716369629, + "num_tokens": 927171.0, + "step": 36 + }, + { + "epoch": 0.004063254996705469, + "ewc_loss": 2.5153212845907547e-12, + "grad_norm": 7.110287189483643, + "learning_rate": 1.3177159590043924e-08, + "loss": 1.4724, + "mean_token_accuracy": 0.6124988794326782, + "num_tokens": 952048.0, + "step": 37 + }, + { + "epoch": 0.00417307269931913, + "ewc_loss": 5.4853899200679734e-12, + "grad_norm": 6.875267505645752, + "learning_rate": 1.3543191800878477e-08, + "loss": 1.5739, + "mean_token_accuracy": 0.597423791885376, + "num_tokens": 981040.0, + "step": 38 + }, + { + "epoch": 0.004282890401932792, + "ewc_loss": 8.355982572538778e-12, + "grad_norm": 6.9071855545043945, + "learning_rate": 1.390922401171303e-08, + "loss": 1.4835, + "mean_token_accuracy": 0.6061628460884094, + "num_tokens": 1008756.0, + "step": 39 + }, + { + "epoch": 0.004392708104546453, + "ewc_loss": 9.72022462519817e-12, + "grad_norm": 7.116220951080322, + "learning_rate": 1.4275256222547585e-08, + "loss": 1.499, + "mean_token_accuracy": 0.601938009262085, + "num_tokens": 1035140.0, + "step": 40 + }, + { + "epoch": 0.004502525807160114, + "ewc_loss": 1.0629719326971099e-11, + "grad_norm": 7.807774543762207, + "learning_rate": 1.4641288433382137e-08, + "loss": 1.525, + "mean_token_accuracy": 0.6019235849380493, + "num_tokens": 1058792.0, + "step": 41 + }, + { + "epoch": 0.004612343509773776, + "ewc_loss": 1.1368683772161603e-11, + "grad_norm": 5.672109127044678, + "learning_rate": 1.500732064421669e-08, + "loss": 1.463, + "mean_token_accuracy": 0.6144424676895142, + "num_tokens": 1096383.0, + "step": 42 + }, + { + "epoch": 0.0047221612123874365, + "ewc_loss": 1.1823431123048067e-11, + "grad_norm": 6.795775890350342, + "learning_rate": 1.5373352855051243e-08, + "loss": 1.5744, + "mean_token_accuracy": 0.591993510723114, + "num_tokens": 1125313.0, + "step": 43 + }, + { + "epoch": 0.004831978915001098, + "ewc_loss": 1.2448708730516955e-11, + "grad_norm": 6.308107376098633, + "learning_rate": 1.57393850658858e-08, + "loss": 1.4453, + "mean_token_accuracy": 0.6225632429122925, + "num_tokens": 1155650.0, + "step": 44 + }, + { + "epoch": 0.00494179661761476, + "ewc_loss": 1.2789769243681803e-11, + "grad_norm": 7.576415538787842, + "learning_rate": 1.610541727672035e-08, + "loss": 1.468, + "mean_token_accuracy": 0.6196162104606628, + "num_tokens": 1178843.0, + "step": 45 + }, + { + "epoch": 0.0050516143202284204, + "ewc_loss": 1.3983481039758772e-11, + "grad_norm": 6.943343162536621, + "learning_rate": 1.6471449487554904e-08, + "loss": 1.5238, + "mean_token_accuracy": 0.6027805805206299, + "num_tokens": 1206271.0, + "step": 46 + }, + { + "epoch": 0.005161432022842082, + "ewc_loss": 1.5347723092418164e-11, + "grad_norm": 5.865456581115723, + "learning_rate": 1.6837481698389456e-08, + "loss": 1.4711, + "mean_token_accuracy": 0.6162903308868408, + "num_tokens": 1241401.0, + "step": 47 + }, + { + "epoch": 0.005271249725455744, + "ewc_loss": 1.5916157281026244e-11, + "grad_norm": 7.210322380065918, + "learning_rate": 1.7203513909224012e-08, + "loss": 1.4964, + "mean_token_accuracy": 0.6106781363487244, + "num_tokens": 1267455.0, + "step": 48 + }, + { + "epoch": 0.005381067428069404, + "ewc_loss": 1.6825651982799172e-11, + "grad_norm": 6.826697826385498, + "learning_rate": 1.7569546120058564e-08, + "loss": 1.3769, + "mean_token_accuracy": 0.6370328664779663, + "num_tokens": 1293955.0, + "step": 49 + }, + { + "epoch": 0.005490885130683066, + "ewc_loss": 1.7848833522293717e-11, + "grad_norm": 7.790121555328369, + "learning_rate": 1.7935578330893117e-08, + "loss": 1.4398, + "mean_token_accuracy": 0.6284909248352051, + "num_tokens": 1316419.0, + "step": 50 + }, + { + "epoch": 0.0056007028332967276, + "ewc_loss": 1.864464138634503e-11, + "grad_norm": 7.802056312561035, + "learning_rate": 1.8301610541727673e-08, + "loss": 1.5097, + "mean_token_accuracy": 0.6069738268852234, + "num_tokens": 1338607.0, + "step": 51 + }, + { + "epoch": 0.005710520535910389, + "ewc_loss": 2.0691004465334117e-11, + "grad_norm": 6.914675235748291, + "learning_rate": 1.8667642752562225e-08, + "loss": 1.4856, + "mean_token_accuracy": 0.6082533597946167, + "num_tokens": 1366081.0, + "step": 52 + }, + { + "epoch": 0.00582033823852405, + "ewc_loss": 3.296918293926865e-11, + "grad_norm": 7.443317413330078, + "learning_rate": 1.9033674963396777e-08, + "loss": 1.4826, + "mean_token_accuracy": 0.6140607595443726, + "num_tokens": 1390090.0, + "step": 53 + }, + { + "epoch": 0.0059301559411377115, + "ewc_loss": 4.888534022029489e-11, + "grad_norm": 7.5155792236328125, + "learning_rate": 1.9399707174231333e-08, + "loss": 1.4925, + "mean_token_accuracy": 0.6111409068107605, + "num_tokens": 1413964.0, + "step": 54 + }, + { + "epoch": 0.006039973643751373, + "ewc_loss": 5.707079253625125e-11, + "grad_norm": 6.622023582458496, + "learning_rate": 1.9765739385065885e-08, + "loss": 1.5151, + "mean_token_accuracy": 0.6089927554130554, + "num_tokens": 1442849.0, + "step": 55 + }, + { + "epoch": 0.006149791346365034, + "ewc_loss": 6.275513442233205e-11, + "grad_norm": 6.224631309509277, + "learning_rate": 2.0131771595900438e-08, + "loss": 1.5721, + "mean_token_accuracy": 0.5891210436820984, + "num_tokens": 1478493.0, + "step": 56 + }, + { + "epoch": 0.006259609048978695, + "ewc_loss": 6.730260793119669e-11, + "grad_norm": 8.227485656738281, + "learning_rate": 2.0497803806734994e-08, + "loss": 1.5253, + "mean_token_accuracy": 0.616789698600769, + "num_tokens": 1498811.0, + "step": 57 + }, + { + "epoch": 0.006369426751592357, + "ewc_loss": 7.139533408917487e-11, + "grad_norm": 6.813781261444092, + "learning_rate": 2.0863836017569546e-08, + "loss": 1.4462, + "mean_token_accuracy": 0.6245748996734619, + "num_tokens": 1525932.0, + "step": 58 + }, + { + "epoch": 0.006479244454206018, + "ewc_loss": 7.457856554538012e-11, + "grad_norm": 7.393991947174072, + "learning_rate": 2.1229868228404098e-08, + "loss": 1.4912, + "mean_token_accuracy": 0.6090189218521118, + "num_tokens": 1550438.0, + "step": 59 + }, + { + "epoch": 0.006589062156819679, + "ewc_loss": 7.73070496506989e-11, + "grad_norm": 8.673684120178223, + "learning_rate": 2.1595900439238654e-08, + "loss": 1.5012, + "mean_token_accuracy": 0.6125199198722839, + "num_tokens": 1569961.0, + "step": 60 + }, + { + "epoch": 0.006698879859433341, + "ewc_loss": 8.003553375601768e-11, + "grad_norm": 6.062989711761475, + "learning_rate": 2.1961932650073206e-08, + "loss": 1.4963, + "mean_token_accuracy": 0.605790913105011, + "num_tokens": 1604631.0, + "step": 61 + }, + { + "epoch": 0.006808697562047002, + "ewc_loss": 8.276401786133647e-11, + "grad_norm": 7.6397528648376465, + "learning_rate": 2.232796486090776e-08, + "loss": 1.4847, + "mean_token_accuracy": 0.6138597726821899, + "num_tokens": 1628098.0, + "step": 62 + }, + { + "epoch": 0.006918515264660663, + "ewc_loss": 8.685674401931465e-11, + "grad_norm": 6.6538472175598145, + "learning_rate": 2.2693997071742314e-08, + "loss": 1.4801, + "mean_token_accuracy": 0.6211367845535278, + "num_tokens": 1656281.0, + "step": 63 + }, + { + "epoch": 0.007028332967274325, + "ewc_loss": 8.958522812463343e-11, + "grad_norm": 7.442558765411377, + "learning_rate": 2.3060029282576867e-08, + "loss": 1.5132, + "mean_token_accuracy": 0.6158978939056396, + "num_tokens": 1680754.0, + "step": 64 + }, + { + "epoch": 0.007138150669887986, + "ewc_loss": 9.595169103704393e-11, + "grad_norm": 7.662631988525391, + "learning_rate": 2.342606149341142e-08, + "loss": 1.5119, + "mean_token_accuracy": 0.6084181070327759, + "num_tokens": 1704446.0, + "step": 65 + }, + { + "epoch": 0.007247968372501647, + "ewc_loss": 1.0413714335300028e-10, + "grad_norm": 6.888387203216553, + "learning_rate": 2.3792093704245972e-08, + "loss": 1.2965, + "mean_token_accuracy": 0.6489673852920532, + "num_tokens": 1728711.0, + "step": 66 + }, + { + "epoch": 0.007357786075115309, + "ewc_loss": 1.0959411156363785e-10, + "grad_norm": 7.938889026641846, + "learning_rate": 2.4158125915080527e-08, + "loss": 1.4218, + "mean_token_accuracy": 0.6310721635818481, + "num_tokens": 1750331.0, + "step": 67 + }, + { + "epoch": 0.0074676037777289695, + "ewc_loss": 1.141415850725025e-10, + "grad_norm": 6.346872329711914, + "learning_rate": 2.452415812591508e-08, + "loss": 1.4363, + "mean_token_accuracy": 0.6184351444244385, + "num_tokens": 1781737.0, + "step": 68 + }, + { + "epoch": 0.007577421480342631, + "ewc_loss": 1.1641532182693481e-10, + "grad_norm": 6.562450408935547, + "learning_rate": 2.4890190336749632e-08, + "loss": 1.4813, + "mean_token_accuracy": 0.6128454208374023, + "num_tokens": 1811021.0, + "step": 69 + }, + { + "epoch": 0.007687239182956293, + "ewc_loss": 1.2096279533579946e-10, + "grad_norm": 7.525042533874512, + "learning_rate": 2.5256222547584188e-08, + "loss": 1.5346, + "mean_token_accuracy": 0.5984227657318115, + "num_tokens": 1835564.0, + "step": 70 + }, + { + "epoch": 0.007797056885569954, + "ewc_loss": 1.2460077414289117e-10, + "grad_norm": 10.631875991821289, + "learning_rate": 2.562225475841874e-08, + "loss": 1.5251, + "mean_token_accuracy": 0.6131327152252197, + "num_tokens": 1850951.0, + "step": 71 + }, + { + "epoch": 0.007906874588183616, + "ewc_loss": 1.2732925824820995e-10, + "grad_norm": 7.125543594360352, + "learning_rate": 2.5988286969253293e-08, + "loss": 1.5375, + "mean_token_accuracy": 0.6028602123260498, + "num_tokens": 1877690.0, + "step": 72 + }, + { + "epoch": 0.008016692290797276, + "ewc_loss": 1.3278622645884752e-10, + "grad_norm": 7.01025390625, + "learning_rate": 2.6354319180087848e-08, + "loss": 1.577, + "mean_token_accuracy": 0.5918654799461365, + "num_tokens": 1904455.0, + "step": 73 + }, + { + "epoch": 0.008126509993410937, + "ewc_loss": 1.7280399333685637e-10, + "grad_norm": 8.018906593322754, + "learning_rate": 2.67203513909224e-08, + "loss": 1.6078, + "mean_token_accuracy": 0.5877993106842041, + "num_tokens": 1928049.0, + "step": 74 + }, + { + "epoch": 0.008236327696024599, + "ewc_loss": 2.4920154828578234e-10, + "grad_norm": 8.211451530456543, + "learning_rate": 2.7086383601756953e-08, + "loss": 1.4997, + "mean_token_accuracy": 0.6137468218803406, + "num_tokens": 1948422.0, + "step": 75 + }, + { + "epoch": 0.00834614539863826, + "ewc_loss": 3.0377123039215803e-10, + "grad_norm": 7.06309175491333, + "learning_rate": 2.745241581259151e-08, + "loss": 1.4678, + "mean_token_accuracy": 0.6176785230636597, + "num_tokens": 1974329.0, + "step": 76 + }, + { + "epoch": 0.008455963101251922, + "ewc_loss": 3.346940502524376e-10, + "grad_norm": 6.8395915031433105, + "learning_rate": 2.781844802342606e-08, + "loss": 1.4232, + "mean_token_accuracy": 0.6302651166915894, + "num_tokens": 2000531.0, + "step": 77 + }, + { + "epoch": 0.008565780803865584, + "ewc_loss": 3.5652192309498787e-10, + "grad_norm": 7.167519569396973, + "learning_rate": 2.8184480234260614e-08, + "loss": 1.4782, + "mean_token_accuracy": 0.6130858063697815, + "num_tokens": 2023823.0, + "step": 78 + }, + { + "epoch": 0.008675598506479244, + "ewc_loss": 3.728928277269006e-10, + "grad_norm": 7.376220703125, + "learning_rate": 2.855051244509517e-08, + "loss": 1.5141, + "mean_token_accuracy": 0.6031714677810669, + "num_tokens": 2048734.0, + "step": 79 + }, + { + "epoch": 0.008785416209092905, + "ewc_loss": 3.9108272176235914e-10, + "grad_norm": 6.702718257904053, + "learning_rate": 2.8916544655929722e-08, + "loss": 1.4142, + "mean_token_accuracy": 0.6270065307617188, + "num_tokens": 2075741.0, + "step": 80 + }, + { + "epoch": 0.008895233911706567, + "ewc_loss": 4.092726157978177e-10, + "grad_norm": 7.416596412658691, + "learning_rate": 2.9282576866764274e-08, + "loss": 1.5165, + "mean_token_accuracy": 0.6026204824447632, + "num_tokens": 2100247.0, + "step": 81 + }, + { + "epoch": 0.009005051614320228, + "ewc_loss": 4.2018655221909285e-10, + "grad_norm": 7.753892421722412, + "learning_rate": 2.964860907759883e-08, + "loss": 1.5993, + "mean_token_accuracy": 0.6029398441314697, + "num_tokens": 2122939.0, + "step": 82 + }, + { + "epoch": 0.00911486931693389, + "ewc_loss": 4.43833414465189e-10, + "grad_norm": 8.896093368530273, + "learning_rate": 3.001464128843338e-08, + "loss": 1.4991, + "mean_token_accuracy": 0.6034150123596191, + "num_tokens": 2141424.0, + "step": 83 + }, + { + "epoch": 0.009224687019547552, + "ewc_loss": 4.656612873077393e-10, + "grad_norm": 6.277707099914551, + "learning_rate": 3.038067349926794e-08, + "loss": 1.3978, + "mean_token_accuracy": 0.6275681853294373, + "num_tokens": 2171269.0, + "step": 84 + }, + { + "epoch": 0.009334504722161213, + "ewc_loss": 4.729372449219227e-10, + "grad_norm": 6.830385684967041, + "learning_rate": 3.074670571010249e-08, + "loss": 1.3793, + "mean_token_accuracy": 0.6453778147697449, + "num_tokens": 2195892.0, + "step": 85 + }, + { + "epoch": 0.009444322424774873, + "ewc_loss": 4.838511813431978e-10, + "grad_norm": 6.088253498077393, + "learning_rate": 3.111273792093704e-08, + "loss": 1.4783, + "mean_token_accuracy": 0.619906485080719, + "num_tokens": 2227992.0, + "step": 86 + }, + { + "epoch": 0.009554140127388535, + "ewc_loss": 4.94765117764473e-10, + "grad_norm": 6.572902202606201, + "learning_rate": 3.14787701317716e-08, + "loss": 1.5347, + "mean_token_accuracy": 0.6118224859237671, + "num_tokens": 2256995.0, + "step": 87 + }, + { + "epoch": 0.009663957830002196, + "ewc_loss": 5.093170329928398e-10, + "grad_norm": 6.212986469268799, + "learning_rate": 3.184480234260615e-08, + "loss": 1.4482, + "mean_token_accuracy": 0.622512936592102, + "num_tokens": 2286492.0, + "step": 88 + }, + { + "epoch": 0.009773775532615858, + "ewc_loss": 5.20230969414115e-10, + "grad_norm": 7.74827766418457, + "learning_rate": 3.22108345534407e-08, + "loss": 1.5043, + "mean_token_accuracy": 0.6076656579971313, + "num_tokens": 2309004.0, + "step": 89 + }, + { + "epoch": 0.00988359323522952, + "ewc_loss": 5.420588422566652e-10, + "grad_norm": 6.380218982696533, + "learning_rate": 3.257686676427526e-08, + "loss": 1.4034, + "mean_token_accuracy": 0.621545135974884, + "num_tokens": 2336480.0, + "step": 90 + }, + { + "epoch": 0.009993410937843181, + "ewc_loss": 5.820766091346741e-10, + "grad_norm": 6.671430587768555, + "learning_rate": 3.294289897510981e-08, + "loss": 1.4479, + "mean_token_accuracy": 0.6150302886962891, + "num_tokens": 2363526.0, + "step": 91 + }, + { + "epoch": 0.010103228640456841, + "ewc_loss": 6.184563972055912e-10, + "grad_norm": 6.265373706817627, + "learning_rate": 3.3308931185944364e-08, + "loss": 1.3989, + "mean_token_accuracy": 0.6285728216171265, + "num_tokens": 2393847.0, + "step": 92 + }, + { + "epoch": 0.010213046343070502, + "ewc_loss": 6.366462912410498e-10, + "grad_norm": 6.643059253692627, + "learning_rate": 3.367496339677891e-08, + "loss": 1.498, + "mean_token_accuracy": 0.6057474613189697, + "num_tokens": 2421639.0, + "step": 93 + }, + { + "epoch": 0.010322864045684164, + "ewc_loss": 6.548361852765083e-10, + "grad_norm": 7.0453782081604, + "learning_rate": 3.404099560761347e-08, + "loss": 1.4513, + "mean_token_accuracy": 0.6231608986854553, + "num_tokens": 2445896.0, + "step": 94 + }, + { + "epoch": 0.010432681748297826, + "ewc_loss": 6.730260793119669e-10, + "grad_norm": 6.57490873336792, + "learning_rate": 3.4407027818448024e-08, + "loss": 1.5578, + "mean_token_accuracy": 0.6007099747657776, + "num_tokens": 2475814.0, + "step": 95 + }, + { + "epoch": 0.010542499450911487, + "ewc_loss": 6.948539521545172e-10, + "grad_norm": 7.18787145614624, + "learning_rate": 3.477306002928257e-08, + "loss": 1.3836, + "mean_token_accuracy": 0.6397630572319031, + "num_tokens": 2499512.0, + "step": 96 + }, + { + "epoch": 0.010652317153525149, + "ewc_loss": 7.166818249970675e-10, + "grad_norm": 6.585120677947998, + "learning_rate": 3.513909224011713e-08, + "loss": 1.56, + "mean_token_accuracy": 0.5938551425933838, + "num_tokens": 2528412.0, + "step": 97 + }, + { + "epoch": 0.010762134856138809, + "ewc_loss": 7.421476766467094e-10, + "grad_norm": 7.1469621658325195, + "learning_rate": 3.5505124450951685e-08, + "loss": 1.5045, + "mean_token_accuracy": 0.6090213060379028, + "num_tokens": 2553855.0, + "step": 98 + }, + { + "epoch": 0.01087195255875247, + "ewc_loss": 7.566995918750763e-10, + "grad_norm": 8.147625923156738, + "learning_rate": 3.5871156661786234e-08, + "loss": 1.5374, + "mean_token_accuracy": 0.6076910495758057, + "num_tokens": 2574996.0, + "step": 99 + }, + { + "epoch": 0.010981770261366132, + "ewc_loss": 7.785274647176266e-10, + "grad_norm": 6.783403396606445, + "learning_rate": 3.623718887262079e-08, + "loss": 1.4907, + "mean_token_accuracy": 0.6125013828277588, + "num_tokens": 2601684.0, + "step": 100 + }, + { + "epoch": 0.011091587963979794, + "ewc_loss": 7.930793799459934e-10, + "grad_norm": 7.0172224044799805, + "learning_rate": 3.6603221083455345e-08, + "loss": 1.5378, + "mean_token_accuracy": 0.6014245748519897, + "num_tokens": 2628612.0, + "step": 101 + }, + { + "epoch": 0.011201405666593455, + "ewc_loss": 8.185452315956354e-10, + "grad_norm": 6.926872730255127, + "learning_rate": 3.6969253294289894e-08, + "loss": 1.4373, + "mean_token_accuracy": 0.6378090381622314, + "num_tokens": 2652776.0, + "step": 102 + }, + { + "epoch": 0.011311223369207117, + "ewc_loss": 8.658389560878277e-10, + "grad_norm": 6.953574180603027, + "learning_rate": 3.733528550512445e-08, + "loss": 1.4063, + "mean_token_accuracy": 0.6306692361831665, + "num_tokens": 2677807.0, + "step": 103 + }, + { + "epoch": 0.011421041071820778, + "ewc_loss": 9.677023626863956e-10, + "grad_norm": 8.456330299377441, + "learning_rate": 3.7701317715959006e-08, + "loss": 1.5545, + "mean_token_accuracy": 0.6001198291778564, + "num_tokens": 2697989.0, + "step": 104 + }, + { + "epoch": 0.011530858774434438, + "ewc_loss": 1.1059455573558807e-09, + "grad_norm": 7.066056728363037, + "learning_rate": 3.8067349926793555e-08, + "loss": 1.541, + "mean_token_accuracy": 0.6018768548965454, + "num_tokens": 2722341.0, + "step": 105 + }, + { + "epoch": 0.0116406764770481, + "ewc_loss": 1.2514647096395493e-09, + "grad_norm": 6.159657001495361, + "learning_rate": 3.843338213762811e-08, + "loss": 1.5188, + "mean_token_accuracy": 0.6022439002990723, + "num_tokens": 2753675.0, + "step": 106 + }, + { + "epoch": 0.011750494179661761, + "ewc_loss": 1.367880031466484e-09, + "grad_norm": 6.796925067901611, + "learning_rate": 3.8799414348462666e-08, + "loss": 1.4658, + "mean_token_accuracy": 0.6133979558944702, + "num_tokens": 2778974.0, + "step": 107 + }, + { + "epoch": 0.011860311882275423, + "ewc_loss": 1.433363649994135e-09, + "grad_norm": 6.242575645446777, + "learning_rate": 3.9165446559297215e-08, + "loss": 1.5176, + "mean_token_accuracy": 0.602668046951294, + "num_tokens": 2808967.0, + "step": 108 + }, + { + "epoch": 0.011970129584889085, + "ewc_loss": 1.4915713109076023e-09, + "grad_norm": 6.6226067543029785, + "learning_rate": 3.953147877013177e-08, + "loss": 1.5115, + "mean_token_accuracy": 0.601077675819397, + "num_tokens": 2835393.0, + "step": 109 + }, + { + "epoch": 0.012079947287502746, + "ewc_loss": 1.5425030142068863e-09, + "grad_norm": 6.03671932220459, + "learning_rate": 3.9897510980966327e-08, + "loss": 1.4724, + "mean_token_accuracy": 0.6121670007705688, + "num_tokens": 2863678.0, + "step": 110 + }, + { + "epoch": 0.012189764990116406, + "ewc_loss": 1.6152625903487206e-09, + "grad_norm": 6.500568866729736, + "learning_rate": 4.0263543191800876e-08, + "loss": 1.4449, + "mean_token_accuracy": 0.6228834986686707, + "num_tokens": 2890376.0, + "step": 111 + }, + { + "epoch": 0.012299582692730068, + "ewc_loss": 1.673470251262188e-09, + "grad_norm": 6.383557319641113, + "learning_rate": 4.062957540263543e-08, + "loss": 1.4413, + "mean_token_accuracy": 0.6233298778533936, + "num_tokens": 2917128.0, + "step": 112 + }, + { + "epoch": 0.01240940039534373, + "ewc_loss": 1.7171259969472885e-09, + "grad_norm": 6.936566352844238, + "learning_rate": 4.099560761346999e-08, + "loss": 1.4759, + "mean_token_accuracy": 0.6101064085960388, + "num_tokens": 2940812.0, + "step": 113 + }, + { + "epoch": 0.01251921809795739, + "ewc_loss": 1.7535057850182056e-09, + "grad_norm": 6.276569366455078, + "learning_rate": 4.1361639824304536e-08, + "loss": 1.5024, + "mean_token_accuracy": 0.6026677489280701, + "num_tokens": 2969881.0, + "step": 114 + }, + { + "epoch": 0.012629035800571052, + "ewc_loss": 1.7971615307033062e-09, + "grad_norm": 6.712355613708496, + "learning_rate": 4.172767203513909e-08, + "loss": 1.5015, + "mean_token_accuracy": 0.6077354550361633, + "num_tokens": 2996049.0, + "step": 115 + }, + { + "epoch": 0.012738853503184714, + "ewc_loss": 1.862645149230957e-09, + "grad_norm": 5.762145042419434, + "learning_rate": 4.209370424597365e-08, + "loss": 1.3714, + "mean_token_accuracy": 0.632451057434082, + "num_tokens": 3025829.0, + "step": 116 + }, + { + "epoch": 0.012848671205798374, + "ewc_loss": 1.9063008949160576e-09, + "grad_norm": 6.8401994705200195, + "learning_rate": 4.2459736456808197e-08, + "loss": 1.4889, + "mean_token_accuracy": 0.6105377674102783, + "num_tokens": 3051012.0, + "step": 117 + }, + { + "epoch": 0.012958488908412035, + "ewc_loss": 1.964508555829525e-09, + "grad_norm": 6.4564409255981445, + "learning_rate": 4.282576866764275e-08, + "loss": 1.5297, + "mean_token_accuracy": 0.6049107909202576, + "num_tokens": 3077750.0, + "step": 118 + }, + { + "epoch": 0.013068306611025697, + "ewc_loss": 2.051820047199726e-09, + "grad_norm": 6.668231010437012, + "learning_rate": 4.319180087847731e-08, + "loss": 1.4797, + "mean_token_accuracy": 0.6087673902511597, + "num_tokens": 3103271.0, + "step": 119 + }, + { + "epoch": 0.013178124313639359, + "ewc_loss": 2.08092387765646e-09, + "grad_norm": 7.399384498596191, + "learning_rate": 4.355783308931186e-08, + "loss": 1.5062, + "mean_token_accuracy": 0.6008899211883545, + "num_tokens": 3124511.0, + "step": 120 + }, + { + "epoch": 0.01328794201625302, + "ewc_loss": 2.1245796233415604e-09, + "grad_norm": 6.481358051300049, + "learning_rate": 4.392386530014641e-08, + "loss": 1.4025, + "mean_token_accuracy": 0.6359106302261353, + "num_tokens": 3150451.0, + "step": 121 + }, + { + "epoch": 0.013397759718866682, + "ewc_loss": 2.168235369026661e-09, + "grad_norm": 6.132983684539795, + "learning_rate": 4.428989751098097e-08, + "loss": 1.5191, + "mean_token_accuracy": 0.6021833419799805, + "num_tokens": 3180151.0, + "step": 122 + }, + { + "epoch": 0.013507577421480343, + "ewc_loss": 2.2264430299401283e-09, + "grad_norm": 6.403000354766846, + "learning_rate": 4.465592972181552e-08, + "loss": 1.5134, + "mean_token_accuracy": 0.6062251329421997, + "num_tokens": 3207702.0, + "step": 123 + }, + { + "epoch": 0.013617395124094003, + "ewc_loss": 2.270098775625229e-09, + "grad_norm": 8.122723579406738, + "learning_rate": 4.502196193265007e-08, + "loss": 1.3943, + "mean_token_accuracy": 0.637534499168396, + "num_tokens": 3226168.0, + "step": 124 + }, + { + "epoch": 0.013727212826707665, + "ewc_loss": 2.35741026699543e-09, + "grad_norm": 6.2556610107421875, + "learning_rate": 4.538799414348463e-08, + "loss": 1.4777, + "mean_token_accuracy": 0.6083693504333496, + "num_tokens": 3255714.0, + "step": 125 + }, + { + "epoch": 0.013837030529321327, + "ewc_loss": 2.444721758365631e-09, + "grad_norm": 6.3704833984375, + "learning_rate": 4.575402635431918e-08, + "loss": 1.5158, + "mean_token_accuracy": 0.6078286170959473, + "num_tokens": 3283574.0, + "step": 126 + }, + { + "epoch": 0.013946848231934988, + "ewc_loss": 2.5174813345074654e-09, + "grad_norm": 6.537268161773682, + "learning_rate": 4.6120058565153734e-08, + "loss": 1.4589, + "mean_token_accuracy": 0.6194194555282593, + "num_tokens": 3309128.0, + "step": 127 + }, + { + "epoch": 0.01405666593454865, + "ewc_loss": 2.6047928258776665e-09, + "grad_norm": 7.40941047668457, + "learning_rate": 4.648609077598828e-08, + "loss": 1.5124, + "mean_token_accuracy": 0.6050661206245422, + "num_tokens": 3331577.0, + "step": 128 + }, + { + "epoch": 0.014166483637162311, + "ewc_loss": 2.7212081477046013e-09, + "grad_norm": 6.531487464904785, + "learning_rate": 4.685212298682284e-08, + "loss": 1.5675, + "mean_token_accuracy": 0.5893762111663818, + "num_tokens": 3358884.0, + "step": 129 + }, + { + "epoch": 0.014276301339775971, + "ewc_loss": 2.8085196390748024e-09, + "grad_norm": 8.324472427368164, + "learning_rate": 4.7218155197657394e-08, + "loss": 1.5932, + "mean_token_accuracy": 0.588712751865387, + "num_tokens": 3378122.0, + "step": 130 + }, + { + "epoch": 0.014386119042389633, + "ewc_loss": 2.9103830456733704e-09, + "grad_norm": 6.8913655281066895, + "learning_rate": 4.7584187408491943e-08, + "loss": 1.5473, + "mean_token_accuracy": 0.601823627948761, + "num_tokens": 3404642.0, + "step": 131 + }, + { + "epoch": 0.014495936745003294, + "ewc_loss": 2.9831426218152046e-09, + "grad_norm": 7.0495100021362305, + "learning_rate": 4.79502196193265e-08, + "loss": 1.4417, + "mean_token_accuracy": 0.6155201196670532, + "num_tokens": 3429186.0, + "step": 132 + }, + { + "epoch": 0.014605754447616956, + "ewc_loss": 3.0704541131854057e-09, + "grad_norm": 5.908855438232422, + "learning_rate": 4.8316251830161055e-08, + "loss": 1.4775, + "mean_token_accuracy": 0.6187652349472046, + "num_tokens": 3460810.0, + "step": 133 + }, + { + "epoch": 0.014715572150230618, + "ewc_loss": 3.157765604555607e-09, + "grad_norm": 5.937631607055664, + "learning_rate": 4.8682284040995604e-08, + "loss": 1.4437, + "mean_token_accuracy": 0.6178042888641357, + "num_tokens": 3491285.0, + "step": 134 + }, + { + "epoch": 0.01482538985284428, + "ewc_loss": 3.2159732654690742e-09, + "grad_norm": 6.403355598449707, + "learning_rate": 4.904831625183016e-08, + "loss": 1.4388, + "mean_token_accuracy": 0.6208891868591309, + "num_tokens": 3517893.0, + "step": 135 + }, + { + "epoch": 0.014935207555457939, + "ewc_loss": 3.3178366720676422e-09, + "grad_norm": 6.474295139312744, + "learning_rate": 4.9414348462664715e-08, + "loss": 1.4895, + "mean_token_accuracy": 0.6069913506507874, + "num_tokens": 3545614.0, + "step": 136 + }, + { + "epoch": 0.0150450252580716, + "ewc_loss": 3.4051481634378433e-09, + "grad_norm": 6.948873519897461, + "learning_rate": 4.9780380673499264e-08, + "loss": 1.5333, + "mean_token_accuracy": 0.5982491374015808, + "num_tokens": 3571156.0, + "step": 137 + }, + { + "epoch": 0.015154842960685262, + "ewc_loss": 3.4924596548080444e-09, + "grad_norm": 6.924042224884033, + "learning_rate": 5.014641288433382e-08, + "loss": 1.4294, + "mean_token_accuracy": 0.6232999563217163, + "num_tokens": 3595596.0, + "step": 138 + }, + { + "epoch": 0.015264660663298924, + "ewc_loss": 3.5797711461782455e-09, + "grad_norm": 6.047142505645752, + "learning_rate": 5.0512445095168376e-08, + "loss": 1.442, + "mean_token_accuracy": 0.6334732174873352, + "num_tokens": 3625562.0, + "step": 139 + }, + { + "epoch": 0.015374478365912585, + "ewc_loss": 3.6670826375484467e-09, + "grad_norm": 7.117671489715576, + "learning_rate": 5.0878477306002925e-08, + "loss": 1.4961, + "mean_token_accuracy": 0.6199361085891724, + "num_tokens": 3651717.0, + "step": 140 + }, + { + "epoch": 0.015484296068526247, + "ewc_loss": 3.7834979593753815e-09, + "grad_norm": 7.224852561950684, + "learning_rate": 5.124450951683748e-08, + "loss": 1.3791, + "mean_token_accuracy": 0.6362611055374146, + "num_tokens": 3674773.0, + "step": 141 + }, + { + "epoch": 0.015594113771139909, + "ewc_loss": 3.92901711165905e-09, + "grad_norm": 7.061892509460449, + "learning_rate": 5.1610541727672036e-08, + "loss": 1.5633, + "mean_token_accuracy": 0.5894743800163269, + "num_tokens": 3701717.0, + "step": 142 + }, + { + "epoch": 0.01570393147375357, + "ewc_loss": 4.045432433485985e-09, + "grad_norm": 6.156381607055664, + "learning_rate": 5.1976573938506585e-08, + "loss": 1.4102, + "mean_token_accuracy": 0.6267715692520142, + "num_tokens": 3729901.0, + "step": 143 + }, + { + "epoch": 0.015813749176367232, + "ewc_loss": 4.220055416226387e-09, + "grad_norm": 6.249547004699707, + "learning_rate": 5.234260614934114e-08, + "loss": 1.5112, + "mean_token_accuracy": 0.5949780941009521, + "num_tokens": 3757405.0, + "step": 144 + }, + { + "epoch": 0.01592356687898089, + "ewc_loss": 4.3655745685100555e-09, + "grad_norm": 5.922534942626953, + "learning_rate": 5.2708638360175697e-08, + "loss": 1.3735, + "mean_token_accuracy": 0.6362641453742981, + "num_tokens": 3787289.0, + "step": 145 + }, + { + "epoch": 0.01603338458159455, + "ewc_loss": 4.423782229423523e-09, + "grad_norm": 6.448569297790527, + "learning_rate": 5.3074670571010246e-08, + "loss": 1.4096, + "mean_token_accuracy": 0.6326770186424255, + "num_tokens": 3811919.0, + "step": 146 + }, + { + "epoch": 0.016143202284208215, + "ewc_loss": 4.598405212163925e-09, + "grad_norm": 6.366332054138184, + "learning_rate": 5.34407027818448e-08, + "loss": 1.3516, + "mean_token_accuracy": 0.6393665075302124, + "num_tokens": 3835967.0, + "step": 147 + }, + { + "epoch": 0.016253019986821875, + "ewc_loss": 4.71482053399086e-09, + "grad_norm": 6.430514335632324, + "learning_rate": 5.380673499267936e-08, + "loss": 1.397, + "mean_token_accuracy": 0.6269119381904602, + "num_tokens": 3859268.0, + "step": 148 + }, + { + "epoch": 0.016362837689435538, + "ewc_loss": 4.889443516731262e-09, + "grad_norm": 5.18072509765625, + "learning_rate": 5.4172767203513906e-08, + "loss": 1.4468, + "mean_token_accuracy": 0.6131888628005981, + "num_tokens": 3892417.0, + "step": 149 + }, + { + "epoch": 0.016472655392049198, + "ewc_loss": 5.180481821298599e-09, + "grad_norm": 7.067427635192871, + "learning_rate": 5.453879941434846e-08, + "loss": 1.3739, + "mean_token_accuracy": 0.6362538933753967, + "num_tokens": 3913323.0, + "step": 150 + }, + { + "epoch": 0.01658247309466286, + "ewc_loss": 5.413312464952469e-09, + "grad_norm": 5.0731072425842285, + "learning_rate": 5.490483162518302e-08, + "loss": 1.2672, + "mean_token_accuracy": 0.6560317277908325, + "num_tokens": 3942390.0, + "step": 151 + }, + { + "epoch": 0.01669229079727652, + "ewc_loss": 5.558831617236137e-09, + "grad_norm": 5.941414833068848, + "learning_rate": 5.527086383601757e-08, + "loss": 1.4524, + "mean_token_accuracy": 0.6137887835502625, + "num_tokens": 3968060.0, + "step": 152 + }, + { + "epoch": 0.01680210849989018, + "ewc_loss": 5.762558430433273e-09, + "grad_norm": 5.179152011871338, + "learning_rate": 5.563689604685212e-08, + "loss": 1.329, + "mean_token_accuracy": 0.6440565586090088, + "num_tokens": 3996149.0, + "step": 153 + }, + { + "epoch": 0.016911926202503844, + "ewc_loss": 5.878973752260208e-09, + "grad_norm": 5.538205146789551, + "learning_rate": 5.600292825768668e-08, + "loss": 1.4636, + "mean_token_accuracy": 0.6076212525367737, + "num_tokens": 4022708.0, + "step": 154 + }, + { + "epoch": 0.017021743905117504, + "ewc_loss": 6.05359673500061e-09, + "grad_norm": 5.9023590087890625, + "learning_rate": 5.636896046852123e-08, + "loss": 1.4981, + "mean_token_accuracy": 0.6097058057785034, + "num_tokens": 4048984.0, + "step": 155 + }, + { + "epoch": 0.017131561607731167, + "ewc_loss": 6.199115887284279e-09, + "grad_norm": 5.328039169311523, + "learning_rate": 5.673499267935578e-08, + "loss": 1.3853, + "mean_token_accuracy": 0.6300330758094788, + "num_tokens": 4075654.0, + "step": 156 + }, + { + "epoch": 0.017241379310344827, + "ewc_loss": 6.373738870024681e-09, + "grad_norm": 5.295948028564453, + "learning_rate": 5.710102489019034e-08, + "loss": 1.5405, + "mean_token_accuracy": 0.5877740383148193, + "num_tokens": 4106959.0, + "step": 157 + }, + { + "epoch": 0.017351197012958487, + "ewc_loss": 6.6356733441352844e-09, + "grad_norm": 5.303983211517334, + "learning_rate": 5.746705710102489e-08, + "loss": 1.43, + "mean_token_accuracy": 0.623774528503418, + "num_tokens": 4135143.0, + "step": 158 + }, + { + "epoch": 0.01746101471557215, + "ewc_loss": 6.83940015733242e-09, + "grad_norm": 5.134082794189453, + "learning_rate": 5.7833089311859443e-08, + "loss": 1.3716, + "mean_token_accuracy": 0.6370586156845093, + "num_tokens": 4162641.0, + "step": 159 + }, + { + "epoch": 0.01757083241818581, + "ewc_loss": 6.984919309616089e-09, + "grad_norm": 5.224027633666992, + "learning_rate": 5.8199121522694e-08, + "loss": 1.3296, + "mean_token_accuracy": 0.6351552605628967, + "num_tokens": 4189265.0, + "step": 160 + }, + { + "epoch": 0.017680650120799474, + "ewc_loss": 7.159542292356491e-09, + "grad_norm": 6.153186798095703, + "learning_rate": 5.856515373352855e-08, + "loss": 1.4976, + "mean_token_accuracy": 0.6058662533760071, + "num_tokens": 4209765.0, + "step": 161 + }, + { + "epoch": 0.017790467823413134, + "ewc_loss": 7.334165275096893e-09, + "grad_norm": 5.896090507507324, + "learning_rate": 5.8931185944363104e-08, + "loss": 1.392, + "mean_token_accuracy": 0.6262022256851196, + "num_tokens": 4232444.0, + "step": 162 + }, + { + "epoch": 0.017900285526026797, + "ewc_loss": 7.450580596923828e-09, + "grad_norm": 4.710315704345703, + "learning_rate": 5.929721815519766e-08, + "loss": 1.3951, + "mean_token_accuracy": 0.6198011040687561, + "num_tokens": 4263314.0, + "step": 163 + }, + { + "epoch": 0.018010103228640457, + "ewc_loss": 7.450580596923828e-09, + "grad_norm": 4.84669828414917, + "learning_rate": 5.966325036603222e-08, + "loss": 1.4032, + "mean_token_accuracy": 0.6265224814414978, + "num_tokens": 4292268.0, + "step": 164 + }, + { + "epoch": 0.018119920931254117, + "ewc_loss": 7.566995918750763e-09, + "grad_norm": 5.347260475158691, + "learning_rate": 6.002928257686676e-08, + "loss": 1.4704, + "mean_token_accuracy": 0.6153043508529663, + "num_tokens": 4317038.0, + "step": 165 + }, + { + "epoch": 0.01822973863386778, + "ewc_loss": 7.799826562404633e-09, + "grad_norm": 5.143074989318848, + "learning_rate": 6.039531478770131e-08, + "loss": 1.4735, + "mean_token_accuracy": 0.6042478084564209, + "num_tokens": 4344439.0, + "step": 166 + }, + { + "epoch": 0.01833955633648144, + "ewc_loss": 7.916241884231567e-09, + "grad_norm": 5.031787395477295, + "learning_rate": 6.076134699853588e-08, + "loss": 1.3819, + "mean_token_accuracy": 0.6277737617492676, + "num_tokens": 4371828.0, + "step": 167 + }, + { + "epoch": 0.018449374039095103, + "ewc_loss": 7.974449545145035e-09, + "grad_norm": 5.740262031555176, + "learning_rate": 6.112737920937042e-08, + "loss": 1.3015, + "mean_token_accuracy": 0.6419751048088074, + "num_tokens": 4393611.0, + "step": 168 + }, + { + "epoch": 0.018559191741708763, + "ewc_loss": 8.149072527885437e-09, + "grad_norm": 5.118387699127197, + "learning_rate": 6.149341142020497e-08, + "loss": 1.3818, + "mean_token_accuracy": 0.6270297765731812, + "num_tokens": 4419458.0, + "step": 169 + }, + { + "epoch": 0.018669009444322426, + "ewc_loss": 8.265487849712372e-09, + "grad_norm": 5.046391487121582, + "learning_rate": 6.185944363103954e-08, + "loss": 1.3353, + "mean_token_accuracy": 0.6409823298454285, + "num_tokens": 4444110.0, + "step": 170 + }, + { + "epoch": 0.018778827146936086, + "ewc_loss": 8.32369551062584e-09, + "grad_norm": 5.875481605529785, + "learning_rate": 6.222547584187409e-08, + "loss": 1.3719, + "mean_token_accuracy": 0.6265662312507629, + "num_tokens": 4465140.0, + "step": 171 + }, + { + "epoch": 0.018888644849549746, + "ewc_loss": 8.440110832452774e-09, + "grad_norm": 5.22693395614624, + "learning_rate": 6.259150805270863e-08, + "loss": 1.3929, + "mean_token_accuracy": 0.6219501495361328, + "num_tokens": 4489605.0, + "step": 172 + }, + { + "epoch": 0.01899846255216341, + "ewc_loss": 8.498318493366241e-09, + "grad_norm": 4.922569274902344, + "learning_rate": 6.29575402635432e-08, + "loss": 1.3405, + "mean_token_accuracy": 0.6354894042015076, + "num_tokens": 4515095.0, + "step": 173 + }, + { + "epoch": 0.01910828025477707, + "ewc_loss": 8.556526154279709e-09, + "grad_norm": 4.791773796081543, + "learning_rate": 6.332357247437773e-08, + "loss": 1.3746, + "mean_token_accuracy": 0.6239596605300903, + "num_tokens": 4541160.0, + "step": 174 + }, + { + "epoch": 0.019218097957390733, + "ewc_loss": 8.672941476106644e-09, + "grad_norm": 5.410389423370361, + "learning_rate": 6.36896046852123e-08, + "loss": 1.3629, + "mean_token_accuracy": 0.62598717212677, + "num_tokens": 4563441.0, + "step": 175 + }, + { + "epoch": 0.019327915660004392, + "ewc_loss": 8.905772119760513e-09, + "grad_norm": 5.117117404937744, + "learning_rate": 6.405563689604684e-08, + "loss": 1.3109, + "mean_token_accuracy": 0.6398118734359741, + "num_tokens": 4586881.0, + "step": 176 + }, + { + "epoch": 0.019437733362618052, + "ewc_loss": 9.080395102500916e-09, + "grad_norm": 4.595563888549805, + "learning_rate": 6.44216691068814e-08, + "loss": 1.3935, + "mean_token_accuracy": 0.6226621270179749, + "num_tokens": 4616811.0, + "step": 177 + }, + { + "epoch": 0.019547551065231716, + "ewc_loss": 9.313225746154785e-09, + "grad_norm": 5.177950382232666, + "learning_rate": 6.478770131771596e-08, + "loss": 1.2355, + "mean_token_accuracy": 0.6584978103637695, + "num_tokens": 4639573.0, + "step": 178 + }, + { + "epoch": 0.019657368767845376, + "ewc_loss": 9.42964106798172e-09, + "grad_norm": 4.818875312805176, + "learning_rate": 6.515373352855052e-08, + "loss": 1.4225, + "mean_token_accuracy": 0.618417501449585, + "num_tokens": 4666392.0, + "step": 179 + }, + { + "epoch": 0.01976718647045904, + "ewc_loss": 9.604264050722122e-09, + "grad_norm": 4.9884033203125, + "learning_rate": 6.551976573938505e-08, + "loss": 1.3214, + "mean_token_accuracy": 0.6358859539031982, + "num_tokens": 4690288.0, + "step": 180 + }, + { + "epoch": 0.0198770041730727, + "ewc_loss": 9.953510016202927e-09, + "grad_norm": 4.664217948913574, + "learning_rate": 6.588579795021962e-08, + "loss": 1.3611, + "mean_token_accuracy": 0.6245070099830627, + "num_tokens": 4716583.0, + "step": 181 + }, + { + "epoch": 0.019986821875686362, + "ewc_loss": 1.0128132998943329e-08, + "grad_norm": 4.856940746307373, + "learning_rate": 6.625183016105416e-08, + "loss": 1.2923, + "mean_token_accuracy": 0.6483598351478577, + "num_tokens": 4741315.0, + "step": 182 + }, + { + "epoch": 0.020096639578300022, + "ewc_loss": 1.0302755981683731e-08, + "grad_norm": 4.546483039855957, + "learning_rate": 6.661786237188873e-08, + "loss": 1.4541, + "mean_token_accuracy": 0.6118334531784058, + "num_tokens": 4771862.0, + "step": 183 + }, + { + "epoch": 0.020206457280913682, + "ewc_loss": 1.0419171303510666e-08, + "grad_norm": 4.730998992919922, + "learning_rate": 6.698389458272328e-08, + "loss": 1.4037, + "mean_token_accuracy": 0.6216364502906799, + "num_tokens": 4798609.0, + "step": 184 + }, + { + "epoch": 0.020316274983527345, + "ewc_loss": 1.0710209608078003e-08, + "grad_norm": 4.782520294189453, + "learning_rate": 6.734992679355783e-08, + "loss": 1.3454, + "mean_token_accuracy": 0.6297258138656616, + "num_tokens": 4821516.0, + "step": 185 + }, + { + "epoch": 0.020426092686141005, + "ewc_loss": 1.0884832590818405e-08, + "grad_norm": 4.862817287445068, + "learning_rate": 6.771595900439237e-08, + "loss": 1.3259, + "mean_token_accuracy": 0.641615629196167, + "num_tokens": 4845487.0, + "step": 186 + }, + { + "epoch": 0.02053591038875467, + "ewc_loss": 1.1117663234472275e-08, + "grad_norm": 5.014048099517822, + "learning_rate": 6.808199121522694e-08, + "loss": 1.3922, + "mean_token_accuracy": 0.6292468309402466, + "num_tokens": 4868981.0, + "step": 187 + }, + { + "epoch": 0.020645728091368328, + "ewc_loss": 1.1408701539039612e-08, + "grad_norm": 4.470716953277588, + "learning_rate": 6.844802342606149e-08, + "loss": 1.3661, + "mean_token_accuracy": 0.6243540644645691, + "num_tokens": 4895095.0, + "step": 188 + }, + { + "epoch": 0.02075554579398199, + "ewc_loss": 1.1525116860866547e-08, + "grad_norm": 4.761801242828369, + "learning_rate": 6.881405563689605e-08, + "loss": 1.4301, + "mean_token_accuracy": 0.6141797304153442, + "num_tokens": 4920215.0, + "step": 189 + }, + { + "epoch": 0.02086536349659565, + "ewc_loss": 1.1699739843606949e-08, + "grad_norm": 4.426476955413818, + "learning_rate": 6.91800878477306e-08, + "loss": 1.3944, + "mean_token_accuracy": 0.6198953986167908, + "num_tokens": 4950113.0, + "step": 190 + }, + { + "epoch": 0.02097518119920931, + "ewc_loss": 1.1932570487260818e-08, + "grad_norm": 5.206794261932373, + "learning_rate": 6.954612005856515e-08, + "loss": 1.3938, + "mean_token_accuracy": 0.6186672449111938, + "num_tokens": 4970901.0, + "step": 191 + }, + { + "epoch": 0.021084998901822975, + "ewc_loss": 1.2398231774568558e-08, + "grad_norm": 4.4039740562438965, + "learning_rate": 6.99121522693997e-08, + "loss": 1.3317, + "mean_token_accuracy": 0.6373385787010193, + "num_tokens": 4996117.0, + "step": 192 + }, + { + "epoch": 0.021194816604436634, + "ewc_loss": 1.2689270079135895e-08, + "grad_norm": 4.500173091888428, + "learning_rate": 7.027818448023426e-08, + "loss": 1.3673, + "mean_token_accuracy": 0.6255741119384766, + "num_tokens": 5020994.0, + "step": 193 + }, + { + "epoch": 0.021304634307050298, + "ewc_loss": 1.2922100722789764e-08, + "grad_norm": 4.317251205444336, + "learning_rate": 7.064421669106881e-08, + "loss": 1.4246, + "mean_token_accuracy": 0.618274450302124, + "num_tokens": 5053438.0, + "step": 194 + }, + { + "epoch": 0.021414452009663958, + "ewc_loss": 1.3213139027357101e-08, + "grad_norm": 5.1979875564575195, + "learning_rate": 7.101024890190337e-08, + "loss": 1.4082, + "mean_token_accuracy": 0.6197792887687683, + "num_tokens": 5077854.0, + "step": 195 + }, + { + "epoch": 0.021524269712277617, + "ewc_loss": 1.3504177331924438e-08, + "grad_norm": 4.780888080596924, + "learning_rate": 7.137628111273792e-08, + "loss": 1.4045, + "mean_token_accuracy": 0.6133818626403809, + "num_tokens": 5103447.0, + "step": 196 + }, + { + "epoch": 0.02163408741489128, + "ewc_loss": 1.3737007975578308e-08, + "grad_norm": 4.717243671417236, + "learning_rate": 7.174231332357247e-08, + "loss": 1.4108, + "mean_token_accuracy": 0.6143902540206909, + "num_tokens": 5129474.0, + "step": 197 + }, + { + "epoch": 0.02174390511750494, + "ewc_loss": 1.4028046280145645e-08, + "grad_norm": 5.207295894622803, + "learning_rate": 7.210834553440702e-08, + "loss": 1.377, + "mean_token_accuracy": 0.6232057213783264, + "num_tokens": 5154331.0, + "step": 198 + }, + { + "epoch": 0.021853722820118604, + "ewc_loss": 1.4319084584712982e-08, + "grad_norm": 4.142381191253662, + "learning_rate": 7.247437774524158e-08, + "loss": 1.3711, + "mean_token_accuracy": 0.6304599046707153, + "num_tokens": 5182997.0, + "step": 199 + }, + { + "epoch": 0.021963540522732264, + "ewc_loss": 1.4551915228366852e-08, + "grad_norm": 4.267331600189209, + "learning_rate": 7.284040995607613e-08, + "loss": 1.3081, + "mean_token_accuracy": 0.6622883081436157, + "num_tokens": 5209941.0, + "step": 200 + }, + { + "epoch": 0.022073358225345927, + "ewc_loss": 1.4784745872020721e-08, + "grad_norm": 4.485095024108887, + "learning_rate": 7.320644216691069e-08, + "loss": 1.2482, + "mean_token_accuracy": 0.6559638977050781, + "num_tokens": 5235749.0, + "step": 201 + }, + { + "epoch": 0.022183175927959587, + "ewc_loss": 1.5133991837501526e-08, + "grad_norm": 4.801198959350586, + "learning_rate": 7.357247437774524e-08, + "loss": 1.3229, + "mean_token_accuracy": 0.6420211791992188, + "num_tokens": 5259439.0, + "step": 202 + }, + { + "epoch": 0.022292993630573247, + "ewc_loss": 1.5366822481155396e-08, + "grad_norm": 4.697139739990234, + "learning_rate": 7.393850658857979e-08, + "loss": 1.335, + "mean_token_accuracy": 0.6307386159896851, + "num_tokens": 5283571.0, + "step": 203 + }, + { + "epoch": 0.02240281133318691, + "ewc_loss": 1.5599653124809265e-08, + "grad_norm": 4.6519927978515625, + "learning_rate": 7.430453879941434e-08, + "loss": 1.2662, + "mean_token_accuracy": 0.650070071220398, + "num_tokens": 5306606.0, + "step": 204 + }, + { + "epoch": 0.02251262903580057, + "ewc_loss": 1.57160684466362e-08, + "grad_norm": 4.418269634246826, + "learning_rate": 7.46705710102489e-08, + "loss": 1.3644, + "mean_token_accuracy": 0.6395562291145325, + "num_tokens": 5333475.0, + "step": 205 + }, + { + "epoch": 0.022622446738414233, + "ewc_loss": 1.594889909029007e-08, + "grad_norm": 4.539560794830322, + "learning_rate": 7.503660322108345e-08, + "loss": 1.4125, + "mean_token_accuracy": 0.614435613155365, + "num_tokens": 5360729.0, + "step": 206 + }, + { + "epoch": 0.022732264441027893, + "ewc_loss": 1.618172973394394e-08, + "grad_norm": 4.836935520172119, + "learning_rate": 7.540263543191801e-08, + "loss": 1.3014, + "mean_token_accuracy": 0.6468950510025024, + "num_tokens": 5384083.0, + "step": 207 + }, + { + "epoch": 0.022842082143641557, + "ewc_loss": 1.6298145055770874e-08, + "grad_norm": 4.624597072601318, + "learning_rate": 7.576866764275256e-08, + "loss": 1.3199, + "mean_token_accuracy": 0.634461522102356, + "num_tokens": 5407198.0, + "step": 208 + }, + { + "epoch": 0.022951899846255217, + "ewc_loss": 1.641456037759781e-08, + "grad_norm": 4.677724361419678, + "learning_rate": 7.613469985358711e-08, + "loss": 1.3242, + "mean_token_accuracy": 0.641832172870636, + "num_tokens": 5433236.0, + "step": 209 + }, + { + "epoch": 0.023061717548868876, + "ewc_loss": 1.664739102125168e-08, + "grad_norm": 4.920680046081543, + "learning_rate": 7.650073206442166e-08, + "loss": 1.3849, + "mean_token_accuracy": 0.6277036666870117, + "num_tokens": 5457306.0, + "step": 210 + }, + { + "epoch": 0.02317153525148254, + "ewc_loss": 1.7113052308559418e-08, + "grad_norm": 4.808106899261475, + "learning_rate": 7.686676427525622e-08, + "loss": 1.31, + "mean_token_accuracy": 0.6407407522201538, + "num_tokens": 5477673.0, + "step": 211 + }, + { + "epoch": 0.0232813529540962, + "ewc_loss": 1.7462298274040222e-08, + "grad_norm": 4.009082317352295, + "learning_rate": 7.723279648609077e-08, + "loss": 1.3774, + "mean_token_accuracy": 0.6283350586891174, + "num_tokens": 5506648.0, + "step": 212 + }, + { + "epoch": 0.023391170656709863, + "ewc_loss": 1.7578713595867157e-08, + "grad_norm": 3.7421905994415283, + "learning_rate": 7.759882869692533e-08, + "loss": 1.4429, + "mean_token_accuracy": 0.6055045127868652, + "num_tokens": 5537209.0, + "step": 213 + }, + { + "epoch": 0.023500988359323523, + "ewc_loss": 1.792795956134796e-08, + "grad_norm": 5.200310707092285, + "learning_rate": 7.796486090775987e-08, + "loss": 1.3909, + "mean_token_accuracy": 0.6314777135848999, + "num_tokens": 5559612.0, + "step": 214 + }, + { + "epoch": 0.023610806061937183, + "ewc_loss": 1.816079020500183e-08, + "grad_norm": 4.355843544006348, + "learning_rate": 7.833089311859443e-08, + "loss": 1.3009, + "mean_token_accuracy": 0.648677408695221, + "num_tokens": 5581853.0, + "step": 215 + }, + { + "epoch": 0.023720623764550846, + "ewc_loss": 1.8510036170482635e-08, + "grad_norm": 4.126903057098389, + "learning_rate": 7.869692532942898e-08, + "loss": 1.3133, + "mean_token_accuracy": 0.6422706246376038, + "num_tokens": 5605429.0, + "step": 216 + }, + { + "epoch": 0.023830441467164506, + "ewc_loss": 1.885928213596344e-08, + "grad_norm": 3.71537709236145, + "learning_rate": 7.906295754026354e-08, + "loss": 1.2862, + "mean_token_accuracy": 0.6428636908531189, + "num_tokens": 5635630.0, + "step": 217 + }, + { + "epoch": 0.02394025916977817, + "ewc_loss": 1.9208528101444244e-08, + "grad_norm": 4.515703201293945, + "learning_rate": 7.942898975109809e-08, + "loss": 1.3956, + "mean_token_accuracy": 0.618374228477478, + "num_tokens": 5661702.0, + "step": 218 + }, + { + "epoch": 0.02405007687239183, + "ewc_loss": 1.9441358745098114e-08, + "grad_norm": 4.852375030517578, + "learning_rate": 7.979502196193265e-08, + "loss": 1.2158, + "mean_token_accuracy": 0.6564237475395203, + "num_tokens": 5682274.0, + "step": 219 + }, + { + "epoch": 0.024159894575005492, + "ewc_loss": 1.9674189388751984e-08, + "grad_norm": 4.071182727813721, + "learning_rate": 8.016105417276719e-08, + "loss": 1.3131, + "mean_token_accuracy": 0.6427499055862427, + "num_tokens": 5715819.0, + "step": 220 + }, + { + "epoch": 0.024269712277619152, + "ewc_loss": 2.0372681319713593e-08, + "grad_norm": 4.568675994873047, + "learning_rate": 8.052708638360175e-08, + "loss": 1.3468, + "mean_token_accuracy": 0.6319066286087036, + "num_tokens": 5740534.0, + "step": 221 + }, + { + "epoch": 0.024379529980232812, + "ewc_loss": 2.0838342607021332e-08, + "grad_norm": 5.515529632568359, + "learning_rate": 8.08931185944363e-08, + "loss": 1.392, + "mean_token_accuracy": 0.622442364692688, + "num_tokens": 5762357.0, + "step": 222 + }, + { + "epoch": 0.024489347682846475, + "ewc_loss": 2.10711732506752e-08, + "grad_norm": 4.571751117706299, + "learning_rate": 8.125915080527086e-08, + "loss": 1.2187, + "mean_token_accuracy": 0.6539255380630493, + "num_tokens": 5787211.0, + "step": 223 + }, + { + "epoch": 0.024599165385460135, + "ewc_loss": 2.130400389432907e-08, + "grad_norm": 4.152260780334473, + "learning_rate": 8.162518301610541e-08, + "loss": 1.3456, + "mean_token_accuracy": 0.6393896341323853, + "num_tokens": 5820989.0, + "step": 224 + }, + { + "epoch": 0.0247089830880738, + "ewc_loss": 2.1653249859809875e-08, + "grad_norm": 5.007293701171875, + "learning_rate": 8.199121522693997e-08, + "loss": 1.2988, + "mean_token_accuracy": 0.6351078748703003, + "num_tokens": 5844903.0, + "step": 225 + }, + { + "epoch": 0.02481880079068746, + "ewc_loss": 2.1886080503463745e-08, + "grad_norm": 4.5635857582092285, + "learning_rate": 8.235724743777451e-08, + "loss": 1.2992, + "mean_token_accuracy": 0.6455297470092773, + "num_tokens": 5867365.0, + "step": 226 + }, + { + "epoch": 0.024928618493301122, + "ewc_loss": 2.223532646894455e-08, + "grad_norm": 4.37521505355835, + "learning_rate": 8.272327964860907e-08, + "loss": 1.3751, + "mean_token_accuracy": 0.621649444103241, + "num_tokens": 5895675.0, + "step": 227 + }, + { + "epoch": 0.02503843619591478, + "ewc_loss": 2.2584572434425354e-08, + "grad_norm": 4.642502784729004, + "learning_rate": 8.308931185944362e-08, + "loss": 1.3309, + "mean_token_accuracy": 0.6391657590866089, + "num_tokens": 5920138.0, + "step": 228 + }, + { + "epoch": 0.02514825389852844, + "ewc_loss": 2.293381839990616e-08, + "grad_norm": 4.595395088195801, + "learning_rate": 8.345534407027818e-08, + "loss": 1.3092, + "mean_token_accuracy": 0.6341997385025024, + "num_tokens": 5945763.0, + "step": 229 + }, + { + "epoch": 0.025258071601142105, + "ewc_loss": 2.3515895009040833e-08, + "grad_norm": 4.8616862297058105, + "learning_rate": 8.382137628111273e-08, + "loss": 1.2262, + "mean_token_accuracy": 0.6626526713371277, + "num_tokens": 5971811.0, + "step": 230 + }, + { + "epoch": 0.025367889303755765, + "ewc_loss": 2.3632310330867767e-08, + "grad_norm": 4.6311211585998535, + "learning_rate": 8.41874084919473e-08, + "loss": 1.3606, + "mean_token_accuracy": 0.6320379972457886, + "num_tokens": 5995852.0, + "step": 231 + }, + { + "epoch": 0.025477707006369428, + "ewc_loss": 2.3981556296348572e-08, + "grad_norm": 4.9345927238464355, + "learning_rate": 8.455344070278183e-08, + "loss": 1.2904, + "mean_token_accuracy": 0.6421390175819397, + "num_tokens": 6018357.0, + "step": 232 + }, + { + "epoch": 0.025587524708983088, + "ewc_loss": 2.4097971618175507e-08, + "grad_norm": 4.57362699508667, + "learning_rate": 8.491947291361639e-08, + "loss": 1.3394, + "mean_token_accuracy": 0.6355322599411011, + "num_tokens": 6046503.0, + "step": 233 + }, + { + "epoch": 0.025697342411596748, + "ewc_loss": 2.421438694000244e-08, + "grad_norm": 5.038322448730469, + "learning_rate": 8.528550512445094e-08, + "loss": 1.2714, + "mean_token_accuracy": 0.6455214023590088, + "num_tokens": 6067210.0, + "step": 234 + }, + { + "epoch": 0.02580716011421041, + "ewc_loss": 2.444721758365631e-08, + "grad_norm": 5.326223373413086, + "learning_rate": 8.56515373352855e-08, + "loss": 1.2486, + "mean_token_accuracy": 0.6502362489700317, + "num_tokens": 6091649.0, + "step": 235 + }, + { + "epoch": 0.02591697781682407, + "ewc_loss": 2.491287887096405e-08, + "grad_norm": 5.812960624694824, + "learning_rate": 8.601756954612005e-08, + "loss": 1.3091, + "mean_token_accuracy": 0.6432743072509766, + "num_tokens": 6119052.0, + "step": 236 + }, + { + "epoch": 0.026026795519437734, + "ewc_loss": 2.5029294192790985e-08, + "grad_norm": 6.074481010437012, + "learning_rate": 8.638360175695462e-08, + "loss": 1.2519, + "mean_token_accuracy": 0.6582497954368591, + "num_tokens": 6141688.0, + "step": 237 + }, + { + "epoch": 0.026136613222051394, + "ewc_loss": 2.5262124836444855e-08, + "grad_norm": 5.335132122039795, + "learning_rate": 8.674963396778915e-08, + "loss": 1.3314, + "mean_token_accuracy": 0.6296399235725403, + "num_tokens": 6168591.0, + "step": 238 + }, + { + "epoch": 0.026246430924665057, + "ewc_loss": 2.5494955480098724e-08, + "grad_norm": 6.59600830078125, + "learning_rate": 8.711566617862371e-08, + "loss": 1.2038, + "mean_token_accuracy": 0.6599462032318115, + "num_tokens": 6186942.0, + "step": 239 + }, + { + "epoch": 0.026356248627278717, + "ewc_loss": 2.5727786123752594e-08, + "grad_norm": 5.3535261154174805, + "learning_rate": 8.748169838945826e-08, + "loss": 1.2993, + "mean_token_accuracy": 0.6413037776947021, + "num_tokens": 6214324.0, + "step": 240 + }, + { + "epoch": 0.026466066329892377, + "ewc_loss": 2.60770320892334e-08, + "grad_norm": 4.158238410949707, + "learning_rate": 8.784773060029283e-08, + "loss": 1.3166, + "mean_token_accuracy": 0.6374633312225342, + "num_tokens": 6240342.0, + "step": 241 + }, + { + "epoch": 0.02657588403250604, + "ewc_loss": 2.6426278054714203e-08, + "grad_norm": 4.510664463043213, + "learning_rate": 8.821376281112737e-08, + "loss": 1.3677, + "mean_token_accuracy": 0.626255989074707, + "num_tokens": 6266070.0, + "step": 242 + }, + { + "epoch": 0.0266857017351197, + "ewc_loss": 2.6659108698368073e-08, + "grad_norm": 4.365664958953857, + "learning_rate": 8.857979502196194e-08, + "loss": 1.3198, + "mean_token_accuracy": 0.6369044184684753, + "num_tokens": 6294087.0, + "step": 243 + }, + { + "epoch": 0.026795519437733364, + "ewc_loss": 2.6891939342021942e-08, + "grad_norm": 4.709228515625, + "learning_rate": 8.894582723279647e-08, + "loss": 1.2847, + "mean_token_accuracy": 0.6402798295021057, + "num_tokens": 6318429.0, + "step": 244 + }, + { + "epoch": 0.026905337140347024, + "ewc_loss": 2.7241185307502747e-08, + "grad_norm": 4.59368371963501, + "learning_rate": 8.931185944363104e-08, + "loss": 1.3166, + "mean_token_accuracy": 0.64130699634552, + "num_tokens": 6348101.0, + "step": 245 + }, + { + "epoch": 0.027015154842960687, + "ewc_loss": 2.759043127298355e-08, + "grad_norm": 5.258339881896973, + "learning_rate": 8.967789165446558e-08, + "loss": 1.2915, + "mean_token_accuracy": 0.6465389132499695, + "num_tokens": 6373529.0, + "step": 246 + }, + { + "epoch": 0.027124972545574347, + "ewc_loss": 2.782326191663742e-08, + "grad_norm": 4.361431121826172, + "learning_rate": 9.004392386530015e-08, + "loss": 1.3992, + "mean_token_accuracy": 0.6174030303955078, + "num_tokens": 6400776.0, + "step": 247 + }, + { + "epoch": 0.027234790248188007, + "ewc_loss": 2.828892320394516e-08, + "grad_norm": 5.153146743774414, + "learning_rate": 9.04099560761347e-08, + "loss": 1.2655, + "mean_token_accuracy": 0.6405037641525269, + "num_tokens": 6425702.0, + "step": 248 + }, + { + "epoch": 0.02734460795080167, + "ewc_loss": 2.852175384759903e-08, + "grad_norm": 5.300673961639404, + "learning_rate": 9.077598828696926e-08, + "loss": 1.271, + "mean_token_accuracy": 0.6487585306167603, + "num_tokens": 6447754.0, + "step": 249 + }, + { + "epoch": 0.02745442565341533, + "ewc_loss": 2.87545844912529e-08, + "grad_norm": 4.114651679992676, + "learning_rate": 9.11420204978038e-08, + "loss": 1.2151, + "mean_token_accuracy": 0.6675628423690796, + "num_tokens": 6477535.0, + "step": 250 + }, + { + "epoch": 0.027564243356028993, + "ewc_loss": 2.922024577856064e-08, + "grad_norm": 4.984722137451172, + "learning_rate": 9.150805270863836e-08, + "loss": 1.2571, + "mean_token_accuracy": 0.6532633304595947, + "num_tokens": 6504584.0, + "step": 251 + }, + { + "epoch": 0.027674061058642653, + "ewc_loss": 2.9685907065868378e-08, + "grad_norm": 5.02840518951416, + "learning_rate": 9.18740849194729e-08, + "loss": 1.1506, + "mean_token_accuracy": 0.6793819665908813, + "num_tokens": 6529726.0, + "step": 252 + }, + { + "epoch": 0.027783878761256313, + "ewc_loss": 2.9802322387695312e-08, + "grad_norm": 6.813133716583252, + "learning_rate": 9.224011713030747e-08, + "loss": 1.3492, + "mean_token_accuracy": 0.6326684951782227, + "num_tokens": 6550265.0, + "step": 253 + }, + { + "epoch": 0.027893696463869976, + "ewc_loss": 3.003515303134918e-08, + "grad_norm": 5.570688247680664, + "learning_rate": 9.260614934114202e-08, + "loss": 1.4213, + "mean_token_accuracy": 0.615261435508728, + "num_tokens": 6579733.0, + "step": 254 + }, + { + "epoch": 0.028003514166483636, + "ewc_loss": 3.026798367500305e-08, + "grad_norm": 3.987008571624756, + "learning_rate": 9.297218155197657e-08, + "loss": 1.3092, + "mean_token_accuracy": 0.6413546800613403, + "num_tokens": 6613781.0, + "step": 255 + }, + { + "epoch": 0.0281133318690973, + "ewc_loss": 3.050081431865692e-08, + "grad_norm": 5.671831130981445, + "learning_rate": 9.333821376281111e-08, + "loss": 1.2218, + "mean_token_accuracy": 0.6488465070724487, + "num_tokens": 6635982.0, + "step": 256 + }, + { + "epoch": 0.02822314957171096, + "ewc_loss": 3.050081431865692e-08, + "grad_norm": 7.1676459312438965, + "learning_rate": 9.370424597364568e-08, + "loss": 1.2867, + "mean_token_accuracy": 0.6369583606719971, + "num_tokens": 6660180.0, + "step": 257 + }, + { + "epoch": 0.028332967274324623, + "ewc_loss": 3.073364496231079e-08, + "grad_norm": 4.499671459197998, + "learning_rate": 9.407027818448023e-08, + "loss": 1.2862, + "mean_token_accuracy": 0.6395137906074524, + "num_tokens": 6687798.0, + "step": 258 + }, + { + "epoch": 0.028442784976938282, + "ewc_loss": 3.096647560596466e-08, + "grad_norm": 4.739867210388184, + "learning_rate": 9.443631039531479e-08, + "loss": 1.3111, + "mean_token_accuracy": 0.641301155090332, + "num_tokens": 6716005.0, + "step": 259 + }, + { + "epoch": 0.028552602679551942, + "ewc_loss": 3.096647560596466e-08, + "grad_norm": 4.198145866394043, + "learning_rate": 9.480234260614934e-08, + "loss": 1.2904, + "mean_token_accuracy": 0.6436617970466614, + "num_tokens": 6744026.0, + "step": 260 + }, + { + "epoch": 0.028662420382165606, + "ewc_loss": 3.119930624961853e-08, + "grad_norm": 7.4228363037109375, + "learning_rate": 9.516837481698389e-08, + "loss": 1.2598, + "mean_token_accuracy": 0.6451791524887085, + "num_tokens": 6765799.0, + "step": 261 + }, + { + "epoch": 0.028772238084779266, + "ewc_loss": 3.14321368932724e-08, + "grad_norm": 6.26281213760376, + "learning_rate": 9.553440702781844e-08, + "loss": 1.3254, + "mean_token_accuracy": 0.6293820738792419, + "num_tokens": 6792636.0, + "step": 262 + }, + { + "epoch": 0.02888205578739293, + "ewc_loss": 3.166496753692627e-08, + "grad_norm": 7.42271089553833, + "learning_rate": 9.5900439238653e-08, + "loss": 1.1266, + "mean_token_accuracy": 0.6819988489151001, + "num_tokens": 6814122.0, + "step": 263 + }, + { + "epoch": 0.02899187349000659, + "ewc_loss": 3.166496753692627e-08, + "grad_norm": 5.52346658706665, + "learning_rate": 9.626647144948755e-08, + "loss": 1.3381, + "mean_token_accuracy": 0.628153920173645, + "num_tokens": 6844162.0, + "step": 264 + }, + { + "epoch": 0.029101691192620252, + "ewc_loss": 3.189779818058014e-08, + "grad_norm": 6.7815656661987305, + "learning_rate": 9.663250366032211e-08, + "loss": 1.1338, + "mean_token_accuracy": 0.6776254177093506, + "num_tokens": 6867856.0, + "step": 265 + }, + { + "epoch": 0.029211508895233912, + "ewc_loss": 3.189779818058014e-08, + "grad_norm": 5.413896083831787, + "learning_rate": 9.699853587115666e-08, + "loss": 1.2346, + "mean_token_accuracy": 0.6619607210159302, + "num_tokens": 6894300.0, + "step": 266 + }, + { + "epoch": 0.029321326597847572, + "ewc_loss": 3.213062882423401e-08, + "grad_norm": 6.473917007446289, + "learning_rate": 9.736456808199121e-08, + "loss": 1.25, + "mean_token_accuracy": 0.6513502597808838, + "num_tokens": 6919880.0, + "step": 267 + }, + { + "epoch": 0.029431144300461235, + "ewc_loss": 3.213062882423401e-08, + "grad_norm": 5.566222667694092, + "learning_rate": 9.773060029282576e-08, + "loss": 1.2932, + "mean_token_accuracy": 0.6449161767959595, + "num_tokens": 6948610.0, + "step": 268 + }, + { + "epoch": 0.029540962003074895, + "ewc_loss": 3.259629011154175e-08, + "grad_norm": 5.366939067840576, + "learning_rate": 9.809663250366032e-08, + "loss": 1.272, + "mean_token_accuracy": 0.6432253122329712, + "num_tokens": 6974058.0, + "step": 269 + }, + { + "epoch": 0.02965077970568856, + "ewc_loss": 3.259629011154175e-08, + "grad_norm": 5.962132930755615, + "learning_rate": 9.846266471449487e-08, + "loss": 1.3245, + "mean_token_accuracy": 0.6343725323677063, + "num_tokens": 7001527.0, + "step": 270 + }, + { + "epoch": 0.029760597408302218, + "ewc_loss": 3.282912075519562e-08, + "grad_norm": 6.8030853271484375, + "learning_rate": 9.882869692532943e-08, + "loss": 1.3006, + "mean_token_accuracy": 0.6332356929779053, + "num_tokens": 7025908.0, + "step": 271 + }, + { + "epoch": 0.029870415110915878, + "ewc_loss": 3.282912075519562e-08, + "grad_norm": 6.724353313446045, + "learning_rate": 9.919472913616398e-08, + "loss": 1.198, + "mean_token_accuracy": 0.6580362319946289, + "num_tokens": 7048417.0, + "step": 272 + }, + { + "epoch": 0.02998023281352954, + "ewc_loss": 3.282912075519562e-08, + "grad_norm": 6.997157096862793, + "learning_rate": 9.956076134699853e-08, + "loss": 1.234, + "mean_token_accuracy": 0.6525381207466125, + "num_tokens": 7072239.0, + "step": 273 + }, + { + "epoch": 0.0300900505161432, + "ewc_loss": 3.3527612686157227e-08, + "grad_norm": 5.49432897567749, + "learning_rate": 9.992679355783308e-08, + "loss": 1.297, + "mean_token_accuracy": 0.6391619443893433, + "num_tokens": 7097960.0, + "step": 274 + }, + { + "epoch": 0.030199868218756865, + "ewc_loss": 3.3993273973464966e-08, + "grad_norm": 6.304009437561035, + "learning_rate": 1.0029282576866764e-07, + "loss": 1.2771, + "mean_token_accuracy": 0.6487125158309937, + "num_tokens": 7129347.0, + "step": 275 + }, + { + "epoch": 0.030309685921370524, + "ewc_loss": 3.4226104617118835e-08, + "grad_norm": 5.56227970123291, + "learning_rate": 1.0065885797950219e-07, + "loss": 1.2684, + "mean_token_accuracy": 0.6436766982078552, + "num_tokens": 7155811.0, + "step": 276 + }, + { + "epoch": 0.030419503623984188, + "ewc_loss": 3.4691765904426575e-08, + "grad_norm": 7.472891807556152, + "learning_rate": 1.0102489019033675e-07, + "loss": 1.1605, + "mean_token_accuracy": 0.6723228693008423, + "num_tokens": 7175864.0, + "step": 277 + }, + { + "epoch": 0.030529321326597848, + "ewc_loss": 3.4924596548080444e-08, + "grad_norm": 7.778210639953613, + "learning_rate": 1.013909224011713e-07, + "loss": 1.191, + "mean_token_accuracy": 0.6594414710998535, + "num_tokens": 7197905.0, + "step": 278 + }, + { + "epoch": 0.030639139029211507, + "ewc_loss": 3.4924596548080444e-08, + "grad_norm": 7.692864418029785, + "learning_rate": 1.0175695461200585e-07, + "loss": 1.2765, + "mean_token_accuracy": 0.6465479135513306, + "num_tokens": 7223710.0, + "step": 279 + }, + { + "epoch": 0.03074895673182517, + "ewc_loss": 3.5157427191734314e-08, + "grad_norm": 7.374377250671387, + "learning_rate": 1.021229868228404e-07, + "loss": 1.2301, + "mean_token_accuracy": 0.6531208753585815, + "num_tokens": 7245870.0, + "step": 280 + }, + { + "epoch": 0.03085877443443883, + "ewc_loss": 3.5390257835388184e-08, + "grad_norm": 6.10888671875, + "learning_rate": 1.0248901903367496e-07, + "loss": 1.2681, + "mean_token_accuracy": 0.6460639238357544, + "num_tokens": 7271138.0, + "step": 281 + }, + { + "epoch": 0.030968592137052494, + "ewc_loss": 3.5390257835388184e-08, + "grad_norm": 5.876791477203369, + "learning_rate": 1.0285505124450951e-07, + "loss": 1.285, + "mean_token_accuracy": 0.6487439870834351, + "num_tokens": 7297052.0, + "step": 282 + }, + { + "epoch": 0.031078409839666154, + "ewc_loss": 3.585591912269592e-08, + "grad_norm": 5.893066883087158, + "learning_rate": 1.0322108345534407e-07, + "loss": 1.3525, + "mean_token_accuracy": 0.6289811134338379, + "num_tokens": 7326733.0, + "step": 283 + }, + { + "epoch": 0.031188227542279817, + "ewc_loss": 3.632158041000366e-08, + "grad_norm": 4.650724411010742, + "learning_rate": 1.0358711566617862e-07, + "loss": 1.2991, + "mean_token_accuracy": 0.6365339159965515, + "num_tokens": 7358626.0, + "step": 284 + }, + { + "epoch": 0.03129804524489348, + "ewc_loss": 3.655441105365753e-08, + "grad_norm": 5.431400299072266, + "learning_rate": 1.0395314787701317e-07, + "loss": 1.2993, + "mean_token_accuracy": 0.6384732127189636, + "num_tokens": 7385058.0, + "step": 285 + }, + { + "epoch": 0.03140786294750714, + "ewc_loss": 3.702007234096527e-08, + "grad_norm": 6.091305732727051, + "learning_rate": 1.0431918008784772e-07, + "loss": 1.2556, + "mean_token_accuracy": 0.6518702507019043, + "num_tokens": 7414433.0, + "step": 286 + }, + { + "epoch": 0.0315176806501208, + "ewc_loss": 3.748573362827301e-08, + "grad_norm": 5.441969394683838, + "learning_rate": 1.0468521229868228e-07, + "loss": 1.3706, + "mean_token_accuracy": 0.6210126876831055, + "num_tokens": 7446196.0, + "step": 287 + }, + { + "epoch": 0.031627498352734464, + "ewc_loss": 3.795139491558075e-08, + "grad_norm": 7.839461803436279, + "learning_rate": 1.0505124450951683e-07, + "loss": 1.2301, + "mean_token_accuracy": 0.6476937532424927, + "num_tokens": 7469003.0, + "step": 288 + }, + { + "epoch": 0.03173731605534812, + "ewc_loss": 3.818422555923462e-08, + "grad_norm": 8.495381355285645, + "learning_rate": 1.0541727672035139e-07, + "loss": 1.2943, + "mean_token_accuracy": 0.6409914493560791, + "num_tokens": 7494794.0, + "step": 289 + }, + { + "epoch": 0.03184713375796178, + "ewc_loss": 3.91155481338501e-08, + "grad_norm": 6.85034704208374, + "learning_rate": 1.0578330893118593e-07, + "loss": 1.2916, + "mean_token_accuracy": 0.6379244327545166, + "num_tokens": 7531151.0, + "step": 290 + }, + { + "epoch": 0.03195695146057544, + "ewc_loss": 3.9814040064811707e-08, + "grad_norm": 6.36924409866333, + "learning_rate": 1.0614934114202049e-07, + "loss": 1.1983, + "mean_token_accuracy": 0.6599891185760498, + "num_tokens": 7556362.0, + "step": 291 + }, + { + "epoch": 0.0320667691631891, + "ewc_loss": 4.0046870708465576e-08, + "grad_norm": 5.554937839508057, + "learning_rate": 1.0651537335285504e-07, + "loss": 1.2865, + "mean_token_accuracy": 0.6388962864875793, + "num_tokens": 7580209.0, + "step": 292 + }, + { + "epoch": 0.03217658686580277, + "ewc_loss": 4.0046870708465576e-08, + "grad_norm": 5.733303070068359, + "learning_rate": 1.068814055636896e-07, + "loss": 1.3436, + "mean_token_accuracy": 0.6281580328941345, + "num_tokens": 7607823.0, + "step": 293 + }, + { + "epoch": 0.03228640456841643, + "ewc_loss": 4.0279701352119446e-08, + "grad_norm": 6.595129489898682, + "learning_rate": 1.0724743777452415e-07, + "loss": 1.3542, + "mean_token_accuracy": 0.628096342086792, + "num_tokens": 7630083.0, + "step": 294 + }, + { + "epoch": 0.03239622227103009, + "ewc_loss": 4.0279701352119446e-08, + "grad_norm": 5.162665367126465, + "learning_rate": 1.0761346998535871e-07, + "loss": 1.3429, + "mean_token_accuracy": 0.6323556900024414, + "num_tokens": 7657339.0, + "step": 295 + }, + { + "epoch": 0.03250603997364375, + "ewc_loss": 4.0745362639427185e-08, + "grad_norm": 6.058937072753906, + "learning_rate": 1.0797950219619325e-07, + "loss": 1.1319, + "mean_token_accuracy": 0.6686737537384033, + "num_tokens": 7678789.0, + "step": 296 + }, + { + "epoch": 0.03261585767625741, + "ewc_loss": 4.1443854570388794e-08, + "grad_norm": 5.799112796783447, + "learning_rate": 1.0834553440702781e-07, + "loss": 1.2134, + "mean_token_accuracy": 0.6620169878005981, + "num_tokens": 7705531.0, + "step": 297 + }, + { + "epoch": 0.032725675378871076, + "ewc_loss": 4.1443854570388794e-08, + "grad_norm": 6.346941947937012, + "learning_rate": 1.0871156661786236e-07, + "loss": 1.2365, + "mean_token_accuracy": 0.6511984467506409, + "num_tokens": 7726907.0, + "step": 298 + }, + { + "epoch": 0.032835493081484736, + "ewc_loss": 4.1676685214042664e-08, + "grad_norm": 4.972687244415283, + "learning_rate": 1.0907759882869692e-07, + "loss": 1.403, + "mean_token_accuracy": 0.6283669471740723, + "num_tokens": 7754917.0, + "step": 299 + }, + { + "epoch": 0.032945310784098396, + "ewc_loss": 4.21423465013504e-08, + "grad_norm": 6.642063140869141, + "learning_rate": 1.0944363103953147e-07, + "loss": 1.2572, + "mean_token_accuracy": 0.6434324383735657, + "num_tokens": 7780922.0, + "step": 300 + }, + { + "epoch": 0.033055128486712056, + "ewc_loss": 4.237517714500427e-08, + "grad_norm": 6.216006278991699, + "learning_rate": 1.0980966325036604e-07, + "loss": 1.2773, + "mean_token_accuracy": 0.6452693939208984, + "num_tokens": 7807004.0, + "step": 301 + }, + { + "epoch": 0.03316494618932572, + "ewc_loss": 4.260800778865814e-08, + "grad_norm": 7.619591236114502, + "learning_rate": 1.1017569546120057e-07, + "loss": 1.157, + "mean_token_accuracy": 0.6653314828872681, + "num_tokens": 7830480.0, + "step": 302 + }, + { + "epoch": 0.03327476389193938, + "ewc_loss": 4.330649971961975e-08, + "grad_norm": 6.6467790603637695, + "learning_rate": 1.1054172767203513e-07, + "loss": 1.2622, + "mean_token_accuracy": 0.643237829208374, + "num_tokens": 7856537.0, + "step": 303 + }, + { + "epoch": 0.03338458159455304, + "ewc_loss": 4.353933036327362e-08, + "grad_norm": 9.4442720413208, + "learning_rate": 1.1090775988286968e-07, + "loss": 1.1903, + "mean_token_accuracy": 0.6615409851074219, + "num_tokens": 7875428.0, + "step": 304 + }, + { + "epoch": 0.0334943992971667, + "ewc_loss": 4.353933036327362e-08, + "grad_norm": 7.164065361022949, + "learning_rate": 1.1127379209370424e-07, + "loss": 1.2212, + "mean_token_accuracy": 0.6511616706848145, + "num_tokens": 7899765.0, + "step": 305 + }, + { + "epoch": 0.03360421699978036, + "ewc_loss": 4.377216100692749e-08, + "grad_norm": 7.930950164794922, + "learning_rate": 1.116398243045388e-07, + "loss": 1.2542, + "mean_token_accuracy": 0.6470091342926025, + "num_tokens": 7922002.0, + "step": 306 + }, + { + "epoch": 0.03371403470239403, + "ewc_loss": 4.44706529378891e-08, + "grad_norm": 6.953316688537598, + "learning_rate": 1.1200585651537336e-07, + "loss": 1.2616, + "mean_token_accuracy": 0.6514612436294556, + "num_tokens": 7945165.0, + "step": 307 + }, + { + "epoch": 0.03382385240500769, + "ewc_loss": 4.44706529378891e-08, + "grad_norm": 6.4906158447265625, + "learning_rate": 1.1237188872620789e-07, + "loss": 1.1835, + "mean_token_accuracy": 0.67215496301651, + "num_tokens": 7967060.0, + "step": 308 + }, + { + "epoch": 0.03393367010762135, + "ewc_loss": 4.493631422519684e-08, + "grad_norm": 5.875910758972168, + "learning_rate": 1.1273792093704245e-07, + "loss": 1.2029, + "mean_token_accuracy": 0.6545925140380859, + "num_tokens": 7988205.0, + "step": 309 + }, + { + "epoch": 0.03404348781023501, + "ewc_loss": 4.493631422519684e-08, + "grad_norm": 5.894249439239502, + "learning_rate": 1.13103953147877e-07, + "loss": 1.1458, + "mean_token_accuracy": 0.6642098426818848, + "num_tokens": 8018516.0, + "step": 310 + }, + { + "epoch": 0.03415330551284867, + "ewc_loss": 4.516914486885071e-08, + "grad_norm": 4.83705472946167, + "learning_rate": 1.1346998535871157e-07, + "loss": 1.2418, + "mean_token_accuracy": 0.6554678678512573, + "num_tokens": 8049713.0, + "step": 311 + }, + { + "epoch": 0.034263123215462335, + "ewc_loss": 4.563480615615845e-08, + "grad_norm": 5.635374069213867, + "learning_rate": 1.1383601756954611e-07, + "loss": 1.2415, + "mean_token_accuracy": 0.6547979116439819, + "num_tokens": 8077271.0, + "step": 312 + }, + { + "epoch": 0.034372940918075995, + "ewc_loss": 4.563480615615845e-08, + "grad_norm": 5.5607733726501465, + "learning_rate": 1.1420204978038068e-07, + "loss": 1.2492, + "mean_token_accuracy": 0.6559541821479797, + "num_tokens": 8100334.0, + "step": 313 + }, + { + "epoch": 0.034482758620689655, + "ewc_loss": 4.563480615615845e-08, + "grad_norm": 6.562287330627441, + "learning_rate": 1.1456808199121521e-07, + "loss": 1.2693, + "mean_token_accuracy": 0.6625165343284607, + "num_tokens": 8127291.0, + "step": 314 + }, + { + "epoch": 0.034592576323303315, + "ewc_loss": 4.6333298087120056e-08, + "grad_norm": 5.174922466278076, + "learning_rate": 1.1493411420204978e-07, + "loss": 1.3021, + "mean_token_accuracy": 0.6378816366195679, + "num_tokens": 8158273.0, + "step": 315 + }, + { + "epoch": 0.034702394025916974, + "ewc_loss": 4.6798959374427795e-08, + "grad_norm": 5.882766246795654, + "learning_rate": 1.1530014641288432e-07, + "loss": 1.1683, + "mean_token_accuracy": 0.6641651391983032, + "num_tokens": 8180056.0, + "step": 316 + }, + { + "epoch": 0.03481221172853064, + "ewc_loss": 4.7031790018081665e-08, + "grad_norm": 7.586732864379883, + "learning_rate": 1.1566617862371889e-07, + "loss": 1.2437, + "mean_token_accuracy": 0.6544003486633301, + "num_tokens": 8203267.0, + "step": 317 + }, + { + "epoch": 0.0349220294311443, + "ewc_loss": 4.7264620661735535e-08, + "grad_norm": 6.00674295425415, + "learning_rate": 1.1603221083455344e-07, + "loss": 1.2338, + "mean_token_accuracy": 0.6574459075927734, + "num_tokens": 8225551.0, + "step": 318 + }, + { + "epoch": 0.03503184713375796, + "ewc_loss": 4.7730281949043274e-08, + "grad_norm": 6.508787631988525, + "learning_rate": 1.16398243045388e-07, + "loss": 1.285, + "mean_token_accuracy": 0.6336483359336853, + "num_tokens": 8249591.0, + "step": 319 + }, + { + "epoch": 0.03514166483637162, + "ewc_loss": 4.7963112592697144e-08, + "grad_norm": 7.147418975830078, + "learning_rate": 1.1676427525622253e-07, + "loss": 1.2538, + "mean_token_accuracy": 0.6428220272064209, + "num_tokens": 8274565.0, + "step": 320 + }, + { + "epoch": 0.03525148253898529, + "ewc_loss": 4.819594323635101e-08, + "grad_norm": 6.206871032714844, + "learning_rate": 1.171303074670571e-07, + "loss": 1.2723, + "mean_token_accuracy": 0.6433445811271667, + "num_tokens": 8301284.0, + "step": 321 + }, + { + "epoch": 0.03536130024159895, + "ewc_loss": 4.842877388000488e-08, + "grad_norm": 7.273705005645752, + "learning_rate": 1.1749633967789165e-07, + "loss": 1.2446, + "mean_token_accuracy": 0.6448398232460022, + "num_tokens": 8322113.0, + "step": 322 + }, + { + "epoch": 0.03547111794421261, + "ewc_loss": 4.866160452365875e-08, + "grad_norm": 4.910754203796387, + "learning_rate": 1.1786237188872621e-07, + "loss": 1.2794, + "mean_token_accuracy": 0.6364221572875977, + "num_tokens": 8349315.0, + "step": 323 + }, + { + "epoch": 0.03558093564682627, + "ewc_loss": 4.866160452365875e-08, + "grad_norm": 8.046136856079102, + "learning_rate": 1.1822840409956076e-07, + "loss": 1.2402, + "mean_token_accuracy": 0.6482524871826172, + "num_tokens": 8369953.0, + "step": 324 + }, + { + "epoch": 0.03569075334943993, + "ewc_loss": 4.889443516731262e-08, + "grad_norm": 5.345475673675537, + "learning_rate": 1.1859443631039532e-07, + "loss": 1.2707, + "mean_token_accuracy": 0.6444668769836426, + "num_tokens": 8394310.0, + "step": 325 + }, + { + "epoch": 0.035800571052053594, + "ewc_loss": 4.936009645462036e-08, + "grad_norm": 6.730803966522217, + "learning_rate": 1.1896046852122986e-07, + "loss": 1.3217, + "mean_token_accuracy": 0.631319522857666, + "num_tokens": 8419318.0, + "step": 326 + }, + { + "epoch": 0.035910388754667254, + "ewc_loss": 4.936009645462036e-08, + "grad_norm": 6.91164493560791, + "learning_rate": 1.1932650073206443e-07, + "loss": 1.1013, + "mean_token_accuracy": 0.679543137550354, + "num_tokens": 8443229.0, + "step": 327 + }, + { + "epoch": 0.036020206457280914, + "ewc_loss": 4.98257577419281e-08, + "grad_norm": 6.792092323303223, + "learning_rate": 1.1969253294289895e-07, + "loss": 1.253, + "mean_token_accuracy": 0.650921642780304, + "num_tokens": 8467929.0, + "step": 328 + }, + { + "epoch": 0.03613002415989457, + "ewc_loss": 5.005858838558197e-08, + "grad_norm": 7.587632179260254, + "learning_rate": 1.2005856515373353e-07, + "loss": 1.2163, + "mean_token_accuracy": 0.6557713150978088, + "num_tokens": 8491470.0, + "step": 329 + }, + { + "epoch": 0.03623984186250823, + "ewc_loss": 5.005858838558197e-08, + "grad_norm": 6.557726860046387, + "learning_rate": 1.2042459736456808e-07, + "loss": 1.1845, + "mean_token_accuracy": 0.6643165946006775, + "num_tokens": 8519715.0, + "step": 330 + }, + { + "epoch": 0.0363496595651219, + "ewc_loss": 5.029141902923584e-08, + "grad_norm": 7.290064334869385, + "learning_rate": 1.2079062957540263e-07, + "loss": 1.2435, + "mean_token_accuracy": 0.6469221115112305, + "num_tokens": 8542376.0, + "step": 331 + }, + { + "epoch": 0.03645947726773556, + "ewc_loss": 5.075708031654358e-08, + "grad_norm": 6.248574733734131, + "learning_rate": 1.2115666178623718e-07, + "loss": 1.2543, + "mean_token_accuracy": 0.6526477336883545, + "num_tokens": 8564513.0, + "step": 332 + }, + { + "epoch": 0.03656929497034922, + "ewc_loss": 5.075708031654358e-08, + "grad_norm": 6.269374370574951, + "learning_rate": 1.2152269399707175e-07, + "loss": 1.1752, + "mean_token_accuracy": 0.6686953902244568, + "num_tokens": 8587957.0, + "step": 333 + }, + { + "epoch": 0.03667911267296288, + "ewc_loss": 5.145557224750519e-08, + "grad_norm": 5.895638942718506, + "learning_rate": 1.2188872620790627e-07, + "loss": 1.1946, + "mean_token_accuracy": 0.6658316850662231, + "num_tokens": 8612155.0, + "step": 334 + }, + { + "epoch": 0.03678893037557654, + "ewc_loss": 5.145557224750519e-08, + "grad_norm": 5.928967475891113, + "learning_rate": 1.2225475841874085e-07, + "loss": 1.2166, + "mean_token_accuracy": 0.6613396406173706, + "num_tokens": 8643677.0, + "step": 335 + }, + { + "epoch": 0.036898748078190206, + "ewc_loss": 5.168840289115906e-08, + "grad_norm": 5.423673629760742, + "learning_rate": 1.226207906295754e-07, + "loss": 1.1721, + "mean_token_accuracy": 0.663331151008606, + "num_tokens": 8672599.0, + "step": 336 + }, + { + "epoch": 0.037008565780803866, + "ewc_loss": 5.192123353481293e-08, + "grad_norm": 6.375072479248047, + "learning_rate": 1.2298682284040995e-07, + "loss": 1.2643, + "mean_token_accuracy": 0.6554862856864929, + "num_tokens": 8696109.0, + "step": 337 + }, + { + "epoch": 0.037118383483417526, + "ewc_loss": 5.21540641784668e-08, + "grad_norm": 4.99992036819458, + "learning_rate": 1.233528550512445e-07, + "loss": 1.1701, + "mean_token_accuracy": 0.6709259152412415, + "num_tokens": 8725617.0, + "step": 338 + }, + { + "epoch": 0.037228201186031186, + "ewc_loss": 5.2386894822120667e-08, + "grad_norm": 4.974957466125488, + "learning_rate": 1.2371888726207907e-07, + "loss": 1.1802, + "mean_token_accuracy": 0.6661860942840576, + "num_tokens": 8753416.0, + "step": 339 + }, + { + "epoch": 0.03733801888864485, + "ewc_loss": 5.2852556109428406e-08, + "grad_norm": 5.523962497711182, + "learning_rate": 1.240849194729136e-07, + "loss": 1.2447, + "mean_token_accuracy": 0.6437151432037354, + "num_tokens": 8781610.0, + "step": 340 + }, + { + "epoch": 0.03744783659125851, + "ewc_loss": 5.2852556109428406e-08, + "grad_norm": 4.8851847648620605, + "learning_rate": 1.2445095168374817e-07, + "loss": 1.1779, + "mean_token_accuracy": 0.6563483476638794, + "num_tokens": 8805386.0, + "step": 341 + }, + { + "epoch": 0.03755765429387217, + "ewc_loss": 5.3085386753082275e-08, + "grad_norm": 6.662985801696777, + "learning_rate": 1.2481698389458272e-07, + "loss": 1.3174, + "mean_token_accuracy": 0.6366109848022461, + "num_tokens": 8830009.0, + "step": 342 + }, + { + "epoch": 0.03766747199648583, + "ewc_loss": 5.3783878684043884e-08, + "grad_norm": 7.117298126220703, + "learning_rate": 1.2518301610541727e-07, + "loss": 1.1867, + "mean_token_accuracy": 0.6590917110443115, + "num_tokens": 8848499.0, + "step": 343 + }, + { + "epoch": 0.03777728969909949, + "ewc_loss": 5.4016709327697754e-08, + "grad_norm": 6.164840221405029, + "learning_rate": 1.2554904831625182e-07, + "loss": 1.1608, + "mean_token_accuracy": 0.6646026372909546, + "num_tokens": 8869974.0, + "step": 344 + }, + { + "epoch": 0.03788710740171316, + "ewc_loss": 5.4249539971351624e-08, + "grad_norm": 5.607041835784912, + "learning_rate": 1.259150805270864e-07, + "loss": 1.2853, + "mean_token_accuracy": 0.6382218599319458, + "num_tokens": 8897952.0, + "step": 345 + }, + { + "epoch": 0.03799692510432682, + "ewc_loss": 5.448237061500549e-08, + "grad_norm": 4.741281509399414, + "learning_rate": 1.2628111273792094e-07, + "loss": 1.209, + "mean_token_accuracy": 0.6606768369674683, + "num_tokens": 8931394.0, + "step": 346 + }, + { + "epoch": 0.03810674280694048, + "ewc_loss": 5.471520125865936e-08, + "grad_norm": 4.885409832000732, + "learning_rate": 1.2664714494875547e-07, + "loss": 1.2596, + "mean_token_accuracy": 0.6483864784240723, + "num_tokens": 8962484.0, + "step": 347 + }, + { + "epoch": 0.03821656050955414, + "ewc_loss": 5.494803190231323e-08, + "grad_norm": 4.8609161376953125, + "learning_rate": 1.2701317715959004e-07, + "loss": 1.2246, + "mean_token_accuracy": 0.6507129669189453, + "num_tokens": 8988810.0, + "step": 348 + }, + { + "epoch": 0.0383263782121678, + "ewc_loss": 5.51808625459671e-08, + "grad_norm": 6.889989852905273, + "learning_rate": 1.273792093704246e-07, + "loss": 1.1704, + "mean_token_accuracy": 0.6685264110565186, + "num_tokens": 9011086.0, + "step": 349 + }, + { + "epoch": 0.038436195914781465, + "ewc_loss": 5.541369318962097e-08, + "grad_norm": 4.8165788650512695, + "learning_rate": 1.2774524158125917e-07, + "loss": 1.2457, + "mean_token_accuracy": 0.6506102085113525, + "num_tokens": 9038964.0, + "step": 350 + }, + { + "epoch": 0.038546013617395125, + "ewc_loss": 5.564652383327484e-08, + "grad_norm": 5.559482574462891, + "learning_rate": 1.281112737920937e-07, + "loss": 1.3156, + "mean_token_accuracy": 0.6299352049827576, + "num_tokens": 9065347.0, + "step": 351 + }, + { + "epoch": 0.038655831320008785, + "ewc_loss": 5.587935447692871e-08, + "grad_norm": 7.261521339416504, + "learning_rate": 1.2847730600292824e-07, + "loss": 1.146, + "mean_token_accuracy": 0.675446093082428, + "num_tokens": 9089414.0, + "step": 352 + }, + { + "epoch": 0.038765649022622445, + "ewc_loss": 5.587935447692871e-08, + "grad_norm": 4.236743927001953, + "learning_rate": 1.288433382137628e-07, + "loss": 1.3101, + "mean_token_accuracy": 0.6368640065193176, + "num_tokens": 9120551.0, + "step": 353 + }, + { + "epoch": 0.038875466725236105, + "ewc_loss": 5.634501576423645e-08, + "grad_norm": 5.813968181610107, + "learning_rate": 1.2920937042459736e-07, + "loss": 1.2122, + "mean_token_accuracy": 0.6574801802635193, + "num_tokens": 9143936.0, + "step": 354 + }, + { + "epoch": 0.03898528442784977, + "ewc_loss": 5.657784640789032e-08, + "grad_norm": 4.883533954620361, + "learning_rate": 1.295754026354319e-07, + "loss": 1.2954, + "mean_token_accuracy": 0.6343373656272888, + "num_tokens": 9171765.0, + "step": 355 + }, + { + "epoch": 0.03909510213046343, + "ewc_loss": 5.657784640789032e-08, + "grad_norm": 5.059711933135986, + "learning_rate": 1.2994143484626646e-07, + "loss": 1.2502, + "mean_token_accuracy": 0.6464051008224487, + "num_tokens": 9198275.0, + "step": 356 + }, + { + "epoch": 0.03920491983307709, + "ewc_loss": 5.704350769519806e-08, + "grad_norm": 5.992596626281738, + "learning_rate": 1.3030746705710104e-07, + "loss": 1.201, + "mean_token_accuracy": 0.6514371633529663, + "num_tokens": 9221218.0, + "step": 357 + }, + { + "epoch": 0.03931473753569075, + "ewc_loss": 5.727633833885193e-08, + "grad_norm": 4.489619255065918, + "learning_rate": 1.3067349926793558e-07, + "loss": 1.2012, + "mean_token_accuracy": 0.6503357887268066, + "num_tokens": 9249309.0, + "step": 358 + }, + { + "epoch": 0.03942455523830442, + "ewc_loss": 5.774199962615967e-08, + "grad_norm": 7.155152797698975, + "learning_rate": 1.310395314787701e-07, + "loss": 1.2318, + "mean_token_accuracy": 0.6490497589111328, + "num_tokens": 9270701.0, + "step": 359 + }, + { + "epoch": 0.03953437294091808, + "ewc_loss": 5.797483026981354e-08, + "grad_norm": 5.320860862731934, + "learning_rate": 1.3140556368960468e-07, + "loss": 1.2643, + "mean_token_accuracy": 0.6473100185394287, + "num_tokens": 9303156.0, + "step": 360 + }, + { + "epoch": 0.03964419064353174, + "ewc_loss": 5.820766091346741e-08, + "grad_norm": 7.14857816696167, + "learning_rate": 1.3177159590043923e-07, + "loss": 1.1607, + "mean_token_accuracy": 0.6719732880592346, + "num_tokens": 9325163.0, + "step": 361 + }, + { + "epoch": 0.0397540083461454, + "ewc_loss": 5.844049155712128e-08, + "grad_norm": 7.828958511352539, + "learning_rate": 1.321376281112738e-07, + "loss": 1.1329, + "mean_token_accuracy": 0.6760522127151489, + "num_tokens": 9349474.0, + "step": 362 + }, + { + "epoch": 0.03986382604875906, + "ewc_loss": 5.8906152844429016e-08, + "grad_norm": 5.273746967315674, + "learning_rate": 1.3250366032210833e-07, + "loss": 1.1975, + "mean_token_accuracy": 0.6584579944610596, + "num_tokens": 9376227.0, + "step": 363 + }, + { + "epoch": 0.039973643751372724, + "ewc_loss": 5.9371814131736755e-08, + "grad_norm": 5.86379337310791, + "learning_rate": 1.3286969253294288e-07, + "loss": 1.1475, + "mean_token_accuracy": 0.6683629155158997, + "num_tokens": 9398802.0, + "step": 364 + }, + { + "epoch": 0.040083461453986384, + "ewc_loss": 5.960464477539063e-08, + "grad_norm": 4.8319783210754395, + "learning_rate": 1.3323572474377745e-07, + "loss": 1.2298, + "mean_token_accuracy": 0.6522361636161804, + "num_tokens": 9425588.0, + "step": 365 + }, + { + "epoch": 0.040193279156600044, + "ewc_loss": 6.007030606269836e-08, + "grad_norm": 4.151319980621338, + "learning_rate": 1.33601756954612e-07, + "loss": 1.2809, + "mean_token_accuracy": 0.6346626281738281, + "num_tokens": 9457907.0, + "step": 366 + }, + { + "epoch": 0.040303096859213704, + "ewc_loss": 6.007030606269836e-08, + "grad_norm": 6.459424018859863, + "learning_rate": 1.3396778916544655e-07, + "loss": 1.1311, + "mean_token_accuracy": 0.683387279510498, + "num_tokens": 9482988.0, + "step": 367 + }, + { + "epoch": 0.040412914561827364, + "ewc_loss": 6.05359673500061e-08, + "grad_norm": 7.595620632171631, + "learning_rate": 1.343338213762811e-07, + "loss": 1.2487, + "mean_token_accuracy": 0.6457054615020752, + "num_tokens": 9503553.0, + "step": 368 + }, + { + "epoch": 0.04052273226444103, + "ewc_loss": 6.05359673500061e-08, + "grad_norm": 5.38472843170166, + "learning_rate": 1.3469985358711565e-07, + "loss": 1.2475, + "mean_token_accuracy": 0.6496913433074951, + "num_tokens": 9528274.0, + "step": 369 + }, + { + "epoch": 0.04063254996705469, + "ewc_loss": 6.05359673500061e-08, + "grad_norm": 4.8229804039001465, + "learning_rate": 1.3506588579795023e-07, + "loss": 1.2081, + "mean_token_accuracy": 0.653639554977417, + "num_tokens": 9563887.0, + "step": 370 + }, + { + "epoch": 0.04074236766966835, + "ewc_loss": 6.05359673500061e-08, + "grad_norm": 6.823788166046143, + "learning_rate": 1.3543191800878475e-07, + "loss": 1.259, + "mean_token_accuracy": 0.6474335789680481, + "num_tokens": 9586556.0, + "step": 371 + }, + { + "epoch": 0.04085218537228201, + "ewc_loss": 6.100162863731384e-08, + "grad_norm": 4.7906365394592285, + "learning_rate": 1.3579795021961932e-07, + "loss": 1.2082, + "mean_token_accuracy": 0.6571387052536011, + "num_tokens": 9616303.0, + "step": 372 + }, + { + "epoch": 0.04096200307489567, + "ewc_loss": 6.100162863731384e-08, + "grad_norm": 4.657477378845215, + "learning_rate": 1.3616398243045387e-07, + "loss": 1.2342, + "mean_token_accuracy": 0.652094841003418, + "num_tokens": 9645693.0, + "step": 373 + }, + { + "epoch": 0.04107182077750934, + "ewc_loss": 6.146728992462158e-08, + "grad_norm": 6.376185417175293, + "learning_rate": 1.3653001464128845e-07, + "loss": 1.2279, + "mean_token_accuracy": 0.6565430760383606, + "num_tokens": 9672371.0, + "step": 374 + }, + { + "epoch": 0.041181638480122996, + "ewc_loss": 6.146728992462158e-08, + "grad_norm": 5.899044513702393, + "learning_rate": 1.3689604685212297e-07, + "loss": 1.1898, + "mean_token_accuracy": 0.6723721027374268, + "num_tokens": 9705264.0, + "step": 375 + }, + { + "epoch": 0.041291456182736656, + "ewc_loss": 6.146728992462158e-08, + "grad_norm": 6.214390277862549, + "learning_rate": 1.3726207906295752e-07, + "loss": 1.132, + "mean_token_accuracy": 0.6756516695022583, + "num_tokens": 9729726.0, + "step": 376 + }, + { + "epoch": 0.041401273885350316, + "ewc_loss": 6.193295121192932e-08, + "grad_norm": 6.263103008270264, + "learning_rate": 1.376281112737921e-07, + "loss": 1.2553, + "mean_token_accuracy": 0.6540525555610657, + "num_tokens": 9756079.0, + "step": 377 + }, + { + "epoch": 0.04151109158796398, + "ewc_loss": 6.193295121192932e-08, + "grad_norm": 5.731513977050781, + "learning_rate": 1.3799414348462665e-07, + "loss": 1.1137, + "mean_token_accuracy": 0.6809456944465637, + "num_tokens": 9778317.0, + "step": 378 + }, + { + "epoch": 0.04162090929057764, + "ewc_loss": 6.239861249923706e-08, + "grad_norm": 5.082241058349609, + "learning_rate": 1.383601756954612e-07, + "loss": 1.2446, + "mean_token_accuracy": 0.6486742496490479, + "num_tokens": 9808114.0, + "step": 379 + }, + { + "epoch": 0.0417307269931913, + "ewc_loss": 6.239861249923706e-08, + "grad_norm": 4.479623794555664, + "learning_rate": 1.3872620790629574e-07, + "loss": 1.2238, + "mean_token_accuracy": 0.6439341306686401, + "num_tokens": 9833909.0, + "step": 380 + }, + { + "epoch": 0.04184054469580496, + "ewc_loss": 6.239861249923706e-08, + "grad_norm": 6.128339767456055, + "learning_rate": 1.390922401171303e-07, + "loss": 1.1871, + "mean_token_accuracy": 0.6647377014160156, + "num_tokens": 9857091.0, + "step": 381 + }, + { + "epoch": 0.04195036239841862, + "ewc_loss": 6.28642737865448e-08, + "grad_norm": 5.9445109367370605, + "learning_rate": 1.3945827232796487e-07, + "loss": 1.1533, + "mean_token_accuracy": 0.6661388874053955, + "num_tokens": 9880395.0, + "step": 382 + }, + { + "epoch": 0.04206018010103229, + "ewc_loss": 6.28642737865448e-08, + "grad_norm": 6.132054328918457, + "learning_rate": 1.398243045387994e-07, + "loss": 1.2755, + "mean_token_accuracy": 0.6410099864006042, + "num_tokens": 9904483.0, + "step": 383 + }, + { + "epoch": 0.04216999780364595, + "ewc_loss": 6.332993507385254e-08, + "grad_norm": 4.792935848236084, + "learning_rate": 1.4019033674963397e-07, + "loss": 1.1565, + "mean_token_accuracy": 0.6655392646789551, + "num_tokens": 9927865.0, + "step": 384 + }, + { + "epoch": 0.04227981550625961, + "ewc_loss": 6.332993507385254e-08, + "grad_norm": 5.613373279571533, + "learning_rate": 1.4055636896046852e-07, + "loss": 1.2378, + "mean_token_accuracy": 0.6561001539230347, + "num_tokens": 9956712.0, + "step": 385 + }, + { + "epoch": 0.04238963320887327, + "ewc_loss": 6.379559636116028e-08, + "grad_norm": 5.384181022644043, + "learning_rate": 1.409224011713031e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6899953484535217, + "num_tokens": 9979222.0, + "step": 386 + }, + { + "epoch": 0.04249945091148693, + "ewc_loss": 6.379559636116028e-08, + "grad_norm": 6.042055130004883, + "learning_rate": 1.4128843338213761e-07, + "loss": 1.2208, + "mean_token_accuracy": 0.6484212875366211, + "num_tokens": 10007509.0, + "step": 387 + }, + { + "epoch": 0.042609268614100596, + "ewc_loss": 6.426125764846802e-08, + "grad_norm": 5.003174781799316, + "learning_rate": 1.4165446559297216e-07, + "loss": 1.257, + "mean_token_accuracy": 0.649585485458374, + "num_tokens": 10037020.0, + "step": 388 + }, + { + "epoch": 0.042719086316714255, + "ewc_loss": 6.426125764846802e-08, + "grad_norm": 4.876445770263672, + "learning_rate": 1.4202049780380674e-07, + "loss": 1.2786, + "mean_token_accuracy": 0.6392292380332947, + "num_tokens": 10062076.0, + "step": 389 + }, + { + "epoch": 0.042828904019327915, + "ewc_loss": 6.426125764846802e-08, + "grad_norm": 4.3956122398376465, + "learning_rate": 1.423865300146413e-07, + "loss": 1.1932, + "mean_token_accuracy": 0.6599547863006592, + "num_tokens": 10093034.0, + "step": 390 + }, + { + "epoch": 0.042938721721941575, + "ewc_loss": 6.426125764846802e-08, + "grad_norm": 5.930423259735107, + "learning_rate": 1.4275256222547584e-07, + "loss": 1.2246, + "mean_token_accuracy": 0.6580770015716553, + "num_tokens": 10117118.0, + "step": 391 + }, + { + "epoch": 0.043048539424555235, + "ewc_loss": 6.472691893577576e-08, + "grad_norm": 4.484592437744141, + "learning_rate": 1.4311859443631039e-07, + "loss": 1.2663, + "mean_token_accuracy": 0.6436982154846191, + "num_tokens": 10151018.0, + "step": 392 + }, + { + "epoch": 0.0431583571271689, + "ewc_loss": 6.472691893577576e-08, + "grad_norm": 7.285355567932129, + "learning_rate": 1.4348462664714493e-07, + "loss": 1.11, + "mean_token_accuracy": 0.677497386932373, + "num_tokens": 10168012.0, + "step": 393 + }, + { + "epoch": 0.04326817482978256, + "ewc_loss": 6.472691893577576e-08, + "grad_norm": 6.279913902282715, + "learning_rate": 1.438506588579795e-07, + "loss": 1.2409, + "mean_token_accuracy": 0.6464107036590576, + "num_tokens": 10193267.0, + "step": 394 + }, + { + "epoch": 0.04337799253239622, + "ewc_loss": 6.51925802230835e-08, + "grad_norm": 5.21528434753418, + "learning_rate": 1.4421669106881403e-07, + "loss": 1.1478, + "mean_token_accuracy": 0.6632053256034851, + "num_tokens": 10219769.0, + "step": 395 + }, + { + "epoch": 0.04348781023500988, + "ewc_loss": 6.565824151039124e-08, + "grad_norm": 7.196839809417725, + "learning_rate": 1.445827232796486e-07, + "loss": 1.2539, + "mean_token_accuracy": 0.6446946859359741, + "num_tokens": 10244899.0, + "step": 396 + }, + { + "epoch": 0.04359762793762355, + "ewc_loss": 6.565824151039124e-08, + "grad_norm": 7.186933517456055, + "learning_rate": 1.4494875549048316e-07, + "loss": 1.1931, + "mean_token_accuracy": 0.6598750352859497, + "num_tokens": 10263877.0, + "step": 397 + }, + { + "epoch": 0.04370744564023721, + "ewc_loss": 6.565824151039124e-08, + "grad_norm": 5.068794250488281, + "learning_rate": 1.4531478770131773e-07, + "loss": 1.2517, + "mean_token_accuracy": 0.6458762288093567, + "num_tokens": 10287450.0, + "step": 398 + }, + { + "epoch": 0.04381726334285087, + "ewc_loss": 6.612390279769897e-08, + "grad_norm": 5.172651767730713, + "learning_rate": 1.4568081991215226e-07, + "loss": 1.2073, + "mean_token_accuracy": 0.6654785871505737, + "num_tokens": 10314939.0, + "step": 399 + }, + { + "epoch": 0.04392708104546453, + "ewc_loss": 6.612390279769897e-08, + "grad_norm": 4.655849933624268, + "learning_rate": 1.460468521229868e-07, + "loss": 1.0969, + "mean_token_accuracy": 0.6854739785194397, + "num_tokens": 10340075.0, + "step": 400 + }, + { + "epoch": 0.04403689874807819, + "ewc_loss": 6.612390279769897e-08, + "grad_norm": 4.927425861358643, + "learning_rate": 1.4641288433382138e-07, + "loss": 1.2501, + "mean_token_accuracy": 0.6518680453300476, + "num_tokens": 10364116.0, + "step": 401 + }, + { + "epoch": 0.044146716450691854, + "ewc_loss": 6.612390279769897e-08, + "grad_norm": 4.459461212158203, + "learning_rate": 1.4677891654465593e-07, + "loss": 1.262, + "mean_token_accuracy": 0.6483991146087646, + "num_tokens": 10391839.0, + "step": 402 + }, + { + "epoch": 0.044256534153305514, + "ewc_loss": 6.658956408500671e-08, + "grad_norm": 4.290105819702148, + "learning_rate": 1.4714494875549048e-07, + "loss": 1.197, + "mean_token_accuracy": 0.6733931303024292, + "num_tokens": 10419954.0, + "step": 403 + }, + { + "epoch": 0.044366351855919174, + "ewc_loss": 6.705522537231445e-08, + "grad_norm": 6.14592170715332, + "learning_rate": 1.4751098096632503e-07, + "loss": 1.2589, + "mean_token_accuracy": 0.6438363790512085, + "num_tokens": 10444143.0, + "step": 404 + }, + { + "epoch": 0.044476169558532834, + "ewc_loss": 6.705522537231445e-08, + "grad_norm": 5.170601844787598, + "learning_rate": 1.4787701317715958e-07, + "loss": 1.1915, + "mean_token_accuracy": 0.6598103642463684, + "num_tokens": 10473697.0, + "step": 405 + }, + { + "epoch": 0.044585987261146494, + "ewc_loss": 6.705522537231445e-08, + "grad_norm": 3.959761142730713, + "learning_rate": 1.4824304538799415e-07, + "loss": 1.1851, + "mean_token_accuracy": 0.663716197013855, + "num_tokens": 10502612.0, + "step": 406 + }, + { + "epoch": 0.04469580496376016, + "ewc_loss": 6.752088665962219e-08, + "grad_norm": 4.944035530090332, + "learning_rate": 1.4860907759882867e-07, + "loss": 1.2287, + "mean_token_accuracy": 0.6500813961029053, + "num_tokens": 10529266.0, + "step": 407 + }, + { + "epoch": 0.04480562266637382, + "ewc_loss": 6.845220923423767e-08, + "grad_norm": 5.3889265060424805, + "learning_rate": 1.4897510980966325e-07, + "loss": 1.2089, + "mean_token_accuracy": 0.6557345390319824, + "num_tokens": 10550616.0, + "step": 408 + }, + { + "epoch": 0.04491544036898748, + "ewc_loss": 6.845220923423767e-08, + "grad_norm": 5.045780181884766, + "learning_rate": 1.493411420204978e-07, + "loss": 1.1727, + "mean_token_accuracy": 0.6648773550987244, + "num_tokens": 10582682.0, + "step": 409 + }, + { + "epoch": 0.04502525807160114, + "ewc_loss": 6.938353180885315e-08, + "grad_norm": 5.345633029937744, + "learning_rate": 1.4970717423133235e-07, + "loss": 1.1019, + "mean_token_accuracy": 0.6807976365089417, + "num_tokens": 10606229.0, + "step": 410 + }, + { + "epoch": 0.0451350757742148, + "ewc_loss": 6.938353180885315e-08, + "grad_norm": 3.908350944519043, + "learning_rate": 1.500732064421669e-07, + "loss": 1.2495, + "mean_token_accuracy": 0.6444616317749023, + "num_tokens": 10639889.0, + "step": 411 + }, + { + "epoch": 0.04524489347682847, + "ewc_loss": 7.031485438346863e-08, + "grad_norm": 4.848712921142578, + "learning_rate": 1.5043923865300145e-07, + "loss": 1.256, + "mean_token_accuracy": 0.639743447303772, + "num_tokens": 10664352.0, + "step": 412 + }, + { + "epoch": 0.04535471117944213, + "ewc_loss": 7.171183824539185e-08, + "grad_norm": 4.157025337219238, + "learning_rate": 1.5080527086383602e-07, + "loss": 1.1646, + "mean_token_accuracy": 0.6700019240379333, + "num_tokens": 10692255.0, + "step": 413 + }, + { + "epoch": 0.04546452888205579, + "ewc_loss": 7.217749953269958e-08, + "grad_norm": 6.460756778717041, + "learning_rate": 1.5117130307467057e-07, + "loss": 1.1896, + "mean_token_accuracy": 0.6657021045684814, + "num_tokens": 10713360.0, + "step": 414 + }, + { + "epoch": 0.045574346584669446, + "ewc_loss": 7.217749953269958e-08, + "grad_norm": 5.036333084106445, + "learning_rate": 1.5153733528550512e-07, + "loss": 1.176, + "mean_token_accuracy": 0.6660938858985901, + "num_tokens": 10743290.0, + "step": 415 + }, + { + "epoch": 0.04568416428728311, + "ewc_loss": 7.264316082000732e-08, + "grad_norm": 5.544648170471191, + "learning_rate": 1.5190336749633967e-07, + "loss": 1.1925, + "mean_token_accuracy": 0.6538996696472168, + "num_tokens": 10771567.0, + "step": 416 + }, + { + "epoch": 0.04579398198989677, + "ewc_loss": 7.264316082000732e-08, + "grad_norm": 4.923274040222168, + "learning_rate": 1.5226939970717422e-07, + "loss": 1.1555, + "mean_token_accuracy": 0.6803246736526489, + "num_tokens": 10796435.0, + "step": 417 + }, + { + "epoch": 0.04590379969251043, + "ewc_loss": 7.264316082000732e-08, + "grad_norm": 4.4296698570251465, + "learning_rate": 1.526354319180088e-07, + "loss": 1.2184, + "mean_token_accuracy": 0.6628931164741516, + "num_tokens": 10826350.0, + "step": 418 + }, + { + "epoch": 0.04601361739512409, + "ewc_loss": 7.310882210731506e-08, + "grad_norm": 4.928363800048828, + "learning_rate": 1.5300146412884332e-07, + "loss": 1.1552, + "mean_token_accuracy": 0.6690644025802612, + "num_tokens": 10851407.0, + "step": 419 + }, + { + "epoch": 0.04612343509773775, + "ewc_loss": 7.35744833946228e-08, + "grad_norm": 4.471861362457275, + "learning_rate": 1.533674963396779e-07, + "loss": 1.3084, + "mean_token_accuracy": 0.6455811262130737, + "num_tokens": 10885508.0, + "step": 420 + }, + { + "epoch": 0.04623325280035142, + "ewc_loss": 7.35744833946228e-08, + "grad_norm": 4.809022903442383, + "learning_rate": 1.5373352855051244e-07, + "loss": 1.1899, + "mean_token_accuracy": 0.6554079055786133, + "num_tokens": 10911158.0, + "step": 421 + }, + { + "epoch": 0.04634307050296508, + "ewc_loss": 7.35744833946228e-08, + "grad_norm": 5.310109615325928, + "learning_rate": 1.54099560761347e-07, + "loss": 1.1248, + "mean_token_accuracy": 0.6767531633377075, + "num_tokens": 10932147.0, + "step": 422 + }, + { + "epoch": 0.04645288820557874, + "ewc_loss": 7.404014468193054e-08, + "grad_norm": 5.324670791625977, + "learning_rate": 1.5446559297218154e-07, + "loss": 1.1785, + "mean_token_accuracy": 0.663377046585083, + "num_tokens": 10952566.0, + "step": 423 + }, + { + "epoch": 0.0465627059081924, + "ewc_loss": 7.450580596923828e-08, + "grad_norm": 6.745604038238525, + "learning_rate": 1.548316251830161e-07, + "loss": 1.1479, + "mean_token_accuracy": 0.6692085266113281, + "num_tokens": 10971181.0, + "step": 424 + }, + { + "epoch": 0.04667252361080606, + "ewc_loss": 7.497146725654602e-08, + "grad_norm": 5.325643062591553, + "learning_rate": 1.5519765739385066e-07, + "loss": 1.1101, + "mean_token_accuracy": 0.6794785261154175, + "num_tokens": 10991562.0, + "step": 425 + }, + { + "epoch": 0.046782341313419726, + "ewc_loss": 7.543712854385376e-08, + "grad_norm": 5.058926582336426, + "learning_rate": 1.5556368960468521e-07, + "loss": 1.1843, + "mean_token_accuracy": 0.6601271629333496, + "num_tokens": 11016836.0, + "step": 426 + }, + { + "epoch": 0.046892159016033386, + "ewc_loss": 7.543712854385376e-08, + "grad_norm": 4.643797874450684, + "learning_rate": 1.5592972181551974e-07, + "loss": 1.2274, + "mean_token_accuracy": 0.6500517129898071, + "num_tokens": 11040431.0, + "step": 427 + }, + { + "epoch": 0.047001976718647046, + "ewc_loss": 7.59027898311615e-08, + "grad_norm": 4.853285312652588, + "learning_rate": 1.562957540263543e-07, + "loss": 1.1742, + "mean_token_accuracy": 0.6655807495117188, + "num_tokens": 11065612.0, + "step": 428 + }, + { + "epoch": 0.047111794421260705, + "ewc_loss": 7.636845111846924e-08, + "grad_norm": 5.942261219024658, + "learning_rate": 1.5666178623718886e-07, + "loss": 1.1227, + "mean_token_accuracy": 0.6813658475875854, + "num_tokens": 11090027.0, + "step": 429 + }, + { + "epoch": 0.047221612123874365, + "ewc_loss": 7.729977369308472e-08, + "grad_norm": 4.2342610359191895, + "learning_rate": 1.5702781844802344e-07, + "loss": 1.2877, + "mean_token_accuracy": 0.6349428296089172, + "num_tokens": 11121097.0, + "step": 430 + }, + { + "epoch": 0.04733142982648803, + "ewc_loss": 7.776543498039246e-08, + "grad_norm": 5.098260402679443, + "learning_rate": 1.5739385065885796e-07, + "loss": 1.2031, + "mean_token_accuracy": 0.6535669565200806, + "num_tokens": 11144282.0, + "step": 431 + }, + { + "epoch": 0.04744124752910169, + "ewc_loss": 7.869675755500793e-08, + "grad_norm": 4.233004570007324, + "learning_rate": 1.5775988286969253e-07, + "loss": 1.2055, + "mean_token_accuracy": 0.6607368588447571, + "num_tokens": 11172508.0, + "step": 432 + }, + { + "epoch": 0.04755106523171535, + "ewc_loss": 7.869675755500793e-08, + "grad_norm": 5.708499431610107, + "learning_rate": 1.5812591508052708e-07, + "loss": 1.1567, + "mean_token_accuracy": 0.6715973615646362, + "num_tokens": 11194936.0, + "step": 433 + }, + { + "epoch": 0.04766088293432901, + "ewc_loss": 7.916241884231567e-08, + "grad_norm": 4.4759063720703125, + "learning_rate": 1.5849194729136163e-07, + "loss": 1.0588, + "mean_token_accuracy": 0.6913940906524658, + "num_tokens": 11220873.0, + "step": 434 + }, + { + "epoch": 0.04777070063694268, + "ewc_loss": 7.962808012962341e-08, + "grad_norm": 4.809643268585205, + "learning_rate": 1.5885797950219618e-07, + "loss": 1.2008, + "mean_token_accuracy": 0.6555190086364746, + "num_tokens": 11245691.0, + "step": 435 + }, + { + "epoch": 0.04788051833955634, + "ewc_loss": 7.962808012962341e-08, + "grad_norm": 4.857936859130859, + "learning_rate": 1.5922401171303073e-07, + "loss": 1.248, + "mean_token_accuracy": 0.6639552116394043, + "num_tokens": 11272386.0, + "step": 436 + }, + { + "epoch": 0.04799033604217, + "ewc_loss": 7.962808012962341e-08, + "grad_norm": 4.001628398895264, + "learning_rate": 1.595900439238653e-07, + "loss": 1.253, + "mean_token_accuracy": 0.6491787433624268, + "num_tokens": 11301211.0, + "step": 437 + }, + { + "epoch": 0.04810015374478366, + "ewc_loss": 7.962808012962341e-08, + "grad_norm": 4.003810882568359, + "learning_rate": 1.5995607613469986e-07, + "loss": 1.2448, + "mean_token_accuracy": 0.6526179313659668, + "num_tokens": 11331253.0, + "step": 438 + }, + { + "epoch": 0.04820997144739732, + "ewc_loss": 7.962808012962341e-08, + "grad_norm": 3.9453837871551514, + "learning_rate": 1.6032210834553438e-07, + "loss": 1.1682, + "mean_token_accuracy": 0.6749414801597595, + "num_tokens": 11360900.0, + "step": 439 + }, + { + "epoch": 0.048319789150010985, + "ewc_loss": 8.009374141693115e-08, + "grad_norm": 4.378568649291992, + "learning_rate": 1.6068814055636895e-07, + "loss": 1.281, + "mean_token_accuracy": 0.6354601979255676, + "num_tokens": 11391599.0, + "step": 440 + }, + { + "epoch": 0.048429606852624645, + "ewc_loss": 8.009374141693115e-08, + "grad_norm": 3.9591493606567383, + "learning_rate": 1.610541727672035e-07, + "loss": 1.206, + "mean_token_accuracy": 0.6588048934936523, + "num_tokens": 11424019.0, + "step": 441 + }, + { + "epoch": 0.048539424555238304, + "ewc_loss": 8.055940270423889e-08, + "grad_norm": 4.693767070770264, + "learning_rate": 1.6142020497803808e-07, + "loss": 1.1777, + "mean_token_accuracy": 0.656975507736206, + "num_tokens": 11446707.0, + "step": 442 + }, + { + "epoch": 0.048649242257851964, + "ewc_loss": 8.149072527885437e-08, + "grad_norm": 5.030144214630127, + "learning_rate": 1.617862371888726e-07, + "loss": 1.1347, + "mean_token_accuracy": 0.6714138984680176, + "num_tokens": 11467310.0, + "step": 443 + }, + { + "epoch": 0.048759059960465624, + "ewc_loss": 8.242204785346985e-08, + "grad_norm": 4.46711540222168, + "learning_rate": 1.6215226939970718e-07, + "loss": 1.1876, + "mean_token_accuracy": 0.659125030040741, + "num_tokens": 11493973.0, + "step": 444 + }, + { + "epoch": 0.04886887766307929, + "ewc_loss": 8.242204785346985e-08, + "grad_norm": 5.255026817321777, + "learning_rate": 1.6251830161054173e-07, + "loss": 1.1267, + "mean_token_accuracy": 0.6838050484657288, + "num_tokens": 11514611.0, + "step": 445 + }, + { + "epoch": 0.04897869536569295, + "ewc_loss": 8.288770914077759e-08, + "grad_norm": 5.449615955352783, + "learning_rate": 1.6288433382137627e-07, + "loss": 1.271, + "mean_token_accuracy": 0.6463444828987122, + "num_tokens": 11538158.0, + "step": 446 + }, + { + "epoch": 0.04908851306830661, + "ewc_loss": 8.335337042808533e-08, + "grad_norm": 4.531970977783203, + "learning_rate": 1.6325036603221082e-07, + "loss": 1.1869, + "mean_token_accuracy": 0.6631996035575867, + "num_tokens": 11570295.0, + "step": 447 + }, + { + "epoch": 0.04919833077092027, + "ewc_loss": 8.381903171539307e-08, + "grad_norm": 4.409245491027832, + "learning_rate": 1.6361639824304537e-07, + "loss": 1.1567, + "mean_token_accuracy": 0.6718332171440125, + "num_tokens": 11597780.0, + "step": 448 + }, + { + "epoch": 0.04930814847353393, + "ewc_loss": 8.381903171539307e-08, + "grad_norm": 4.003989219665527, + "learning_rate": 1.6398243045387995e-07, + "loss": 1.2127, + "mean_token_accuracy": 0.6606179475784302, + "num_tokens": 11623199.0, + "step": 449 + }, + { + "epoch": 0.0494179661761476, + "ewc_loss": 8.42846930027008e-08, + "grad_norm": 4.185800075531006, + "learning_rate": 1.643484626647145e-07, + "loss": 1.2482, + "mean_token_accuracy": 0.6408044695854187, + "num_tokens": 11651657.0, + "step": 450 + }, + { + "epoch": 0.04952778387876126, + "ewc_loss": 8.42846930027008e-08, + "grad_norm": 4.038951396942139, + "learning_rate": 1.6471449487554902e-07, + "loss": 1.1868, + "mean_token_accuracy": 0.6585921049118042, + "num_tokens": 11680381.0, + "step": 451 + }, + { + "epoch": 0.04963760158137492, + "ewc_loss": 8.475035429000854e-08, + "grad_norm": 4.395724296569824, + "learning_rate": 1.650805270863836e-07, + "loss": 1.1448, + "mean_token_accuracy": 0.6710017919540405, + "num_tokens": 11705013.0, + "step": 452 + }, + { + "epoch": 0.04974741928398858, + "ewc_loss": 8.475035429000854e-08, + "grad_norm": 3.974034070968628, + "learning_rate": 1.6544655929721814e-07, + "loss": 1.0866, + "mean_token_accuracy": 0.6853500008583069, + "num_tokens": 11731140.0, + "step": 453 + }, + { + "epoch": 0.049857236986602244, + "ewc_loss": 8.475035429000854e-08, + "grad_norm": 3.8814785480499268, + "learning_rate": 1.6581259150805272e-07, + "loss": 1.2944, + "mean_token_accuracy": 0.636090099811554, + "num_tokens": 11761261.0, + "step": 454 + }, + { + "epoch": 0.0499670546892159, + "ewc_loss": 8.475035429000854e-08, + "grad_norm": 4.034365177154541, + "learning_rate": 1.6617862371888724e-07, + "loss": 1.1658, + "mean_token_accuracy": 0.6738075613975525, + "num_tokens": 11787241.0, + "step": 455 + }, + { + "epoch": 0.05007687239182956, + "ewc_loss": 8.521601557731628e-08, + "grad_norm": 4.963100910186768, + "learning_rate": 1.6654465592972182e-07, + "loss": 1.2001, + "mean_token_accuracy": 0.6594251394271851, + "num_tokens": 11808564.0, + "step": 456 + }, + { + "epoch": 0.05018669009444322, + "ewc_loss": 8.568167686462402e-08, + "grad_norm": 3.9192469120025635, + "learning_rate": 1.6691068814055637e-07, + "loss": 1.1707, + "mean_token_accuracy": 0.6680961847305298, + "num_tokens": 11835170.0, + "step": 457 + }, + { + "epoch": 0.05029650779705688, + "ewc_loss": 8.568167686462402e-08, + "grad_norm": 3.7621068954467773, + "learning_rate": 1.6727672035139092e-07, + "loss": 1.1726, + "mean_token_accuracy": 0.6673965454101562, + "num_tokens": 11862838.0, + "step": 458 + }, + { + "epoch": 0.05040632549967055, + "ewc_loss": 8.614733815193176e-08, + "grad_norm": 3.9450318813323975, + "learning_rate": 1.6764275256222547e-07, + "loss": 1.303, + "mean_token_accuracy": 0.640945553779602, + "num_tokens": 11892865.0, + "step": 459 + }, + { + "epoch": 0.05051614320228421, + "ewc_loss": 8.614733815193176e-08, + "grad_norm": 5.023293495178223, + "learning_rate": 1.6800878477306001e-07, + "loss": 1.1196, + "mean_token_accuracy": 0.6805577278137207, + "num_tokens": 11913385.0, + "step": 460 + }, + { + "epoch": 0.05062596090489787, + "ewc_loss": 8.66129994392395e-08, + "grad_norm": 4.182340621948242, + "learning_rate": 1.683748169838946e-07, + "loss": 1.2154, + "mean_token_accuracy": 0.6545276045799255, + "num_tokens": 11942433.0, + "step": 461 + }, + { + "epoch": 0.05073577860751153, + "ewc_loss": 8.66129994392395e-08, + "grad_norm": 4.453800201416016, + "learning_rate": 1.6874084919472914e-07, + "loss": 1.1874, + "mean_token_accuracy": 0.6575649976730347, + "num_tokens": 11968602.0, + "step": 462 + }, + { + "epoch": 0.05084559631012519, + "ewc_loss": 8.707866072654724e-08, + "grad_norm": 4.854336738586426, + "learning_rate": 1.6910688140556366e-07, + "loss": 1.1756, + "mean_token_accuracy": 0.6668816208839417, + "num_tokens": 11989652.0, + "step": 463 + }, + { + "epoch": 0.050955414012738856, + "ewc_loss": 8.754432201385498e-08, + "grad_norm": 3.946658134460449, + "learning_rate": 1.6947291361639824e-07, + "loss": 1.2063, + "mean_token_accuracy": 0.6493210792541504, + "num_tokens": 12014428.0, + "step": 464 + }, + { + "epoch": 0.051065231715352516, + "ewc_loss": 8.754432201385498e-08, + "grad_norm": 4.391969203948975, + "learning_rate": 1.6983894582723279e-07, + "loss": 1.1062, + "mean_token_accuracy": 0.6784700751304626, + "num_tokens": 12041211.0, + "step": 465 + }, + { + "epoch": 0.051175049417966176, + "ewc_loss": 8.800998330116272e-08, + "grad_norm": 3.759336471557617, + "learning_rate": 1.7020497803806736e-07, + "loss": 1.2361, + "mean_token_accuracy": 0.6604752540588379, + "num_tokens": 12071813.0, + "step": 466 + }, + { + "epoch": 0.051284867120579836, + "ewc_loss": 8.800998330116272e-08, + "grad_norm": 4.394277095794678, + "learning_rate": 1.7057101024890188e-07, + "loss": 1.2371, + "mean_token_accuracy": 0.6498492360115051, + "num_tokens": 12099495.0, + "step": 467 + }, + { + "epoch": 0.051394684823193496, + "ewc_loss": 8.800998330116272e-08, + "grad_norm": 4.384762287139893, + "learning_rate": 1.7093704245973643e-07, + "loss": 1.2356, + "mean_token_accuracy": 0.6446406841278076, + "num_tokens": 12125207.0, + "step": 468 + }, + { + "epoch": 0.05150450252580716, + "ewc_loss": 8.800998330116272e-08, + "grad_norm": 4.594635009765625, + "learning_rate": 1.71303074670571e-07, + "loss": 1.1912, + "mean_token_accuracy": 0.6522799730300903, + "num_tokens": 12146398.0, + "step": 469 + }, + { + "epoch": 0.05161432022842082, + "ewc_loss": 8.847564458847046e-08, + "grad_norm": 3.8296473026275635, + "learning_rate": 1.7166910688140556e-07, + "loss": 1.0916, + "mean_token_accuracy": 0.6794878244400024, + "num_tokens": 12170215.0, + "step": 470 + }, + { + "epoch": 0.05172413793103448, + "ewc_loss": 8.940696716308594e-08, + "grad_norm": 4.811753749847412, + "learning_rate": 1.720351390922401e-07, + "loss": 1.243, + "mean_token_accuracy": 0.6449528336524963, + "num_tokens": 12192610.0, + "step": 471 + }, + { + "epoch": 0.05183395563364814, + "ewc_loss": 8.940696716308594e-08, + "grad_norm": 4.336169719696045, + "learning_rate": 1.7240117130307466e-07, + "loss": 1.1871, + "mean_token_accuracy": 0.6652578711509705, + "num_tokens": 12220138.0, + "step": 472 + }, + { + "epoch": 0.05194377333626181, + "ewc_loss": 8.940696716308594e-08, + "grad_norm": 3.994051933288574, + "learning_rate": 1.7276720351390923e-07, + "loss": 1.195, + "mean_token_accuracy": 0.6606504917144775, + "num_tokens": 12251326.0, + "step": 473 + }, + { + "epoch": 0.05205359103887547, + "ewc_loss": 8.940696716308594e-08, + "grad_norm": 3.61814284324646, + "learning_rate": 1.7313323572474378e-07, + "loss": 1.2527, + "mean_token_accuracy": 0.6479777097702026, + "num_tokens": 12280837.0, + "step": 474 + }, + { + "epoch": 0.05216340874148913, + "ewc_loss": 9.033828973770142e-08, + "grad_norm": 3.7805819511413574, + "learning_rate": 1.734992679355783e-07, + "loss": 1.0743, + "mean_token_accuracy": 0.6895675659179688, + "num_tokens": 12304158.0, + "step": 475 + }, + { + "epoch": 0.05227322644410279, + "ewc_loss": 9.080395102500916e-08, + "grad_norm": 3.6713786125183105, + "learning_rate": 1.7386530014641288e-07, + "loss": 1.2496, + "mean_token_accuracy": 0.6441269516944885, + "num_tokens": 12331193.0, + "step": 476 + }, + { + "epoch": 0.05238304414671645, + "ewc_loss": 9.173527359962463e-08, + "grad_norm": 3.798166275024414, + "learning_rate": 1.7423133235724743e-07, + "loss": 1.2302, + "mean_token_accuracy": 0.6545511484146118, + "num_tokens": 12357251.0, + "step": 477 + }, + { + "epoch": 0.052492861849330115, + "ewc_loss": 9.173527359962463e-08, + "grad_norm": 3.9415454864501953, + "learning_rate": 1.74597364568082e-07, + "loss": 1.2456, + "mean_token_accuracy": 0.6461343765258789, + "num_tokens": 12384863.0, + "step": 478 + }, + { + "epoch": 0.052602679551943775, + "ewc_loss": 9.266659617424011e-08, + "grad_norm": 3.696479320526123, + "learning_rate": 1.7496339677891653e-07, + "loss": 1.1051, + "mean_token_accuracy": 0.6810677647590637, + "num_tokens": 12415065.0, + "step": 479 + }, + { + "epoch": 0.052712497254557435, + "ewc_loss": 9.266659617424011e-08, + "grad_norm": 4.31052303314209, + "learning_rate": 1.7532942898975108e-07, + "loss": 1.119, + "mean_token_accuracy": 0.6756763458251953, + "num_tokens": 12439060.0, + "step": 480 + }, + { + "epoch": 0.052822314957171095, + "ewc_loss": 9.359791874885559e-08, + "grad_norm": 3.9091248512268066, + "learning_rate": 1.7569546120058565e-07, + "loss": 1.1274, + "mean_token_accuracy": 0.6708647012710571, + "num_tokens": 12462856.0, + "step": 481 + }, + { + "epoch": 0.052932132659784754, + "ewc_loss": 9.359791874885559e-08, + "grad_norm": 4.191350936889648, + "learning_rate": 1.760614934114202e-07, + "loss": 1.0941, + "mean_token_accuracy": 0.6795573234558105, + "num_tokens": 12487168.0, + "step": 482 + }, + { + "epoch": 0.05304195036239842, + "ewc_loss": 9.452924132347107e-08, + "grad_norm": 5.625139236450195, + "learning_rate": 1.7642752562225475e-07, + "loss": 1.252, + "mean_token_accuracy": 0.6432894468307495, + "num_tokens": 12516282.0, + "step": 483 + }, + { + "epoch": 0.05315176806501208, + "ewc_loss": 9.452924132347107e-08, + "grad_norm": 4.151322841644287, + "learning_rate": 1.767935578330893e-07, + "loss": 1.1118, + "mean_token_accuracy": 0.6790677309036255, + "num_tokens": 12541665.0, + "step": 484 + }, + { + "epoch": 0.05326158576762574, + "ewc_loss": 9.452924132347107e-08, + "grad_norm": 3.9968154430389404, + "learning_rate": 1.7715959004392387e-07, + "loss": 1.2418, + "mean_token_accuracy": 0.6540231704711914, + "num_tokens": 12573589.0, + "step": 485 + }, + { + "epoch": 0.0533714034702394, + "ewc_loss": 9.499490261077881e-08, + "grad_norm": 3.5681331157684326, + "learning_rate": 1.7752562225475842e-07, + "loss": 1.1353, + "mean_token_accuracy": 0.672529935836792, + "num_tokens": 12601106.0, + "step": 486 + }, + { + "epoch": 0.05348122117285306, + "ewc_loss": 9.546056389808655e-08, + "grad_norm": 3.7172157764434814, + "learning_rate": 1.7789165446559295e-07, + "loss": 1.1588, + "mean_token_accuracy": 0.67023104429245, + "num_tokens": 12631770.0, + "step": 487 + }, + { + "epoch": 0.05359103887546673, + "ewc_loss": 9.592622518539429e-08, + "grad_norm": 4.431347370147705, + "learning_rate": 1.7825768667642752e-07, + "loss": 1.1435, + "mean_token_accuracy": 0.6626328229904175, + "num_tokens": 12655234.0, + "step": 488 + }, + { + "epoch": 0.05370085657808039, + "ewc_loss": 9.685754776000977e-08, + "grad_norm": 3.5096302032470703, + "learning_rate": 1.7862371888726207e-07, + "loss": 1.1923, + "mean_token_accuracy": 0.6592422127723694, + "num_tokens": 12682179.0, + "step": 489 + }, + { + "epoch": 0.05381067428069405, + "ewc_loss": 9.73232090473175e-08, + "grad_norm": 3.9026060104370117, + "learning_rate": 1.7898975109809665e-07, + "loss": 1.1534, + "mean_token_accuracy": 0.6674436926841736, + "num_tokens": 12711843.0, + "step": 490 + }, + { + "epoch": 0.05392049198330771, + "ewc_loss": 9.73232090473175e-08, + "grad_norm": 3.967731237411499, + "learning_rate": 1.7935578330893117e-07, + "loss": 1.1956, + "mean_token_accuracy": 0.6570665240287781, + "num_tokens": 12741567.0, + "step": 491 + }, + { + "epoch": 0.054030309685921374, + "ewc_loss": 9.73232090473175e-08, + "grad_norm": 3.8897740840911865, + "learning_rate": 1.7972181551976572e-07, + "loss": 1.1946, + "mean_token_accuracy": 0.661945104598999, + "num_tokens": 12767598.0, + "step": 492 + }, + { + "epoch": 0.054140127388535034, + "ewc_loss": 9.778887033462524e-08, + "grad_norm": 3.976118803024292, + "learning_rate": 1.800878477306003e-07, + "loss": 1.1483, + "mean_token_accuracy": 0.6748595237731934, + "num_tokens": 12790026.0, + "step": 493 + }, + { + "epoch": 0.054249945091148694, + "ewc_loss": 9.872019290924072e-08, + "grad_norm": 3.7596230506896973, + "learning_rate": 1.8045387994143484e-07, + "loss": 1.1542, + "mean_token_accuracy": 0.6725994348526001, + "num_tokens": 12818197.0, + "step": 494 + }, + { + "epoch": 0.05435976279376235, + "ewc_loss": 9.872019290924072e-08, + "grad_norm": 4.19468879699707, + "learning_rate": 1.808199121522694e-07, + "loss": 1.1641, + "mean_token_accuracy": 0.6633111238479614, + "num_tokens": 12840299.0, + "step": 495 + }, + { + "epoch": 0.05446958049637601, + "ewc_loss": 9.96515154838562e-08, + "grad_norm": 4.187453269958496, + "learning_rate": 1.8118594436310394e-07, + "loss": 1.2402, + "mean_token_accuracy": 0.646398663520813, + "num_tokens": 12865428.0, + "step": 496 + }, + { + "epoch": 0.05457939819898968, + "ewc_loss": 1.0058283805847168e-07, + "grad_norm": 3.249796152114868, + "learning_rate": 1.8155197657393852e-07, + "loss": 1.1551, + "mean_token_accuracy": 0.6756320595741272, + "num_tokens": 12892978.0, + "step": 497 + }, + { + "epoch": 0.05468921590160334, + "ewc_loss": 1.0058283805847168e-07, + "grad_norm": 3.7397725582122803, + "learning_rate": 1.8191800878477307e-07, + "loss": 1.2003, + "mean_token_accuracy": 0.6684126853942871, + "num_tokens": 12921515.0, + "step": 498 + }, + { + "epoch": 0.054799033604217, + "ewc_loss": 1.0058283805847168e-07, + "grad_norm": 4.27509069442749, + "learning_rate": 1.822840409956076e-07, + "loss": 1.1835, + "mean_token_accuracy": 0.6629720330238342, + "num_tokens": 12947231.0, + "step": 499 + }, + { + "epoch": 0.05490885130683066, + "ewc_loss": 1.0058283805847168e-07, + "grad_norm": 4.514637470245361, + "learning_rate": 1.8265007320644216e-07, + "loss": 1.1933, + "mean_token_accuracy": 0.6570484042167664, + "num_tokens": 12975796.0, + "step": 500 + }, + { + "epoch": 0.05501866900944432, + "ewc_loss": 1.0058283805847168e-07, + "grad_norm": 3.3871443271636963, + "learning_rate": 1.830161054172767e-07, + "loss": 1.1297, + "mean_token_accuracy": 0.6785916686058044, + "num_tokens": 13005740.0, + "step": 501 + }, + { + "epoch": 0.055128486712057986, + "ewc_loss": 1.0104849934577942e-07, + "grad_norm": 3.889449119567871, + "learning_rate": 1.833821376281113e-07, + "loss": 1.1633, + "mean_token_accuracy": 0.6598455309867859, + "num_tokens": 13029608.0, + "step": 502 + }, + { + "epoch": 0.055238304414671646, + "ewc_loss": 1.0104849934577942e-07, + "grad_norm": 4.483788013458252, + "learning_rate": 1.837481698389458e-07, + "loss": 1.0729, + "mean_token_accuracy": 0.6899316310882568, + "num_tokens": 13050822.0, + "step": 503 + }, + { + "epoch": 0.055348122117285306, + "ewc_loss": 1.0104849934577942e-07, + "grad_norm": 3.7295114994049072, + "learning_rate": 1.8411420204978036e-07, + "loss": 0.9972, + "mean_token_accuracy": 0.70363450050354, + "num_tokens": 13076150.0, + "step": 504 + }, + { + "epoch": 0.055457939819898966, + "ewc_loss": 1.0151416063308716e-07, + "grad_norm": 3.359771251678467, + "learning_rate": 1.8448023426061494e-07, + "loss": 1.2143, + "mean_token_accuracy": 0.6511077880859375, + "num_tokens": 13103581.0, + "step": 505 + }, + { + "epoch": 0.055567757522512626, + "ewc_loss": 1.0244548320770264e-07, + "grad_norm": 3.381442070007324, + "learning_rate": 1.8484626647144948e-07, + "loss": 1.1471, + "mean_token_accuracy": 0.6627415418624878, + "num_tokens": 13131904.0, + "step": 506 + }, + { + "epoch": 0.05567757522512629, + "ewc_loss": 1.0291114449501038e-07, + "grad_norm": 4.383810043334961, + "learning_rate": 1.8521229868228403e-07, + "loss": 1.054, + "mean_token_accuracy": 0.6901146769523621, + "num_tokens": 13155934.0, + "step": 507 + }, + { + "epoch": 0.05578739292773995, + "ewc_loss": 1.0291114449501038e-07, + "grad_norm": 3.930464029312134, + "learning_rate": 1.8557833089311858e-07, + "loss": 1.1097, + "mean_token_accuracy": 0.6770781874656677, + "num_tokens": 13179650.0, + "step": 508 + }, + { + "epoch": 0.05589721063035361, + "ewc_loss": 1.0291114449501038e-07, + "grad_norm": 3.9483463764190674, + "learning_rate": 1.8594436310395313e-07, + "loss": 1.2012, + "mean_token_accuracy": 0.6556391716003418, + "num_tokens": 13205174.0, + "step": 509 + }, + { + "epoch": 0.05600702833296727, + "ewc_loss": 1.0291114449501038e-07, + "grad_norm": 3.708503246307373, + "learning_rate": 1.863103953147877e-07, + "loss": 1.1427, + "mean_token_accuracy": 0.666342556476593, + "num_tokens": 13231544.0, + "step": 510 + }, + { + "epoch": 0.05611684603558094, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 4.193999290466309, + "learning_rate": 1.8667642752562223e-07, + "loss": 1.1717, + "mean_token_accuracy": 0.662018895149231, + "num_tokens": 13252845.0, + "step": 511 + }, + { + "epoch": 0.0562266637381946, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 3.4331016540527344, + "learning_rate": 1.870424597364568e-07, + "loss": 1.2658, + "mean_token_accuracy": 0.6383774280548096, + "num_tokens": 13285009.0, + "step": 512 + }, + { + "epoch": 0.05633648144080826, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 3.1821043491363525, + "learning_rate": 1.8740849194729135e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6884359121322632, + "num_tokens": 13312518.0, + "step": 513 + }, + { + "epoch": 0.05644629914342192, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 3.489478349685669, + "learning_rate": 1.8777452415812593e-07, + "loss": 1.0851, + "mean_token_accuracy": 0.6892057657241821, + "num_tokens": 13335578.0, + "step": 514 + }, + { + "epoch": 0.05655611684603558, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 3.2487807273864746, + "learning_rate": 1.8814055636896045e-07, + "loss": 1.1914, + "mean_token_accuracy": 0.6666512489318848, + "num_tokens": 13368449.0, + "step": 515 + }, + { + "epoch": 0.056665934548649245, + "ewc_loss": 1.0384246706962585e-07, + "grad_norm": 3.755046844482422, + "learning_rate": 1.88506588579795e-07, + "loss": 1.2205, + "mean_token_accuracy": 0.6547521948814392, + "num_tokens": 13397250.0, + "step": 516 + }, + { + "epoch": 0.056775752251262905, + "ewc_loss": 1.043081283569336e-07, + "grad_norm": 4.3575544357299805, + "learning_rate": 1.8887262079062958e-07, + "loss": 1.161, + "mean_token_accuracy": 0.6584144234657288, + "num_tokens": 13422857.0, + "step": 517 + }, + { + "epoch": 0.056885569953876565, + "ewc_loss": 1.043081283569336e-07, + "grad_norm": 3.5851938724517822, + "learning_rate": 1.8923865300146413e-07, + "loss": 1.0862, + "mean_token_accuracy": 0.6770681142807007, + "num_tokens": 13447180.0, + "step": 518 + }, + { + "epoch": 0.056995387656490225, + "ewc_loss": 1.043081283569336e-07, + "grad_norm": 4.005226135253906, + "learning_rate": 1.8960468521229868e-07, + "loss": 1.2382, + "mean_token_accuracy": 0.6408832669258118, + "num_tokens": 13469860.0, + "step": 519 + }, + { + "epoch": 0.057105205359103885, + "ewc_loss": 1.0523945093154907e-07, + "grad_norm": 3.653468370437622, + "learning_rate": 1.8997071742313322e-07, + "loss": 1.1862, + "mean_token_accuracy": 0.6684273481369019, + "num_tokens": 13495606.0, + "step": 520 + }, + { + "epoch": 0.05721502306171755, + "ewc_loss": 1.0570511221885681e-07, + "grad_norm": 3.5834453105926514, + "learning_rate": 1.9033674963396777e-07, + "loss": 1.1675, + "mean_token_accuracy": 0.6653974056243896, + "num_tokens": 13527790.0, + "step": 521 + }, + { + "epoch": 0.05732484076433121, + "ewc_loss": 1.0570511221885681e-07, + "grad_norm": 3.647390127182007, + "learning_rate": 1.9070278184480235e-07, + "loss": 1.1494, + "mean_token_accuracy": 0.679023027420044, + "num_tokens": 13551986.0, + "step": 522 + }, + { + "epoch": 0.05743465846694487, + "ewc_loss": 1.0570511221885681e-07, + "grad_norm": 3.5365028381347656, + "learning_rate": 1.9106881405563687e-07, + "loss": 1.1957, + "mean_token_accuracy": 0.6570612788200378, + "num_tokens": 13577954.0, + "step": 523 + }, + { + "epoch": 0.05754447616955853, + "ewc_loss": 1.0570511221885681e-07, + "grad_norm": 3.5762338638305664, + "learning_rate": 1.9143484626647145e-07, + "loss": 1.1951, + "mean_token_accuracy": 0.6541205048561096, + "num_tokens": 13604870.0, + "step": 524 + }, + { + "epoch": 0.05765429387217219, + "ewc_loss": 1.0570511221885681e-07, + "grad_norm": 3.0418760776519775, + "learning_rate": 1.91800878477306e-07, + "loss": 1.2111, + "mean_token_accuracy": 0.6586421728134155, + "num_tokens": 13637200.0, + "step": 525 + }, + { + "epoch": 0.05776411157478586, + "ewc_loss": 1.0617077350616455e-07, + "grad_norm": 4.904503345489502, + "learning_rate": 1.9216691068814057e-07, + "loss": 1.2036, + "mean_token_accuracy": 0.65028315782547, + "num_tokens": 13656718.0, + "step": 526 + }, + { + "epoch": 0.05787392927739952, + "ewc_loss": 1.0617077350616455e-07, + "grad_norm": 4.87709903717041, + "learning_rate": 1.925329428989751e-07, + "loss": 1.1189, + "mean_token_accuracy": 0.6757477521896362, + "num_tokens": 13680498.0, + "step": 527 + }, + { + "epoch": 0.05798374698001318, + "ewc_loss": 1.0617077350616455e-07, + "grad_norm": 4.665056228637695, + "learning_rate": 1.9289897510980964e-07, + "loss": 1.0911, + "mean_token_accuracy": 0.6854211091995239, + "num_tokens": 13701339.0, + "step": 528 + }, + { + "epoch": 0.05809356468262684, + "ewc_loss": 1.0617077350616455e-07, + "grad_norm": 3.831965208053589, + "learning_rate": 1.9326500732064422e-07, + "loss": 1.1731, + "mean_token_accuracy": 0.6646527051925659, + "num_tokens": 13724927.0, + "step": 529 + }, + { + "epoch": 0.058203382385240504, + "ewc_loss": 1.0617077350616455e-07, + "grad_norm": 4.206918239593506, + "learning_rate": 1.9363103953147877e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6874370574951172, + "num_tokens": 13745791.0, + "step": 530 + }, + { + "epoch": 0.058313200087854164, + "ewc_loss": 1.0663643479347229e-07, + "grad_norm": 3.6930999755859375, + "learning_rate": 1.9399707174231332e-07, + "loss": 1.0975, + "mean_token_accuracy": 0.6783806681632996, + "num_tokens": 13770891.0, + "step": 531 + }, + { + "epoch": 0.058423017790467824, + "ewc_loss": 1.0663643479347229e-07, + "grad_norm": 4.694607257843018, + "learning_rate": 1.9436310395314787e-07, + "loss": 1.123, + "mean_token_accuracy": 0.6683223247528076, + "num_tokens": 13790625.0, + "step": 532 + }, + { + "epoch": 0.058532835493081484, + "ewc_loss": 1.0710209608078003e-07, + "grad_norm": 3.484440326690674, + "learning_rate": 1.9472913616398242e-07, + "loss": 1.0871, + "mean_token_accuracy": 0.6800087690353394, + "num_tokens": 13813346.0, + "step": 533 + }, + { + "epoch": 0.058642653195695144, + "ewc_loss": 1.0710209608078003e-07, + "grad_norm": 3.459143877029419, + "learning_rate": 1.95095168374817e-07, + "loss": 1.2015, + "mean_token_accuracy": 0.6562846302986145, + "num_tokens": 13837250.0, + "step": 534 + }, + { + "epoch": 0.05875247089830881, + "ewc_loss": 1.0710209608078003e-07, + "grad_norm": 3.3715691566467285, + "learning_rate": 1.9546120058565151e-07, + "loss": 1.1256, + "mean_token_accuracy": 0.6777064800262451, + "num_tokens": 13865730.0, + "step": 535 + }, + { + "epoch": 0.05886228860092247, + "ewc_loss": 1.0756775736808777e-07, + "grad_norm": 3.827408790588379, + "learning_rate": 1.958272327964861e-07, + "loss": 1.1132, + "mean_token_accuracy": 0.6750330924987793, + "num_tokens": 13893215.0, + "step": 536 + }, + { + "epoch": 0.05897210630353613, + "ewc_loss": 1.0803341865539551e-07, + "grad_norm": 3.3361692428588867, + "learning_rate": 1.9619326500732064e-07, + "loss": 1.1629, + "mean_token_accuracy": 0.6644637584686279, + "num_tokens": 13919232.0, + "step": 537 + }, + { + "epoch": 0.05908192400614979, + "ewc_loss": 1.0849907994270325e-07, + "grad_norm": 3.5054965019226074, + "learning_rate": 1.9655929721815521e-07, + "loss": 1.1707, + "mean_token_accuracy": 0.6641207933425903, + "num_tokens": 13943661.0, + "step": 538 + }, + { + "epoch": 0.05919174170876345, + "ewc_loss": 1.0896474123001099e-07, + "grad_norm": 3.953789234161377, + "learning_rate": 1.9692532942898974e-07, + "loss": 1.174, + "mean_token_accuracy": 0.657228410243988, + "num_tokens": 13967209.0, + "step": 539 + }, + { + "epoch": 0.05930155941137712, + "ewc_loss": 1.0896474123001099e-07, + "grad_norm": 3.834622383117676, + "learning_rate": 1.9729136163982429e-07, + "loss": 1.0371, + "mean_token_accuracy": 0.6963911652565002, + "num_tokens": 13987948.0, + "step": 540 + }, + { + "epoch": 0.059411377113990776, + "ewc_loss": 1.0896474123001099e-07, + "grad_norm": 3.520174026489258, + "learning_rate": 1.9765739385065886e-07, + "loss": 1.2027, + "mean_token_accuracy": 0.659548282623291, + "num_tokens": 14016862.0, + "step": 541 + }, + { + "epoch": 0.059521194816604436, + "ewc_loss": 1.0943040251731873e-07, + "grad_norm": 3.4923973083496094, + "learning_rate": 1.980234260614934e-07, + "loss": 1.1856, + "mean_token_accuracy": 0.6639961004257202, + "num_tokens": 14041242.0, + "step": 542 + }, + { + "epoch": 0.059631012519218096, + "ewc_loss": 1.0943040251731873e-07, + "grad_norm": 3.3542373180389404, + "learning_rate": 1.9838945827232796e-07, + "loss": 1.088, + "mean_token_accuracy": 0.6863478422164917, + "num_tokens": 14064732.0, + "step": 543 + }, + { + "epoch": 0.059740830221831756, + "ewc_loss": 1.0943040251731873e-07, + "grad_norm": 3.2417356967926025, + "learning_rate": 1.987554904831625e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6868901252746582, + "num_tokens": 14090186.0, + "step": 544 + }, + { + "epoch": 0.05985064792444542, + "ewc_loss": 1.0989606380462646e-07, + "grad_norm": 2.9082791805267334, + "learning_rate": 1.9912152269399706e-07, + "loss": 1.2148, + "mean_token_accuracy": 0.6549419164657593, + "num_tokens": 14121747.0, + "step": 545 + }, + { + "epoch": 0.05996046562705908, + "ewc_loss": 1.1082738637924194e-07, + "grad_norm": 2.964714527130127, + "learning_rate": 1.9948755490483163e-07, + "loss": 1.2256, + "mean_token_accuracy": 0.650661826133728, + "num_tokens": 14152053.0, + "step": 546 + }, + { + "epoch": 0.06007028332967274, + "ewc_loss": 1.1082738637924194e-07, + "grad_norm": 5.176568984985352, + "learning_rate": 1.9985358711566616e-07, + "loss": 1.1255, + "mean_token_accuracy": 0.6811807155609131, + "num_tokens": 14170890.0, + "step": 547 + }, + { + "epoch": 0.0601801010322864, + "ewc_loss": 1.1175870895385742e-07, + "grad_norm": 3.120346784591675, + "learning_rate": 2.0021961932650073e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.676670253276825, + "num_tokens": 14200001.0, + "step": 548 + }, + { + "epoch": 0.06028991873490007, + "ewc_loss": 1.1175870895385742e-07, + "grad_norm": 3.209749460220337, + "learning_rate": 2.0058565153733528e-07, + "loss": 1.1537, + "mean_token_accuracy": 0.6706610918045044, + "num_tokens": 14225062.0, + "step": 549 + }, + { + "epoch": 0.06039973643751373, + "ewc_loss": 1.1175870895385742e-07, + "grad_norm": 3.400712013244629, + "learning_rate": 2.0095168374816983e-07, + "loss": 1.1251, + "mean_token_accuracy": 0.6723026037216187, + "num_tokens": 14248169.0, + "step": 550 + }, + { + "epoch": 0.06050955414012739, + "ewc_loss": 1.1175870895385742e-07, + "grad_norm": 3.2912232875823975, + "learning_rate": 2.0131771595900438e-07, + "loss": 1.146, + "mean_token_accuracy": 0.6751861572265625, + "num_tokens": 14273309.0, + "step": 551 + }, + { + "epoch": 0.06061937184274105, + "ewc_loss": 1.1222437024116516e-07, + "grad_norm": 3.287330150604248, + "learning_rate": 2.0168374816983893e-07, + "loss": 1.1054, + "mean_token_accuracy": 0.6805760860443115, + "num_tokens": 14300494.0, + "step": 552 + }, + { + "epoch": 0.06072918954535471, + "ewc_loss": 1.126900315284729e-07, + "grad_norm": 3.3339076042175293, + "learning_rate": 2.020497803806735e-07, + "loss": 1.1404, + "mean_token_accuracy": 0.6652939319610596, + "num_tokens": 14327576.0, + "step": 553 + }, + { + "epoch": 0.060839007247968376, + "ewc_loss": 1.126900315284729e-07, + "grad_norm": 2.822925090789795, + "learning_rate": 2.0241581259150805e-07, + "loss": 1.1642, + "mean_token_accuracy": 0.6606607437133789, + "num_tokens": 14359413.0, + "step": 554 + }, + { + "epoch": 0.060948824950582035, + "ewc_loss": 1.1408701539039612e-07, + "grad_norm": 3.3725943565368652, + "learning_rate": 2.027818448023426e-07, + "loss": 1.283, + "mean_token_accuracy": 0.6491679549217224, + "num_tokens": 14389772.0, + "step": 555 + }, + { + "epoch": 0.061058642653195695, + "ewc_loss": 1.1455267667770386e-07, + "grad_norm": 3.179894208908081, + "learning_rate": 2.0314787701317715e-07, + "loss": 1.1972, + "mean_token_accuracy": 0.6620694398880005, + "num_tokens": 14416775.0, + "step": 556 + }, + { + "epoch": 0.061168460355809355, + "ewc_loss": 1.1548399925231934e-07, + "grad_norm": 3.3495140075683594, + "learning_rate": 2.035139092240117e-07, + "loss": 1.2381, + "mean_token_accuracy": 0.6433632969856262, + "num_tokens": 14445547.0, + "step": 557 + }, + { + "epoch": 0.061278278058423015, + "ewc_loss": 1.1548399925231934e-07, + "grad_norm": 3.534414291381836, + "learning_rate": 2.0387994143484627e-07, + "loss": 1.2102, + "mean_token_accuracy": 0.6513330936431885, + "num_tokens": 14469951.0, + "step": 558 + }, + { + "epoch": 0.06138809576103668, + "ewc_loss": 1.1641532182693481e-07, + "grad_norm": 3.406728744506836, + "learning_rate": 2.042459736456808e-07, + "loss": 1.1694, + "mean_token_accuracy": 0.665431559085846, + "num_tokens": 14492322.0, + "step": 559 + }, + { + "epoch": 0.06149791346365034, + "ewc_loss": 1.1641532182693481e-07, + "grad_norm": 3.0219104290008545, + "learning_rate": 2.0461200585651537e-07, + "loss": 1.1605, + "mean_token_accuracy": 0.6663335561752319, + "num_tokens": 14517881.0, + "step": 560 + }, + { + "epoch": 0.061607731166264, + "ewc_loss": 1.1734664440155029e-07, + "grad_norm": 3.741159677505493, + "learning_rate": 2.0497803806734992e-07, + "loss": 1.1902, + "mean_token_accuracy": 0.6648459434509277, + "num_tokens": 14543018.0, + "step": 561 + }, + { + "epoch": 0.06171754886887766, + "ewc_loss": 1.1827796697616577e-07, + "grad_norm": 3.560257911682129, + "learning_rate": 2.0534407027818447e-07, + "loss": 1.1699, + "mean_token_accuracy": 0.6602627635002136, + "num_tokens": 14570418.0, + "step": 562 + }, + { + "epoch": 0.06182736657149132, + "ewc_loss": 1.1920928955078125e-07, + "grad_norm": 3.8037002086639404, + "learning_rate": 2.0571010248901902e-07, + "loss": 1.0272, + "mean_token_accuracy": 0.6955463886260986, + "num_tokens": 14589713.0, + "step": 563 + }, + { + "epoch": 0.06193718427410499, + "ewc_loss": 1.1920928955078125e-07, + "grad_norm": 3.1651268005371094, + "learning_rate": 2.0607613469985357e-07, + "loss": 1.1703, + "mean_token_accuracy": 0.6623156666755676, + "num_tokens": 14617418.0, + "step": 564 + }, + { + "epoch": 0.06204700197671865, + "ewc_loss": 1.2014061212539673e-07, + "grad_norm": 3.7551462650299072, + "learning_rate": 2.0644216691068814e-07, + "loss": 1.1425, + "mean_token_accuracy": 0.6752495765686035, + "num_tokens": 14645074.0, + "step": 565 + }, + { + "epoch": 0.06215681967933231, + "ewc_loss": 1.2014061212539673e-07, + "grad_norm": 3.1198172569274902, + "learning_rate": 2.068081991215227e-07, + "loss": 1.2151, + "mean_token_accuracy": 0.6540681719779968, + "num_tokens": 14671552.0, + "step": 566 + }, + { + "epoch": 0.06226663738194597, + "ewc_loss": 1.210719347000122e-07, + "grad_norm": 3.4915385246276855, + "learning_rate": 2.0717423133235724e-07, + "loss": 1.1367, + "mean_token_accuracy": 0.6719883680343628, + "num_tokens": 14694223.0, + "step": 567 + }, + { + "epoch": 0.062376455084559634, + "ewc_loss": 1.2200325727462769e-07, + "grad_norm": 3.3863046169281006, + "learning_rate": 2.075402635431918e-07, + "loss": 1.0837, + "mean_token_accuracy": 0.6901718378067017, + "num_tokens": 14719184.0, + "step": 568 + }, + { + "epoch": 0.062486272787173294, + "ewc_loss": 1.2200325727462769e-07, + "grad_norm": 3.1630611419677734, + "learning_rate": 2.0790629575402634e-07, + "loss": 1.1428, + "mean_token_accuracy": 0.6718358397483826, + "num_tokens": 14742235.0, + "step": 569 + }, + { + "epoch": 0.06259609048978695, + "ewc_loss": 1.2200325727462769e-07, + "grad_norm": 3.292558193206787, + "learning_rate": 2.0827232796486092e-07, + "loss": 1.2201, + "mean_token_accuracy": 0.6531417369842529, + "num_tokens": 14769348.0, + "step": 570 + }, + { + "epoch": 0.06270590819240061, + "ewc_loss": 1.2200325727462769e-07, + "grad_norm": 3.167306661605835, + "learning_rate": 2.0863836017569544e-07, + "loss": 1.0511, + "mean_token_accuracy": 0.6917932629585266, + "num_tokens": 14794555.0, + "step": 571 + }, + { + "epoch": 0.06281572589501427, + "ewc_loss": 1.2200325727462769e-07, + "grad_norm": 3.268923282623291, + "learning_rate": 2.0900439238653001e-07, + "loss": 1.1408, + "mean_token_accuracy": 0.6693302392959595, + "num_tokens": 14817364.0, + "step": 572 + }, + { + "epoch": 0.06292554359762793, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 3.6721839904785156, + "learning_rate": 2.0937042459736456e-07, + "loss": 1.1735, + "mean_token_accuracy": 0.6674728393554688, + "num_tokens": 14837472.0, + "step": 573 + }, + { + "epoch": 0.0630353613002416, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.798621892929077, + "learning_rate": 2.097364568081991e-07, + "loss": 1.116, + "mean_token_accuracy": 0.6779206991195679, + "num_tokens": 14865112.0, + "step": 574 + }, + { + "epoch": 0.06314517900285525, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 3.2041070461273193, + "learning_rate": 2.1010248901903366e-07, + "loss": 1.1526, + "mean_token_accuracy": 0.6633718609809875, + "num_tokens": 14894808.0, + "step": 575 + }, + { + "epoch": 0.06325499670546893, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 3.925204038619995, + "learning_rate": 2.104685212298682e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6956723928451538, + "num_tokens": 14913195.0, + "step": 576 + }, + { + "epoch": 0.06336481440808259, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.879476547241211, + "learning_rate": 2.1083455344070279e-07, + "loss": 1.189, + "mean_token_accuracy": 0.6603565812110901, + "num_tokens": 14943929.0, + "step": 577 + }, + { + "epoch": 0.06347463211069625, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.9931533336639404, + "learning_rate": 2.1120058565153734e-07, + "loss": 1.1705, + "mean_token_accuracy": 0.6654261350631714, + "num_tokens": 14969696.0, + "step": 578 + }, + { + "epoch": 0.0635844498133099, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 3.58992600440979, + "learning_rate": 2.1156661786237186e-07, + "loss": 1.0125, + "mean_token_accuracy": 0.6973406672477722, + "num_tokens": 14989320.0, + "step": 579 + }, + { + "epoch": 0.06369426751592357, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.9582552909851074, + "learning_rate": 2.1193265007320643e-07, + "loss": 1.0774, + "mean_token_accuracy": 0.6837865114212036, + "num_tokens": 15014145.0, + "step": 580 + }, + { + "epoch": 0.06380408521853723, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.8779501914978027, + "learning_rate": 2.1229868228404098e-07, + "loss": 1.1572, + "mean_token_accuracy": 0.6615336537361145, + "num_tokens": 15040827.0, + "step": 581 + }, + { + "epoch": 0.06391390292115089, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 2.8072164058685303, + "learning_rate": 2.1266471449487556e-07, + "loss": 1.2056, + "mean_token_accuracy": 0.6578577160835266, + "num_tokens": 15069707.0, + "step": 582 + }, + { + "epoch": 0.06402372062376455, + "ewc_loss": 1.2293457984924316e-07, + "grad_norm": 3.604595422744751, + "learning_rate": 2.1303074670571008e-07, + "loss": 1.1427, + "mean_token_accuracy": 0.6684044599533081, + "num_tokens": 15091567.0, + "step": 583 + }, + { + "epoch": 0.0641335383263782, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.5932137966156006, + "learning_rate": 2.1339677891654466e-07, + "loss": 1.2179, + "mean_token_accuracy": 0.6530357599258423, + "num_tokens": 15117176.0, + "step": 584 + }, + { + "epoch": 0.06424335602899188, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 2.698054313659668, + "learning_rate": 2.137628111273792e-07, + "loss": 1.1608, + "mean_token_accuracy": 0.6680794954299927, + "num_tokens": 15151656.0, + "step": 585 + }, + { + "epoch": 0.06435317373160554, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.1010444164276123, + "learning_rate": 2.1412884333821375e-07, + "loss": 1.0639, + "mean_token_accuracy": 0.6953916549682617, + "num_tokens": 15176031.0, + "step": 586 + }, + { + "epoch": 0.0644629914342192, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.081023693084717, + "learning_rate": 2.144948755490483e-07, + "loss": 1.1201, + "mean_token_accuracy": 0.669176459312439, + "num_tokens": 15197744.0, + "step": 587 + }, + { + "epoch": 0.06457280913683286, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.6625096797943115, + "learning_rate": 2.1486090775988285e-07, + "loss": 1.0988, + "mean_token_accuracy": 0.6811056137084961, + "num_tokens": 15219377.0, + "step": 588 + }, + { + "epoch": 0.06468262683944652, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.019169807434082, + "learning_rate": 2.1522693997071743e-07, + "loss": 1.1132, + "mean_token_accuracy": 0.6753664016723633, + "num_tokens": 15244196.0, + "step": 589 + }, + { + "epoch": 0.06479244454206018, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 2.819765329360962, + "learning_rate": 2.1559297218155198e-07, + "loss": 1.1652, + "mean_token_accuracy": 0.6742969751358032, + "num_tokens": 15273093.0, + "step": 590 + }, + { + "epoch": 0.06490226224467384, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.0821754932403564, + "learning_rate": 2.159590043923865e-07, + "loss": 1.1215, + "mean_token_accuracy": 0.677141547203064, + "num_tokens": 15299076.0, + "step": 591 + }, + { + "epoch": 0.0650120799472875, + "ewc_loss": 1.2386590242385864e-07, + "grad_norm": 3.1920061111450195, + "learning_rate": 2.1632503660322108e-07, + "loss": 1.0408, + "mean_token_accuracy": 0.7004088163375854, + "num_tokens": 15323632.0, + "step": 592 + }, + { + "epoch": 0.06512189764990116, + "ewc_loss": 1.257285475730896e-07, + "grad_norm": 2.704221487045288, + "learning_rate": 2.1669106881405563e-07, + "loss": 1.1379, + "mean_token_accuracy": 0.6719347834587097, + "num_tokens": 15356700.0, + "step": 593 + }, + { + "epoch": 0.06523171535251482, + "ewc_loss": 1.257285475730896e-07, + "grad_norm": 2.957132577896118, + "learning_rate": 2.170571010248902e-07, + "loss": 1.0454, + "mean_token_accuracy": 0.6970637440681458, + "num_tokens": 15382486.0, + "step": 594 + }, + { + "epoch": 0.06534153305512849, + "ewc_loss": 1.2759119272232056e-07, + "grad_norm": 3.7799758911132812, + "learning_rate": 2.1742313323572472e-07, + "loss": 1.0713, + "mean_token_accuracy": 0.6910166144371033, + "num_tokens": 15403037.0, + "step": 595 + }, + { + "epoch": 0.06545135075774215, + "ewc_loss": 1.2759119272232056e-07, + "grad_norm": 3.278299570083618, + "learning_rate": 2.177891654465593e-07, + "loss": 1.0671, + "mean_token_accuracy": 0.6902358531951904, + "num_tokens": 15427884.0, + "step": 596 + }, + { + "epoch": 0.06556116846035581, + "ewc_loss": 1.2759119272232056e-07, + "grad_norm": 2.5324230194091797, + "learning_rate": 2.1815519765739385e-07, + "loss": 1.1402, + "mean_token_accuracy": 0.6673785448074341, + "num_tokens": 15461669.0, + "step": 597 + }, + { + "epoch": 0.06567098616296947, + "ewc_loss": 1.2852251529693604e-07, + "grad_norm": 3.5406670570373535, + "learning_rate": 2.185212298682284e-07, + "loss": 1.0826, + "mean_token_accuracy": 0.6793152093887329, + "num_tokens": 15484624.0, + "step": 598 + }, + { + "epoch": 0.06578080386558313, + "ewc_loss": 1.2852251529693604e-07, + "grad_norm": 3.012824535369873, + "learning_rate": 2.1888726207906295e-07, + "loss": 1.1781, + "mean_token_accuracy": 0.6691081523895264, + "num_tokens": 15508390.0, + "step": 599 + }, + { + "epoch": 0.06589062156819679, + "ewc_loss": 1.2945383787155151e-07, + "grad_norm": 2.9984130859375, + "learning_rate": 2.192532942898975e-07, + "loss": 1.1044, + "mean_token_accuracy": 0.6796809434890747, + "num_tokens": 15535744.0, + "step": 600 + }, + { + "epoch": 0.06600043927081045, + "ewc_loss": 1.2945383787155151e-07, + "grad_norm": 2.77679181098938, + "learning_rate": 2.1961932650073207e-07, + "loss": 1.108, + "mean_token_accuracy": 0.6796026229858398, + "num_tokens": 15565036.0, + "step": 601 + }, + { + "epoch": 0.06611025697342411, + "ewc_loss": 1.2945383787155151e-07, + "grad_norm": 2.3727433681488037, + "learning_rate": 2.1998535871156662e-07, + "loss": 1.2022, + "mean_token_accuracy": 0.6585608720779419, + "num_tokens": 15597806.0, + "step": 602 + }, + { + "epoch": 0.06622007467603777, + "ewc_loss": 1.2945383787155151e-07, + "grad_norm": 3.1493453979492188, + "learning_rate": 2.2035139092240114e-07, + "loss": 1.1158, + "mean_token_accuracy": 0.6729243993759155, + "num_tokens": 15623941.0, + "step": 603 + }, + { + "epoch": 0.06632989237865144, + "ewc_loss": 1.30385160446167e-07, + "grad_norm": 3.2955148220062256, + "learning_rate": 2.2071742313323572e-07, + "loss": 1.1554, + "mean_token_accuracy": 0.6621822118759155, + "num_tokens": 15649843.0, + "step": 604 + }, + { + "epoch": 0.0664397100812651, + "ewc_loss": 1.3131648302078247e-07, + "grad_norm": 2.8885421752929688, + "learning_rate": 2.2108345534407027e-07, + "loss": 1.0493, + "mean_token_accuracy": 0.6892386674880981, + "num_tokens": 15677720.0, + "step": 605 + }, + { + "epoch": 0.06654952778387876, + "ewc_loss": 1.3131648302078247e-07, + "grad_norm": 3.1841490268707275, + "learning_rate": 2.2144948755490484e-07, + "loss": 1.0588, + "mean_token_accuracy": 0.688636302947998, + "num_tokens": 15704251.0, + "step": 606 + }, + { + "epoch": 0.06665934548649242, + "ewc_loss": 1.3224780559539795e-07, + "grad_norm": 3.298861026763916, + "learning_rate": 2.2181551976573937e-07, + "loss": 1.1387, + "mean_token_accuracy": 0.6674466729164124, + "num_tokens": 15729100.0, + "step": 607 + }, + { + "epoch": 0.06676916318910608, + "ewc_loss": 1.3317912817001343e-07, + "grad_norm": 3.2391934394836426, + "learning_rate": 2.2218155197657394e-07, + "loss": 1.165, + "mean_token_accuracy": 0.6554292440414429, + "num_tokens": 15753408.0, + "step": 608 + }, + { + "epoch": 0.06687898089171974, + "ewc_loss": 1.3317912817001343e-07, + "grad_norm": 3.131502151489258, + "learning_rate": 2.225475841874085e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6908332109451294, + "num_tokens": 15777400.0, + "step": 609 + }, + { + "epoch": 0.0669887985943334, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 2.795825719833374, + "learning_rate": 2.2291361639824304e-07, + "loss": 1.1924, + "mean_token_accuracy": 0.6642647981643677, + "num_tokens": 15808367.0, + "step": 610 + }, + { + "epoch": 0.06709861629694706, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 3.110978603363037, + "learning_rate": 2.232796486090776e-07, + "loss": 1.0218, + "mean_token_accuracy": 0.7075163722038269, + "num_tokens": 15831105.0, + "step": 611 + }, + { + "epoch": 0.06720843399956072, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 3.0038604736328125, + "learning_rate": 2.2364568081991214e-07, + "loss": 1.1306, + "mean_token_accuracy": 0.6736431121826172, + "num_tokens": 15854816.0, + "step": 612 + }, + { + "epoch": 0.06731825170217438, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 3.4330368041992188, + "learning_rate": 2.240117130307467e-07, + "loss": 1.118, + "mean_token_accuracy": 0.6746511459350586, + "num_tokens": 15873146.0, + "step": 613 + }, + { + "epoch": 0.06742806940478806, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 2.7414770126342773, + "learning_rate": 2.2437774524158126e-07, + "loss": 1.1666, + "mean_token_accuracy": 0.6693072319030762, + "num_tokens": 15898651.0, + "step": 614 + }, + { + "epoch": 0.06753788710740172, + "ewc_loss": 1.341104507446289e-07, + "grad_norm": 3.1385061740875244, + "learning_rate": 2.2474377745241578e-07, + "loss": 1.1651, + "mean_token_accuracy": 0.6573776006698608, + "num_tokens": 15921202.0, + "step": 615 + }, + { + "epoch": 0.06764770481001538, + "ewc_loss": 1.3504177331924438e-07, + "grad_norm": 3.3201887607574463, + "learning_rate": 2.2510980966325036e-07, + "loss": 1.0155, + "mean_token_accuracy": 0.6958348751068115, + "num_tokens": 15940488.0, + "step": 616 + }, + { + "epoch": 0.06775752251262904, + "ewc_loss": 1.3597309589385986e-07, + "grad_norm": 2.8872628211975098, + "learning_rate": 2.254758418740849e-07, + "loss": 1.1001, + "mean_token_accuracy": 0.684607744216919, + "num_tokens": 15967477.0, + "step": 617 + }, + { + "epoch": 0.0678673402152427, + "ewc_loss": 1.3597309589385986e-07, + "grad_norm": 2.859930992126465, + "learning_rate": 2.2584187408491948e-07, + "loss": 1.1159, + "mean_token_accuracy": 0.6740992069244385, + "num_tokens": 15995250.0, + "step": 618 + }, + { + "epoch": 0.06797715791785636, + "ewc_loss": 1.3690441846847534e-07, + "grad_norm": 2.9358911514282227, + "learning_rate": 2.26207906295754e-07, + "loss": 1.2263, + "mean_token_accuracy": 0.6544902324676514, + "num_tokens": 16020883.0, + "step": 619 + }, + { + "epoch": 0.06808697562047002, + "ewc_loss": 1.3690441846847534e-07, + "grad_norm": 2.7160823345184326, + "learning_rate": 2.2657393850658856e-07, + "loss": 1.2404, + "mean_token_accuracy": 0.6501462459564209, + "num_tokens": 16047600.0, + "step": 620 + }, + { + "epoch": 0.06819679332308368, + "ewc_loss": 1.387670636177063e-07, + "grad_norm": 3.297534942626953, + "learning_rate": 2.2693997071742313e-07, + "loss": 1.1319, + "mean_token_accuracy": 0.6716499328613281, + "num_tokens": 16070115.0, + "step": 621 + }, + { + "epoch": 0.06830661102569734, + "ewc_loss": 1.387670636177063e-07, + "grad_norm": 2.9325273036956787, + "learning_rate": 2.2730600292825768e-07, + "loss": 1.1621, + "mean_token_accuracy": 0.6680539846420288, + "num_tokens": 16096102.0, + "step": 622 + }, + { + "epoch": 0.06841642872831101, + "ewc_loss": 1.387670636177063e-07, + "grad_norm": 2.9754624366760254, + "learning_rate": 2.2767203513909223e-07, + "loss": 1.1969, + "mean_token_accuracy": 0.6616511940956116, + "num_tokens": 16122671.0, + "step": 623 + }, + { + "epoch": 0.06852624643092467, + "ewc_loss": 1.387670636177063e-07, + "grad_norm": 2.851780891418457, + "learning_rate": 2.2803806734992678e-07, + "loss": 1.0472, + "mean_token_accuracy": 0.6874707937240601, + "num_tokens": 16149145.0, + "step": 624 + }, + { + "epoch": 0.06863606413353833, + "ewc_loss": 1.387670636177063e-07, + "grad_norm": 2.593566656112671, + "learning_rate": 2.2840409956076135e-07, + "loss": 1.1313, + "mean_token_accuracy": 0.6814534664154053, + "num_tokens": 16178336.0, + "step": 625 + }, + { + "epoch": 0.06874588183615199, + "ewc_loss": 1.3969838619232178e-07, + "grad_norm": 3.060110330581665, + "learning_rate": 2.287701317715959e-07, + "loss": 1.134, + "mean_token_accuracy": 0.673925518989563, + "num_tokens": 16200808.0, + "step": 626 + }, + { + "epoch": 0.06885569953876565, + "ewc_loss": 1.3969838619232178e-07, + "grad_norm": 2.684286594390869, + "learning_rate": 2.2913616398243043e-07, + "loss": 1.2545, + "mean_token_accuracy": 0.6425625085830688, + "num_tokens": 16228780.0, + "step": 627 + }, + { + "epoch": 0.06896551724137931, + "ewc_loss": 1.3969838619232178e-07, + "grad_norm": 2.8380134105682373, + "learning_rate": 2.29502196193265e-07, + "loss": 1.1885, + "mean_token_accuracy": 0.6528246998786926, + "num_tokens": 16256783.0, + "step": 628 + }, + { + "epoch": 0.06907533494399297, + "ewc_loss": 1.3969838619232178e-07, + "grad_norm": 2.936497211456299, + "learning_rate": 2.2986822840409955e-07, + "loss": 1.1635, + "mean_token_accuracy": 0.6587975025177002, + "num_tokens": 16281018.0, + "step": 629 + }, + { + "epoch": 0.06918515264660663, + "ewc_loss": 1.4062970876693726e-07, + "grad_norm": 2.96618914604187, + "learning_rate": 2.3023426061493413e-07, + "loss": 1.0589, + "mean_token_accuracy": 0.6895112991333008, + "num_tokens": 16306575.0, + "step": 630 + }, + { + "epoch": 0.06929497034922029, + "ewc_loss": 1.424923539161682e-07, + "grad_norm": 2.601144552230835, + "learning_rate": 2.3060029282576865e-07, + "loss": 1.1959, + "mean_token_accuracy": 0.656477153301239, + "num_tokens": 16337454.0, + "step": 631 + }, + { + "epoch": 0.06940478805183395, + "ewc_loss": 1.424923539161682e-07, + "grad_norm": 2.9670045375823975, + "learning_rate": 2.309663250366032e-07, + "loss": 1.1331, + "mean_token_accuracy": 0.6657751798629761, + "num_tokens": 16360648.0, + "step": 632 + }, + { + "epoch": 0.06951460575444762, + "ewc_loss": 1.434236764907837e-07, + "grad_norm": 2.4942848682403564, + "learning_rate": 2.3133235724743777e-07, + "loss": 1.1312, + "mean_token_accuracy": 0.6700342893600464, + "num_tokens": 16389325.0, + "step": 633 + }, + { + "epoch": 0.06962442345706128, + "ewc_loss": 1.4435499906539917e-07, + "grad_norm": 3.277761220932007, + "learning_rate": 2.3169838945827232e-07, + "loss": 1.1629, + "mean_token_accuracy": 0.6662276983261108, + "num_tokens": 16412500.0, + "step": 634 + }, + { + "epoch": 0.06973424115967494, + "ewc_loss": 1.4435499906539917e-07, + "grad_norm": 2.297250509262085, + "learning_rate": 2.3206442166910687e-07, + "loss": 1.2102, + "mean_token_accuracy": 0.6467113494873047, + "num_tokens": 16449528.0, + "step": 635 + }, + { + "epoch": 0.0698440588622886, + "ewc_loss": 1.4435499906539917e-07, + "grad_norm": 2.653740644454956, + "learning_rate": 2.3243045387994142e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.7055431008338928, + "num_tokens": 16475464.0, + "step": 636 + }, + { + "epoch": 0.06995387656490226, + "ewc_loss": 1.4528632164001465e-07, + "grad_norm": 2.6837716102600098, + "learning_rate": 2.32796486090776e-07, + "loss": 1.1205, + "mean_token_accuracy": 0.6732784509658813, + "num_tokens": 16503191.0, + "step": 637 + }, + { + "epoch": 0.07006369426751592, + "ewc_loss": 1.4528632164001465e-07, + "grad_norm": 2.6396379470825195, + "learning_rate": 2.3316251830161055e-07, + "loss": 1.1446, + "mean_token_accuracy": 0.6643758416175842, + "num_tokens": 16532650.0, + "step": 638 + }, + { + "epoch": 0.07017351197012958, + "ewc_loss": 1.4621764421463013e-07, + "grad_norm": 2.763871908187866, + "learning_rate": 2.3352855051244507e-07, + "loss": 1.2165, + "mean_token_accuracy": 0.659692645072937, + "num_tokens": 16560255.0, + "step": 639 + }, + { + "epoch": 0.07028332967274324, + "ewc_loss": 1.4808028936386108e-07, + "grad_norm": 2.397930383682251, + "learning_rate": 2.3389458272327964e-07, + "loss": 1.173, + "mean_token_accuracy": 0.6586405038833618, + "num_tokens": 16591859.0, + "step": 640 + }, + { + "epoch": 0.0703931473753569, + "ewc_loss": 1.4808028936386108e-07, + "grad_norm": 2.6434102058410645, + "learning_rate": 2.342606149341142e-07, + "loss": 1.1653, + "mean_token_accuracy": 0.6660199761390686, + "num_tokens": 16620782.0, + "step": 641 + }, + { + "epoch": 0.07050296507797058, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 2.9149789810180664, + "learning_rate": 2.3462664714494877e-07, + "loss": 1.1728, + "mean_token_accuracy": 0.6551752686500549, + "num_tokens": 16648051.0, + "step": 642 + }, + { + "epoch": 0.07061278278058424, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 2.687584161758423, + "learning_rate": 2.349926793557833e-07, + "loss": 1.2095, + "mean_token_accuracy": 0.6578634977340698, + "num_tokens": 16677294.0, + "step": 643 + }, + { + "epoch": 0.0707226004831979, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 2.6386220455169678, + "learning_rate": 2.3535871156661784e-07, + "loss": 1.2419, + "mean_token_accuracy": 0.6461822390556335, + "num_tokens": 16707122.0, + "step": 644 + }, + { + "epoch": 0.07083241818581155, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 2.540253162384033, + "learning_rate": 2.3572474377745242e-07, + "loss": 1.1451, + "mean_token_accuracy": 0.6658681631088257, + "num_tokens": 16735039.0, + "step": 645 + }, + { + "epoch": 0.07094223588842521, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 3.1346333026885986, + "learning_rate": 2.3609077598828696e-07, + "loss": 1.0867, + "mean_token_accuracy": 0.6860034465789795, + "num_tokens": 16759433.0, + "step": 646 + }, + { + "epoch": 0.07105205359103887, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 2.6800239086151123, + "learning_rate": 2.3645680819912151e-07, + "loss": 1.0951, + "mean_token_accuracy": 0.6770033836364746, + "num_tokens": 16785559.0, + "step": 647 + }, + { + "epoch": 0.07116187129365253, + "ewc_loss": 1.4901161193847656e-07, + "grad_norm": 3.1487371921539307, + "learning_rate": 2.3682284040995606e-07, + "loss": 1.0795, + "mean_token_accuracy": 0.683826208114624, + "num_tokens": 16805432.0, + "step": 648 + }, + { + "epoch": 0.0712716889962662, + "ewc_loss": 1.4994293451309204e-07, + "grad_norm": 2.9379501342773438, + "learning_rate": 2.3718887262079064e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.679700493812561, + "num_tokens": 16828225.0, + "step": 649 + }, + { + "epoch": 0.07138150669887985, + "ewc_loss": 1.4994293451309204e-07, + "grad_norm": 2.9331765174865723, + "learning_rate": 2.375549048316252e-07, + "loss": 1.1397, + "mean_token_accuracy": 0.6758203506469727, + "num_tokens": 16853283.0, + "step": 650 + }, + { + "epoch": 0.07149132440149351, + "ewc_loss": 1.4994293451309204e-07, + "grad_norm": 2.6584153175354004, + "learning_rate": 2.379209370424597e-07, + "loss": 1.1569, + "mean_token_accuracy": 0.6655356884002686, + "num_tokens": 16880661.0, + "step": 651 + }, + { + "epoch": 0.07160114210410719, + "ewc_loss": 1.4994293451309204e-07, + "grad_norm": 2.823420763015747, + "learning_rate": 2.3828696925329429e-07, + "loss": 1.1954, + "mean_token_accuracy": 0.6548274755477905, + "num_tokens": 16905765.0, + "step": 652 + }, + { + "epoch": 0.07171095980672085, + "ewc_loss": 1.5087425708770752e-07, + "grad_norm": 3.3209593296051025, + "learning_rate": 2.3865300146412886e-07, + "loss": 1.129, + "mean_token_accuracy": 0.6732141971588135, + "num_tokens": 16926151.0, + "step": 653 + }, + { + "epoch": 0.07182077750933451, + "ewc_loss": 1.5087425708770752e-07, + "grad_norm": 2.5825107097625732, + "learning_rate": 2.390190336749634e-07, + "loss": 1.1505, + "mean_token_accuracy": 0.6674333810806274, + "num_tokens": 16955098.0, + "step": 654 + }, + { + "epoch": 0.07193059521194817, + "ewc_loss": 1.5087425708770752e-07, + "grad_norm": 3.284862518310547, + "learning_rate": 2.393850658857979e-07, + "loss": 1.0793, + "mean_token_accuracy": 0.6848053336143494, + "num_tokens": 16975049.0, + "step": 655 + }, + { + "epoch": 0.07204041291456183, + "ewc_loss": 1.5087425708770752e-07, + "grad_norm": 3.4881763458251953, + "learning_rate": 2.397510980966325e-07, + "loss": 1.0046, + "mean_token_accuracy": 0.7014285326004028, + "num_tokens": 16993119.0, + "step": 656 + }, + { + "epoch": 0.07215023061717549, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.90044903755188, + "learning_rate": 2.4011713030746706e-07, + "loss": 1.0827, + "mean_token_accuracy": 0.6827858686447144, + "num_tokens": 17016760.0, + "step": 657 + }, + { + "epoch": 0.07226004831978915, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.299220085144043, + "learning_rate": 2.404831625183016e-07, + "loss": 1.1439, + "mean_token_accuracy": 0.6645978689193726, + "num_tokens": 17046841.0, + "step": 658 + }, + { + "epoch": 0.0723698660224028, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.7012784481048584, + "learning_rate": 2.4084919472913616e-07, + "loss": 1.2295, + "mean_token_accuracy": 0.6481427550315857, + "num_tokens": 17073782.0, + "step": 659 + }, + { + "epoch": 0.07247968372501647, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.5405917167663574, + "learning_rate": 2.412152269399707e-07, + "loss": 1.1664, + "mean_token_accuracy": 0.6670801043510437, + "num_tokens": 17105514.0, + "step": 660 + }, + { + "epoch": 0.07258950142763014, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.892519474029541, + "learning_rate": 2.4158125915080525e-07, + "loss": 1.1094, + "mean_token_accuracy": 0.6753530502319336, + "num_tokens": 17128183.0, + "step": 661 + }, + { + "epoch": 0.0726993191302438, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 3.1064395904541016, + "learning_rate": 2.419472913616398e-07, + "loss": 1.1259, + "mean_token_accuracy": 0.6699349880218506, + "num_tokens": 17149575.0, + "step": 662 + }, + { + "epoch": 0.07280913683285746, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.772256374359131, + "learning_rate": 2.4231332357247435e-07, + "loss": 1.0799, + "mean_token_accuracy": 0.6869295835494995, + "num_tokens": 17175446.0, + "step": 663 + }, + { + "epoch": 0.07291895453547112, + "ewc_loss": 1.51805579662323e-07, + "grad_norm": 2.5445547103881836, + "learning_rate": 2.426793557833089e-07, + "loss": 1.0795, + "mean_token_accuracy": 0.6830497980117798, + "num_tokens": 17206437.0, + "step": 664 + }, + { + "epoch": 0.07302877223808478, + "ewc_loss": 1.5273690223693848e-07, + "grad_norm": 2.659824848175049, + "learning_rate": 2.430453879941435e-07, + "loss": 1.1819, + "mean_token_accuracy": 0.6527514457702637, + "num_tokens": 17232640.0, + "step": 665 + }, + { + "epoch": 0.07313858994069844, + "ewc_loss": 1.5273690223693848e-07, + "grad_norm": 2.6496050357818604, + "learning_rate": 2.4341142020497805e-07, + "loss": 1.1701, + "mean_token_accuracy": 0.6622365713119507, + "num_tokens": 17260421.0, + "step": 666 + }, + { + "epoch": 0.0732484076433121, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 2.9083313941955566, + "learning_rate": 2.4377745241581255e-07, + "loss": 0.9737, + "mean_token_accuracy": 0.7122673988342285, + "num_tokens": 17283253.0, + "step": 667 + }, + { + "epoch": 0.07335822534592576, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 3.1165883541107178, + "learning_rate": 2.4414348462664715e-07, + "loss": 1.1338, + "mean_token_accuracy": 0.663633406162262, + "num_tokens": 17303932.0, + "step": 668 + }, + { + "epoch": 0.07346804304853942, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 2.9897279739379883, + "learning_rate": 2.445095168374817e-07, + "loss": 1.0161, + "mean_token_accuracy": 0.6991400122642517, + "num_tokens": 17324634.0, + "step": 669 + }, + { + "epoch": 0.07357786075115308, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 2.745620012283325, + "learning_rate": 2.4487554904831625e-07, + "loss": 0.9912, + "mean_token_accuracy": 0.7049219608306885, + "num_tokens": 17347927.0, + "step": 670 + }, + { + "epoch": 0.07368767845376675, + "ewc_loss": 1.5459954738616943e-07, + "grad_norm": 2.735863447189331, + "learning_rate": 2.452415812591508e-07, + "loss": 1.0491, + "mean_token_accuracy": 0.6952067613601685, + "num_tokens": 17374596.0, + "step": 671 + }, + { + "epoch": 0.07379749615638041, + "ewc_loss": 1.555308699607849e-07, + "grad_norm": 3.1521685123443604, + "learning_rate": 2.4560761346998535e-07, + "loss": 1.0401, + "mean_token_accuracy": 0.6854103803634644, + "num_tokens": 17397546.0, + "step": 672 + }, + { + "epoch": 0.07390731385899407, + "ewc_loss": 1.555308699607849e-07, + "grad_norm": 2.9903273582458496, + "learning_rate": 2.459736456808199e-07, + "loss": 1.2238, + "mean_token_accuracy": 0.6551464796066284, + "num_tokens": 17420746.0, + "step": 673 + }, + { + "epoch": 0.07401713156160773, + "ewc_loss": 1.555308699607849e-07, + "grad_norm": 3.1167569160461426, + "learning_rate": 2.4633967789165444e-07, + "loss": 1.0766, + "mean_token_accuracy": 0.6918143033981323, + "num_tokens": 17441694.0, + "step": 674 + }, + { + "epoch": 0.07412694926422139, + "ewc_loss": 1.564621925354004e-07, + "grad_norm": 2.516944169998169, + "learning_rate": 2.46705710102489e-07, + "loss": 1.1537, + "mean_token_accuracy": 0.6698046326637268, + "num_tokens": 17468898.0, + "step": 675 + }, + { + "epoch": 0.07423676696683505, + "ewc_loss": 1.564621925354004e-07, + "grad_norm": 3.0917983055114746, + "learning_rate": 2.4707174231332354e-07, + "loss": 1.0356, + "mean_token_accuracy": 0.6937268972396851, + "num_tokens": 17488143.0, + "step": 676 + }, + { + "epoch": 0.07434658466944871, + "ewc_loss": 1.5739351511001587e-07, + "grad_norm": 2.7008864879608154, + "learning_rate": 2.4743777452415815e-07, + "loss": 1.0558, + "mean_token_accuracy": 0.6897485256195068, + "num_tokens": 17514042.0, + "step": 677 + }, + { + "epoch": 0.07445640237206237, + "ewc_loss": 1.5739351511001587e-07, + "grad_norm": 2.816481351852417, + "learning_rate": 2.478038067349927e-07, + "loss": 1.0434, + "mean_token_accuracy": 0.6898455619812012, + "num_tokens": 17538766.0, + "step": 678 + }, + { + "epoch": 0.07456622007467603, + "ewc_loss": 1.5832483768463135e-07, + "grad_norm": 2.451542615890503, + "learning_rate": 2.481698389458272e-07, + "loss": 1.1587, + "mean_token_accuracy": 0.6802065968513489, + "num_tokens": 17567824.0, + "step": 679 + }, + { + "epoch": 0.0746760377772897, + "ewc_loss": 1.5832483768463135e-07, + "grad_norm": 2.4214398860931396, + "learning_rate": 2.485358711566618e-07, + "loss": 1.1821, + "mean_token_accuracy": 0.6600780487060547, + "num_tokens": 17599355.0, + "step": 680 + }, + { + "epoch": 0.07478585547990337, + "ewc_loss": 1.5832483768463135e-07, + "grad_norm": 2.571739912033081, + "learning_rate": 2.4890190336749634e-07, + "loss": 1.1535, + "mean_token_accuracy": 0.6641680598258972, + "num_tokens": 17625251.0, + "step": 681 + }, + { + "epoch": 0.07489567318251703, + "ewc_loss": 1.5832483768463135e-07, + "grad_norm": 2.5234854221343994, + "learning_rate": 2.492679355783309e-07, + "loss": 1.1515, + "mean_token_accuracy": 0.6638996601104736, + "num_tokens": 17654107.0, + "step": 682 + }, + { + "epoch": 0.07500549088513069, + "ewc_loss": 1.601874828338623e-07, + "grad_norm": 2.9947903156280518, + "learning_rate": 2.4963396778916544e-07, + "loss": 1.0533, + "mean_token_accuracy": 0.694914698600769, + "num_tokens": 17677403.0, + "step": 683 + }, + { + "epoch": 0.07511530858774434, + "ewc_loss": 1.601874828338623e-07, + "grad_norm": 2.5261566638946533, + "learning_rate": 2.5e-07, + "loss": 1.1711, + "mean_token_accuracy": 0.6636015772819519, + "num_tokens": 17706274.0, + "step": 684 + }, + { + "epoch": 0.075225126290358, + "ewc_loss": 1.6205012798309326e-07, + "grad_norm": 2.681138277053833, + "learning_rate": 2.5036603221083454e-07, + "loss": 1.1437, + "mean_token_accuracy": 0.6580157279968262, + "num_tokens": 17732923.0, + "step": 685 + }, + { + "epoch": 0.07533494399297166, + "ewc_loss": 1.6205012798309326e-07, + "grad_norm": 2.650453805923462, + "learning_rate": 2.507320644216691e-07, + "loss": 1.0878, + "mean_token_accuracy": 0.6858391761779785, + "num_tokens": 17758057.0, + "step": 686 + }, + { + "epoch": 0.07544476169558532, + "ewc_loss": 1.6205012798309326e-07, + "grad_norm": 2.4743971824645996, + "learning_rate": 2.5109809663250364e-07, + "loss": 1.0876, + "mean_token_accuracy": 0.6796285510063171, + "num_tokens": 17787037.0, + "step": 687 + }, + { + "epoch": 0.07555457939819898, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 3.2741634845733643, + "learning_rate": 2.514641288433382e-07, + "loss": 1.1438, + "mean_token_accuracy": 0.6720909476280212, + "num_tokens": 17807531.0, + "step": 688 + }, + { + "epoch": 0.07566439710081264, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 3.5065770149230957, + "learning_rate": 2.518301610541728e-07, + "loss": 1.0343, + "mean_token_accuracy": 0.6903947591781616, + "num_tokens": 17825910.0, + "step": 689 + }, + { + "epoch": 0.07577421480342632, + "ewc_loss": 1.6391277313232422e-07, + "grad_norm": 3.0358078479766846, + "learning_rate": 2.521961932650073e-07, + "loss": 1.1923, + "mean_token_accuracy": 0.650958240032196, + "num_tokens": 17849674.0, + "step": 690 + }, + { + "epoch": 0.07588403250603998, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 2.3469979763031006, + "learning_rate": 2.525622254758419e-07, + "loss": 1.2063, + "mean_token_accuracy": 0.6604977250099182, + "num_tokens": 17884281.0, + "step": 691 + }, + { + "epoch": 0.07599385020865364, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 2.7087230682373047, + "learning_rate": 2.5292825768667643e-07, + "loss": 1.122, + "mean_token_accuracy": 0.6758402585983276, + "num_tokens": 17907327.0, + "step": 692 + }, + { + "epoch": 0.0761036679112673, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 2.655954599380493, + "learning_rate": 2.5329428989751093e-07, + "loss": 1.1176, + "mean_token_accuracy": 0.677223801612854, + "num_tokens": 17931212.0, + "step": 693 + }, + { + "epoch": 0.07621348561388096, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 2.8443045616149902, + "learning_rate": 2.5366032210834553e-07, + "loss": 1.1816, + "mean_token_accuracy": 0.658890426158905, + "num_tokens": 17957452.0, + "step": 694 + }, + { + "epoch": 0.07632330331649462, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 3.587148904800415, + "learning_rate": 2.540263543191801e-07, + "loss": 1.0586, + "mean_token_accuracy": 0.6827383637428284, + "num_tokens": 17974591.0, + "step": 695 + }, + { + "epoch": 0.07643312101910828, + "ewc_loss": 1.6298145055770874e-07, + "grad_norm": 2.7852165699005127, + "learning_rate": 2.543923865300146e-07, + "loss": 1.1039, + "mean_token_accuracy": 0.6853533983230591, + "num_tokens": 18002528.0, + "step": 696 + }, + { + "epoch": 0.07654293872172194, + "ewc_loss": 1.6391277313232422e-07, + "grad_norm": 2.8456547260284424, + "learning_rate": 2.547584187408492e-07, + "loss": 1.1215, + "mean_token_accuracy": 0.6691709756851196, + "num_tokens": 18027367.0, + "step": 697 + }, + { + "epoch": 0.0766527564243356, + "ewc_loss": 1.6391277313232422e-07, + "grad_norm": 2.7023096084594727, + "learning_rate": 2.5512445095168373e-07, + "loss": 1.1525, + "mean_token_accuracy": 0.6629804372787476, + "num_tokens": 18052873.0, + "step": 698 + }, + { + "epoch": 0.07676257412694927, + "ewc_loss": 1.6391277313232422e-07, + "grad_norm": 2.458958148956299, + "learning_rate": 2.5549048316251833e-07, + "loss": 1.1394, + "mean_token_accuracy": 0.6659390330314636, + "num_tokens": 18086772.0, + "step": 699 + }, + { + "epoch": 0.07687239182956293, + "ewc_loss": 1.6391277313232422e-07, + "grad_norm": 2.589984655380249, + "learning_rate": 2.5585651537335283e-07, + "loss": 1.1758, + "mean_token_accuracy": 0.6595163941383362, + "num_tokens": 18115267.0, + "step": 700 + }, + { + "epoch": 0.07698220953217659, + "ewc_loss": 1.6577541828155518e-07, + "grad_norm": 2.7382142543792725, + "learning_rate": 2.562225475841874e-07, + "loss": 1.1313, + "mean_token_accuracy": 0.6707941889762878, + "num_tokens": 18144689.0, + "step": 701 + }, + { + "epoch": 0.07709202723479025, + "ewc_loss": 1.6577541828155518e-07, + "grad_norm": 2.987204074859619, + "learning_rate": 2.56588579795022e-07, + "loss": 1.1423, + "mean_token_accuracy": 0.6741464138031006, + "num_tokens": 18168911.0, + "step": 702 + }, + { + "epoch": 0.07720184493740391, + "ewc_loss": 1.6577541828155518e-07, + "grad_norm": 2.979400634765625, + "learning_rate": 2.569546120058565e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.676014244556427, + "num_tokens": 18193758.0, + "step": 703 + }, + { + "epoch": 0.07731166264001757, + "ewc_loss": 1.6577541828155518e-07, + "grad_norm": 2.850217342376709, + "learning_rate": 2.57320644216691e-07, + "loss": 1.1017, + "mean_token_accuracy": 0.6741418242454529, + "num_tokens": 18216265.0, + "step": 704 + }, + { + "epoch": 0.07742148034263123, + "ewc_loss": 1.6577541828155518e-07, + "grad_norm": 2.6114501953125, + "learning_rate": 2.576866764275256e-07, + "loss": 1.0711, + "mean_token_accuracy": 0.6824700236320496, + "num_tokens": 18240261.0, + "step": 705 + }, + { + "epoch": 0.07753129804524489, + "ewc_loss": 1.6763806343078613e-07, + "grad_norm": 2.5393881797790527, + "learning_rate": 2.580527086383602e-07, + "loss": 1.2074, + "mean_token_accuracy": 0.6508747339248657, + "num_tokens": 18267819.0, + "step": 706 + }, + { + "epoch": 0.07764111574785855, + "ewc_loss": 1.6763806343078613e-07, + "grad_norm": 2.633901596069336, + "learning_rate": 2.584187408491947e-07, + "loss": 1.1139, + "mean_token_accuracy": 0.6789953708648682, + "num_tokens": 18295498.0, + "step": 707 + }, + { + "epoch": 0.07775093345047221, + "ewc_loss": 1.685693860054016e-07, + "grad_norm": 2.4847307205200195, + "learning_rate": 2.5878477306002927e-07, + "loss": 1.1474, + "mean_token_accuracy": 0.6614297032356262, + "num_tokens": 18324344.0, + "step": 708 + }, + { + "epoch": 0.07786075115308588, + "ewc_loss": 1.685693860054016e-07, + "grad_norm": 3.122530698776245, + "learning_rate": 2.591508052708638e-07, + "loss": 1.1141, + "mean_token_accuracy": 0.6753584146499634, + "num_tokens": 18345249.0, + "step": 709 + }, + { + "epoch": 0.07797056885569954, + "ewc_loss": 1.695007085800171e-07, + "grad_norm": 2.7253870964050293, + "learning_rate": 2.5951683748169837e-07, + "loss": 1.2175, + "mean_token_accuracy": 0.6525812149047852, + "num_tokens": 18369976.0, + "step": 710 + }, + { + "epoch": 0.0780803865583132, + "ewc_loss": 1.695007085800171e-07, + "grad_norm": 2.8930602073669434, + "learning_rate": 2.598828696925329e-07, + "loss": 1.0488, + "mean_token_accuracy": 0.6888777017593384, + "num_tokens": 18389828.0, + "step": 711 + }, + { + "epoch": 0.07819020426092686, + "ewc_loss": 1.695007085800171e-07, + "grad_norm": 2.455997943878174, + "learning_rate": 2.6024890190336747e-07, + "loss": 1.1244, + "mean_token_accuracy": 0.6793985962867737, + "num_tokens": 18418228.0, + "step": 712 + }, + { + "epoch": 0.07830002196354052, + "ewc_loss": 1.695007085800171e-07, + "grad_norm": 2.701829195022583, + "learning_rate": 2.6061493411420207e-07, + "loss": 1.1239, + "mean_token_accuracy": 0.6635980606079102, + "num_tokens": 18442229.0, + "step": 713 + }, + { + "epoch": 0.07840983966615418, + "ewc_loss": 1.7043203115463257e-07, + "grad_norm": 2.520317316055298, + "learning_rate": 2.6098096632503657e-07, + "loss": 1.1547, + "mean_token_accuracy": 0.6648585796356201, + "num_tokens": 18469419.0, + "step": 714 + }, + { + "epoch": 0.07851965736876784, + "ewc_loss": 1.7043203115463257e-07, + "grad_norm": 2.592540740966797, + "learning_rate": 2.6134699853587117e-07, + "loss": 0.9564, + "mean_token_accuracy": 0.7130157947540283, + "num_tokens": 18493762.0, + "step": 715 + }, + { + "epoch": 0.0786294750713815, + "ewc_loss": 1.7136335372924805e-07, + "grad_norm": 3.1164658069610596, + "learning_rate": 2.617130307467057e-07, + "loss": 1.1223, + "mean_token_accuracy": 0.6712024211883545, + "num_tokens": 18516052.0, + "step": 716 + }, + { + "epoch": 0.07873929277399516, + "ewc_loss": 1.7136335372924805e-07, + "grad_norm": 2.4279356002807617, + "learning_rate": 2.620790629575402e-07, + "loss": 1.1112, + "mean_token_accuracy": 0.6804686784744263, + "num_tokens": 18542974.0, + "step": 717 + }, + { + "epoch": 0.07884911047660884, + "ewc_loss": 1.7136335372924805e-07, + "grad_norm": 3.144777774810791, + "learning_rate": 2.624450951683748e-07, + "loss": 1.0487, + "mean_token_accuracy": 0.6864203214645386, + "num_tokens": 18561988.0, + "step": 718 + }, + { + "epoch": 0.0789589281792225, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 2.652195930480957, + "learning_rate": 2.6281112737920937e-07, + "loss": 1.1678, + "mean_token_accuracy": 0.6630251407623291, + "num_tokens": 18588546.0, + "step": 719 + }, + { + "epoch": 0.07906874588183616, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 2.6898605823516846, + "learning_rate": 2.6317715959004386e-07, + "loss": 1.195, + "mean_token_accuracy": 0.6530371308326721, + "num_tokens": 18615106.0, + "step": 720 + }, + { + "epoch": 0.07917856358444982, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 2.419036388397217, + "learning_rate": 2.6354319180087846e-07, + "loss": 1.111, + "mean_token_accuracy": 0.6759470701217651, + "num_tokens": 18642433.0, + "step": 721 + }, + { + "epoch": 0.07928838128706348, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 2.4187123775482178, + "learning_rate": 2.63909224011713e-07, + "loss": 1.1058, + "mean_token_accuracy": 0.6736469864845276, + "num_tokens": 18670567.0, + "step": 722 + }, + { + "epoch": 0.07939819898967714, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 3.059765100479126, + "learning_rate": 2.642752562225476e-07, + "loss": 1.1157, + "mean_token_accuracy": 0.6712362170219421, + "num_tokens": 18691814.0, + "step": 723 + }, + { + "epoch": 0.0795080166922908, + "ewc_loss": 1.7229467630386353e-07, + "grad_norm": 2.5231969356536865, + "learning_rate": 2.646412884333821e-07, + "loss": 1.1509, + "mean_token_accuracy": 0.6662156581878662, + "num_tokens": 18717182.0, + "step": 724 + }, + { + "epoch": 0.07961783439490445, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.3908026218414307, + "learning_rate": 2.6500732064421666e-07, + "loss": 1.1307, + "mean_token_accuracy": 0.6651685833930969, + "num_tokens": 18746831.0, + "step": 725 + }, + { + "epoch": 0.07972765209751811, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.413386821746826, + "learning_rate": 2.6537335285505126e-07, + "loss": 1.072, + "mean_token_accuracy": 0.6837024688720703, + "num_tokens": 18777233.0, + "step": 726 + }, + { + "epoch": 0.07983746980013177, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.6925876140594482, + "learning_rate": 2.6573938506588576e-07, + "loss": 1.1228, + "mean_token_accuracy": 0.6767383217811584, + "num_tokens": 18802479.0, + "step": 727 + }, + { + "epoch": 0.07994728750274545, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.343503952026367, + "learning_rate": 2.661054172767203e-07, + "loss": 1.1458, + "mean_token_accuracy": 0.662580668926239, + "num_tokens": 18831630.0, + "step": 728 + }, + { + "epoch": 0.08005710520535911, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.9535133838653564, + "learning_rate": 2.664714494875549e-07, + "loss": 0.9646, + "mean_token_accuracy": 0.7155031561851501, + "num_tokens": 18853499.0, + "step": 729 + }, + { + "epoch": 0.08016692290797277, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.7766056060791016, + "learning_rate": 2.6683748169838946e-07, + "loss": 1.1806, + "mean_token_accuracy": 0.6645959615707397, + "num_tokens": 18875616.0, + "step": 730 + }, + { + "epoch": 0.08027674061058643, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.9452662467956543, + "learning_rate": 2.67203513909224e-07, + "loss": 1.1073, + "mean_token_accuracy": 0.6783711314201355, + "num_tokens": 18895943.0, + "step": 731 + }, + { + "epoch": 0.08038655831320009, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.1604835987091064, + "learning_rate": 2.6756954612005856e-07, + "loss": 1.1632, + "mean_token_accuracy": 0.6631931662559509, + "num_tokens": 18930183.0, + "step": 732 + }, + { + "epoch": 0.08049637601581375, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.798745632171631, + "learning_rate": 2.679355783308931e-07, + "loss": 1.1611, + "mean_token_accuracy": 0.6644535064697266, + "num_tokens": 18953193.0, + "step": 733 + }, + { + "epoch": 0.08060619371842741, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.4439854621887207, + "learning_rate": 2.6830161054172765e-07, + "loss": 1.1927, + "mean_token_accuracy": 0.6578404903411865, + "num_tokens": 18981251.0, + "step": 734 + }, + { + "epoch": 0.08071601142104107, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.492180824279785, + "learning_rate": 2.686676427525622e-07, + "loss": 1.1424, + "mean_token_accuracy": 0.6769749522209167, + "num_tokens": 19009491.0, + "step": 735 + }, + { + "epoch": 0.08082582912365473, + "ewc_loss": 1.73225998878479e-07, + "grad_norm": 2.4412262439727783, + "learning_rate": 2.6903367496339675e-07, + "loss": 1.0732, + "mean_token_accuracy": 0.6854243278503418, + "num_tokens": 19037182.0, + "step": 736 + }, + { + "epoch": 0.0809356468262684, + "ewc_loss": 1.7415732145309448e-07, + "grad_norm": 2.871915102005005, + "learning_rate": 2.693997071742313e-07, + "loss": 1.0853, + "mean_token_accuracy": 0.6815794110298157, + "num_tokens": 19059901.0, + "step": 737 + }, + { + "epoch": 0.08104546452888206, + "ewc_loss": 1.7508864402770996e-07, + "grad_norm": 2.621892213821411, + "learning_rate": 2.6976573938506585e-07, + "loss": 1.1687, + "mean_token_accuracy": 0.6621853709220886, + "num_tokens": 19085117.0, + "step": 738 + }, + { + "epoch": 0.08115528223149572, + "ewc_loss": 1.7415732145309448e-07, + "grad_norm": 2.54357647895813, + "learning_rate": 2.7013177159590045e-07, + "loss": 1.0758, + "mean_token_accuracy": 0.6817019581794739, + "num_tokens": 19111093.0, + "step": 739 + }, + { + "epoch": 0.08126509993410938, + "ewc_loss": 1.7508864402770996e-07, + "grad_norm": 2.698091506958008, + "learning_rate": 2.70497803806735e-07, + "loss": 1.0896, + "mean_token_accuracy": 0.6760367751121521, + "num_tokens": 19135550.0, + "step": 740 + }, + { + "epoch": 0.08137491763672304, + "ewc_loss": 1.7508864402770996e-07, + "grad_norm": 2.5144906044006348, + "learning_rate": 2.708638360175695e-07, + "loss": 1.0616, + "mean_token_accuracy": 0.6852490305900574, + "num_tokens": 19162633.0, + "step": 741 + }, + { + "epoch": 0.0814847353393367, + "ewc_loss": 1.7601996660232544e-07, + "grad_norm": 2.518202781677246, + "learning_rate": 2.712298682284041e-07, + "loss": 1.054, + "mean_token_accuracy": 0.6967436075210571, + "num_tokens": 19186813.0, + "step": 742 + }, + { + "epoch": 0.08159455304195036, + "ewc_loss": 1.7601996660232544e-07, + "grad_norm": 2.349273681640625, + "learning_rate": 2.7159590043923865e-07, + "loss": 1.1024, + "mean_token_accuracy": 0.6741527318954468, + "num_tokens": 19217664.0, + "step": 743 + }, + { + "epoch": 0.08170437074456402, + "ewc_loss": 1.7695128917694092e-07, + "grad_norm": 2.946650981903076, + "learning_rate": 2.7196193265007315e-07, + "loss": 1.0819, + "mean_token_accuracy": 0.6768238544464111, + "num_tokens": 19239127.0, + "step": 744 + }, + { + "epoch": 0.08181418844717768, + "ewc_loss": 1.778826117515564e-07, + "grad_norm": 2.690995931625366, + "learning_rate": 2.7232796486090775e-07, + "loss": 1.1032, + "mean_token_accuracy": 0.6784749031066895, + "num_tokens": 19263028.0, + "step": 745 + }, + { + "epoch": 0.08192400614979134, + "ewc_loss": 1.7881393432617188e-07, + "grad_norm": 2.3555636405944824, + "learning_rate": 2.726939970717423e-07, + "loss": 1.1843, + "mean_token_accuracy": 0.6530269384384155, + "num_tokens": 19292808.0, + "step": 746 + }, + { + "epoch": 0.08203382385240501, + "ewc_loss": 1.7881393432617188e-07, + "grad_norm": 2.6926839351654053, + "learning_rate": 2.730600292825769e-07, + "loss": 1.1159, + "mean_token_accuracy": 0.6743430495262146, + "num_tokens": 19314443.0, + "step": 747 + }, + { + "epoch": 0.08214364155501867, + "ewc_loss": 1.7974525690078735e-07, + "grad_norm": 2.642151355743408, + "learning_rate": 2.734260614934114e-07, + "loss": 1.0643, + "mean_token_accuracy": 0.6901077628135681, + "num_tokens": 19339945.0, + "step": 748 + }, + { + "epoch": 0.08225345925763233, + "ewc_loss": 1.7974525690078735e-07, + "grad_norm": 2.3964121341705322, + "learning_rate": 2.7379209370424594e-07, + "loss": 1.1836, + "mean_token_accuracy": 0.6590663194656372, + "num_tokens": 19370487.0, + "step": 749 + }, + { + "epoch": 0.08236327696024599, + "ewc_loss": 1.7974525690078735e-07, + "grad_norm": 2.4788835048675537, + "learning_rate": 2.7415812591508055e-07, + "loss": 1.0845, + "mean_token_accuracy": 0.6805879473686218, + "num_tokens": 19395644.0, + "step": 750 + }, + { + "epoch": 0.08247309466285965, + "ewc_loss": 1.7974525690078735e-07, + "grad_norm": 2.2289018630981445, + "learning_rate": 2.7452415812591504e-07, + "loss": 1.1652, + "mean_token_accuracy": 0.6598275303840637, + "num_tokens": 19426218.0, + "step": 751 + }, + { + "epoch": 0.08258291236547331, + "ewc_loss": 1.7974525690078735e-07, + "grad_norm": 2.540152072906494, + "learning_rate": 2.748901903367496e-07, + "loss": 1.1626, + "mean_token_accuracy": 0.666319727897644, + "num_tokens": 19451846.0, + "step": 752 + }, + { + "epoch": 0.08269273006808697, + "ewc_loss": 1.816079020500183e-07, + "grad_norm": 2.58980655670166, + "learning_rate": 2.752562225475842e-07, + "loss": 1.0854, + "mean_token_accuracy": 0.6864382028579712, + "num_tokens": 19476296.0, + "step": 753 + }, + { + "epoch": 0.08280254777070063, + "ewc_loss": 1.816079020500183e-07, + "grad_norm": 2.430241346359253, + "learning_rate": 2.756222547584187e-07, + "loss": 1.1251, + "mean_token_accuracy": 0.6789584159851074, + "num_tokens": 19504200.0, + "step": 754 + }, + { + "epoch": 0.08291236547331429, + "ewc_loss": 1.816079020500183e-07, + "grad_norm": 2.7521989345550537, + "learning_rate": 2.759882869692533e-07, + "loss": 1.2028, + "mean_token_accuracy": 0.6557236909866333, + "num_tokens": 19527134.0, + "step": 755 + }, + { + "epoch": 0.08302218317592797, + "ewc_loss": 1.825392246246338e-07, + "grad_norm": 2.4392471313476562, + "learning_rate": 2.7635431918008784e-07, + "loss": 1.0866, + "mean_token_accuracy": 0.6845049262046814, + "num_tokens": 19556429.0, + "step": 756 + }, + { + "epoch": 0.08313200087854163, + "ewc_loss": 1.8347054719924927e-07, + "grad_norm": 2.1911425590515137, + "learning_rate": 2.767203513909224e-07, + "loss": 1.1595, + "mean_token_accuracy": 0.6618449687957764, + "num_tokens": 19587759.0, + "step": 757 + }, + { + "epoch": 0.08324181858115529, + "ewc_loss": 1.8347054719924927e-07, + "grad_norm": 3.101276397705078, + "learning_rate": 2.7708638360175694e-07, + "loss": 1.1092, + "mean_token_accuracy": 0.6820889115333557, + "num_tokens": 19607412.0, + "step": 758 + }, + { + "epoch": 0.08335163628376895, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.6694414615631104, + "learning_rate": 2.774524158125915e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.7105387449264526, + "num_tokens": 19630184.0, + "step": 759 + }, + { + "epoch": 0.0834614539863826, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.6369102001190186, + "learning_rate": 2.7781844802342604e-07, + "loss": 1.0871, + "mean_token_accuracy": 0.6825014352798462, + "num_tokens": 19655787.0, + "step": 760 + }, + { + "epoch": 0.08357127168899627, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.416717290878296, + "learning_rate": 2.781844802342606e-07, + "loss": 1.0524, + "mean_token_accuracy": 0.6927522420883179, + "num_tokens": 19683813.0, + "step": 761 + }, + { + "epoch": 0.08368108939160993, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.730959892272949, + "learning_rate": 2.7855051244509513e-07, + "loss": 1.0805, + "mean_token_accuracy": 0.6813210248947144, + "num_tokens": 19706512.0, + "step": 762 + }, + { + "epoch": 0.08379090709422359, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.359360694885254, + "learning_rate": 2.7891654465592974e-07, + "loss": 1.1501, + "mean_token_accuracy": 0.6707624793052673, + "num_tokens": 19737097.0, + "step": 763 + }, + { + "epoch": 0.08390072479683724, + "ewc_loss": 1.8440186977386475e-07, + "grad_norm": 2.532769203186035, + "learning_rate": 2.792825768667643e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.7026410698890686, + "num_tokens": 19762815.0, + "step": 764 + }, + { + "epoch": 0.0840105424994509, + "ewc_loss": 1.8533319234848022e-07, + "grad_norm": 2.8030121326446533, + "learning_rate": 2.796486090775988e-07, + "loss": 1.1177, + "mean_token_accuracy": 0.6752446889877319, + "num_tokens": 19785444.0, + "step": 765 + }, + { + "epoch": 0.08412036020206458, + "ewc_loss": 1.8533319234848022e-07, + "grad_norm": 2.6845951080322266, + "learning_rate": 2.800146412884334e-07, + "loss": 1.1253, + "mean_token_accuracy": 0.6716697812080383, + "num_tokens": 19809678.0, + "step": 766 + }, + { + "epoch": 0.08423017790467824, + "ewc_loss": 1.8533319234848022e-07, + "grad_norm": 2.497102737426758, + "learning_rate": 2.8038067349926793e-07, + "loss": 1.0991, + "mean_token_accuracy": 0.6743587851524353, + "num_tokens": 19834658.0, + "step": 767 + }, + { + "epoch": 0.0843399956072919, + "ewc_loss": 1.862645149230957e-07, + "grad_norm": 2.4419264793395996, + "learning_rate": 2.8074670571010243e-07, + "loss": 1.185, + "mean_token_accuracy": 0.6609646081924438, + "num_tokens": 19860076.0, + "step": 768 + }, + { + "epoch": 0.08444981330990556, + "ewc_loss": 1.862645149230957e-07, + "grad_norm": 2.692838430404663, + "learning_rate": 2.8111273792093703e-07, + "loss": 1.1399, + "mean_token_accuracy": 0.6686279773712158, + "num_tokens": 19885069.0, + "step": 769 + }, + { + "epoch": 0.08455963101251922, + "ewc_loss": 1.862645149230957e-07, + "grad_norm": 2.6871848106384277, + "learning_rate": 2.814787701317716e-07, + "loss": 1.1999, + "mean_token_accuracy": 0.660123348236084, + "num_tokens": 19909921.0, + "step": 770 + }, + { + "epoch": 0.08466944871513288, + "ewc_loss": 1.862645149230957e-07, + "grad_norm": 2.478097915649414, + "learning_rate": 2.818448023426062e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6979023218154907, + "num_tokens": 19935788.0, + "step": 771 + }, + { + "epoch": 0.08477926641774654, + "ewc_loss": 1.8719583749771118e-07, + "grad_norm": 2.4230847358703613, + "learning_rate": 2.822108345534407e-07, + "loss": 1.1328, + "mean_token_accuracy": 0.667827844619751, + "num_tokens": 19965350.0, + "step": 772 + }, + { + "epoch": 0.0848890841203602, + "ewc_loss": 1.8719583749771118e-07, + "grad_norm": 2.6704137325286865, + "learning_rate": 2.8257686676427523e-07, + "loss": 1.0059, + "mean_token_accuracy": 0.7048922777175903, + "num_tokens": 19985991.0, + "step": 773 + }, + { + "epoch": 0.08499890182297386, + "ewc_loss": 1.8719583749771118e-07, + "grad_norm": 2.4515786170959473, + "learning_rate": 2.8294289897510983e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6831092834472656, + "num_tokens": 20013789.0, + "step": 774 + }, + { + "epoch": 0.08510871952558753, + "ewc_loss": 1.8998980522155762e-07, + "grad_norm": 2.7693867683410645, + "learning_rate": 2.833089311859443e-07, + "loss": 1.0242, + "mean_token_accuracy": 0.6923285722732544, + "num_tokens": 20034001.0, + "step": 775 + }, + { + "epoch": 0.08521853722820119, + "ewc_loss": 1.9185245037078857e-07, + "grad_norm": 2.6186351776123047, + "learning_rate": 2.836749633967789e-07, + "loss": 1.1448, + "mean_token_accuracy": 0.6641409397125244, + "num_tokens": 20060222.0, + "step": 776 + }, + { + "epoch": 0.08532835493081485, + "ewc_loss": 1.9278377294540405e-07, + "grad_norm": 2.361137866973877, + "learning_rate": 2.840409956076135e-07, + "loss": 1.0791, + "mean_token_accuracy": 0.6841327548027039, + "num_tokens": 20093804.0, + "step": 777 + }, + { + "epoch": 0.08543817263342851, + "ewc_loss": 1.94646418094635e-07, + "grad_norm": 2.2466769218444824, + "learning_rate": 2.8440702781844797e-07, + "loss": 1.2166, + "mean_token_accuracy": 0.6423548460006714, + "num_tokens": 20126639.0, + "step": 778 + }, + { + "epoch": 0.08554799033604217, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.4807655811309814, + "learning_rate": 2.847730600292826e-07, + "loss": 1.0144, + "mean_token_accuracy": 0.7049177289009094, + "num_tokens": 20154205.0, + "step": 779 + }, + { + "epoch": 0.08565780803865583, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.2778968811035156, + "learning_rate": 2.851390922401171e-07, + "loss": 1.2067, + "mean_token_accuracy": 0.6646120548248291, + "num_tokens": 20185261.0, + "step": 780 + }, + { + "epoch": 0.08576762574126949, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.837930679321289, + "learning_rate": 2.8550512445095167e-07, + "loss": 1.0085, + "mean_token_accuracy": 0.6950445771217346, + "num_tokens": 20205777.0, + "step": 781 + }, + { + "epoch": 0.08587744344388315, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.2587413787841797, + "learning_rate": 2.858711566617862e-07, + "loss": 1.0526, + "mean_token_accuracy": 0.6892855167388916, + "num_tokens": 20237386.0, + "step": 782 + }, + { + "epoch": 0.08598726114649681, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.567570924758911, + "learning_rate": 2.8623718887262077e-07, + "loss": 1.1107, + "mean_token_accuracy": 0.6774088144302368, + "num_tokens": 20261696.0, + "step": 783 + }, + { + "epoch": 0.08609707884911047, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.464510440826416, + "learning_rate": 2.866032210834553e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6807329654693604, + "num_tokens": 20287351.0, + "step": 784 + }, + { + "epoch": 0.08620689655172414, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.636361598968506, + "learning_rate": 2.8696925329428987e-07, + "loss": 0.9998, + "mean_token_accuracy": 0.7027122974395752, + "num_tokens": 20309391.0, + "step": 785 + }, + { + "epoch": 0.0863167142543378, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.695484161376953, + "learning_rate": 2.873352855051244e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6821368336677551, + "num_tokens": 20332515.0, + "step": 786 + }, + { + "epoch": 0.08642653195695146, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.5393567085266113, + "learning_rate": 2.87701317715959e-07, + "loss": 1.0967, + "mean_token_accuracy": 0.6863274574279785, + "num_tokens": 20359379.0, + "step": 787 + }, + { + "epoch": 0.08653634965956512, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.6482996940612793, + "learning_rate": 2.8806734992679357e-07, + "loss": 1.1269, + "mean_token_accuracy": 0.6667490601539612, + "num_tokens": 20383436.0, + "step": 788 + }, + { + "epoch": 0.08664616736217878, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.227149248123169, + "learning_rate": 2.8843338213762807e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.6763156652450562, + "num_tokens": 20414443.0, + "step": 789 + }, + { + "epoch": 0.08675598506479244, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.7543866634368896, + "learning_rate": 2.8879941434846267e-07, + "loss": 1.0642, + "mean_token_accuracy": 0.6814924478530884, + "num_tokens": 20437080.0, + "step": 790 + }, + { + "epoch": 0.0868658027674061, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.586895227432251, + "learning_rate": 2.891654465592972e-07, + "loss": 0.9616, + "mean_token_accuracy": 0.7151600122451782, + "num_tokens": 20461294.0, + "step": 791 + }, + { + "epoch": 0.08697562047001976, + "ewc_loss": 1.955777406692505e-07, + "grad_norm": 2.5667970180511475, + "learning_rate": 2.895314787701317e-07, + "loss": 1.1001, + "mean_token_accuracy": 0.6735426187515259, + "num_tokens": 20486377.0, + "step": 792 + }, + { + "epoch": 0.08708543817263342, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.789679527282715, + "learning_rate": 2.898975109809663e-07, + "loss": 1.0529, + "mean_token_accuracy": 0.6910876035690308, + "num_tokens": 20508898.0, + "step": 793 + }, + { + "epoch": 0.0871952558752471, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.2569644451141357, + "learning_rate": 2.9026354319180086e-07, + "loss": 1.1311, + "mean_token_accuracy": 0.66994708776474, + "num_tokens": 20538441.0, + "step": 794 + }, + { + "epoch": 0.08730507357786076, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.518000602722168, + "learning_rate": 2.9062957540263547e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.683864176273346, + "num_tokens": 20561412.0, + "step": 795 + }, + { + "epoch": 0.08741489128047442, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.661813259124756, + "learning_rate": 2.9099560761346996e-07, + "loss": 1.0776, + "mean_token_accuracy": 0.6815458536148071, + "num_tokens": 20586226.0, + "step": 796 + }, + { + "epoch": 0.08752470898308808, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.293977737426758, + "learning_rate": 2.913616398243045e-07, + "loss": 1.2007, + "mean_token_accuracy": 0.655354917049408, + "num_tokens": 20618182.0, + "step": 797 + }, + { + "epoch": 0.08763452668570174, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.410973310470581, + "learning_rate": 2.917276720351391e-07, + "loss": 1.1653, + "mean_token_accuracy": 0.6710770726203918, + "num_tokens": 20647672.0, + "step": 798 + }, + { + "epoch": 0.0877443443883154, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.5638575553894043, + "learning_rate": 2.920937042459736e-07, + "loss": 1.1033, + "mean_token_accuracy": 0.689970850944519, + "num_tokens": 20671724.0, + "step": 799 + }, + { + "epoch": 0.08785416209092906, + "ewc_loss": 1.9650906324386597e-07, + "grad_norm": 2.3067591190338135, + "learning_rate": 2.9245973645680816e-07, + "loss": 1.1446, + "mean_token_accuracy": 0.6832021474838257, + "num_tokens": 20700193.0, + "step": 800 + }, + { + "epoch": 0.08796397979354272, + "ewc_loss": 1.9744038581848145e-07, + "grad_norm": 2.4902443885803223, + "learning_rate": 2.9282576866764276e-07, + "loss": 1.0701, + "mean_token_accuracy": 0.6811820864677429, + "num_tokens": 20724905.0, + "step": 801 + }, + { + "epoch": 0.08807379749615638, + "ewc_loss": 1.9837170839309692e-07, + "grad_norm": 2.5192253589630127, + "learning_rate": 2.9319180087847726e-07, + "loss": 1.129, + "mean_token_accuracy": 0.6705811619758606, + "num_tokens": 20749749.0, + "step": 802 + }, + { + "epoch": 0.08818361519877004, + "ewc_loss": 1.9744038581848145e-07, + "grad_norm": 2.3017423152923584, + "learning_rate": 2.9355783308931186e-07, + "loss": 1.0028, + "mean_token_accuracy": 0.6986328959465027, + "num_tokens": 20778139.0, + "step": 803 + }, + { + "epoch": 0.08829343290138371, + "ewc_loss": 1.9837170839309692e-07, + "grad_norm": 2.505979299545288, + "learning_rate": 2.939238653001464e-07, + "loss": 1.1147, + "mean_token_accuracy": 0.6860924959182739, + "num_tokens": 20801934.0, + "step": 804 + }, + { + "epoch": 0.08840325060399737, + "ewc_loss": 1.9837170839309692e-07, + "grad_norm": 2.2757794857025146, + "learning_rate": 2.9428989751098096e-07, + "loss": 1.1125, + "mean_token_accuracy": 0.6766706705093384, + "num_tokens": 20832336.0, + "step": 805 + }, + { + "epoch": 0.08851306830661103, + "ewc_loss": 1.993030309677124e-07, + "grad_norm": 2.4269397258758545, + "learning_rate": 2.946559297218155e-07, + "loss": 1.1164, + "mean_token_accuracy": 0.6707425713539124, + "num_tokens": 20858646.0, + "step": 806 + }, + { + "epoch": 0.08862288600922469, + "ewc_loss": 1.993030309677124e-07, + "grad_norm": 2.3769161701202393, + "learning_rate": 2.9502196193265006e-07, + "loss": 1.1409, + "mean_token_accuracy": 0.6680508852005005, + "num_tokens": 20887602.0, + "step": 807 + }, + { + "epoch": 0.08873270371183835, + "ewc_loss": 2.0023435354232788e-07, + "grad_norm": 2.6607825756073, + "learning_rate": 2.953879941434846e-07, + "loss": 1.1665, + "mean_token_accuracy": 0.6598958969116211, + "num_tokens": 20912488.0, + "step": 808 + }, + { + "epoch": 0.08884252141445201, + "ewc_loss": 2.0023435354232788e-07, + "grad_norm": 2.341560125350952, + "learning_rate": 2.9575402635431915e-07, + "loss": 1.1897, + "mean_token_accuracy": 0.6533458828926086, + "num_tokens": 20940624.0, + "step": 809 + }, + { + "epoch": 0.08895233911706567, + "ewc_loss": 2.0209699869155884e-07, + "grad_norm": 2.225205421447754, + "learning_rate": 2.961200585651537e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.683159589767456, + "num_tokens": 20970873.0, + "step": 810 + }, + { + "epoch": 0.08906215681967933, + "ewc_loss": 2.0209699869155884e-07, + "grad_norm": 2.658785820007324, + "learning_rate": 2.964860907759883e-07, + "loss": 1.0231, + "mean_token_accuracy": 0.6953030824661255, + "num_tokens": 20993869.0, + "step": 811 + }, + { + "epoch": 0.08917197452229299, + "ewc_loss": 2.0302832126617432e-07, + "grad_norm": 2.4002389907836914, + "learning_rate": 2.9685212298682285e-07, + "loss": 1.0288, + "mean_token_accuracy": 0.6932934522628784, + "num_tokens": 21019651.0, + "step": 812 + }, + { + "epoch": 0.08928179222490666, + "ewc_loss": 2.0302832126617432e-07, + "grad_norm": 2.2890236377716064, + "learning_rate": 2.9721815519765735e-07, + "loss": 1.0597, + "mean_token_accuracy": 0.689145028591156, + "num_tokens": 21048736.0, + "step": 813 + }, + { + "epoch": 0.08939160992752032, + "ewc_loss": 2.0489096641540527e-07, + "grad_norm": 2.7293484210968018, + "learning_rate": 2.9758418740849195e-07, + "loss": 1.1903, + "mean_token_accuracy": 0.6668885350227356, + "num_tokens": 21071505.0, + "step": 814 + }, + { + "epoch": 0.08950142763013398, + "ewc_loss": 2.0675361156463623e-07, + "grad_norm": 2.7096354961395264, + "learning_rate": 2.979502196193265e-07, + "loss": 1.0229, + "mean_token_accuracy": 0.6973192691802979, + "num_tokens": 21094803.0, + "step": 815 + }, + { + "epoch": 0.08961124533274764, + "ewc_loss": 2.0675361156463623e-07, + "grad_norm": 2.5541603565216064, + "learning_rate": 2.98316251830161e-07, + "loss": 1.0978, + "mean_token_accuracy": 0.681185245513916, + "num_tokens": 21120328.0, + "step": 816 + }, + { + "epoch": 0.0897210630353613, + "ewc_loss": 2.086162567138672e-07, + "grad_norm": 2.4226555824279785, + "learning_rate": 2.986822840409956e-07, + "loss": 1.1128, + "mean_token_accuracy": 0.6720547676086426, + "num_tokens": 21146162.0, + "step": 817 + }, + { + "epoch": 0.08983088073797496, + "ewc_loss": 2.086162567138672e-07, + "grad_norm": 2.547788381576538, + "learning_rate": 2.9904831625183015e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6890726089477539, + "num_tokens": 21169041.0, + "step": 818 + }, + { + "epoch": 0.08994069844058862, + "ewc_loss": 2.0954757928848267e-07, + "grad_norm": 2.4236600399017334, + "learning_rate": 2.994143484626647e-07, + "loss": 1.0971, + "mean_token_accuracy": 0.6816102266311646, + "num_tokens": 21195693.0, + "step": 819 + }, + { + "epoch": 0.09005051614320228, + "ewc_loss": 2.1047890186309814e-07, + "grad_norm": 2.7634336948394775, + "learning_rate": 2.9978038067349925e-07, + "loss": 1.0601, + "mean_token_accuracy": 0.6920616030693054, + "num_tokens": 21217125.0, + "step": 820 + }, + { + "epoch": 0.09016033384581594, + "ewc_loss": 2.1047890186309814e-07, + "grad_norm": 2.4940438270568848, + "learning_rate": 3.001464128843338e-07, + "loss": 1.1605, + "mean_token_accuracy": 0.6595374345779419, + "num_tokens": 21243201.0, + "step": 821 + }, + { + "epoch": 0.0902701515484296, + "ewc_loss": 2.1047890186309814e-07, + "grad_norm": 2.264469623565674, + "learning_rate": 3.005124450951684e-07, + "loss": 1.0631, + "mean_token_accuracy": 0.6905690431594849, + "num_tokens": 21272069.0, + "step": 822 + }, + { + "epoch": 0.09037996925104327, + "ewc_loss": 2.1047890186309814e-07, + "grad_norm": 2.4167773723602295, + "learning_rate": 3.008784773060029e-07, + "loss": 1.1493, + "mean_token_accuracy": 0.6638721823692322, + "num_tokens": 21299739.0, + "step": 823 + }, + { + "epoch": 0.09048978695365693, + "ewc_loss": 2.1141022443771362e-07, + "grad_norm": 2.975052833557129, + "learning_rate": 3.0124450951683744e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6770355105400085, + "num_tokens": 21318516.0, + "step": 824 + }, + { + "epoch": 0.0905996046562706, + "ewc_loss": 2.1141022443771362e-07, + "grad_norm": 2.857703685760498, + "learning_rate": 3.0161054172767204e-07, + "loss": 1.0801, + "mean_token_accuracy": 0.6750919818878174, + "num_tokens": 21337893.0, + "step": 825 + }, + { + "epoch": 0.09070942235888425, + "ewc_loss": 2.1141022443771362e-07, + "grad_norm": 2.3511409759521484, + "learning_rate": 3.0197657393850654e-07, + "loss": 1.1949, + "mean_token_accuracy": 0.6479693651199341, + "num_tokens": 21366728.0, + "step": 826 + }, + { + "epoch": 0.09081924006149791, + "ewc_loss": 2.1141022443771362e-07, + "grad_norm": 2.151120901107788, + "learning_rate": 3.0234260614934114e-07, + "loss": 1.1858, + "mean_token_accuracy": 0.6516772508621216, + "num_tokens": 21398742.0, + "step": 827 + }, + { + "epoch": 0.09092905776411157, + "ewc_loss": 2.1141022443771362e-07, + "grad_norm": 2.6164188385009766, + "learning_rate": 3.027086383601757e-07, + "loss": 0.9931, + "mean_token_accuracy": 0.7046070694923401, + "num_tokens": 21422256.0, + "step": 828 + }, + { + "epoch": 0.09103887546672523, + "ewc_loss": 2.1327286958694458e-07, + "grad_norm": 2.4956979751586914, + "learning_rate": 3.0307467057101024e-07, + "loss": 1.126, + "mean_token_accuracy": 0.6715692281723022, + "num_tokens": 21445882.0, + "step": 829 + }, + { + "epoch": 0.09114869316933889, + "ewc_loss": 2.1327286958694458e-07, + "grad_norm": 2.4181571006774902, + "learning_rate": 3.034407027818448e-07, + "loss": 1.103, + "mean_token_accuracy": 0.6730440855026245, + "num_tokens": 21472863.0, + "step": 830 + }, + { + "epoch": 0.09125851087195255, + "ewc_loss": 2.1327286958694458e-07, + "grad_norm": 2.726024866104126, + "learning_rate": 3.0380673499267934e-07, + "loss": 1.0401, + "mean_token_accuracy": 0.6922026872634888, + "num_tokens": 21496899.0, + "step": 831 + }, + { + "epoch": 0.09136832857456623, + "ewc_loss": 2.1327286958694458e-07, + "grad_norm": 2.7593917846679688, + "learning_rate": 3.041727672035139e-07, + "loss": 1.1141, + "mean_token_accuracy": 0.6769224405288696, + "num_tokens": 21519107.0, + "step": 832 + }, + { + "epoch": 0.09147814627717989, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.586045265197754, + "learning_rate": 3.0453879941434844e-07, + "loss": 1.1499, + "mean_token_accuracy": 0.6648916602134705, + "num_tokens": 21544059.0, + "step": 833 + }, + { + "epoch": 0.09158796397979355, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.5696983337402344, + "learning_rate": 3.04904831625183e-07, + "loss": 1.0736, + "mean_token_accuracy": 0.6867415904998779, + "num_tokens": 21568468.0, + "step": 834 + }, + { + "epoch": 0.0916977816824072, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.5032553672790527, + "learning_rate": 3.052708638360176e-07, + "loss": 1.0918, + "mean_token_accuracy": 0.6742088794708252, + "num_tokens": 21592891.0, + "step": 835 + }, + { + "epoch": 0.09180759938502087, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.8075997829437256, + "learning_rate": 3.056368960468521e-07, + "loss": 1.0842, + "mean_token_accuracy": 0.6840050220489502, + "num_tokens": 21614569.0, + "step": 836 + }, + { + "epoch": 0.09191741708763453, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.571122884750366, + "learning_rate": 3.0600292825768663e-07, + "loss": 1.151, + "mean_token_accuracy": 0.6579828858375549, + "num_tokens": 21638089.0, + "step": 837 + }, + { + "epoch": 0.09202723479024819, + "ewc_loss": 2.1420419216156006e-07, + "grad_norm": 2.1108381748199463, + "learning_rate": 3.0636896046852124e-07, + "loss": 1.096, + "mean_token_accuracy": 0.6872513294219971, + "num_tokens": 21668891.0, + "step": 838 + }, + { + "epoch": 0.09213705249286185, + "ewc_loss": 2.1513551473617554e-07, + "grad_norm": 2.403212070465088, + "learning_rate": 3.067349926793558e-07, + "loss": 1.1474, + "mean_token_accuracy": 0.6689990758895874, + "num_tokens": 21696713.0, + "step": 839 + }, + { + "epoch": 0.0922468701954755, + "ewc_loss": 2.1513551473617554e-07, + "grad_norm": 2.2728254795074463, + "learning_rate": 3.071010248901903e-07, + "loss": 1.0089, + "mean_token_accuracy": 0.7035396099090576, + "num_tokens": 21724588.0, + "step": 840 + }, + { + "epoch": 0.09235668789808917, + "ewc_loss": 2.1606683731079102e-07, + "grad_norm": 2.8767316341400146, + "learning_rate": 3.074670571010249e-07, + "loss": 1.0715, + "mean_token_accuracy": 0.6935055255889893, + "num_tokens": 21745178.0, + "step": 841 + }, + { + "epoch": 0.09246650560070284, + "ewc_loss": 2.1606683731079102e-07, + "grad_norm": 2.5585010051727295, + "learning_rate": 3.0783308931185943e-07, + "loss": 1.0719, + "mean_token_accuracy": 0.6883265972137451, + "num_tokens": 21768870.0, + "step": 842 + }, + { + "epoch": 0.0925763233033165, + "ewc_loss": 2.169981598854065e-07, + "grad_norm": 2.3591620922088623, + "learning_rate": 3.08199121522694e-07, + "loss": 1.0568, + "mean_token_accuracy": 0.6867681741714478, + "num_tokens": 21796552.0, + "step": 843 + }, + { + "epoch": 0.09268614100593016, + "ewc_loss": 2.169981598854065e-07, + "grad_norm": 2.2541768550872803, + "learning_rate": 3.0856515373352853e-07, + "loss": 1.129, + "mean_token_accuracy": 0.6735903024673462, + "num_tokens": 21825989.0, + "step": 844 + }, + { + "epoch": 0.09279595870854382, + "ewc_loss": 2.169981598854065e-07, + "grad_norm": 2.6468749046325684, + "learning_rate": 3.089311859443631e-07, + "loss": 1.0362, + "mean_token_accuracy": 0.6918623447418213, + "num_tokens": 21848075.0, + "step": 845 + }, + { + "epoch": 0.09290577641115748, + "ewc_loss": 2.169981598854065e-07, + "grad_norm": 2.3169870376586914, + "learning_rate": 3.092972181551977e-07, + "loss": 1.1331, + "mean_token_accuracy": 0.6735628843307495, + "num_tokens": 21877666.0, + "step": 846 + }, + { + "epoch": 0.09301559411377114, + "ewc_loss": 2.1792948246002197e-07, + "grad_norm": 2.2683897018432617, + "learning_rate": 3.096632503660322e-07, + "loss": 1.0819, + "mean_token_accuracy": 0.6774631142616272, + "num_tokens": 21904463.0, + "step": 847 + }, + { + "epoch": 0.0931254118163848, + "ewc_loss": 2.1886080503463745e-07, + "grad_norm": 2.5325522422790527, + "learning_rate": 3.1002928257686673e-07, + "loss": 1.0576, + "mean_token_accuracy": 0.6824655532836914, + "num_tokens": 21927852.0, + "step": 848 + }, + { + "epoch": 0.09323522951899846, + "ewc_loss": 2.1886080503463745e-07, + "grad_norm": 2.348140001296997, + "learning_rate": 3.1039531478770133e-07, + "loss": 1.0527, + "mean_token_accuracy": 0.6952584385871887, + "num_tokens": 21956317.0, + "step": 849 + }, + { + "epoch": 0.09334504722161212, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.3842601776123047, + "learning_rate": 3.107613469985358e-07, + "loss": 1.0533, + "mean_token_accuracy": 0.6846904158592224, + "num_tokens": 21981920.0, + "step": 850 + }, + { + "epoch": 0.09345486492422579, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.7431252002716064, + "learning_rate": 3.1112737920937043e-07, + "loss": 1.1161, + "mean_token_accuracy": 0.6720759868621826, + "num_tokens": 22006163.0, + "step": 851 + }, + { + "epoch": 0.09356468262683945, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.558820962905884, + "learning_rate": 3.11493411420205e-07, + "loss": 1.0872, + "mean_token_accuracy": 0.676538348197937, + "num_tokens": 22029906.0, + "step": 852 + }, + { + "epoch": 0.09367450032945311, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.4616031646728516, + "learning_rate": 3.1185944363103947e-07, + "loss": 1.1264, + "mean_token_accuracy": 0.6742088794708252, + "num_tokens": 22055611.0, + "step": 853 + }, + { + "epoch": 0.09378431803206677, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.3641343116760254, + "learning_rate": 3.122254758418741e-07, + "loss": 1.0748, + "mean_token_accuracy": 0.6755555272102356, + "num_tokens": 22081614.0, + "step": 854 + }, + { + "epoch": 0.09389413573468043, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.5719006061553955, + "learning_rate": 3.125915080527086e-07, + "loss": 1.0584, + "mean_token_accuracy": 0.6831886768341064, + "num_tokens": 22104956.0, + "step": 855 + }, + { + "epoch": 0.09400395343729409, + "ewc_loss": 2.207234501838684e-07, + "grad_norm": 2.227792978286743, + "learning_rate": 3.1295754026354317e-07, + "loss": 1.1056, + "mean_token_accuracy": 0.6720327138900757, + "num_tokens": 22135119.0, + "step": 856 + }, + { + "epoch": 0.09411377113990775, + "ewc_loss": 2.2258609533309937e-07, + "grad_norm": 2.685560703277588, + "learning_rate": 3.133235724743777e-07, + "loss": 1.1198, + "mean_token_accuracy": 0.6757140159606934, + "num_tokens": 22157625.0, + "step": 857 + }, + { + "epoch": 0.09422358884252141, + "ewc_loss": 2.2258609533309937e-07, + "grad_norm": 2.228656053543091, + "learning_rate": 3.1368960468521227e-07, + "loss": 1.094, + "mean_token_accuracy": 0.6819137334823608, + "num_tokens": 22188178.0, + "step": 858 + }, + { + "epoch": 0.09433340654513507, + "ewc_loss": 2.2258609533309937e-07, + "grad_norm": 2.3894219398498535, + "learning_rate": 3.1405563689604687e-07, + "loss": 1.0972, + "mean_token_accuracy": 0.6732995510101318, + "num_tokens": 22213762.0, + "step": 859 + }, + { + "epoch": 0.09444322424774873, + "ewc_loss": 2.2258609533309937e-07, + "grad_norm": 2.7222254276275635, + "learning_rate": 3.1442166910688137e-07, + "loss": 1.1038, + "mean_token_accuracy": 0.6763510704040527, + "num_tokens": 22235680.0, + "step": 860 + }, + { + "epoch": 0.0945530419503624, + "ewc_loss": 2.2351741790771484e-07, + "grad_norm": 2.6429693698883057, + "learning_rate": 3.147877013177159e-07, + "loss": 1.0331, + "mean_token_accuracy": 0.6928789615631104, + "num_tokens": 22258638.0, + "step": 861 + }, + { + "epoch": 0.09466285965297606, + "ewc_loss": 2.2351741790771484e-07, + "grad_norm": 2.5291643142700195, + "learning_rate": 3.151537335285505e-07, + "loss": 1.0763, + "mean_token_accuracy": 0.6820829510688782, + "num_tokens": 22282129.0, + "step": 862 + }, + { + "epoch": 0.09477267735558972, + "ewc_loss": 2.2258609533309937e-07, + "grad_norm": 2.1909310817718506, + "learning_rate": 3.1551976573938507e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.7021774053573608, + "num_tokens": 22312222.0, + "step": 863 + }, + { + "epoch": 0.09488249505820338, + "ewc_loss": 2.2444874048233032e-07, + "grad_norm": 2.7139956951141357, + "learning_rate": 3.1588579795021956e-07, + "loss": 1.0893, + "mean_token_accuracy": 0.675153374671936, + "num_tokens": 22335476.0, + "step": 864 + }, + { + "epoch": 0.09499231276081704, + "ewc_loss": 2.2444874048233032e-07, + "grad_norm": 2.4368319511413574, + "learning_rate": 3.1625183016105417e-07, + "loss": 1.1429, + "mean_token_accuracy": 0.6857519149780273, + "num_tokens": 22362628.0, + "step": 865 + }, + { + "epoch": 0.0951021304634307, + "ewc_loss": 2.253800630569458e-07, + "grad_norm": 2.569307565689087, + "learning_rate": 3.166178623718887e-07, + "loss": 1.1432, + "mean_token_accuracy": 0.6698716282844543, + "num_tokens": 22388402.0, + "step": 866 + }, + { + "epoch": 0.09521194816604436, + "ewc_loss": 2.253800630569458e-07, + "grad_norm": 2.320150136947632, + "learning_rate": 3.1698389458272327e-07, + "loss": 1.1232, + "mean_token_accuracy": 0.6762195229530334, + "num_tokens": 22417436.0, + "step": 867 + }, + { + "epoch": 0.09532176586865802, + "ewc_loss": 2.253800630569458e-07, + "grad_norm": 2.699371337890625, + "learning_rate": 3.173499267935578e-07, + "loss": 1.151, + "mean_token_accuracy": 0.6636670827865601, + "num_tokens": 22440356.0, + "step": 868 + }, + { + "epoch": 0.09543158357127168, + "ewc_loss": 2.2631138563156128e-07, + "grad_norm": 2.8311455249786377, + "learning_rate": 3.1771595900439236e-07, + "loss": 1.1455, + "mean_token_accuracy": 0.6726667881011963, + "num_tokens": 22462445.0, + "step": 869 + }, + { + "epoch": 0.09554140127388536, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.164168119430542, + "learning_rate": 3.1808199121522697e-07, + "loss": 1.1419, + "mean_token_accuracy": 0.6688045263290405, + "num_tokens": 22495714.0, + "step": 870 + }, + { + "epoch": 0.09565121897649902, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.4447731971740723, + "learning_rate": 3.1844802342606146e-07, + "loss": 1.1267, + "mean_token_accuracy": 0.6681946516036987, + "num_tokens": 22520890.0, + "step": 871 + }, + { + "epoch": 0.09576103667911268, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.279940605163574, + "learning_rate": 3.18814055636896e-07, + "loss": 1.0985, + "mean_token_accuracy": 0.6843520998954773, + "num_tokens": 22548125.0, + "step": 872 + }, + { + "epoch": 0.09587085438172634, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.5886526107788086, + "learning_rate": 3.191800878477306e-07, + "loss": 1.0437, + "mean_token_accuracy": 0.6967799663543701, + "num_tokens": 22571203.0, + "step": 873 + }, + { + "epoch": 0.09598067208434, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.332472801208496, + "learning_rate": 3.195461200585651e-07, + "loss": 1.0684, + "mean_token_accuracy": 0.6805777549743652, + "num_tokens": 22600078.0, + "step": 874 + }, + { + "epoch": 0.09609048978695366, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.5007164478302, + "learning_rate": 3.199121522693997e-07, + "loss": 1.1092, + "mean_token_accuracy": 0.680762529373169, + "num_tokens": 22626236.0, + "step": 875 + }, + { + "epoch": 0.09620030748956732, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.282320022583008, + "learning_rate": 3.2027818448023426e-07, + "loss": 1.0968, + "mean_token_accuracy": 0.673709511756897, + "num_tokens": 22653233.0, + "step": 876 + }, + { + "epoch": 0.09631012519218098, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.3619189262390137, + "learning_rate": 3.2064421669106876e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.6703174710273743, + "num_tokens": 22679272.0, + "step": 877 + }, + { + "epoch": 0.09641994289479464, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.5259482860565186, + "learning_rate": 3.2101024890190336e-07, + "loss": 1.0515, + "mean_token_accuracy": 0.6905306577682495, + "num_tokens": 22704003.0, + "step": 878 + }, + { + "epoch": 0.0965297605974083, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.4144198894500732, + "learning_rate": 3.213762811127379e-07, + "loss": 1.037, + "mean_token_accuracy": 0.6956803202629089, + "num_tokens": 22729119.0, + "step": 879 + }, + { + "epoch": 0.09663957830002197, + "ewc_loss": 2.2724270820617676e-07, + "grad_norm": 2.380537748336792, + "learning_rate": 3.2174231332357246e-07, + "loss": 1.1155, + "mean_token_accuracy": 0.674277663230896, + "num_tokens": 22755678.0, + "step": 880 + }, + { + "epoch": 0.09674939600263563, + "ewc_loss": 2.2817403078079224e-07, + "grad_norm": 2.2215373516082764, + "learning_rate": 3.22108345534407e-07, + "loss": 1.0922, + "mean_token_accuracy": 0.6805331707000732, + "num_tokens": 22788600.0, + "step": 881 + }, + { + "epoch": 0.09685921370524929, + "ewc_loss": 2.2910535335540771e-07, + "grad_norm": 2.257017135620117, + "learning_rate": 3.2247437774524155e-07, + "loss": 1.0139, + "mean_token_accuracy": 0.6981251835823059, + "num_tokens": 22814636.0, + "step": 882 + }, + { + "epoch": 0.09696903140786295, + "ewc_loss": 2.2910535335540771e-07, + "grad_norm": 2.598191022872925, + "learning_rate": 3.2284040995607616e-07, + "loss": 1.1279, + "mean_token_accuracy": 0.6697138547897339, + "num_tokens": 22839289.0, + "step": 883 + }, + { + "epoch": 0.09707884911047661, + "ewc_loss": 2.2910535335540771e-07, + "grad_norm": 2.4634664058685303, + "learning_rate": 3.2320644216691065e-07, + "loss": 1.0779, + "mean_token_accuracy": 0.6780736446380615, + "num_tokens": 22864607.0, + "step": 884 + }, + { + "epoch": 0.09718866681309027, + "ewc_loss": 2.2910535335540771e-07, + "grad_norm": 2.5377109050750732, + "learning_rate": 3.235724743777452e-07, + "loss": 1.1275, + "mean_token_accuracy": 0.6821534633636475, + "num_tokens": 22892406.0, + "step": 885 + }, + { + "epoch": 0.09729848451570393, + "ewc_loss": 2.2910535335540771e-07, + "grad_norm": 2.5349485874176025, + "learning_rate": 3.239385065885798e-07, + "loss": 1.1023, + "mean_token_accuracy": 0.6722571849822998, + "num_tokens": 22917563.0, + "step": 886 + }, + { + "epoch": 0.09740830221831759, + "ewc_loss": 2.3096799850463867e-07, + "grad_norm": 2.3549530506134033, + "learning_rate": 3.2430453879941435e-07, + "loss": 1.078, + "mean_token_accuracy": 0.6819044351577759, + "num_tokens": 22944690.0, + "step": 887 + }, + { + "epoch": 0.09751811992093125, + "ewc_loss": 2.3096799850463867e-07, + "grad_norm": 2.5173327922821045, + "learning_rate": 3.2467057101024885e-07, + "loss": 1.0661, + "mean_token_accuracy": 0.6867697238922119, + "num_tokens": 22970435.0, + "step": 888 + }, + { + "epoch": 0.09762793762354492, + "ewc_loss": 2.3096799850463867e-07, + "grad_norm": 2.4291043281555176, + "learning_rate": 3.2503660322108345e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.696887731552124, + "num_tokens": 22996469.0, + "step": 889 + }, + { + "epoch": 0.09773775532615858, + "ewc_loss": 2.3283064365386963e-07, + "grad_norm": 2.5980451107025146, + "learning_rate": 3.25402635431918e-07, + "loss": 1.1583, + "mean_token_accuracy": 0.6698218584060669, + "num_tokens": 23019491.0, + "step": 890 + }, + { + "epoch": 0.09784757302877224, + "ewc_loss": 2.3283064365386963e-07, + "grad_norm": 2.281923770904541, + "learning_rate": 3.2576866764275255e-07, + "loss": 1.0759, + "mean_token_accuracy": 0.6882656812667847, + "num_tokens": 23048481.0, + "step": 891 + }, + { + "epoch": 0.0979573907313859, + "ewc_loss": 2.3283064365386963e-07, + "grad_norm": 2.4274742603302, + "learning_rate": 3.261346998535871e-07, + "loss": 1.0775, + "mean_token_accuracy": 0.6815576553344727, + "num_tokens": 23074453.0, + "step": 892 + }, + { + "epoch": 0.09806720843399956, + "ewc_loss": 2.337619662284851e-07, + "grad_norm": 2.2199738025665283, + "learning_rate": 3.2650073206442165e-07, + "loss": 0.9889, + "mean_token_accuracy": 0.705241858959198, + "num_tokens": 23103151.0, + "step": 893 + }, + { + "epoch": 0.09817702613661322, + "ewc_loss": 2.3469328880310059e-07, + "grad_norm": 2.179995059967041, + "learning_rate": 3.2686676427525625e-07, + "loss": 1.0995, + "mean_token_accuracy": 0.6749718189239502, + "num_tokens": 23134289.0, + "step": 894 + }, + { + "epoch": 0.09828684383922688, + "ewc_loss": 2.3469328880310059e-07, + "grad_norm": 2.638089418411255, + "learning_rate": 3.2723279648609075e-07, + "loss": 1.0286, + "mean_token_accuracy": 0.7006914019584656, + "num_tokens": 23156101.0, + "step": 895 + }, + { + "epoch": 0.09839666154184054, + "ewc_loss": 2.3562461137771606e-07, + "grad_norm": 2.5214784145355225, + "learning_rate": 3.275988286969253e-07, + "loss": 1.0025, + "mean_token_accuracy": 0.699367105960846, + "num_tokens": 23179178.0, + "step": 896 + }, + { + "epoch": 0.0985064792444542, + "ewc_loss": 2.3655593395233154e-07, + "grad_norm": 2.651360511779785, + "learning_rate": 3.279648609077599e-07, + "loss": 1.1026, + "mean_token_accuracy": 0.6787664294242859, + "num_tokens": 23200455.0, + "step": 897 + }, + { + "epoch": 0.09861629694706786, + "ewc_loss": 2.3655593395233154e-07, + "grad_norm": 2.4726667404174805, + "learning_rate": 3.283308931185944e-07, + "loss": 1.0956, + "mean_token_accuracy": 0.693148136138916, + "num_tokens": 23226237.0, + "step": 898 + }, + { + "epoch": 0.09872611464968153, + "ewc_loss": 2.3748725652694702e-07, + "grad_norm": 2.334319829940796, + "learning_rate": 3.28696925329429e-07, + "loss": 1.0828, + "mean_token_accuracy": 0.6874325275421143, + "num_tokens": 23251813.0, + "step": 899 + }, + { + "epoch": 0.0988359323522952, + "ewc_loss": 2.3748725652694702e-07, + "grad_norm": 2.3736791610717773, + "learning_rate": 3.2906295754026354e-07, + "loss": 1.1455, + "mean_token_accuracy": 0.66002357006073, + "num_tokens": 23278810.0, + "step": 900 + }, + { + "epoch": 0.09894575005490885, + "ewc_loss": 2.3748725652694702e-07, + "grad_norm": 2.226653575897217, + "learning_rate": 3.2942898975109804e-07, + "loss": 1.1323, + "mean_token_accuracy": 0.6731486320495605, + "num_tokens": 23307162.0, + "step": 901 + }, + { + "epoch": 0.09905556775752251, + "ewc_loss": 2.3748725652694702e-07, + "grad_norm": 2.2695765495300293, + "learning_rate": 3.2979502196193264e-07, + "loss": 1.1437, + "mean_token_accuracy": 0.6653493046760559, + "num_tokens": 23335308.0, + "step": 902 + }, + { + "epoch": 0.09916538546013617, + "ewc_loss": 2.3748725652694702e-07, + "grad_norm": 2.4704902172088623, + "learning_rate": 3.301610541727672e-07, + "loss": 1.0443, + "mean_token_accuracy": 0.6935670375823975, + "num_tokens": 23359068.0, + "step": 903 + }, + { + "epoch": 0.09927520316274983, + "ewc_loss": 2.384185791015625e-07, + "grad_norm": 2.455481767654419, + "learning_rate": 3.3052708638360174e-07, + "loss": 1.0081, + "mean_token_accuracy": 0.7059117555618286, + "num_tokens": 23381987.0, + "step": 904 + }, + { + "epoch": 0.0993850208653635, + "ewc_loss": 2.4028122425079346e-07, + "grad_norm": 1.9711048603057861, + "learning_rate": 3.308931185944363e-07, + "loss": 1.1213, + "mean_token_accuracy": 0.6711671352386475, + "num_tokens": 23420367.0, + "step": 905 + }, + { + "epoch": 0.09949483856797715, + "ewc_loss": 2.4028122425079346e-07, + "grad_norm": 2.6709370613098145, + "learning_rate": 3.3125915080527084e-07, + "loss": 1.0629, + "mean_token_accuracy": 0.6887120008468628, + "num_tokens": 23442587.0, + "step": 906 + }, + { + "epoch": 0.09960465627059081, + "ewc_loss": 2.4028122425079346e-07, + "grad_norm": 2.4146907329559326, + "learning_rate": 3.3162518301610544e-07, + "loss": 1.1286, + "mean_token_accuracy": 0.6738859415054321, + "num_tokens": 23469268.0, + "step": 907 + }, + { + "epoch": 0.09971447397320449, + "ewc_loss": 2.4028122425079346e-07, + "grad_norm": 2.1675570011138916, + "learning_rate": 3.3199121522693994e-07, + "loss": 1.1901, + "mean_token_accuracy": 0.6461153030395508, + "num_tokens": 23500886.0, + "step": 908 + }, + { + "epoch": 0.09982429167581815, + "ewc_loss": 2.421438694000244e-07, + "grad_norm": 2.6467702388763428, + "learning_rate": 3.323572474377745e-07, + "loss": 1.0925, + "mean_token_accuracy": 0.6821655631065369, + "num_tokens": 23523429.0, + "step": 909 + }, + { + "epoch": 0.0999341093784318, + "ewc_loss": 2.421438694000244e-07, + "grad_norm": 2.510216236114502, + "learning_rate": 3.327232796486091e-07, + "loss": 1.0416, + "mean_token_accuracy": 0.6980821490287781, + "num_tokens": 23545940.0, + "step": 910 + }, + { + "epoch": 0.10004392708104547, + "ewc_loss": 2.421438694000244e-07, + "grad_norm": 2.5523765087127686, + "learning_rate": 3.3308931185944364e-07, + "loss": 1.1855, + "mean_token_accuracy": 0.6607937812805176, + "num_tokens": 23572000.0, + "step": 911 + }, + { + "epoch": 0.10015374478365913, + "ewc_loss": 2.4400651454925537e-07, + "grad_norm": 2.3868892192840576, + "learning_rate": 3.3345534407027813e-07, + "loss": 1.1093, + "mean_token_accuracy": 0.6784772276878357, + "num_tokens": 23600798.0, + "step": 912 + }, + { + "epoch": 0.10026356248627279, + "ewc_loss": 2.4400651454925537e-07, + "grad_norm": 2.494737386703491, + "learning_rate": 3.3382137628111273e-07, + "loss": 1.0233, + "mean_token_accuracy": 0.6953369379043579, + "num_tokens": 23624264.0, + "step": 913 + }, + { + "epoch": 0.10037338018888645, + "ewc_loss": 2.4586915969848633e-07, + "grad_norm": 2.1721351146698, + "learning_rate": 3.341874084919473e-07, + "loss": 1.0628, + "mean_token_accuracy": 0.6851119995117188, + "num_tokens": 23655171.0, + "step": 914 + }, + { + "epoch": 0.1004831978915001, + "ewc_loss": 2.4586915969848633e-07, + "grad_norm": 2.260953187942505, + "learning_rate": 3.3455344070278183e-07, + "loss": 1.1237, + "mean_token_accuracy": 0.6724157929420471, + "num_tokens": 23683903.0, + "step": 915 + }, + { + "epoch": 0.10059301559411377, + "ewc_loss": 2.4586915969848633e-07, + "grad_norm": 2.6085269451141357, + "learning_rate": 3.349194729136164e-07, + "loss": 1.0574, + "mean_token_accuracy": 0.6929042339324951, + "num_tokens": 23706406.0, + "step": 916 + }, + { + "epoch": 0.10070283329672743, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.400172233581543, + "learning_rate": 3.3528550512445093e-07, + "loss": 1.1264, + "mean_token_accuracy": 0.6668390035629272, + "num_tokens": 23732433.0, + "step": 917 + }, + { + "epoch": 0.1008126509993411, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.204061269760132, + "learning_rate": 3.356515373352855e-07, + "loss": 1.0316, + "mean_token_accuracy": 0.69246506690979, + "num_tokens": 23759236.0, + "step": 918 + }, + { + "epoch": 0.10092246870195476, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.4556334018707275, + "learning_rate": 3.3601756954612003e-07, + "loss": 1.0222, + "mean_token_accuracy": 0.6973530650138855, + "num_tokens": 23784398.0, + "step": 919 + }, + { + "epoch": 0.10103228640456842, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.3470025062561035, + "learning_rate": 3.363836017569546e-07, + "loss": 1.0054, + "mean_token_accuracy": 0.6988651752471924, + "num_tokens": 23812650.0, + "step": 920 + }, + { + "epoch": 0.10114210410718208, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.1923532485961914, + "learning_rate": 3.367496339677892e-07, + "loss": 1.0928, + "mean_token_accuracy": 0.6812169551849365, + "num_tokens": 23841776.0, + "step": 921 + }, + { + "epoch": 0.10125192180979574, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.5397303104400635, + "learning_rate": 3.371156661786237e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6777700185775757, + "num_tokens": 23865229.0, + "step": 922 + }, + { + "epoch": 0.1013617395124094, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.426581382751465, + "learning_rate": 3.374816983894583e-07, + "loss": 1.0829, + "mean_token_accuracy": 0.6746689081192017, + "num_tokens": 23889843.0, + "step": 923 + }, + { + "epoch": 0.10147155721502306, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.3983469009399414, + "learning_rate": 3.3784773060029283e-07, + "loss": 1.1604, + "mean_token_accuracy": 0.6657359600067139, + "num_tokens": 23917810.0, + "step": 924 + }, + { + "epoch": 0.10158137491763672, + "ewc_loss": 2.477318048477173e-07, + "grad_norm": 2.6606483459472656, + "learning_rate": 3.382137628111273e-07, + "loss": 1.0835, + "mean_token_accuracy": 0.6782929301261902, + "num_tokens": 23940531.0, + "step": 925 + }, + { + "epoch": 0.10169119262025038, + "ewc_loss": 2.4959444999694824e-07, + "grad_norm": 2.229956865310669, + "learning_rate": 3.385797950219619e-07, + "loss": 1.0747, + "mean_token_accuracy": 0.6817840337753296, + "num_tokens": 23968313.0, + "step": 926 + }, + { + "epoch": 0.10180101032286405, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.245011806488037, + "learning_rate": 3.389458272327965e-07, + "loss": 1.1175, + "mean_token_accuracy": 0.6757511496543884, + "num_tokens": 23999841.0, + "step": 927 + }, + { + "epoch": 0.10191082802547771, + "ewc_loss": 2.4959444999694824e-07, + "grad_norm": 2.1145904064178467, + "learning_rate": 3.39311859443631e-07, + "loss": 1.1391, + "mean_token_accuracy": 0.664097785949707, + "num_tokens": 24031207.0, + "step": 928 + }, + { + "epoch": 0.10202064572809137, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.1793715953826904, + "learning_rate": 3.3967789165446557e-07, + "loss": 1.1248, + "mean_token_accuracy": 0.6722428798675537, + "num_tokens": 24060419.0, + "step": 929 + }, + { + "epoch": 0.10213046343070503, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.3112049102783203, + "learning_rate": 3.400439238653001e-07, + "loss": 1.0916, + "mean_token_accuracy": 0.6853973865509033, + "num_tokens": 24086752.0, + "step": 930 + }, + { + "epoch": 0.10224028113331869, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.4820053577423096, + "learning_rate": 3.404099560761347e-07, + "loss": 1.0404, + "mean_token_accuracy": 0.6879026889801025, + "num_tokens": 24110671.0, + "step": 931 + }, + { + "epoch": 0.10235009883593235, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.222703218460083, + "learning_rate": 3.407759882869692e-07, + "loss": 1.0853, + "mean_token_accuracy": 0.6778155565261841, + "num_tokens": 24142021.0, + "step": 932 + }, + { + "epoch": 0.10245991653854601, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.7422895431518555, + "learning_rate": 3.4114202049780377e-07, + "loss": 1.0889, + "mean_token_accuracy": 0.6788204908370972, + "num_tokens": 24163153.0, + "step": 933 + }, + { + "epoch": 0.10256973424115967, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.217874526977539, + "learning_rate": 3.4150805270863837e-07, + "loss": 1.1531, + "mean_token_accuracy": 0.6629955768585205, + "num_tokens": 24194395.0, + "step": 934 + }, + { + "epoch": 0.10267955194377333, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.349916696548462, + "learning_rate": 3.4187408491947287e-07, + "loss": 1.1367, + "mean_token_accuracy": 0.6648615598678589, + "num_tokens": 24220717.0, + "step": 935 + }, + { + "epoch": 0.10278936964638699, + "ewc_loss": 2.514570951461792e-07, + "grad_norm": 2.435659170150757, + "learning_rate": 3.422401171303074e-07, + "loss": 1.0986, + "mean_token_accuracy": 0.6818246841430664, + "num_tokens": 24245109.0, + "step": 936 + }, + { + "epoch": 0.10289918734900066, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.8707945346832275, + "learning_rate": 3.42606149341142e-07, + "loss": 1.1672, + "mean_token_accuracy": 0.6682864427566528, + "num_tokens": 24264983.0, + "step": 937 + }, + { + "epoch": 0.10300900505161432, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.58062744140625, + "learning_rate": 3.4297218155197657e-07, + "loss": 1.0906, + "mean_token_accuracy": 0.6876718997955322, + "num_tokens": 24288101.0, + "step": 938 + }, + { + "epoch": 0.10311882275422798, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.255220651626587, + "learning_rate": 3.433382137628111e-07, + "loss": 1.0603, + "mean_token_accuracy": 0.6855002045631409, + "num_tokens": 24314589.0, + "step": 939 + }, + { + "epoch": 0.10322864045684164, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.595597743988037, + "learning_rate": 3.4370424597364567e-07, + "loss": 1.1418, + "mean_token_accuracy": 0.6759061217308044, + "num_tokens": 24337109.0, + "step": 940 + }, + { + "epoch": 0.1033384581594553, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.2612173557281494, + "learning_rate": 3.440702781844802e-07, + "loss": 1.128, + "mean_token_accuracy": 0.6745239496231079, + "num_tokens": 24367221.0, + "step": 941 + }, + { + "epoch": 0.10344827586206896, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.433119058609009, + "learning_rate": 3.4443631039531476e-07, + "loss": 1.1473, + "mean_token_accuracy": 0.6642795205116272, + "num_tokens": 24394210.0, + "step": 942 + }, + { + "epoch": 0.10355809356468262, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.360907793045044, + "learning_rate": 3.448023426061493e-07, + "loss": 1.0135, + "mean_token_accuracy": 0.6963386535644531, + "num_tokens": 24419005.0, + "step": 943 + }, + { + "epoch": 0.10366791126729628, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.258082389831543, + "learning_rate": 3.4516837481698386e-07, + "loss": 1.0964, + "mean_token_accuracy": 0.6809926629066467, + "num_tokens": 24448427.0, + "step": 944 + }, + { + "epoch": 0.10377772896990994, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.153463363647461, + "learning_rate": 3.4553440702781846e-07, + "loss": 1.2008, + "mean_token_accuracy": 0.651748538017273, + "num_tokens": 24480209.0, + "step": 945 + }, + { + "epoch": 0.10388754667252362, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.5487236976623535, + "learning_rate": 3.4590043923865296e-07, + "loss": 1.0261, + "mean_token_accuracy": 0.6974385976791382, + "num_tokens": 24503484.0, + "step": 946 + }, + { + "epoch": 0.10399736437513728, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.387099027633667, + "learning_rate": 3.4626647144948756e-07, + "loss": 1.1004, + "mean_token_accuracy": 0.6730628609657288, + "num_tokens": 24528827.0, + "step": 947 + }, + { + "epoch": 0.10410718207775094, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.714660406112671, + "learning_rate": 3.466325036603221e-07, + "loss": 1.0327, + "mean_token_accuracy": 0.6981990337371826, + "num_tokens": 24550278.0, + "step": 948 + }, + { + "epoch": 0.1042169997803646, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.440001964569092, + "learning_rate": 3.469985358711566e-07, + "loss": 1.0065, + "mean_token_accuracy": 0.6977548599243164, + "num_tokens": 24573147.0, + "step": 949 + }, + { + "epoch": 0.10432681748297826, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.489570379257202, + "learning_rate": 3.473645680819912e-07, + "loss": 1.0731, + "mean_token_accuracy": 0.6765297651290894, + "num_tokens": 24597279.0, + "step": 950 + }, + { + "epoch": 0.10443663518559192, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.448230504989624, + "learning_rate": 3.4773060029282576e-07, + "loss": 1.1229, + "mean_token_accuracy": 0.6611580848693848, + "num_tokens": 24623328.0, + "step": 951 + }, + { + "epoch": 0.10454645288820558, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.739229202270508, + "learning_rate": 3.480966325036603e-07, + "loss": 1.0707, + "mean_token_accuracy": 0.6934840679168701, + "num_tokens": 24644301.0, + "step": 952 + }, + { + "epoch": 0.10465627059081924, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.0723369121551514, + "learning_rate": 3.4846266471449486e-07, + "loss": 1.1338, + "mean_token_accuracy": 0.6701886057853699, + "num_tokens": 24678615.0, + "step": 953 + }, + { + "epoch": 0.1047660882934329, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.639450788497925, + "learning_rate": 3.488286969253294e-07, + "loss": 0.9737, + "mean_token_accuracy": 0.7070872187614441, + "num_tokens": 24701336.0, + "step": 954 + }, + { + "epoch": 0.10487590599604656, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.2791478633880615, + "learning_rate": 3.49194729136164e-07, + "loss": 1.0318, + "mean_token_accuracy": 0.6942238211631775, + "num_tokens": 24727917.0, + "step": 955 + }, + { + "epoch": 0.10498572369866023, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.742004871368408, + "learning_rate": 3.495607613469985e-07, + "loss": 1.0261, + "mean_token_accuracy": 0.6904377937316895, + "num_tokens": 24748147.0, + "step": 956 + }, + { + "epoch": 0.10509554140127389, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.449594736099243, + "learning_rate": 3.4992679355783305e-07, + "loss": 1.0651, + "mean_token_accuracy": 0.6894740462303162, + "num_tokens": 24773384.0, + "step": 957 + }, + { + "epoch": 0.10520535910388755, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.2808399200439453, + "learning_rate": 3.5029282576866766e-07, + "loss": 1.1, + "mean_token_accuracy": 0.6730175018310547, + "num_tokens": 24801521.0, + "step": 958 + }, + { + "epoch": 0.10531517680650121, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.3111560344696045, + "learning_rate": 3.5065885797950215e-07, + "loss": 1.1217, + "mean_token_accuracy": 0.6742634177207947, + "num_tokens": 24830015.0, + "step": 959 + }, + { + "epoch": 0.10542499450911487, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.572208881378174, + "learning_rate": 3.510248901903367e-07, + "loss": 1.0833, + "mean_token_accuracy": 0.6775624752044678, + "num_tokens": 24853232.0, + "step": 960 + }, + { + "epoch": 0.10553481221172853, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.438093900680542, + "learning_rate": 3.513909224011713e-07, + "loss": 1.1421, + "mean_token_accuracy": 0.6691529750823975, + "num_tokens": 24877653.0, + "step": 961 + }, + { + "epoch": 0.10564462991434219, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.4505422115325928, + "learning_rate": 3.5175695461200585e-07, + "loss": 1.1237, + "mean_token_accuracy": 0.6702213287353516, + "num_tokens": 24903434.0, + "step": 962 + }, + { + "epoch": 0.10575444761695585, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.393813133239746, + "learning_rate": 3.521229868228404e-07, + "loss": 1.1588, + "mean_token_accuracy": 0.6688419580459595, + "num_tokens": 24931347.0, + "step": 963 + }, + { + "epoch": 0.10586426531956951, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.563016414642334, + "learning_rate": 3.5248901903367495e-07, + "loss": 1.0591, + "mean_token_accuracy": 0.6885991096496582, + "num_tokens": 24955365.0, + "step": 964 + }, + { + "epoch": 0.10597408302218318, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.6899020671844482, + "learning_rate": 3.528550512445095e-07, + "loss": 1.1475, + "mean_token_accuracy": 0.6634778380393982, + "num_tokens": 24977249.0, + "step": 965 + }, + { + "epoch": 0.10608390072479684, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.2031939029693604, + "learning_rate": 3.5322108345534405e-07, + "loss": 1.1577, + "mean_token_accuracy": 0.665228545665741, + "num_tokens": 25008578.0, + "step": 966 + }, + { + "epoch": 0.1061937184274105, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.71093487739563, + "learning_rate": 3.535871156661786e-07, + "loss": 1.0203, + "mean_token_accuracy": 0.6896509528160095, + "num_tokens": 25028258.0, + "step": 967 + }, + { + "epoch": 0.10630353613002416, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.5269198417663574, + "learning_rate": 3.5395314787701315e-07, + "loss": 1.0497, + "mean_token_accuracy": 0.6887959837913513, + "num_tokens": 25052108.0, + "step": 968 + }, + { + "epoch": 0.10641335383263782, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.0452022552490234, + "learning_rate": 3.5431918008784775e-07, + "loss": 1.1111, + "mean_token_accuracy": 0.6784629225730896, + "num_tokens": 25085379.0, + "step": 969 + }, + { + "epoch": 0.10652317153525148, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.571787118911743, + "learning_rate": 3.5468521229868224e-07, + "loss": 1.0889, + "mean_token_accuracy": 0.6809313893318176, + "num_tokens": 25108319.0, + "step": 970 + }, + { + "epoch": 0.10663298923786514, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.2561066150665283, + "learning_rate": 3.5505124450951685e-07, + "loss": 1.1968, + "mean_token_accuracy": 0.6574782729148865, + "num_tokens": 25139744.0, + "step": 971 + }, + { + "epoch": 0.1067428069404788, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.3448097705841064, + "learning_rate": 3.554172767203514e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6801738739013672, + "num_tokens": 25166647.0, + "step": 972 + }, + { + "epoch": 0.10685262464309246, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.233900547027588, + "learning_rate": 3.557833089311859e-07, + "loss": 1.0637, + "mean_token_accuracy": 0.6882697343826294, + "num_tokens": 25195503.0, + "step": 973 + }, + { + "epoch": 0.10696244234570612, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.457890033721924, + "learning_rate": 3.561493411420205e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.690291166305542, + "num_tokens": 25220138.0, + "step": 974 + }, + { + "epoch": 0.1070722600483198, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.4386630058288574, + "learning_rate": 3.5651537335285504e-07, + "loss": 1.0581, + "mean_token_accuracy": 0.6843925714492798, + "num_tokens": 25244278.0, + "step": 975 + }, + { + "epoch": 0.10718207775093345, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.1818153858184814, + "learning_rate": 3.5688140556368954e-07, + "loss": 1.0973, + "mean_token_accuracy": 0.6703822612762451, + "num_tokens": 25275205.0, + "step": 976 + }, + { + "epoch": 0.10729189545354711, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.5884201526641846, + "learning_rate": 3.5724743777452414e-07, + "loss": 1.0333, + "mean_token_accuracy": 0.6920098662376404, + "num_tokens": 25296551.0, + "step": 977 + }, + { + "epoch": 0.10740171315616077, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.474515199661255, + "learning_rate": 3.576134699853587e-07, + "loss": 0.9426, + "mean_token_accuracy": 0.7188918590545654, + "num_tokens": 25319702.0, + "step": 978 + }, + { + "epoch": 0.10751153085877443, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.3936939239501953, + "learning_rate": 3.579795021961933e-07, + "loss": 1.1333, + "mean_token_accuracy": 0.6668248176574707, + "num_tokens": 25345390.0, + "step": 979 + }, + { + "epoch": 0.1076213485613881, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.33931565284729, + "learning_rate": 3.583455344070278e-07, + "loss": 1.0582, + "mean_token_accuracy": 0.6855006217956543, + "num_tokens": 25371403.0, + "step": 980 + }, + { + "epoch": 0.10773116626400175, + "ewc_loss": 2.5331974029541016e-07, + "grad_norm": 2.7748196125030518, + "learning_rate": 3.5871156661786234e-07, + "loss": 1.0102, + "mean_token_accuracy": 0.6978576183319092, + "num_tokens": 25389771.0, + "step": 981 + }, + { + "epoch": 0.10784098396661541, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.6314499378204346, + "learning_rate": 3.5907759882869694e-07, + "loss": 1.0461, + "mean_token_accuracy": 0.6865831017494202, + "num_tokens": 25410123.0, + "step": 982 + }, + { + "epoch": 0.10795080166922907, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.6521379947662354, + "learning_rate": 3.5944363103953144e-07, + "loss": 1.0562, + "mean_token_accuracy": 0.6825577616691589, + "num_tokens": 25432542.0, + "step": 983 + }, + { + "epoch": 0.10806061937184275, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.3372344970703125, + "learning_rate": 3.59809663250366e-07, + "loss": 1.1056, + "mean_token_accuracy": 0.671484112739563, + "num_tokens": 25458481.0, + "step": 984 + }, + { + "epoch": 0.10817043707445641, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.4026308059692383, + "learning_rate": 3.601756954612006e-07, + "loss": 1.0238, + "mean_token_accuracy": 0.7016074657440186, + "num_tokens": 25483261.0, + "step": 985 + }, + { + "epoch": 0.10828025477707007, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.378826856613159, + "learning_rate": 3.6054172767203514e-07, + "loss": 1.0265, + "mean_token_accuracy": 0.6978694796562195, + "num_tokens": 25511391.0, + "step": 986 + }, + { + "epoch": 0.10839007247968373, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.110153913497925, + "learning_rate": 3.609077598828697e-07, + "loss": 1.102, + "mean_token_accuracy": 0.6754989624023438, + "num_tokens": 25541179.0, + "step": 987 + }, + { + "epoch": 0.10849989018229739, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.5502254962921143, + "learning_rate": 3.6127379209370423e-07, + "loss": 1.0123, + "mean_token_accuracy": 0.6995914578437805, + "num_tokens": 25564057.0, + "step": 988 + }, + { + "epoch": 0.10860970788491105, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.3666086196899414, + "learning_rate": 3.616398243045388e-07, + "loss": 1.0502, + "mean_token_accuracy": 0.6943042278289795, + "num_tokens": 25590617.0, + "step": 989 + }, + { + "epoch": 0.1087195255875247, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.8653829097747803, + "learning_rate": 3.6200585651537333e-07, + "loss": 1.0554, + "mean_token_accuracy": 0.6897038817405701, + "num_tokens": 25610733.0, + "step": 990 + }, + { + "epoch": 0.10882934329013837, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.4294793605804443, + "learning_rate": 3.623718887262079e-07, + "loss": 1.0848, + "mean_token_accuracy": 0.6796572208404541, + "num_tokens": 25636827.0, + "step": 991 + }, + { + "epoch": 0.10893916099275203, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.559621810913086, + "learning_rate": 3.6273792093704243e-07, + "loss": 1.1173, + "mean_token_accuracy": 0.6759195923805237, + "num_tokens": 25659218.0, + "step": 992 + }, + { + "epoch": 0.10904897869536569, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.5623598098754883, + "learning_rate": 3.6310395314787703e-07, + "loss": 1.0702, + "mean_token_accuracy": 0.6840873956680298, + "num_tokens": 25682816.0, + "step": 993 + }, + { + "epoch": 0.10915879639797936, + "ewc_loss": 2.551823854446411e-07, + "grad_norm": 2.167097330093384, + "learning_rate": 3.6346998535871153e-07, + "loss": 1.0909, + "mean_token_accuracy": 0.679238498210907, + "num_tokens": 25712238.0, + "step": 994 + }, + { + "epoch": 0.10926861410059302, + "ewc_loss": 2.5704503059387207e-07, + "grad_norm": 2.4062325954437256, + "learning_rate": 3.6383601756954613e-07, + "loss": 0.9847, + "mean_token_accuracy": 0.6981366872787476, + "num_tokens": 25737964.0, + "step": 995 + }, + { + "epoch": 0.10937843180320668, + "ewc_loss": 2.5704503059387207e-07, + "grad_norm": 2.25211763381958, + "learning_rate": 3.642020497803807e-07, + "loss": 1.0113, + "mean_token_accuracy": 0.698669970035553, + "num_tokens": 25765993.0, + "step": 996 + }, + { + "epoch": 0.10948824950582034, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.6170079708099365, + "learning_rate": 3.645680819912152e-07, + "loss": 1.0562, + "mean_token_accuracy": 0.689565122127533, + "num_tokens": 25789007.0, + "step": 997 + }, + { + "epoch": 0.109598067208434, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.3558757305145264, + "learning_rate": 3.649341142020498e-07, + "loss": 0.9862, + "mean_token_accuracy": 0.705655574798584, + "num_tokens": 25814004.0, + "step": 998 + }, + { + "epoch": 0.10970788491104766, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.5958194732666016, + "learning_rate": 3.653001464128843e-07, + "loss": 1.0607, + "mean_token_accuracy": 0.6878606677055359, + "num_tokens": 25835515.0, + "step": 999 + }, + { + "epoch": 0.10981770261366132, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.490121364593506, + "learning_rate": 3.656661786237188e-07, + "loss": 1.1005, + "mean_token_accuracy": 0.6813092827796936, + "num_tokens": 25859300.0, + "step": 1000 + }, + { + "epoch": 0.10992752031627498, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.2788617610931396, + "learning_rate": 3.660322108345534e-07, + "loss": 1.0388, + "mean_token_accuracy": 0.6896860003471375, + "num_tokens": 25888053.0, + "step": 1001 + }, + { + "epoch": 0.11003733801888864, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.40690016746521, + "learning_rate": 3.66398243045388e-07, + "loss": 1.1533, + "mean_token_accuracy": 0.6610209941864014, + "num_tokens": 25913789.0, + "step": 1002 + }, + { + "epoch": 0.11014715572150231, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.5653703212738037, + "learning_rate": 3.667642752562226e-07, + "loss": 1.1631, + "mean_token_accuracy": 0.6697422862052917, + "num_tokens": 25936587.0, + "step": 1003 + }, + { + "epoch": 0.11025697342411597, + "ewc_loss": 2.5890767574310303e-07, + "grad_norm": 2.1489195823669434, + "learning_rate": 3.6713030746705707e-07, + "loss": 1.0481, + "mean_token_accuracy": 0.6889292001724243, + "num_tokens": 25965940.0, + "step": 1004 + }, + { + "epoch": 0.11036679112672963, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.24777889251709, + "learning_rate": 3.674963396778916e-07, + "loss": 0.9894, + "mean_token_accuracy": 0.7072572708129883, + "num_tokens": 25993188.0, + "step": 1005 + }, + { + "epoch": 0.11047660882934329, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.4507505893707275, + "learning_rate": 3.678623718887262e-07, + "loss": 1.0269, + "mean_token_accuracy": 0.6931265592575073, + "num_tokens": 26016160.0, + "step": 1006 + }, + { + "epoch": 0.11058642653195695, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.4456560611724854, + "learning_rate": 3.682284040995607e-07, + "loss": 1.0367, + "mean_token_accuracy": 0.6899846792221069, + "num_tokens": 26039892.0, + "step": 1007 + }, + { + "epoch": 0.11069624423457061, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.552196741104126, + "learning_rate": 3.6859443631039527e-07, + "loss": 1.0675, + "mean_token_accuracy": 0.6824735403060913, + "num_tokens": 26061680.0, + "step": 1008 + }, + { + "epoch": 0.11080606193718427, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.3217172622680664, + "learning_rate": 3.6896046852122987e-07, + "loss": 1.1628, + "mean_token_accuracy": 0.6575155258178711, + "num_tokens": 26088811.0, + "step": 1009 + }, + { + "epoch": 0.11091587963979793, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.4947500228881836, + "learning_rate": 3.693265007320644e-07, + "loss": 1.1232, + "mean_token_accuracy": 0.6704931259155273, + "num_tokens": 26113462.0, + "step": 1010 + }, + { + "epoch": 0.11102569734241159, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.345402956008911, + "learning_rate": 3.6969253294289897e-07, + "loss": 1.1501, + "mean_token_accuracy": 0.6633199453353882, + "num_tokens": 26140867.0, + "step": 1011 + }, + { + "epoch": 0.11113551504502525, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.2707157135009766, + "learning_rate": 3.700585651537335e-07, + "loss": 1.07, + "mean_token_accuracy": 0.6774628162384033, + "num_tokens": 26168121.0, + "step": 1012 + }, + { + "epoch": 0.11124533274763893, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.7222604751586914, + "learning_rate": 3.7042459736456807e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.6907088160514832, + "num_tokens": 26189626.0, + "step": 1013 + }, + { + "epoch": 0.11135515045025259, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.559987783432007, + "learning_rate": 3.707906295754026e-07, + "loss": 1.099, + "mean_token_accuracy": 0.6756689548492432, + "num_tokens": 26213658.0, + "step": 1014 + }, + { + "epoch": 0.11146496815286625, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.111220598220825, + "learning_rate": 3.7115666178623716e-07, + "loss": 1.0379, + "mean_token_accuracy": 0.6895763874053955, + "num_tokens": 26243731.0, + "step": 1015 + }, + { + "epoch": 0.1115747858554799, + "ewc_loss": 2.60770320892334e-07, + "grad_norm": 2.6016440391540527, + "learning_rate": 3.715226939970717e-07, + "loss": 1.0609, + "mean_token_accuracy": 0.6831398606300354, + "num_tokens": 26266656.0, + "step": 1016 + }, + { + "epoch": 0.11168460355809356, + "ewc_loss": 2.6263296604156494e-07, + "grad_norm": 2.9584999084472656, + "learning_rate": 3.7188872620790626e-07, + "loss": 1.0296, + "mean_token_accuracy": 0.6938802599906921, + "num_tokens": 26285162.0, + "step": 1017 + }, + { + "epoch": 0.11179442126070722, + "ewc_loss": 2.6263296604156494e-07, + "grad_norm": 2.4947385787963867, + "learning_rate": 3.722547584187408e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6906590461730957, + "num_tokens": 26309229.0, + "step": 1018 + }, + { + "epoch": 0.11190423896332088, + "ewc_loss": 2.6263296604156494e-07, + "grad_norm": 2.5212013721466064, + "learning_rate": 3.726207906295754e-07, + "loss": 1.1019, + "mean_token_accuracy": 0.6778509616851807, + "num_tokens": 26334702.0, + "step": 1019 + }, + { + "epoch": 0.11201405666593454, + "ewc_loss": 2.6263296604156494e-07, + "grad_norm": 2.5421082973480225, + "learning_rate": 3.7298682284040996e-07, + "loss": 1.0626, + "mean_token_accuracy": 0.6941300630569458, + "num_tokens": 26359954.0, + "step": 1020 + }, + { + "epoch": 0.1121238743685482, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.412365674972534, + "learning_rate": 3.7335285505124446e-07, + "loss": 1.0194, + "mean_token_accuracy": 0.6969408392906189, + "num_tokens": 26383642.0, + "step": 1021 + }, + { + "epoch": 0.11223369207116188, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.1734933853149414, + "learning_rate": 3.7371888726207906e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6975249648094177, + "num_tokens": 26413718.0, + "step": 1022 + }, + { + "epoch": 0.11234350977377554, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.029906749725342, + "learning_rate": 3.740849194729136e-07, + "loss": 1.1476, + "mean_token_accuracy": 0.6665374040603638, + "num_tokens": 26448070.0, + "step": 1023 + }, + { + "epoch": 0.1124533274763892, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.2715115547180176, + "learning_rate": 3.744509516837481e-07, + "loss": 1.1557, + "mean_token_accuracy": 0.6634907722473145, + "num_tokens": 26477876.0, + "step": 1024 + }, + { + "epoch": 0.11256314517900286, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.4737374782562256, + "learning_rate": 3.748169838945827e-07, + "loss": 1.1648, + "mean_token_accuracy": 0.6552870273590088, + "num_tokens": 26503463.0, + "step": 1025 + }, + { + "epoch": 0.11267296288161652, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.325744152069092, + "learning_rate": 3.7518301610541726e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6867421865463257, + "num_tokens": 26530910.0, + "step": 1026 + }, + { + "epoch": 0.11278278058423018, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.33929181098938, + "learning_rate": 3.7554904831625186e-07, + "loss": 1.1271, + "mean_token_accuracy": 0.6651004552841187, + "num_tokens": 26556511.0, + "step": 1027 + }, + { + "epoch": 0.11289259828684384, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.7312536239624023, + "learning_rate": 3.7591508052708636e-07, + "loss": 1.0846, + "mean_token_accuracy": 0.6743237972259521, + "num_tokens": 26577702.0, + "step": 1028 + }, + { + "epoch": 0.1130024159894575, + "ewc_loss": 2.644956111907959e-07, + "grad_norm": 2.437556266784668, + "learning_rate": 3.762811127379209e-07, + "loss": 1.1557, + "mean_token_accuracy": 0.669174313545227, + "num_tokens": 26602465.0, + "step": 1029 + }, + { + "epoch": 0.11311223369207116, + "ewc_loss": 2.6635825634002686e-07, + "grad_norm": 2.7545297145843506, + "learning_rate": 3.766471449487555e-07, + "loss": 1.0554, + "mean_token_accuracy": 0.6833453178405762, + "num_tokens": 26622436.0, + "step": 1030 + }, + { + "epoch": 0.11322205139468482, + "ewc_loss": 2.6635825634002686e-07, + "grad_norm": 2.4661812782287598, + "learning_rate": 3.7701317715959e-07, + "loss": 1.066, + "mean_token_accuracy": 0.6865464448928833, + "num_tokens": 26646386.0, + "step": 1031 + }, + { + "epoch": 0.11333186909729849, + "ewc_loss": 2.6635825634002686e-07, + "grad_norm": 2.3587915897369385, + "learning_rate": 3.7737920937042455e-07, + "loss": 1.1618, + "mean_token_accuracy": 0.6602511405944824, + "num_tokens": 26674495.0, + "step": 1032 + }, + { + "epoch": 0.11344168679991215, + "ewc_loss": 2.6635825634002686e-07, + "grad_norm": 2.261625051498413, + "learning_rate": 3.7774524158125915e-07, + "loss": 1.039, + "mean_token_accuracy": 0.6930997967720032, + "num_tokens": 26701395.0, + "step": 1033 + }, + { + "epoch": 0.11355150450252581, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.568697690963745, + "learning_rate": 3.781112737920937e-07, + "loss": 1.1266, + "mean_token_accuracy": 0.6762019395828247, + "num_tokens": 26724537.0, + "step": 1034 + }, + { + "epoch": 0.11366132220513947, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.444236993789673, + "learning_rate": 3.7847730600292825e-07, + "loss": 1.1025, + "mean_token_accuracy": 0.6779873371124268, + "num_tokens": 26747889.0, + "step": 1035 + }, + { + "epoch": 0.11377113990775313, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.938499689102173, + "learning_rate": 3.788433382137628e-07, + "loss": 1.0109, + "mean_token_accuracy": 0.6982024908065796, + "num_tokens": 26767707.0, + "step": 1036 + }, + { + "epoch": 0.11388095761036679, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.3420698642730713, + "learning_rate": 3.7920937042459735e-07, + "loss": 1.1279, + "mean_token_accuracy": 0.6812542676925659, + "num_tokens": 26793032.0, + "step": 1037 + }, + { + "epoch": 0.11399077531298045, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.4137637615203857, + "learning_rate": 3.795754026354319e-07, + "loss": 1.0382, + "mean_token_accuracy": 0.6897546648979187, + "num_tokens": 26819051.0, + "step": 1038 + }, + { + "epoch": 0.11410059301559411, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.438558340072632, + "learning_rate": 3.7994143484626645e-07, + "loss": 1.0429, + "mean_token_accuracy": 0.6929799318313599, + "num_tokens": 26842616.0, + "step": 1039 + }, + { + "epoch": 0.11421041071820777, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.6026291847229004, + "learning_rate": 3.80307467057101e-07, + "loss": 1.1131, + "mean_token_accuracy": 0.6831699013710022, + "num_tokens": 26865703.0, + "step": 1040 + }, + { + "epoch": 0.11432022842082144, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.909860610961914, + "learning_rate": 3.8067349926793555e-07, + "loss": 1.1048, + "mean_token_accuracy": 0.6682363748550415, + "num_tokens": 26884984.0, + "step": 1041 + }, + { + "epoch": 0.1144300461234351, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.34451961517334, + "learning_rate": 3.810395314787701e-07, + "loss": 1.0822, + "mean_token_accuracy": 0.6832364201545715, + "num_tokens": 26912097.0, + "step": 1042 + }, + { + "epoch": 0.11453986382604876, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.7208962440490723, + "learning_rate": 3.814055636896047e-07, + "loss": 1.0167, + "mean_token_accuracy": 0.6992607712745667, + "num_tokens": 26932033.0, + "step": 1043 + }, + { + "epoch": 0.11464968152866242, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.535529851913452, + "learning_rate": 3.8177159590043925e-07, + "loss": 1.1007, + "mean_token_accuracy": 0.674881100654602, + "num_tokens": 26954418.0, + "step": 1044 + }, + { + "epoch": 0.11475949923127608, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.141385793685913, + "learning_rate": 3.8213762811127374e-07, + "loss": 1.0652, + "mean_token_accuracy": 0.6868683099746704, + "num_tokens": 26984406.0, + "step": 1045 + }, + { + "epoch": 0.11486931693388974, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.396526575088501, + "learning_rate": 3.8250366032210835e-07, + "loss": 1.0651, + "mean_token_accuracy": 0.6898679137229919, + "num_tokens": 27012181.0, + "step": 1046 + }, + { + "epoch": 0.1149791346365034, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.231154680252075, + "learning_rate": 3.828696925329429e-07, + "loss": 1.0896, + "mean_token_accuracy": 0.6782320141792297, + "num_tokens": 27040817.0, + "step": 1047 + }, + { + "epoch": 0.11508895233911706, + "ewc_loss": 2.7008354663848877e-07, + "grad_norm": 2.4756672382354736, + "learning_rate": 3.832357247437774e-07, + "loss": 1.0022, + "mean_token_accuracy": 0.7107162475585938, + "num_tokens": 27063531.0, + "step": 1048 + }, + { + "epoch": 0.11519877004173072, + "ewc_loss": 2.682209014892578e-07, + "grad_norm": 2.509164333343506, + "learning_rate": 3.83601756954612e-07, + "loss": 1.0394, + "mean_token_accuracy": 0.6865130662918091, + "num_tokens": 27085193.0, + "step": 1049 + }, + { + "epoch": 0.11530858774434438, + "ewc_loss": 2.7008354663848877e-07, + "grad_norm": 2.3053019046783447, + "learning_rate": 3.8396778916544654e-07, + "loss": 1.0837, + "mean_token_accuracy": 0.6746861934661865, + "num_tokens": 27112026.0, + "step": 1050 + }, + { + "epoch": 0.11541840544695806, + "ewc_loss": 2.7194619178771973e-07, + "grad_norm": 2.592014789581299, + "learning_rate": 3.8433382137628114e-07, + "loss": 1.0797, + "mean_token_accuracy": 0.6861237287521362, + "num_tokens": 27134736.0, + "step": 1051 + }, + { + "epoch": 0.11552822314957172, + "ewc_loss": 2.7194619178771973e-07, + "grad_norm": 2.0982980728149414, + "learning_rate": 3.8469985358711564e-07, + "loss": 1.1908, + "mean_token_accuracy": 0.6516324281692505, + "num_tokens": 27167154.0, + "step": 1052 + }, + { + "epoch": 0.11563804085218538, + "ewc_loss": 2.738088369369507e-07, + "grad_norm": 2.3223490715026855, + "learning_rate": 3.850658857979502e-07, + "loss": 1.0509, + "mean_token_accuracy": 0.6931691765785217, + "num_tokens": 27191779.0, + "step": 1053 + }, + { + "epoch": 0.11574785855479904, + "ewc_loss": 2.738088369369507e-07, + "grad_norm": 2.66975998878479, + "learning_rate": 3.854319180087848e-07, + "loss": 1.0672, + "mean_token_accuracy": 0.6810593605041504, + "num_tokens": 27214207.0, + "step": 1054 + }, + { + "epoch": 0.1158576762574127, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.389519691467285, + "learning_rate": 3.857979502196193e-07, + "loss": 1.1012, + "mean_token_accuracy": 0.6806203126907349, + "num_tokens": 27240519.0, + "step": 1055 + }, + { + "epoch": 0.11596749396002635, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.380765914916992, + "learning_rate": 3.8616398243045384e-07, + "loss": 1.028, + "mean_token_accuracy": 0.6976205110549927, + "num_tokens": 27265773.0, + "step": 1056 + }, + { + "epoch": 0.11607731166264001, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.2599520683288574, + "learning_rate": 3.8653001464128844e-07, + "loss": 1.1263, + "mean_token_accuracy": 0.6671304702758789, + "num_tokens": 27294062.0, + "step": 1057 + }, + { + "epoch": 0.11618712936525367, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.5607247352600098, + "learning_rate": 3.8689604685212293e-07, + "loss": 1.0597, + "mean_token_accuracy": 0.6870712637901306, + "num_tokens": 27317200.0, + "step": 1058 + }, + { + "epoch": 0.11629694706786733, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.2470390796661377, + "learning_rate": 3.8726207906295754e-07, + "loss": 1.1154, + "mean_token_accuracy": 0.6720963716506958, + "num_tokens": 27345129.0, + "step": 1059 + }, + { + "epoch": 0.11640676477048101, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.1619150638580322, + "learning_rate": 3.876281112737921e-07, + "loss": 1.1304, + "mean_token_accuracy": 0.6697942614555359, + "num_tokens": 27376500.0, + "step": 1060 + }, + { + "epoch": 0.11651658247309467, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.439807891845703, + "learning_rate": 3.8799414348462663e-07, + "loss": 1.1114, + "mean_token_accuracy": 0.6691320538520813, + "num_tokens": 27405136.0, + "step": 1061 + }, + { + "epoch": 0.11662640017570833, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.373379707336426, + "learning_rate": 3.883601756954612e-07, + "loss": 1.0668, + "mean_token_accuracy": 0.6798038482666016, + "num_tokens": 27430922.0, + "step": 1062 + }, + { + "epoch": 0.11673621787832199, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.6742496490478516, + "learning_rate": 3.8872620790629573e-07, + "loss": 1.0343, + "mean_token_accuracy": 0.6936311721801758, + "num_tokens": 27452160.0, + "step": 1063 + }, + { + "epoch": 0.11684603558093565, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.375192642211914, + "learning_rate": 3.890922401171303e-07, + "loss": 1.0957, + "mean_token_accuracy": 0.6797502040863037, + "num_tokens": 27477265.0, + "step": 1064 + }, + { + "epoch": 0.11695585328354931, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.3934385776519775, + "learning_rate": 3.8945827232796483e-07, + "loss": 1.1276, + "mean_token_accuracy": 0.6696563959121704, + "num_tokens": 27503052.0, + "step": 1065 + }, + { + "epoch": 0.11706567098616297, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.5360522270202637, + "learning_rate": 3.898243045387994e-07, + "loss": 1.0514, + "mean_token_accuracy": 0.6935774087905884, + "num_tokens": 27526173.0, + "step": 1066 + }, + { + "epoch": 0.11717548868877663, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.162466287612915, + "learning_rate": 3.90190336749634e-07, + "loss": 1.0969, + "mean_token_accuracy": 0.6858571767807007, + "num_tokens": 27556183.0, + "step": 1067 + }, + { + "epoch": 0.11728530639139029, + "ewc_loss": 2.7567148208618164e-07, + "grad_norm": 2.2541794776916504, + "learning_rate": 3.9055636896046853e-07, + "loss": 1.0699, + "mean_token_accuracy": 0.6846442222595215, + "num_tokens": 27583306.0, + "step": 1068 + }, + { + "epoch": 0.11739512409400395, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.386408805847168, + "learning_rate": 3.9092240117130303e-07, + "loss": 1.0284, + "mean_token_accuracy": 0.6931525468826294, + "num_tokens": 27607080.0, + "step": 1069 + }, + { + "epoch": 0.11750494179661762, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.128997564315796, + "learning_rate": 3.9128843338213763e-07, + "loss": 1.13, + "mean_token_accuracy": 0.6692540049552917, + "num_tokens": 27639131.0, + "step": 1070 + }, + { + "epoch": 0.11761475949923128, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.554544448852539, + "learning_rate": 3.916544655929722e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6822199821472168, + "num_tokens": 27660915.0, + "step": 1071 + }, + { + "epoch": 0.11772457720184494, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.167696714401245, + "learning_rate": 3.920204978038067e-07, + "loss": 1.107, + "mean_token_accuracy": 0.6737392544746399, + "num_tokens": 27693707.0, + "step": 1072 + }, + { + "epoch": 0.1178343949044586, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.337773561477661, + "learning_rate": 3.923865300146413e-07, + "loss": 1.1862, + "mean_token_accuracy": 0.656191885471344, + "num_tokens": 27721535.0, + "step": 1073 + }, + { + "epoch": 0.11794421260707226, + "ewc_loss": 2.775341272354126e-07, + "grad_norm": 2.5526347160339355, + "learning_rate": 3.927525622254758e-07, + "loss": 0.9868, + "mean_token_accuracy": 0.7012396454811096, + "num_tokens": 27744277.0, + "step": 1074 + }, + { + "epoch": 0.11805403030968592, + "ewc_loss": 2.7939677238464355e-07, + "grad_norm": 2.235827684402466, + "learning_rate": 3.9311859443631043e-07, + "loss": 1.1022, + "mean_token_accuracy": 0.6749128103256226, + "num_tokens": 27773010.0, + "step": 1075 + }, + { + "epoch": 0.11816384801229958, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.7045788764953613, + "learning_rate": 3.934846266471449e-07, + "loss": 1.1288, + "mean_token_accuracy": 0.680859386920929, + "num_tokens": 27794162.0, + "step": 1076 + }, + { + "epoch": 0.11827366571491324, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.47088623046875, + "learning_rate": 3.9385065885797947e-07, + "loss": 1.0844, + "mean_token_accuracy": 0.6782389879226685, + "num_tokens": 27817334.0, + "step": 1077 + }, + { + "epoch": 0.1183834834175269, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.5157294273376465, + "learning_rate": 3.942166910688141e-07, + "loss": 1.1484, + "mean_token_accuracy": 0.6597185730934143, + "num_tokens": 27841083.0, + "step": 1078 + }, + { + "epoch": 0.11849330112014057, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.6560006141662598, + "learning_rate": 3.9458272327964857e-07, + "loss": 1.0135, + "mean_token_accuracy": 0.6925668716430664, + "num_tokens": 27861503.0, + "step": 1079 + }, + { + "epoch": 0.11860311882275423, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.331015110015869, + "learning_rate": 3.949487554904831e-07, + "loss": 1.0554, + "mean_token_accuracy": 0.6958146095275879, + "num_tokens": 27888130.0, + "step": 1080 + }, + { + "epoch": 0.1187129365253679, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.3803958892822266, + "learning_rate": 3.953147877013177e-07, + "loss": 1.0637, + "mean_token_accuracy": 0.6821346282958984, + "num_tokens": 27913796.0, + "step": 1081 + }, + { + "epoch": 0.11882275422798155, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.4618582725524902, + "learning_rate": 3.956808199121522e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.6811489462852478, + "num_tokens": 27938790.0, + "step": 1082 + }, + { + "epoch": 0.11893257193059521, + "ewc_loss": 2.812594175338745e-07, + "grad_norm": 2.3089168071746826, + "learning_rate": 3.960468521229868e-07, + "loss": 1.0608, + "mean_token_accuracy": 0.68440842628479, + "num_tokens": 27964042.0, + "step": 1083 + }, + { + "epoch": 0.11904238963320887, + "ewc_loss": 2.8312206268310547e-07, + "grad_norm": 2.330278158187866, + "learning_rate": 3.9641288433382137e-07, + "loss": 1.0787, + "mean_token_accuracy": 0.6869414448738098, + "num_tokens": 27990871.0, + "step": 1084 + }, + { + "epoch": 0.11915220733582253, + "ewc_loss": 2.8312206268310547e-07, + "grad_norm": 2.249501943588257, + "learning_rate": 3.967789165446559e-07, + "loss": 1.0788, + "mean_token_accuracy": 0.6727038621902466, + "num_tokens": 28019249.0, + "step": 1085 + }, + { + "epoch": 0.11926202503843619, + "ewc_loss": 2.8312206268310547e-07, + "grad_norm": 2.276423931121826, + "learning_rate": 3.9714494875549047e-07, + "loss": 1.0797, + "mean_token_accuracy": 0.674125075340271, + "num_tokens": 28047209.0, + "step": 1086 + }, + { + "epoch": 0.11937184274104985, + "ewc_loss": 2.8312206268310547e-07, + "grad_norm": 2.2182278633117676, + "learning_rate": 3.97510980966325e-07, + "loss": 1.1579, + "mean_token_accuracy": 0.6707438230514526, + "num_tokens": 28076909.0, + "step": 1087 + }, + { + "epoch": 0.11948166044366351, + "ewc_loss": 2.8312206268310547e-07, + "grad_norm": 2.2796356678009033, + "learning_rate": 3.9787701317715957e-07, + "loss": 1.1597, + "mean_token_accuracy": 0.6603395342826843, + "num_tokens": 28105066.0, + "step": 1088 + }, + { + "epoch": 0.11959147814627719, + "ewc_loss": 2.849847078323364e-07, + "grad_norm": 2.779376983642578, + "learning_rate": 3.982430453879941e-07, + "loss": 1.0631, + "mean_token_accuracy": 0.67694091796875, + "num_tokens": 28126388.0, + "step": 1089 + }, + { + "epoch": 0.11970129584889085, + "ewc_loss": 2.849847078323364e-07, + "grad_norm": 2.3864529132843018, + "learning_rate": 3.9860907759882866e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.692380428314209, + "num_tokens": 28152525.0, + "step": 1090 + }, + { + "epoch": 0.1198111135515045, + "ewc_loss": 2.849847078323364e-07, + "grad_norm": 2.4317362308502197, + "learning_rate": 3.9897510980966327e-07, + "loss": 0.993, + "mean_token_accuracy": 0.7057559490203857, + "num_tokens": 28176649.0, + "step": 1091 + }, + { + "epoch": 0.11992093125411817, + "ewc_loss": 2.849847078323364e-07, + "grad_norm": 2.81455659866333, + "learning_rate": 3.993411420204978e-07, + "loss": 0.9767, + "mean_token_accuracy": 0.7150428295135498, + "num_tokens": 28194986.0, + "step": 1092 + }, + { + "epoch": 0.12003074895673183, + "ewc_loss": 2.868473529815674e-07, + "grad_norm": 2.1926193237304688, + "learning_rate": 3.997071742313323e-07, + "loss": 1.0696, + "mean_token_accuracy": 0.6848346590995789, + "num_tokens": 28225155.0, + "step": 1093 + }, + { + "epoch": 0.12014056665934549, + "ewc_loss": 2.868473529815674e-07, + "grad_norm": 2.2487218379974365, + "learning_rate": 4.000732064421669e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6809661388397217, + "num_tokens": 28254599.0, + "step": 1094 + }, + { + "epoch": 0.12025038436195915, + "ewc_loss": 2.868473529815674e-07, + "grad_norm": 2.577089786529541, + "learning_rate": 4.0043923865300146e-07, + "loss": 1.0067, + "mean_token_accuracy": 0.6953848600387573, + "num_tokens": 28277178.0, + "step": 1095 + }, + { + "epoch": 0.1203602020645728, + "ewc_loss": 2.8870999813079834e-07, + "grad_norm": 2.4112043380737305, + "learning_rate": 4.0080527086383596e-07, + "loss": 1.0737, + "mean_token_accuracy": 0.6839895248413086, + "num_tokens": 28304301.0, + "step": 1096 + }, + { + "epoch": 0.12047001976718646, + "ewc_loss": 2.8870999813079834e-07, + "grad_norm": 2.315288782119751, + "learning_rate": 4.0117130307467056e-07, + "loss": 0.96, + "mean_token_accuracy": 0.7123810052871704, + "num_tokens": 28328572.0, + "step": 1097 + }, + { + "epoch": 0.12057983746980014, + "ewc_loss": 2.8870999813079834e-07, + "grad_norm": 2.5263028144836426, + "learning_rate": 4.015373352855051e-07, + "loss": 1.0519, + "mean_token_accuracy": 0.6878246665000916, + "num_tokens": 28352705.0, + "step": 1098 + }, + { + "epoch": 0.1206896551724138, + "ewc_loss": 2.8870999813079834e-07, + "grad_norm": 2.3919425010681152, + "learning_rate": 4.0190336749633966e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.6789271831512451, + "num_tokens": 28379284.0, + "step": 1099 + }, + { + "epoch": 0.12079947287502746, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 1.9819698333740234, + "learning_rate": 4.022693997071742e-07, + "loss": 1.0982, + "mean_token_accuracy": 0.6783109903335571, + "num_tokens": 28412910.0, + "step": 1100 + }, + { + "epoch": 0.12090929057764112, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.2150421142578125, + "learning_rate": 4.0263543191800876e-07, + "loss": 1.0163, + "mean_token_accuracy": 0.7034117579460144, + "num_tokens": 28439611.0, + "step": 1101 + }, + { + "epoch": 0.12101910828025478, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.4251856803894043, + "learning_rate": 4.0300146412884336e-07, + "loss": 1.1239, + "mean_token_accuracy": 0.6739267110824585, + "num_tokens": 28464781.0, + "step": 1102 + }, + { + "epoch": 0.12112892598286844, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.6105668544769287, + "learning_rate": 4.0336749633967785e-07, + "loss": 1.029, + "mean_token_accuracy": 0.6913641095161438, + "num_tokens": 28486031.0, + "step": 1103 + }, + { + "epoch": 0.1212387436854821, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.5680928230285645, + "learning_rate": 4.037335285505124e-07, + "loss": 0.9755, + "mean_token_accuracy": 0.7061038017272949, + "num_tokens": 28507282.0, + "step": 1104 + }, + { + "epoch": 0.12134856138809576, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.4518980979919434, + "learning_rate": 4.04099560761347e-07, + "loss": 1.056, + "mean_token_accuracy": 0.6821998357772827, + "num_tokens": 28532491.0, + "step": 1105 + }, + { + "epoch": 0.12145837909070942, + "ewc_loss": 2.905726432800293e-07, + "grad_norm": 2.215630054473877, + "learning_rate": 4.044655929721815e-07, + "loss": 1.0156, + "mean_token_accuracy": 0.7058395147323608, + "num_tokens": 28559439.0, + "step": 1106 + }, + { + "epoch": 0.12156819679332308, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.3353524208068848, + "learning_rate": 4.048316251830161e-07, + "loss": 1.1147, + "mean_token_accuracy": 0.6719588041305542, + "num_tokens": 28584914.0, + "step": 1107 + }, + { + "epoch": 0.12167801449593675, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.2341697216033936, + "learning_rate": 4.0519765739385065e-07, + "loss": 1.0782, + "mean_token_accuracy": 0.6861667633056641, + "num_tokens": 28613358.0, + "step": 1108 + }, + { + "epoch": 0.12178783219855041, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.4026594161987305, + "learning_rate": 4.055636896046852e-07, + "loss": 1.1093, + "mean_token_accuracy": 0.6860647797584534, + "num_tokens": 28637827.0, + "step": 1109 + }, + { + "epoch": 0.12189764990116407, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.338654041290283, + "learning_rate": 4.0592972181551975e-07, + "loss": 1.1619, + "mean_token_accuracy": 0.6633582711219788, + "num_tokens": 28663784.0, + "step": 1110 + }, + { + "epoch": 0.12200746760377773, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.4504997730255127, + "learning_rate": 4.062957540263543e-07, + "loss": 1.1238, + "mean_token_accuracy": 0.6689052581787109, + "num_tokens": 28686562.0, + "step": 1111 + }, + { + "epoch": 0.12211728530639139, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.428074598312378, + "learning_rate": 4.0666178623718885e-07, + "loss": 1.1198, + "mean_token_accuracy": 0.6788098812103271, + "num_tokens": 28713361.0, + "step": 1112 + }, + { + "epoch": 0.12222710300900505, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.5226051807403564, + "learning_rate": 4.070278184480234e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.689945638179779, + "num_tokens": 28736221.0, + "step": 1113 + }, + { + "epoch": 0.12233692071161871, + "ewc_loss": 2.9243528842926025e-07, + "grad_norm": 2.0618598461151123, + "learning_rate": 4.0739385065885795e-07, + "loss": 1.02, + "mean_token_accuracy": 0.6938720941543579, + "num_tokens": 28767480.0, + "step": 1114 + }, + { + "epoch": 0.12244673841423237, + "ewc_loss": 2.9616057872772217e-07, + "grad_norm": 2.6168911457061768, + "learning_rate": 4.0775988286969255e-07, + "loss": 1.0392, + "mean_token_accuracy": 0.6953830718994141, + "num_tokens": 28790622.0, + "step": 1115 + }, + { + "epoch": 0.12255655611684603, + "ewc_loss": 2.9616057872772217e-07, + "grad_norm": 2.2747583389282227, + "learning_rate": 4.081259150805271e-07, + "loss": 1.079, + "mean_token_accuracy": 0.69065260887146, + "num_tokens": 28816316.0, + "step": 1116 + }, + { + "epoch": 0.1226663738194597, + "ewc_loss": 2.9616057872772217e-07, + "grad_norm": 2.450950860977173, + "learning_rate": 4.084919472913616e-07, + "loss": 1.0499, + "mean_token_accuracy": 0.6872431039810181, + "num_tokens": 28838421.0, + "step": 1117 + }, + { + "epoch": 0.12277619152207336, + "ewc_loss": 2.9616057872772217e-07, + "grad_norm": 2.6115005016326904, + "learning_rate": 4.088579795021962e-07, + "loss": 1.1656, + "mean_token_accuracy": 0.6631600260734558, + "num_tokens": 28863527.0, + "step": 1118 + }, + { + "epoch": 0.12288600922468702, + "ewc_loss": 2.9616057872772217e-07, + "grad_norm": 2.716798782348633, + "learning_rate": 4.0922401171303075e-07, + "loss": 0.9153, + "mean_token_accuracy": 0.7247400283813477, + "num_tokens": 28882731.0, + "step": 1119 + }, + { + "epoch": 0.12299582692730068, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 2.552525520324707, + "learning_rate": 4.0959004392386524e-07, + "loss": 1.0841, + "mean_token_accuracy": 0.6782088279724121, + "num_tokens": 28906336.0, + "step": 1120 + }, + { + "epoch": 0.12310564462991434, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 2.5243377685546875, + "learning_rate": 4.0995607613469984e-07, + "loss": 0.9893, + "mean_token_accuracy": 0.7072951793670654, + "num_tokens": 28927718.0, + "step": 1121 + }, + { + "epoch": 0.123215462332528, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 2.32084059715271, + "learning_rate": 4.103221083455344e-07, + "loss": 1.1502, + "mean_token_accuracy": 0.6613510847091675, + "num_tokens": 28956062.0, + "step": 1122 + }, + { + "epoch": 0.12332528003514166, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.3938486576080322, + "learning_rate": 4.1068814055636894e-07, + "loss": 1.0912, + "mean_token_accuracy": 0.6809349656105042, + "num_tokens": 28980636.0, + "step": 1123 + }, + { + "epoch": 0.12343509773775532, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.4236724376678467, + "learning_rate": 4.110541727672035e-07, + "loss": 1.1204, + "mean_token_accuracy": 0.663216769695282, + "num_tokens": 29005838.0, + "step": 1124 + }, + { + "epoch": 0.12354491544036898, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.3096649646759033, + "learning_rate": 4.1142020497803804e-07, + "loss": 1.1352, + "mean_token_accuracy": 0.6629905104637146, + "num_tokens": 29032584.0, + "step": 1125 + }, + { + "epoch": 0.12365473314298264, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.5454678535461426, + "learning_rate": 4.1178623718887264e-07, + "loss": 1.0701, + "mean_token_accuracy": 0.6831828355789185, + "num_tokens": 29054249.0, + "step": 1126 + }, + { + "epoch": 0.12376455084559632, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 2.2155206203460693, + "learning_rate": 4.1215226939970714e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6779799461364746, + "num_tokens": 29083825.0, + "step": 1127 + }, + { + "epoch": 0.12387436854820998, + "ewc_loss": 2.980232238769531e-07, + "grad_norm": 2.1790337562561035, + "learning_rate": 4.125183016105417e-07, + "loss": 1.0941, + "mean_token_accuracy": 0.686408281326294, + "num_tokens": 29113619.0, + "step": 1128 + }, + { + "epoch": 0.12398418625082364, + "ewc_loss": 2.998858690261841e-07, + "grad_norm": 2.2646541595458984, + "learning_rate": 4.128843338213763e-07, + "loss": 1.1343, + "mean_token_accuracy": 0.6716299653053284, + "num_tokens": 29141354.0, + "step": 1129 + }, + { + "epoch": 0.1240940039534373, + "ewc_loss": 2.998858690261841e-07, + "grad_norm": 2.5253617763519287, + "learning_rate": 4.132503660322108e-07, + "loss": 1.0739, + "mean_token_accuracy": 0.6777341365814209, + "num_tokens": 29163113.0, + "step": 1130 + }, + { + "epoch": 0.12420382165605096, + "ewc_loss": 2.998858690261841e-07, + "grad_norm": 2.229062557220459, + "learning_rate": 4.136163982430454e-07, + "loss": 1.0303, + "mean_token_accuracy": 0.6932666301727295, + "num_tokens": 29190986.0, + "step": 1131 + }, + { + "epoch": 0.12431363935866462, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.1452219486236572, + "learning_rate": 4.1398243045387994e-07, + "loss": 0.9628, + "mean_token_accuracy": 0.7109065055847168, + "num_tokens": 29218126.0, + "step": 1132 + }, + { + "epoch": 0.12442345706127828, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.276613235473633, + "learning_rate": 4.143484626647145e-07, + "loss": 1.0822, + "mean_token_accuracy": 0.676474928855896, + "num_tokens": 29246177.0, + "step": 1133 + }, + { + "epoch": 0.12453327476389194, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.417905569076538, + "learning_rate": 4.1471449487554904e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6971054077148438, + "num_tokens": 29270132.0, + "step": 1134 + }, + { + "epoch": 0.1246430924665056, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.2071659564971924, + "learning_rate": 4.150805270863836e-07, + "loss": 1.1171, + "mean_token_accuracy": 0.6695162057876587, + "num_tokens": 29298680.0, + "step": 1135 + }, + { + "epoch": 0.12475291016911927, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.586236000061035, + "learning_rate": 4.1544655929721813e-07, + "loss": 1.0213, + "mean_token_accuracy": 0.6971168518066406, + "num_tokens": 29319341.0, + "step": 1136 + }, + { + "epoch": 0.12486272787173293, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.3258020877838135, + "learning_rate": 4.158125915080527e-07, + "loss": 1.0645, + "mean_token_accuracy": 0.6976291537284851, + "num_tokens": 29344645.0, + "step": 1137 + }, + { + "epoch": 0.12497254557434659, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.284945011138916, + "learning_rate": 4.1617862371888723e-07, + "loss": 1.0363, + "mean_token_accuracy": 0.6910718679428101, + "num_tokens": 29370995.0, + "step": 1138 + }, + { + "epoch": 0.12508236327696023, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.428189277648926, + "learning_rate": 4.1654465592972183e-07, + "loss": 1.1481, + "mean_token_accuracy": 0.6667397022247314, + "num_tokens": 29396692.0, + "step": 1139 + }, + { + "epoch": 0.1251921809795739, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.249202251434326, + "learning_rate": 4.1691068814055633e-07, + "loss": 1.0954, + "mean_token_accuracy": 0.6729789972305298, + "num_tokens": 29426669.0, + "step": 1140 + }, + { + "epoch": 0.12530199868218758, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.3310537338256836, + "learning_rate": 4.172767203513909e-07, + "loss": 1.0039, + "mean_token_accuracy": 0.699266791343689, + "num_tokens": 29451681.0, + "step": 1141 + }, + { + "epoch": 0.12541181638480123, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.027576446533203, + "learning_rate": 4.176427525622255e-07, + "loss": 1.1301, + "mean_token_accuracy": 0.6721799373626709, + "num_tokens": 29484702.0, + "step": 1142 + }, + { + "epoch": 0.1255216340874149, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.655876874923706, + "learning_rate": 4.1800878477306003e-07, + "loss": 0.9788, + "mean_token_accuracy": 0.7049292325973511, + "num_tokens": 29504341.0, + "step": 1143 + }, + { + "epoch": 0.12563145179002855, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.2875423431396484, + "learning_rate": 4.183748169838945e-07, + "loss": 1.0092, + "mean_token_accuracy": 0.6916366815567017, + "num_tokens": 29530130.0, + "step": 1144 + }, + { + "epoch": 0.12574126949264222, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.237114191055298, + "learning_rate": 4.1874084919472913e-07, + "loss": 1.1087, + "mean_token_accuracy": 0.6725659370422363, + "num_tokens": 29559283.0, + "step": 1145 + }, + { + "epoch": 0.12585108719525587, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.3923261165618896, + "learning_rate": 4.191068814055637e-07, + "loss": 1.1017, + "mean_token_accuracy": 0.6756297945976257, + "num_tokens": 29584911.0, + "step": 1146 + }, + { + "epoch": 0.12596090489786954, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.301988124847412, + "learning_rate": 4.194729136163982e-07, + "loss": 0.9868, + "mean_token_accuracy": 0.7018899321556091, + "num_tokens": 29610399.0, + "step": 1147 + }, + { + "epoch": 0.1260707226004832, + "ewc_loss": 3.0174851417541504e-07, + "grad_norm": 2.4534752368927, + "learning_rate": 4.198389458272328e-07, + "loss": 0.9572, + "mean_token_accuracy": 0.707625687122345, + "num_tokens": 29633032.0, + "step": 1148 + }, + { + "epoch": 0.12618054030309686, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 2.3708763122558594, + "learning_rate": 4.202049780380673e-07, + "loss": 1.0667, + "mean_token_accuracy": 0.679577112197876, + "num_tokens": 29657214.0, + "step": 1149 + }, + { + "epoch": 0.1262903580057105, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 2.0907669067382812, + "learning_rate": 4.205710102489019e-07, + "loss": 1.074, + "mean_token_accuracy": 0.6817725896835327, + "num_tokens": 29688013.0, + "step": 1150 + }, + { + "epoch": 0.12640017570832418, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 2.5629658699035645, + "learning_rate": 4.209370424597364e-07, + "loss": 1.0128, + "mean_token_accuracy": 0.6948901414871216, + "num_tokens": 29710451.0, + "step": 1151 + }, + { + "epoch": 0.12650999341093785, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 2.3891818523406982, + "learning_rate": 4.2130307467057097e-07, + "loss": 1.0868, + "mean_token_accuracy": 0.6773709058761597, + "num_tokens": 29735654.0, + "step": 1152 + }, + { + "epoch": 0.1266198111135515, + "ewc_loss": 3.03611159324646e-07, + "grad_norm": 2.368337869644165, + "learning_rate": 4.2166910688140557e-07, + "loss": 1.1351, + "mean_token_accuracy": 0.6621489524841309, + "num_tokens": 29764641.0, + "step": 1153 + }, + { + "epoch": 0.12672962881616517, + "ewc_loss": 3.073364496231079e-07, + "grad_norm": 2.460428237915039, + "learning_rate": 4.2203513909224007e-07, + "loss": 1.1448, + "mean_token_accuracy": 0.6695868968963623, + "num_tokens": 29790458.0, + "step": 1154 + }, + { + "epoch": 0.12683944651877882, + "ewc_loss": 3.073364496231079e-07, + "grad_norm": 2.39630389213562, + "learning_rate": 4.2240117130307467e-07, + "loss": 1.0513, + "mean_token_accuracy": 0.687741756439209, + "num_tokens": 29816208.0, + "step": 1155 + }, + { + "epoch": 0.1269492642213925, + "ewc_loss": 3.073364496231079e-07, + "grad_norm": 2.3760852813720703, + "learning_rate": 4.227672035139092e-07, + "loss": 1.0825, + "mean_token_accuracy": 0.6882033348083496, + "num_tokens": 29840722.0, + "step": 1156 + }, + { + "epoch": 0.12705908192400614, + "ewc_loss": 3.073364496231079e-07, + "grad_norm": 2.2721059322357178, + "learning_rate": 4.231332357247437e-07, + "loss": 1.1731, + "mean_token_accuracy": 0.6574234962463379, + "num_tokens": 29869138.0, + "step": 1157 + }, + { + "epoch": 0.1271688996266198, + "ewc_loss": 3.073364496231079e-07, + "grad_norm": 2.5248944759368896, + "learning_rate": 4.234992679355783e-07, + "loss": 1.0452, + "mean_token_accuracy": 0.6860103607177734, + "num_tokens": 29891804.0, + "step": 1158 + }, + { + "epoch": 0.12727871732923346, + "ewc_loss": 3.0919909477233887e-07, + "grad_norm": 2.3121337890625, + "learning_rate": 4.2386530014641287e-07, + "loss": 1.1321, + "mean_token_accuracy": 0.6738755702972412, + "num_tokens": 29918368.0, + "step": 1159 + }, + { + "epoch": 0.12738853503184713, + "ewc_loss": 3.0919909477233887e-07, + "grad_norm": 2.262883424758911, + "learning_rate": 4.242313323572474e-07, + "loss": 1.0934, + "mean_token_accuracy": 0.6830145120620728, + "num_tokens": 29947243.0, + "step": 1160 + }, + { + "epoch": 0.1274983527344608, + "ewc_loss": 3.110617399215698e-07, + "grad_norm": 2.193201780319214, + "learning_rate": 4.2459736456808197e-07, + "loss": 1.1248, + "mean_token_accuracy": 0.6719908118247986, + "num_tokens": 29976185.0, + "step": 1161 + }, + { + "epoch": 0.12760817043707445, + "ewc_loss": 3.110617399215698e-07, + "grad_norm": 2.229963541030884, + "learning_rate": 4.249633967789165e-07, + "loss": 1.0911, + "mean_token_accuracy": 0.6762000322341919, + "num_tokens": 30004893.0, + "step": 1162 + }, + { + "epoch": 0.12771798813968813, + "ewc_loss": 3.110617399215698e-07, + "grad_norm": 2.179255247116089, + "learning_rate": 4.253294289897511e-07, + "loss": 1.1823, + "mean_token_accuracy": 0.6574799418449402, + "num_tokens": 30033645.0, + "step": 1163 + }, + { + "epoch": 0.12782780584230177, + "ewc_loss": 3.110617399215698e-07, + "grad_norm": 2.52707576751709, + "learning_rate": 4.256954612005856e-07, + "loss": 1.0834, + "mean_token_accuracy": 0.6853324174880981, + "num_tokens": 30055734.0, + "step": 1164 + }, + { + "epoch": 0.12793762354491545, + "ewc_loss": 3.1478703022003174e-07, + "grad_norm": 2.539313554763794, + "learning_rate": 4.2606149341142016e-07, + "loss": 1.0044, + "mean_token_accuracy": 0.6952272057533264, + "num_tokens": 30076609.0, + "step": 1165 + }, + { + "epoch": 0.1280474412475291, + "ewc_loss": 3.1478703022003174e-07, + "grad_norm": 2.51570725440979, + "learning_rate": 4.2642752562225476e-07, + "loss": 1.1208, + "mean_token_accuracy": 0.6633903980255127, + "num_tokens": 30101695.0, + "step": 1166 + }, + { + "epoch": 0.12815725895014277, + "ewc_loss": 3.1478703022003174e-07, + "grad_norm": 2.3832199573516846, + "learning_rate": 4.267935578330893e-07, + "loss": 1.1278, + "mean_token_accuracy": 0.6735670566558838, + "num_tokens": 30128385.0, + "step": 1167 + }, + { + "epoch": 0.1282670766527564, + "ewc_loss": 3.1851232051849365e-07, + "grad_norm": 2.5795602798461914, + "learning_rate": 4.271595900439238e-07, + "loss": 0.9089, + "mean_token_accuracy": 0.7276938557624817, + "num_tokens": 30149434.0, + "step": 1168 + }, + { + "epoch": 0.12837689435537009, + "ewc_loss": 3.1851232051849365e-07, + "grad_norm": 2.2801525592803955, + "learning_rate": 4.275256222547584e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.696382999420166, + "num_tokens": 30175072.0, + "step": 1169 + }, + { + "epoch": 0.12848671205798376, + "ewc_loss": 3.203749656677246e-07, + "grad_norm": 2.183316707611084, + "learning_rate": 4.2789165446559296e-07, + "loss": 1.0785, + "mean_token_accuracy": 0.6836440563201904, + "num_tokens": 30204221.0, + "step": 1170 + }, + { + "epoch": 0.1285965297605974, + "ewc_loss": 3.203749656677246e-07, + "grad_norm": 2.214705467224121, + "learning_rate": 4.282576866764275e-07, + "loss": 1.0998, + "mean_token_accuracy": 0.6751775741577148, + "num_tokens": 30232624.0, + "step": 1171 + }, + { + "epoch": 0.12870634746321108, + "ewc_loss": 3.203749656677246e-07, + "grad_norm": 2.2381107807159424, + "learning_rate": 4.2862371888726206e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6859757900238037, + "num_tokens": 30259309.0, + "step": 1172 + }, + { + "epoch": 0.12881616516582473, + "ewc_loss": 3.203749656677246e-07, + "grad_norm": 2.2762486934661865, + "learning_rate": 4.289897510980966e-07, + "loss": 1.0786, + "mean_token_accuracy": 0.6853881478309631, + "num_tokens": 30286533.0, + "step": 1173 + }, + { + "epoch": 0.1289259828684384, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.436288356781006, + "learning_rate": 4.293557833089312e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.6883550882339478, + "num_tokens": 30311150.0, + "step": 1174 + }, + { + "epoch": 0.12903580057105205, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.4542715549468994, + "learning_rate": 4.297218155197657e-07, + "loss": 1.0055, + "mean_token_accuracy": 0.6964558959007263, + "num_tokens": 30334858.0, + "step": 1175 + }, + { + "epoch": 0.12914561827366572, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.5613229274749756, + "learning_rate": 4.3008784773060026e-07, + "loss": 0.9823, + "mean_token_accuracy": 0.7095253467559814, + "num_tokens": 30358502.0, + "step": 1176 + }, + { + "epoch": 0.12925543597627936, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.327855348587036, + "learning_rate": 4.3045387994143486e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.6703947186470032, + "num_tokens": 30384783.0, + "step": 1177 + }, + { + "epoch": 0.12936525367889304, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.442732095718384, + "learning_rate": 4.3081991215226935e-07, + "loss": 1.0847, + "mean_token_accuracy": 0.6897014379501343, + "num_tokens": 30408913.0, + "step": 1178 + }, + { + "epoch": 0.1294750713815067, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.369884490966797, + "learning_rate": 4.3118594436310396e-07, + "loss": 1.0378, + "mean_token_accuracy": 0.6940097808837891, + "num_tokens": 30433900.0, + "step": 1179 + }, + { + "epoch": 0.12958488908412036, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.4981629848480225, + "learning_rate": 4.315519765739385e-07, + "loss": 1.0018, + "mean_token_accuracy": 0.7037510871887207, + "num_tokens": 30461190.0, + "step": 1180 + }, + { + "epoch": 0.12969470678673403, + "ewc_loss": 3.2223761081695557e-07, + "grad_norm": 2.4492225646972656, + "learning_rate": 4.31918008784773e-07, + "loss": 1.1009, + "mean_token_accuracy": 0.675300657749176, + "num_tokens": 30489651.0, + "step": 1181 + }, + { + "epoch": 0.12980452448934768, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.368274688720703, + "learning_rate": 4.322840409956076e-07, + "loss": 1.0772, + "mean_token_accuracy": 0.6943134665489197, + "num_tokens": 30519052.0, + "step": 1182 + }, + { + "epoch": 0.12991434219196135, + "ewc_loss": 3.259629011154175e-07, + "grad_norm": 2.3047754764556885, + "learning_rate": 4.3265007320644215e-07, + "loss": 0.985, + "mean_token_accuracy": 0.7033411860466003, + "num_tokens": 30543441.0, + "step": 1183 + }, + { + "epoch": 0.130024159894575, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.474172353744507, + "learning_rate": 4.330161054172767e-07, + "loss": 1.1329, + "mean_token_accuracy": 0.6706018447875977, + "num_tokens": 30567401.0, + "step": 1184 + }, + { + "epoch": 0.13013397759718867, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.3312621116638184, + "learning_rate": 4.3338213762811125e-07, + "loss": 1.1313, + "mean_token_accuracy": 0.6666909456253052, + "num_tokens": 30594937.0, + "step": 1185 + }, + { + "epoch": 0.13024379529980232, + "ewc_loss": 3.241002559661865e-07, + "grad_norm": 2.1912639141082764, + "learning_rate": 4.337481698389458e-07, + "loss": 0.9848, + "mean_token_accuracy": 0.708782434463501, + "num_tokens": 30623647.0, + "step": 1186 + }, + { + "epoch": 0.130353613002416, + "ewc_loss": 3.259629011154175e-07, + "grad_norm": 2.2952768802642822, + "learning_rate": 4.341142020497804e-07, + "loss": 1.0933, + "mean_token_accuracy": 0.6830665469169617, + "num_tokens": 30650386.0, + "step": 1187 + }, + { + "epoch": 0.13046343070502964, + "ewc_loss": 3.259629011154175e-07, + "grad_norm": 2.563913106918335, + "learning_rate": 4.344802342606149e-07, + "loss": 1.0672, + "mean_token_accuracy": 0.6811067461967468, + "num_tokens": 30673298.0, + "step": 1188 + }, + { + "epoch": 0.1305732484076433, + "ewc_loss": 3.259629011154175e-07, + "grad_norm": 2.156575918197632, + "learning_rate": 4.3484626647144945e-07, + "loss": 1.0915, + "mean_token_accuracy": 0.6843942403793335, + "num_tokens": 30702162.0, + "step": 1189 + }, + { + "epoch": 0.13068306611025698, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.373262882232666, + "learning_rate": 4.3521229868228405e-07, + "loss": 1.1213, + "mean_token_accuracy": 0.6759973764419556, + "num_tokens": 30728125.0, + "step": 1190 + }, + { + "epoch": 0.13079288381287063, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.365495204925537, + "learning_rate": 4.355783308931186e-07, + "loss": 1.0018, + "mean_token_accuracy": 0.7013760805130005, + "num_tokens": 30753050.0, + "step": 1191 + }, + { + "epoch": 0.1309027015154843, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.4253761768341064, + "learning_rate": 4.359443631039531e-07, + "loss": 1.0523, + "mean_token_accuracy": 0.68732088804245, + "num_tokens": 30778703.0, + "step": 1192 + }, + { + "epoch": 0.13101251921809795, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.062438726425171, + "learning_rate": 4.363103953147877e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.686754584312439, + "num_tokens": 30810814.0, + "step": 1193 + }, + { + "epoch": 0.13112233692071162, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.2893824577331543, + "learning_rate": 4.3667642752562224e-07, + "loss": 1.0576, + "mean_token_accuracy": 0.6981539726257324, + "num_tokens": 30836427.0, + "step": 1194 + }, + { + "epoch": 0.13123215462332527, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.2139980792999268, + "learning_rate": 4.370424597364568e-07, + "loss": 1.1229, + "mean_token_accuracy": 0.6687817573547363, + "num_tokens": 30865532.0, + "step": 1195 + }, + { + "epoch": 0.13134197232593894, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.598695755004883, + "learning_rate": 4.3740849194729134e-07, + "loss": 1.0458, + "mean_token_accuracy": 0.6863061189651489, + "num_tokens": 30885416.0, + "step": 1196 + }, + { + "epoch": 0.1314517900285526, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.407085418701172, + "learning_rate": 4.377745241581259e-07, + "loss": 1.0195, + "mean_token_accuracy": 0.6917608976364136, + "num_tokens": 30909752.0, + "step": 1197 + }, + { + "epoch": 0.13156160773116626, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.7822372913360596, + "learning_rate": 4.381405563689605e-07, + "loss": 1.0023, + "mean_token_accuracy": 0.7020604610443115, + "num_tokens": 30928463.0, + "step": 1198 + }, + { + "epoch": 0.13167142543377994, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.628868341445923, + "learning_rate": 4.38506588579795e-07, + "loss": 1.0436, + "mean_token_accuracy": 0.6901741623878479, + "num_tokens": 30950612.0, + "step": 1199 + }, + { + "epoch": 0.13178124313639358, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.3042891025543213, + "learning_rate": 4.3887262079062954e-07, + "loss": 1.0413, + "mean_token_accuracy": 0.6972565650939941, + "num_tokens": 30975408.0, + "step": 1200 + }, + { + "epoch": 0.13189106083900726, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.4508254528045654, + "learning_rate": 4.3923865300146414e-07, + "loss": 0.9498, + "mean_token_accuracy": 0.7128008604049683, + "num_tokens": 30998382.0, + "step": 1201 + }, + { + "epoch": 0.1320008785416209, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.200939416885376, + "learning_rate": 4.3960468521229864e-07, + "loss": 1.0031, + "mean_token_accuracy": 0.7028053998947144, + "num_tokens": 31024645.0, + "step": 1202 + }, + { + "epoch": 0.13211069624423458, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.432567596435547, + "learning_rate": 4.3997071742313324e-07, + "loss": 1.0602, + "mean_token_accuracy": 0.6933074593544006, + "num_tokens": 31050392.0, + "step": 1203 + }, + { + "epoch": 0.13222051394684822, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.442427396774292, + "learning_rate": 4.403367496339678e-07, + "loss": 0.9948, + "mean_token_accuracy": 0.7022653818130493, + "num_tokens": 31072099.0, + "step": 1204 + }, + { + "epoch": 0.1323303316494619, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.2121546268463135, + "learning_rate": 4.407027818448023e-07, + "loss": 1.0306, + "mean_token_accuracy": 0.6892139911651611, + "num_tokens": 31099348.0, + "step": 1205 + }, + { + "epoch": 0.13244014935207554, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.3024327754974365, + "learning_rate": 4.410688140556369e-07, + "loss": 1.0303, + "mean_token_accuracy": 0.6903156042098999, + "num_tokens": 31127674.0, + "step": 1206 + }, + { + "epoch": 0.13254996705468922, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.6730008125305176, + "learning_rate": 4.4143484626647144e-07, + "loss": 1.0225, + "mean_token_accuracy": 0.7025163769721985, + "num_tokens": 31147826.0, + "step": 1207 + }, + { + "epoch": 0.1326597847573029, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.658405303955078, + "learning_rate": 4.41800878477306e-07, + "loss": 0.9704, + "mean_token_accuracy": 0.7067067623138428, + "num_tokens": 31168192.0, + "step": 1208 + }, + { + "epoch": 0.13276960245991654, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.779555559158325, + "learning_rate": 4.4216691068814053e-07, + "loss": 0.8644, + "mean_token_accuracy": 0.7248732447624207, + "num_tokens": 31186404.0, + "step": 1209 + }, + { + "epoch": 0.1328794201625302, + "ewc_loss": 3.2782554626464844e-07, + "grad_norm": 2.1357669830322266, + "learning_rate": 4.425329428989751e-07, + "loss": 1.118, + "mean_token_accuracy": 0.6651798486709595, + "num_tokens": 31216377.0, + "step": 1210 + }, + { + "epoch": 0.13298923786514386, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 2.4585909843444824, + "learning_rate": 4.428989751098097e-07, + "loss": 1.0709, + "mean_token_accuracy": 0.6822711229324341, + "num_tokens": 31241083.0, + "step": 1211 + }, + { + "epoch": 0.13309905556775753, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 2.319610834121704, + "learning_rate": 4.432650073206442e-07, + "loss": 1.1048, + "mean_token_accuracy": 0.6765035390853882, + "num_tokens": 31269362.0, + "step": 1212 + }, + { + "epoch": 0.13320887327037118, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 1.999867558479309, + "learning_rate": 4.4363103953147873e-07, + "loss": 1.0071, + "mean_token_accuracy": 0.696864128112793, + "num_tokens": 31301121.0, + "step": 1213 + }, + { + "epoch": 0.13331869097298485, + "ewc_loss": 3.296881914138794e-07, + "grad_norm": 2.6072354316711426, + "learning_rate": 4.4399707174231333e-07, + "loss": 1.1048, + "mean_token_accuracy": 0.6698843240737915, + "num_tokens": 31323850.0, + "step": 1214 + }, + { + "epoch": 0.1334285086755985, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 2.340765953063965, + "learning_rate": 4.443631039531479e-07, + "loss": 1.0896, + "mean_token_accuracy": 0.6852232217788696, + "num_tokens": 31349231.0, + "step": 1215 + }, + { + "epoch": 0.13353832637821217, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 2.7058463096618652, + "learning_rate": 4.447291361639824e-07, + "loss": 1.1013, + "mean_token_accuracy": 0.6779396533966064, + "num_tokens": 31368491.0, + "step": 1216 + }, + { + "epoch": 0.13364814408082584, + "ewc_loss": 3.3155083656311035e-07, + "grad_norm": 2.413104772567749, + "learning_rate": 4.45095168374817e-07, + "loss": 0.9682, + "mean_token_accuracy": 0.7034428715705872, + "num_tokens": 31392465.0, + "step": 1217 + }, + { + "epoch": 0.1337579617834395, + "ewc_loss": 3.334134817123413e-07, + "grad_norm": 2.442087173461914, + "learning_rate": 4.4546120058565153e-07, + "loss": 1.0733, + "mean_token_accuracy": 0.6766226291656494, + "num_tokens": 31417587.0, + "step": 1218 + }, + { + "epoch": 0.13386777948605316, + "ewc_loss": 3.334134817123413e-07, + "grad_norm": 2.0952632427215576, + "learning_rate": 4.458272327964861e-07, + "loss": 1.0264, + "mean_token_accuracy": 0.6965035200119019, + "num_tokens": 31448313.0, + "step": 1219 + }, + { + "epoch": 0.1339775971886668, + "ewc_loss": 3.334134817123413e-07, + "grad_norm": 2.41644024848938, + "learning_rate": 4.4619326500732063e-07, + "loss": 1.03, + "mean_token_accuracy": 0.6895737648010254, + "num_tokens": 31472179.0, + "step": 1220 + }, + { + "epoch": 0.13408741489128048, + "ewc_loss": 3.334134817123413e-07, + "grad_norm": 2.1838436126708984, + "learning_rate": 4.465592972181552e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6851388216018677, + "num_tokens": 31501837.0, + "step": 1221 + }, + { + "epoch": 0.13419723259389413, + "ewc_loss": 3.334134817123413e-07, + "grad_norm": 2.4749908447265625, + "learning_rate": 4.469253294289897e-07, + "loss": 1.0111, + "mean_token_accuracy": 0.6988595724105835, + "num_tokens": 31523276.0, + "step": 1222 + }, + { + "epoch": 0.1343070502965078, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.5048794746398926, + "learning_rate": 4.472913616398243e-07, + "loss": 1.1594, + "mean_token_accuracy": 0.6631960868835449, + "num_tokens": 31547514.0, + "step": 1223 + }, + { + "epoch": 0.13441686799912145, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.172142267227173, + "learning_rate": 4.476573938506588e-07, + "loss": 0.993, + "mean_token_accuracy": 0.6999794840812683, + "num_tokens": 31575007.0, + "step": 1224 + }, + { + "epoch": 0.13452668570173512, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.3259260654449463, + "learning_rate": 4.480234260614934e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.6889675855636597, + "num_tokens": 31600483.0, + "step": 1225 + }, + { + "epoch": 0.13463650340434877, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.48455810546875, + "learning_rate": 4.483894582723279e-07, + "loss": 1.0249, + "mean_token_accuracy": 0.699196994304657, + "num_tokens": 31622639.0, + "step": 1226 + }, + { + "epoch": 0.13474632110696244, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.4737465381622314, + "learning_rate": 4.487554904831625e-07, + "loss": 1.046, + "mean_token_accuracy": 0.6990828514099121, + "num_tokens": 31647351.0, + "step": 1227 + }, + { + "epoch": 0.13485613880957611, + "ewc_loss": 3.371387720108032e-07, + "grad_norm": 2.749929189682007, + "learning_rate": 4.4912152269399707e-07, + "loss": 1.1054, + "mean_token_accuracy": 0.6763304471969604, + "num_tokens": 31667220.0, + "step": 1228 + }, + { + "epoch": 0.13496595651218976, + "ewc_loss": 3.4086406230926514e-07, + "grad_norm": 2.542647361755371, + "learning_rate": 4.4948755490483157e-07, + "loss": 1.1163, + "mean_token_accuracy": 0.6794881820678711, + "num_tokens": 31689044.0, + "step": 1229 + }, + { + "epoch": 0.13507577421480343, + "ewc_loss": 3.4086406230926514e-07, + "grad_norm": 2.388230085372925, + "learning_rate": 4.4985358711566617e-07, + "loss": 1.2476, + "mean_token_accuracy": 0.6466350555419922, + "num_tokens": 31717794.0, + "step": 1230 + }, + { + "epoch": 0.13518559191741708, + "ewc_loss": 3.4086406230926514e-07, + "grad_norm": 2.3726449012756348, + "learning_rate": 4.502196193265007e-07, + "loss": 1.1119, + "mean_token_accuracy": 0.6730738878250122, + "num_tokens": 31743349.0, + "step": 1231 + }, + { + "epoch": 0.13529540962003075, + "ewc_loss": 3.4086406230926514e-07, + "grad_norm": 2.81247878074646, + "learning_rate": 4.5058565153733527e-07, + "loss": 0.9946, + "mean_token_accuracy": 0.695925235748291, + "num_tokens": 31761698.0, + "step": 1232 + }, + { + "epoch": 0.1354052273226444, + "ewc_loss": 3.427267074584961e-07, + "grad_norm": 2.885014057159424, + "learning_rate": 4.509516837481698e-07, + "loss": 0.99, + "mean_token_accuracy": 0.7070716023445129, + "num_tokens": 31780060.0, + "step": 1233 + }, + { + "epoch": 0.13551504502525807, + "ewc_loss": 3.427267074584961e-07, + "grad_norm": 2.141798973083496, + "learning_rate": 4.5131771595900437e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.7015175819396973, + "num_tokens": 31810081.0, + "step": 1234 + }, + { + "epoch": 0.13562486272787172, + "ewc_loss": 3.427267074584961e-07, + "grad_norm": 2.21370005607605, + "learning_rate": 4.5168374816983897e-07, + "loss": 1.0718, + "mean_token_accuracy": 0.6814737915992737, + "num_tokens": 31838983.0, + "step": 1235 + }, + { + "epoch": 0.1357346804304854, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.292619466781616, + "learning_rate": 4.5204978038067347e-07, + "loss": 1.08, + "mean_token_accuracy": 0.6874602437019348, + "num_tokens": 31864058.0, + "step": 1236 + }, + { + "epoch": 0.13584449813309907, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.39823579788208, + "learning_rate": 4.52415812591508e-07, + "loss": 1.0584, + "mean_token_accuracy": 0.6838827133178711, + "num_tokens": 31889614.0, + "step": 1237 + }, + { + "epoch": 0.1359543158357127, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.3567540645599365, + "learning_rate": 4.527818448023426e-07, + "loss": 1.1617, + "mean_token_accuracy": 0.6702755093574524, + "num_tokens": 31915897.0, + "step": 1238 + }, + { + "epoch": 0.1360641335383264, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.206468105316162, + "learning_rate": 4.531478770131771e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6918978095054626, + "num_tokens": 31943588.0, + "step": 1239 + }, + { + "epoch": 0.13617395124094003, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.2936465740203857, + "learning_rate": 4.5351390922401166e-07, + "loss": 1.0771, + "mean_token_accuracy": 0.6849609613418579, + "num_tokens": 31970532.0, + "step": 1240 + }, + { + "epoch": 0.1362837689435537, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.3054983615875244, + "learning_rate": 4.5387994143484626e-07, + "loss": 1.0418, + "mean_token_accuracy": 0.6964243054389954, + "num_tokens": 31996697.0, + "step": 1241 + }, + { + "epoch": 0.13639358664616735, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.6018893718719482, + "learning_rate": 4.542459736456808e-07, + "loss": 0.9954, + "mean_token_accuracy": 0.7062177062034607, + "num_tokens": 32018303.0, + "step": 1242 + }, + { + "epoch": 0.13650340434878103, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.1482040882110596, + "learning_rate": 4.5461200585651536e-07, + "loss": 1.1184, + "mean_token_accuracy": 0.6674323678016663, + "num_tokens": 32046571.0, + "step": 1243 + }, + { + "epoch": 0.13661322205139467, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.4314632415771484, + "learning_rate": 4.549780380673499e-07, + "loss": 1.1416, + "mean_token_accuracy": 0.6697298884391785, + "num_tokens": 32072281.0, + "step": 1244 + }, + { + "epoch": 0.13672303975400835, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.4124720096588135, + "learning_rate": 4.5534407027818446e-07, + "loss": 1.1776, + "mean_token_accuracy": 0.661048948764801, + "num_tokens": 32099451.0, + "step": 1245 + }, + { + "epoch": 0.13683285745662202, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.3644680976867676, + "learning_rate": 4.55710102489019e-07, + "loss": 1.043, + "mean_token_accuracy": 0.6924142837524414, + "num_tokens": 32124080.0, + "step": 1246 + }, + { + "epoch": 0.13694267515923567, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.2365829944610596, + "learning_rate": 4.5607613469985356e-07, + "loss": 1.145, + "mean_token_accuracy": 0.663707971572876, + "num_tokens": 32153434.0, + "step": 1247 + }, + { + "epoch": 0.13705249286184934, + "ewc_loss": 3.4458935260772705e-07, + "grad_norm": 2.5013933181762695, + "learning_rate": 4.564421669106881e-07, + "loss": 1.0212, + "mean_token_accuracy": 0.6974015235900879, + "num_tokens": 32176510.0, + "step": 1248 + }, + { + "epoch": 0.13716231056446299, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.059539794921875, + "learning_rate": 4.568081991215227e-07, + "loss": 1.0739, + "mean_token_accuracy": 0.6787518262863159, + "num_tokens": 32206030.0, + "step": 1249 + }, + { + "epoch": 0.13727212826707666, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.2664923667907715, + "learning_rate": 4.571742313323572e-07, + "loss": 1.0512, + "mean_token_accuracy": 0.6907749772071838, + "num_tokens": 32233255.0, + "step": 1250 + }, + { + "epoch": 0.1373819459696903, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.337721109390259, + "learning_rate": 4.575402635431918e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6941601037979126, + "num_tokens": 32257818.0, + "step": 1251 + }, + { + "epoch": 0.13749176367230398, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.222499132156372, + "learning_rate": 4.5790629575402636e-07, + "loss": 1.0455, + "mean_token_accuracy": 0.688899576663971, + "num_tokens": 32286826.0, + "step": 1252 + }, + { + "epoch": 0.13760158137491763, + "ewc_loss": 3.4831464290618896e-07, + "grad_norm": 2.447460651397705, + "learning_rate": 4.5827232796486085e-07, + "loss": 1.0526, + "mean_token_accuracy": 0.6863454580307007, + "num_tokens": 32310447.0, + "step": 1253 + }, + { + "epoch": 0.1377113990775313, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.332693338394165, + "learning_rate": 4.5863836017569545e-07, + "loss": 1.0958, + "mean_token_accuracy": 0.6914648413658142, + "num_tokens": 32335713.0, + "step": 1254 + }, + { + "epoch": 0.13782121678014497, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.276650905609131, + "learning_rate": 4.5900439238653e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.6883683204650879, + "num_tokens": 32361259.0, + "step": 1255 + }, + { + "epoch": 0.13793103448275862, + "ewc_loss": 3.4831464290618896e-07, + "grad_norm": 2.5905797481536865, + "learning_rate": 4.593704245973645e-07, + "loss": 1.0483, + "mean_token_accuracy": 0.6884999871253967, + "num_tokens": 32383138.0, + "step": 1256 + }, + { + "epoch": 0.1380408521853723, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.255748748779297, + "learning_rate": 4.597364568081991e-07, + "loss": 1.0493, + "mean_token_accuracy": 0.6863201856613159, + "num_tokens": 32409879.0, + "step": 1257 + }, + { + "epoch": 0.13815066988798594, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.6906869411468506, + "learning_rate": 4.6010248901903365e-07, + "loss": 1.0296, + "mean_token_accuracy": 0.6898888349533081, + "num_tokens": 32430039.0, + "step": 1258 + }, + { + "epoch": 0.1382604875905996, + "ewc_loss": 3.46451997756958e-07, + "grad_norm": 2.337177038192749, + "learning_rate": 4.6046852122986825e-07, + "loss": 1.0324, + "mean_token_accuracy": 0.69096839427948, + "num_tokens": 32456043.0, + "step": 1259 + }, + { + "epoch": 0.13837030529321326, + "ewc_loss": 3.501772880554199e-07, + "grad_norm": 2.0337421894073486, + "learning_rate": 4.6083455344070275e-07, + "loss": 1.1408, + "mean_token_accuracy": 0.6656345129013062, + "num_tokens": 32489703.0, + "step": 1260 + }, + { + "epoch": 0.13848012299582693, + "ewc_loss": 3.5390257835388184e-07, + "grad_norm": 1.945270299911499, + "learning_rate": 4.612005856515373e-07, + "loss": 1.0945, + "mean_token_accuracy": 0.6760909557342529, + "num_tokens": 32523481.0, + "step": 1261 + }, + { + "epoch": 0.13858994069844058, + "ewc_loss": 3.501772880554199e-07, + "grad_norm": 2.668523073196411, + "learning_rate": 4.615666178623719e-07, + "loss": 1.0313, + "mean_token_accuracy": 0.6948156356811523, + "num_tokens": 32544230.0, + "step": 1262 + }, + { + "epoch": 0.13869975840105425, + "ewc_loss": 3.501772880554199e-07, + "grad_norm": 2.6698412895202637, + "learning_rate": 4.619326500732064e-07, + "loss": 1.0026, + "mean_token_accuracy": 0.6983612179756165, + "num_tokens": 32563592.0, + "step": 1263 + }, + { + "epoch": 0.1388095761036679, + "ewc_loss": 3.5390257835388184e-07, + "grad_norm": 2.2898876667022705, + "learning_rate": 4.6229868228404095e-07, + "loss": 1.1587, + "mean_token_accuracy": 0.6728583574295044, + "num_tokens": 32592096.0, + "step": 1264 + }, + { + "epoch": 0.13891939380628157, + "ewc_loss": 3.520399332046509e-07, + "grad_norm": 2.4401583671569824, + "learning_rate": 4.6266471449487555e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.6907638311386108, + "num_tokens": 32615477.0, + "step": 1265 + }, + { + "epoch": 0.13902921150889525, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.5790069103240967, + "learning_rate": 4.630307467057101e-07, + "loss": 1.0234, + "mean_token_accuracy": 0.6931868195533752, + "num_tokens": 32637654.0, + "step": 1266 + }, + { + "epoch": 0.1391390292115089, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.3288254737854004, + "learning_rate": 4.6339677891654465e-07, + "loss": 0.9993, + "mean_token_accuracy": 0.6999181509017944, + "num_tokens": 32661199.0, + "step": 1267 + }, + { + "epoch": 0.13924884691412256, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.5272786617279053, + "learning_rate": 4.637628111273792e-07, + "loss": 1.0944, + "mean_token_accuracy": 0.6761510968208313, + "num_tokens": 32684302.0, + "step": 1268 + }, + { + "epoch": 0.1393586646167362, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.1687302589416504, + "learning_rate": 4.6412884333821374e-07, + "loss": 1.1302, + "mean_token_accuracy": 0.6670973300933838, + "num_tokens": 32711546.0, + "step": 1269 + }, + { + "epoch": 0.13946848231934988, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.1642544269561768, + "learning_rate": 4.644948755490483e-07, + "loss": 1.1447, + "mean_token_accuracy": 0.6644021272659302, + "num_tokens": 32741165.0, + "step": 1270 + }, + { + "epoch": 0.13957830002196353, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.0879249572753906, + "learning_rate": 4.6486090775988284e-07, + "loss": 1.0579, + "mean_token_accuracy": 0.6866640448570251, + "num_tokens": 32774701.0, + "step": 1271 + }, + { + "epoch": 0.1396881177245772, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.411409616470337, + "learning_rate": 4.652269399707174e-07, + "loss": 1.1436, + "mean_token_accuracy": 0.6668411493301392, + "num_tokens": 32798129.0, + "step": 1272 + }, + { + "epoch": 0.13979793542719085, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.066380500793457, + "learning_rate": 4.65592972181552e-07, + "loss": 1.0612, + "mean_token_accuracy": 0.6894196271896362, + "num_tokens": 32829323.0, + "step": 1273 + }, + { + "epoch": 0.13990775312980452, + "ewc_loss": 3.557652235031128e-07, + "grad_norm": 2.248445510864258, + "learning_rate": 4.659590043923865e-07, + "loss": 0.9854, + "mean_token_accuracy": 0.7054144144058228, + "num_tokens": 32856609.0, + "step": 1274 + }, + { + "epoch": 0.1400175708324182, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.3870649337768555, + "learning_rate": 4.663250366032211e-07, + "loss": 1.0388, + "mean_token_accuracy": 0.686050295829773, + "num_tokens": 32881060.0, + "step": 1275 + }, + { + "epoch": 0.14012738853503184, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.336240291595459, + "learning_rate": 4.6669106881405564e-07, + "loss": 1.0046, + "mean_token_accuracy": 0.7129158973693848, + "num_tokens": 32904678.0, + "step": 1276 + }, + { + "epoch": 0.14023720623764552, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.4900753498077393, + "learning_rate": 4.6705710102489014e-07, + "loss": 1.0139, + "mean_token_accuracy": 0.6909196376800537, + "num_tokens": 32928285.0, + "step": 1277 + }, + { + "epoch": 0.14034702394025916, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.4080939292907715, + "learning_rate": 4.6742313323572474e-07, + "loss": 1.0457, + "mean_token_accuracy": 0.6915991902351379, + "num_tokens": 32953717.0, + "step": 1278 + }, + { + "epoch": 0.14045684164287284, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.0037059783935547, + "learning_rate": 4.677891654465593e-07, + "loss": 1.0235, + "mean_token_accuracy": 0.6949365139007568, + "num_tokens": 32985152.0, + "step": 1279 + }, + { + "epoch": 0.14056665934548648, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.4417850971221924, + "learning_rate": 4.681551976573938e-07, + "loss": 1.0381, + "mean_token_accuracy": 0.6905432939529419, + "num_tokens": 33009420.0, + "step": 1280 + }, + { + "epoch": 0.14067647704810016, + "ewc_loss": 3.5762786865234375e-07, + "grad_norm": 2.1192104816436768, + "learning_rate": 4.685212298682284e-07, + "loss": 1.1672, + "mean_token_accuracy": 0.6542723178863525, + "num_tokens": 33042302.0, + "step": 1281 + }, + { + "epoch": 0.1407862947507138, + "ewc_loss": 3.594905138015747e-07, + "grad_norm": 2.078932523727417, + "learning_rate": 4.6888726207906293e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6833467483520508, + "num_tokens": 33072431.0, + "step": 1282 + }, + { + "epoch": 0.14089611245332748, + "ewc_loss": 3.594905138015747e-07, + "grad_norm": 2.2716848850250244, + "learning_rate": 4.6925329428989754e-07, + "loss": 1.1043, + "mean_token_accuracy": 0.6692934036254883, + "num_tokens": 33098031.0, + "step": 1283 + }, + { + "epoch": 0.14100593015594115, + "ewc_loss": 3.594905138015747e-07, + "grad_norm": 2.4003427028656006, + "learning_rate": 4.6961932650073203e-07, + "loss": 1.088, + "mean_token_accuracy": 0.6843992471694946, + "num_tokens": 33121574.0, + "step": 1284 + }, + { + "epoch": 0.1411157478585548, + "ewc_loss": 3.6135315895080566e-07, + "grad_norm": 2.5180904865264893, + "learning_rate": 4.699853587115666e-07, + "loss": 1.1012, + "mean_token_accuracy": 0.6830777525901794, + "num_tokens": 33145414.0, + "step": 1285 + }, + { + "epoch": 0.14122556556116847, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.5122194290161133, + "learning_rate": 4.703513909224012e-07, + "loss": 0.999, + "mean_token_accuracy": 0.6993212699890137, + "num_tokens": 33166555.0, + "step": 1286 + }, + { + "epoch": 0.14133538326378212, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.010274887084961, + "learning_rate": 4.707174231332357e-07, + "loss": 1.1243, + "mean_token_accuracy": 0.6703701019287109, + "num_tokens": 33200074.0, + "step": 1287 + }, + { + "epoch": 0.1414452009663958, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.667937994003296, + "learning_rate": 4.7108345534407023e-07, + "loss": 0.8997, + "mean_token_accuracy": 0.7295776009559631, + "num_tokens": 33220090.0, + "step": 1288 + }, + { + "epoch": 0.14155501866900944, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.226081371307373, + "learning_rate": 4.7144948755490483e-07, + "loss": 1.0176, + "mean_token_accuracy": 0.6891142725944519, + "num_tokens": 33247080.0, + "step": 1289 + }, + { + "epoch": 0.1416648363716231, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.0826449394226074, + "learning_rate": 4.718155197657394e-07, + "loss": 1.1153, + "mean_token_accuracy": 0.6777825355529785, + "num_tokens": 33276588.0, + "step": 1290 + }, + { + "epoch": 0.14177465407423676, + "ewc_loss": 3.632158041000366e-07, + "grad_norm": 2.6746954917907715, + "learning_rate": 4.7218155197657393e-07, + "loss": 0.9709, + "mean_token_accuracy": 0.7058554291725159, + "num_tokens": 33296291.0, + "step": 1291 + }, + { + "epoch": 0.14188447177685043, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.1563217639923096, + "learning_rate": 4.725475841874085e-07, + "loss": 1.1211, + "mean_token_accuracy": 0.6684224009513855, + "num_tokens": 33327544.0, + "step": 1292 + }, + { + "epoch": 0.1419942894794641, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.0002505779266357, + "learning_rate": 4.7291361639824303e-07, + "loss": 1.0467, + "mean_token_accuracy": 0.6921347379684448, + "num_tokens": 33360733.0, + "step": 1293 + }, + { + "epoch": 0.14210410718207775, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.2530508041381836, + "learning_rate": 4.732796486090776e-07, + "loss": 1.0174, + "mean_token_accuracy": 0.6979411840438843, + "num_tokens": 33386974.0, + "step": 1294 + }, + { + "epoch": 0.14221392488469142, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.452096939086914, + "learning_rate": 4.736456808199121e-07, + "loss": 1.0455, + "mean_token_accuracy": 0.6991584300994873, + "num_tokens": 33410360.0, + "step": 1295 + }, + { + "epoch": 0.14232374258730507, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.309957981109619, + "learning_rate": 4.740117130307467e-07, + "loss": 1.0447, + "mean_token_accuracy": 0.6886961460113525, + "num_tokens": 33435322.0, + "step": 1296 + }, + { + "epoch": 0.14243356028991874, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.2593679428100586, + "learning_rate": 4.743777452415813e-07, + "loss": 1.2055, + "mean_token_accuracy": 0.6538686752319336, + "num_tokens": 33464229.0, + "step": 1297 + }, + { + "epoch": 0.1425433779925324, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.498460054397583, + "learning_rate": 4.7474377745241577e-07, + "loss": 0.9519, + "mean_token_accuracy": 0.7190637588500977, + "num_tokens": 33485248.0, + "step": 1298 + }, + { + "epoch": 0.14265319569514606, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.2464468479156494, + "learning_rate": 4.751098096632504e-07, + "loss": 1.0543, + "mean_token_accuracy": 0.6868091225624084, + "num_tokens": 33516065.0, + "step": 1299 + }, + { + "epoch": 0.1427630133977597, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 1.990065574645996, + "learning_rate": 4.754758418740849e-07, + "loss": 0.9966, + "mean_token_accuracy": 0.7065740823745728, + "num_tokens": 33548790.0, + "step": 1300 + }, + { + "epoch": 0.14287283110037338, + "ewc_loss": 3.650784492492676e-07, + "grad_norm": 2.1775972843170166, + "learning_rate": 4.758418740849194e-07, + "loss": 1.0662, + "mean_token_accuracy": 0.6876839399337769, + "num_tokens": 33577167.0, + "step": 1301 + }, + { + "epoch": 0.14298264880298703, + "ewc_loss": 3.6694109439849854e-07, + "grad_norm": 2.447714328765869, + "learning_rate": 4.76207906295754e-07, + "loss": 1.0381, + "mean_token_accuracy": 0.6963865756988525, + "num_tokens": 33599933.0, + "step": 1302 + }, + { + "epoch": 0.1430924665056007, + "ewc_loss": 3.6694109439849854e-07, + "grad_norm": 2.5784566402435303, + "learning_rate": 4.7657393850658857e-07, + "loss": 0.9516, + "mean_token_accuracy": 0.7086700201034546, + "num_tokens": 33618648.0, + "step": 1303 + }, + { + "epoch": 0.14320228420821438, + "ewc_loss": 3.6694109439849854e-07, + "grad_norm": 2.2726967334747314, + "learning_rate": 4.769399707174231e-07, + "loss": 1.0233, + "mean_token_accuracy": 0.6930227875709534, + "num_tokens": 33643320.0, + "step": 1304 + }, + { + "epoch": 0.14331210191082802, + "ewc_loss": 3.6694109439849854e-07, + "grad_norm": 2.380638837814331, + "learning_rate": 4.773060029282577e-07, + "loss": 1.1087, + "mean_token_accuracy": 0.6686092615127563, + "num_tokens": 33667215.0, + "step": 1305 + }, + { + "epoch": 0.1434219196134417, + "ewc_loss": 3.6694109439849854e-07, + "grad_norm": 2.4496257305145264, + "learning_rate": 4.776720351390922e-07, + "loss": 1.1182, + "mean_token_accuracy": 0.6679407358169556, + "num_tokens": 33691314.0, + "step": 1306 + }, + { + "epoch": 0.14353173731605534, + "ewc_loss": 3.7066638469696045e-07, + "grad_norm": 2.4278759956359863, + "learning_rate": 4.780380673499268e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.682830810546875, + "num_tokens": 33716231.0, + "step": 1307 + }, + { + "epoch": 0.14364155501866901, + "ewc_loss": 3.7066638469696045e-07, + "grad_norm": 1.9541898965835571, + "learning_rate": 4.784040995607613e-07, + "loss": 1.0657, + "mean_token_accuracy": 0.6827534437179565, + "num_tokens": 33750223.0, + "step": 1308 + }, + { + "epoch": 0.14375137272128266, + "ewc_loss": 3.7066638469696045e-07, + "grad_norm": 2.4840235710144043, + "learning_rate": 4.787701317715958e-07, + "loss": 0.9959, + "mean_token_accuracy": 0.7058376669883728, + "num_tokens": 33772320.0, + "step": 1309 + }, + { + "epoch": 0.14386119042389633, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.5714776515960693, + "learning_rate": 4.791361639824304e-07, + "loss": 1.0454, + "mean_token_accuracy": 0.6903203725814819, + "num_tokens": 33793342.0, + "step": 1310 + }, + { + "epoch": 0.14397100812650998, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.5893313884735107, + "learning_rate": 4.79502196193265e-07, + "loss": 0.9613, + "mean_token_accuracy": 0.7056151032447815, + "num_tokens": 33813186.0, + "step": 1311 + }, + { + "epoch": 0.14408082582912365, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.273754596710205, + "learning_rate": 4.798682284040995e-07, + "loss": 1.01, + "mean_token_accuracy": 0.6915172934532166, + "num_tokens": 33840214.0, + "step": 1312 + }, + { + "epoch": 0.14419064353173733, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.5876705646514893, + "learning_rate": 4.802342606149341e-07, + "loss": 1.0281, + "mean_token_accuracy": 0.6892035007476807, + "num_tokens": 33863647.0, + "step": 1313 + }, + { + "epoch": 0.14430046123435097, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.0322976112365723, + "learning_rate": 4.806002928257686e-07, + "loss": 1.1009, + "mean_token_accuracy": 0.6787475347518921, + "num_tokens": 33896827.0, + "step": 1314 + }, + { + "epoch": 0.14441027893696465, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.407804250717163, + "learning_rate": 4.809663250366032e-07, + "loss": 1.0614, + "mean_token_accuracy": 0.6804494857788086, + "num_tokens": 33920697.0, + "step": 1315 + }, + { + "epoch": 0.1445200966395783, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.258113145828247, + "learning_rate": 4.813323572474377e-07, + "loss": 1.018, + "mean_token_accuracy": 0.6992595791816711, + "num_tokens": 33945998.0, + "step": 1316 + }, + { + "epoch": 0.14462991434219197, + "ewc_loss": 3.7439167499542236e-07, + "grad_norm": 2.448758602142334, + "learning_rate": 4.816983894582723e-07, + "loss": 1.0377, + "mean_token_accuracy": 0.6968326568603516, + "num_tokens": 33969471.0, + "step": 1317 + }, + { + "epoch": 0.1447397320448056, + "ewc_loss": 3.781169652938843e-07, + "grad_norm": 2.2887678146362305, + "learning_rate": 4.820644216691069e-07, + "loss": 1.0712, + "mean_token_accuracy": 0.6788781881332397, + "num_tokens": 33995373.0, + "step": 1318 + }, + { + "epoch": 0.1448495497474193, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.4973552227020264, + "learning_rate": 4.824304538799414e-07, + "loss": 0.9729, + "mean_token_accuracy": 0.7079947590827942, + "num_tokens": 34019470.0, + "step": 1319 + }, + { + "epoch": 0.14495936745003293, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.3011841773986816, + "learning_rate": 4.827964860907759e-07, + "loss": 1.0926, + "mean_token_accuracy": 0.6780270338058472, + "num_tokens": 34046792.0, + "step": 1320 + }, + { + "epoch": 0.1450691851526466, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.2740488052368164, + "learning_rate": 4.831625183016105e-07, + "loss": 1.0709, + "mean_token_accuracy": 0.6793798208236694, + "num_tokens": 34073908.0, + "step": 1321 + }, + { + "epoch": 0.14517900285526028, + "ewc_loss": 3.818422555923462e-07, + "grad_norm": 2.158437728881836, + "learning_rate": 4.835285505124451e-07, + "loss": 0.9326, + "mean_token_accuracy": 0.7145960927009583, + "num_tokens": 34101665.0, + "step": 1322 + }, + { + "epoch": 0.14528882055787393, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.3679471015930176, + "learning_rate": 4.838945827232796e-07, + "loss": 0.9592, + "mean_token_accuracy": 0.7102530002593994, + "num_tokens": 34123530.0, + "step": 1323 + }, + { + "epoch": 0.1453986382604876, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.342266798019409, + "learning_rate": 4.842606149341142e-07, + "loss": 1.0506, + "mean_token_accuracy": 0.6774305701255798, + "num_tokens": 34148038.0, + "step": 1324 + }, + { + "epoch": 0.14550845596310125, + "ewc_loss": 3.7997961044311523e-07, + "grad_norm": 2.398165464401245, + "learning_rate": 4.846266471449487e-07, + "loss": 1.0555, + "mean_token_accuracy": 0.6820700764656067, + "num_tokens": 34171039.0, + "step": 1325 + }, + { + "epoch": 0.14561827366571492, + "ewc_loss": 3.818422555923462e-07, + "grad_norm": 2.3134732246398926, + "learning_rate": 4.849926793557833e-07, + "loss": 1.0583, + "mean_token_accuracy": 0.6827102899551392, + "num_tokens": 34197557.0, + "step": 1326 + }, + { + "epoch": 0.14572809136832857, + "ewc_loss": 3.818422555923462e-07, + "grad_norm": 2.4879508018493652, + "learning_rate": 4.853587115666178e-07, + "loss": 0.9974, + "mean_token_accuracy": 0.697208046913147, + "num_tokens": 34219659.0, + "step": 1327 + }, + { + "epoch": 0.14583790907094224, + "ewc_loss": 3.8370490074157715e-07, + "grad_norm": 2.542975902557373, + "learning_rate": 4.857247437774524e-07, + "loss": 1.0617, + "mean_token_accuracy": 0.6816272139549255, + "num_tokens": 34241179.0, + "step": 1328 + }, + { + "epoch": 0.14594772677355589, + "ewc_loss": 3.8370490074157715e-07, + "grad_norm": 2.5356504917144775, + "learning_rate": 4.86090775988287e-07, + "loss": 0.9418, + "mean_token_accuracy": 0.7162377834320068, + "num_tokens": 34261000.0, + "step": 1329 + }, + { + "epoch": 0.14605754447616956, + "ewc_loss": 3.8370490074157715e-07, + "grad_norm": 2.445807933807373, + "learning_rate": 4.864568081991215e-07, + "loss": 1.049, + "mean_token_accuracy": 0.6859423518180847, + "num_tokens": 34285817.0, + "step": 1330 + }, + { + "epoch": 0.14616736217878323, + "ewc_loss": 3.855675458908081e-07, + "grad_norm": 2.1785833835601807, + "learning_rate": 4.868228404099561e-07, + "loss": 1.1329, + "mean_token_accuracy": 0.6649094820022583, + "num_tokens": 34315537.0, + "step": 1331 + }, + { + "epoch": 0.14627717988139688, + "ewc_loss": 3.855675458908081e-07, + "grad_norm": 2.331216812133789, + "learning_rate": 4.871888726207906e-07, + "loss": 1.0567, + "mean_token_accuracy": 0.6864405870437622, + "num_tokens": 34338741.0, + "step": 1332 + }, + { + "epoch": 0.14638699758401055, + "ewc_loss": 3.855675458908081e-07, + "grad_norm": 2.3982489109039307, + "learning_rate": 4.875549048316251e-07, + "loss": 1.0358, + "mean_token_accuracy": 0.6924277544021606, + "num_tokens": 34362546.0, + "step": 1333 + }, + { + "epoch": 0.1464968152866242, + "ewc_loss": 3.855675458908081e-07, + "grad_norm": 1.9907981157302856, + "learning_rate": 4.879209370424597e-07, + "loss": 1.1285, + "mean_token_accuracy": 0.6660377979278564, + "num_tokens": 34394885.0, + "step": 1334 + }, + { + "epoch": 0.14660663298923787, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.023299217224121, + "learning_rate": 4.882869692532943e-07, + "loss": 1.0773, + "mean_token_accuracy": 0.6740445494651794, + "num_tokens": 34425299.0, + "step": 1335 + }, + { + "epoch": 0.14671645069185152, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.3918330669403076, + "learning_rate": 4.886530014641288e-07, + "loss": 1.0356, + "mean_token_accuracy": 0.6922577619552612, + "num_tokens": 34448680.0, + "step": 1336 + }, + { + "epoch": 0.1468262683944652, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.28395676612854, + "learning_rate": 4.890190336749634e-07, + "loss": 1.0849, + "mean_token_accuracy": 0.6814135313034058, + "num_tokens": 34474831.0, + "step": 1337 + }, + { + "epoch": 0.14693608609707884, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.3462345600128174, + "learning_rate": 4.893850658857979e-07, + "loss": 0.9188, + "mean_token_accuracy": 0.7207353115081787, + "num_tokens": 34496472.0, + "step": 1338 + }, + { + "epoch": 0.1470459037996925, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.451141834259033, + "learning_rate": 4.897510980966325e-07, + "loss": 0.9868, + "mean_token_accuracy": 0.7138317823410034, + "num_tokens": 34519372.0, + "step": 1339 + }, + { + "epoch": 0.14715572150230616, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.288193702697754, + "learning_rate": 4.90117130307467e-07, + "loss": 1.0797, + "mean_token_accuracy": 0.6924029588699341, + "num_tokens": 34545661.0, + "step": 1340 + }, + { + "epoch": 0.14726553920491983, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.1701996326446533, + "learning_rate": 4.904831625183016e-07, + "loss": 1.1629, + "mean_token_accuracy": 0.6641274690628052, + "num_tokens": 34576126.0, + "step": 1341 + }, + { + "epoch": 0.1473753569075335, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.4138691425323486, + "learning_rate": 4.908491947291362e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6812481880187988, + "num_tokens": 34600086.0, + "step": 1342 + }, + { + "epoch": 0.14748517461014715, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.24820876121521, + "learning_rate": 4.912152269399707e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.7047157287597656, + "num_tokens": 34626401.0, + "step": 1343 + }, + { + "epoch": 0.14759499231276083, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.2749361991882324, + "learning_rate": 4.915812591508052e-07, + "loss": 1.0921, + "mean_token_accuracy": 0.6769858598709106, + "num_tokens": 34653201.0, + "step": 1344 + }, + { + "epoch": 0.14770481001537447, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.263550281524658, + "learning_rate": 4.919472913616398e-07, + "loss": 1.0748, + "mean_token_accuracy": 0.678353488445282, + "num_tokens": 34680307.0, + "step": 1345 + }, + { + "epoch": 0.14781462771798815, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.4798288345336914, + "learning_rate": 4.923133235724744e-07, + "loss": 1.0309, + "mean_token_accuracy": 0.6963150501251221, + "num_tokens": 34702531.0, + "step": 1346 + }, + { + "epoch": 0.1479244454206018, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.0447487831115723, + "learning_rate": 4.926793557833089e-07, + "loss": 1.0598, + "mean_token_accuracy": 0.6845646500587463, + "num_tokens": 34733708.0, + "step": 1347 + }, + { + "epoch": 0.14803426312321546, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.4055628776550293, + "learning_rate": 4.930453879941435e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.6763646006584167, + "num_tokens": 34758625.0, + "step": 1348 + }, + { + "epoch": 0.1481440808258291, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.2033567428588867, + "learning_rate": 4.93411420204978e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6952714920043945, + "num_tokens": 34786952.0, + "step": 1349 + }, + { + "epoch": 0.14825389852844278, + "ewc_loss": 3.8743019104003906e-07, + "grad_norm": 2.2337207794189453, + "learning_rate": 4.937774524158126e-07, + "loss": 1.1146, + "mean_token_accuracy": 0.6721510887145996, + "num_tokens": 34813181.0, + "step": 1350 + }, + { + "epoch": 0.14836371623105646, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.1040802001953125, + "learning_rate": 4.941434846266471e-07, + "loss": 1.0097, + "mean_token_accuracy": 0.6949975490570068, + "num_tokens": 34845742.0, + "step": 1351 + }, + { + "epoch": 0.1484735339336701, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.2669553756713867, + "learning_rate": 4.945095168374817e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6852606534957886, + "num_tokens": 34872515.0, + "step": 1352 + }, + { + "epoch": 0.14858335163628378, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.3749516010284424, + "learning_rate": 4.948755490483163e-07, + "loss": 1.111, + "mean_token_accuracy": 0.6678732633590698, + "num_tokens": 34899778.0, + "step": 1353 + }, + { + "epoch": 0.14869316933889742, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.249826192855835, + "learning_rate": 4.952415812591508e-07, + "loss": 0.995, + "mean_token_accuracy": 0.6944044232368469, + "num_tokens": 34926103.0, + "step": 1354 + }, + { + "epoch": 0.1488029870415111, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.1725168228149414, + "learning_rate": 4.956076134699854e-07, + "loss": 0.9651, + "mean_token_accuracy": 0.7076855897903442, + "num_tokens": 34952471.0, + "step": 1355 + }, + { + "epoch": 0.14891280474412474, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.1367719173431396, + "learning_rate": 4.959736456808199e-07, + "loss": 1.0939, + "mean_token_accuracy": 0.6713327765464783, + "num_tokens": 34981653.0, + "step": 1356 + }, + { + "epoch": 0.14902262244673842, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.4335904121398926, + "learning_rate": 4.963396778916544e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7088159918785095, + "num_tokens": 35004160.0, + "step": 1357 + }, + { + "epoch": 0.14913244014935206, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.1553189754486084, + "learning_rate": 4.96705710102489e-07, + "loss": 1.0759, + "mean_token_accuracy": 0.6849902868270874, + "num_tokens": 35033695.0, + "step": 1358 + }, + { + "epoch": 0.14924225785196574, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.2758948802948, + "learning_rate": 4.970717423133236e-07, + "loss": 1.0074, + "mean_token_accuracy": 0.6912810206413269, + "num_tokens": 35058960.0, + "step": 1359 + }, + { + "epoch": 0.1493520755545794, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.292804718017578, + "learning_rate": 4.974377745241581e-07, + "loss": 1.029, + "mean_token_accuracy": 0.693785548210144, + "num_tokens": 35082458.0, + "step": 1360 + }, + { + "epoch": 0.14946189325719306, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.1532301902770996, + "learning_rate": 4.978038067349927e-07, + "loss": 1.0706, + "mean_token_accuracy": 0.6908861398696899, + "num_tokens": 35111320.0, + "step": 1361 + }, + { + "epoch": 0.14957171095980673, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.491584300994873, + "learning_rate": 4.981698389458272e-07, + "loss": 1.0978, + "mean_token_accuracy": 0.6688292026519775, + "num_tokens": 35136034.0, + "step": 1362 + }, + { + "epoch": 0.14968152866242038, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.394618272781372, + "learning_rate": 4.985358711566618e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6921576261520386, + "num_tokens": 35160498.0, + "step": 1363 + }, + { + "epoch": 0.14979134636503405, + "ewc_loss": 3.8929283618927e-07, + "grad_norm": 2.0749292373657227, + "learning_rate": 4.989019033674963e-07, + "loss": 1.0785, + "mean_token_accuracy": 0.6839332580566406, + "num_tokens": 35191074.0, + "step": 1364 + }, + { + "epoch": 0.1499011640676477, + "ewc_loss": 3.91155481338501e-07, + "grad_norm": 2.329113245010376, + "learning_rate": 4.992679355783309e-07, + "loss": 1.0492, + "mean_token_accuracy": 0.6914902329444885, + "num_tokens": 35216079.0, + "step": 1365 + }, + { + "epoch": 0.15001098177026137, + "ewc_loss": 3.91155481338501e-07, + "grad_norm": 2.4596824645996094, + "learning_rate": 4.996339677891655e-07, + "loss": 1.0525, + "mean_token_accuracy": 0.6946204900741577, + "num_tokens": 35239149.0, + "step": 1366 + }, + { + "epoch": 0.15012079947287502, + "ewc_loss": 3.91155481338501e-07, + "grad_norm": 2.205862522125244, + "learning_rate": 5e-07, + "loss": 1.1017, + "mean_token_accuracy": 0.6785522699356079, + "num_tokens": 35267266.0, + "step": 1367 + }, + { + "epoch": 0.1502306171754887, + "ewc_loss": 3.91155481338501e-07, + "grad_norm": 2.206070899963379, + "learning_rate": 5.003660322108345e-07, + "loss": 1.1083, + "mean_token_accuracy": 0.6761370897293091, + "num_tokens": 35295774.0, + "step": 1368 + }, + { + "epoch": 0.15034043487810236, + "ewc_loss": 3.9301812648773193e-07, + "grad_norm": 2.669877529144287, + "learning_rate": 5.007320644216691e-07, + "loss": 1.0367, + "mean_token_accuracy": 0.688030481338501, + "num_tokens": 35315866.0, + "step": 1369 + }, + { + "epoch": 0.150450252580716, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.367973804473877, + "learning_rate": 5.010980966325037e-07, + "loss": 1.0713, + "mean_token_accuracy": 0.6881433725357056, + "num_tokens": 35340461.0, + "step": 1370 + }, + { + "epoch": 0.15056007028332968, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.0981626510620117, + "learning_rate": 5.014641288433382e-07, + "loss": 0.9817, + "mean_token_accuracy": 0.7033863067626953, + "num_tokens": 35367551.0, + "step": 1371 + }, + { + "epoch": 0.15066988798594333, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.137873649597168, + "learning_rate": 5.018301610541728e-07, + "loss": 1.1115, + "mean_token_accuracy": 0.6659072637557983, + "num_tokens": 35397737.0, + "step": 1372 + }, + { + "epoch": 0.150779705688557, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.4454615116119385, + "learning_rate": 5.021961932650073e-07, + "loss": 1.0272, + "mean_token_accuracy": 0.6937390565872192, + "num_tokens": 35422047.0, + "step": 1373 + }, + { + "epoch": 0.15088952339117065, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.1487343311309814, + "learning_rate": 5.025622254758418e-07, + "loss": 1.0491, + "mean_token_accuracy": 0.6884806156158447, + "num_tokens": 35450974.0, + "step": 1374 + }, + { + "epoch": 0.15099934109378432, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.1779191493988037, + "learning_rate": 5.029282576866764e-07, + "loss": 0.9849, + "mean_token_accuracy": 0.7041106820106506, + "num_tokens": 35478196.0, + "step": 1375 + }, + { + "epoch": 0.15110915879639797, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.0923573970794678, + "learning_rate": 5.03294289897511e-07, + "loss": 1.1129, + "mean_token_accuracy": 0.6687540411949158, + "num_tokens": 35510077.0, + "step": 1376 + }, + { + "epoch": 0.15121897649901164, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.272195816040039, + "learning_rate": 5.036603221083456e-07, + "loss": 1.0611, + "mean_token_accuracy": 0.6844807863235474, + "num_tokens": 35538238.0, + "step": 1377 + }, + { + "epoch": 0.1513287942016253, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.153862237930298, + "learning_rate": 5.040263543191801e-07, + "loss": 1.1084, + "mean_token_accuracy": 0.6695359945297241, + "num_tokens": 35569282.0, + "step": 1378 + }, + { + "epoch": 0.15143861190423896, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.343590021133423, + "learning_rate": 5.043923865300146e-07, + "loss": 1.1516, + "mean_token_accuracy": 0.672206699848175, + "num_tokens": 35595162.0, + "step": 1379 + }, + { + "epoch": 0.15154842960685264, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.3194451332092285, + "learning_rate": 5.047584187408492e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.6981818675994873, + "num_tokens": 35620173.0, + "step": 1380 + }, + { + "epoch": 0.15165824730946628, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.5476975440979004, + "learning_rate": 5.051244509516838e-07, + "loss": 0.9836, + "mean_token_accuracy": 0.7041811943054199, + "num_tokens": 35641699.0, + "step": 1381 + }, + { + "epoch": 0.15176806501207996, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.163841485977173, + "learning_rate": 5.054904831625183e-07, + "loss": 1.0926, + "mean_token_accuracy": 0.6737962961196899, + "num_tokens": 35672048.0, + "step": 1382 + }, + { + "epoch": 0.1518778827146936, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.3990159034729004, + "learning_rate": 5.058565153733529e-07, + "loss": 1.0878, + "mean_token_accuracy": 0.6790768504142761, + "num_tokens": 35697483.0, + "step": 1383 + }, + { + "epoch": 0.15198770041730728, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.2629013061523438, + "learning_rate": 5.062225475841874e-07, + "loss": 1.058, + "mean_token_accuracy": 0.6970082521438599, + "num_tokens": 35723408.0, + "step": 1384 + }, + { + "epoch": 0.15209751811992092, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.324385404586792, + "learning_rate": 5.065885797950219e-07, + "loss": 1.0286, + "mean_token_accuracy": 0.6912847757339478, + "num_tokens": 35747189.0, + "step": 1385 + }, + { + "epoch": 0.1522073358225346, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.1059255599975586, + "learning_rate": 5.069546120058566e-07, + "loss": 1.0572, + "mean_token_accuracy": 0.6826111078262329, + "num_tokens": 35777084.0, + "step": 1386 + }, + { + "epoch": 0.15231715352514824, + "ewc_loss": 3.948807716369629e-07, + "grad_norm": 2.3511855602264404, + "learning_rate": 5.073206442166911e-07, + "loss": 0.9718, + "mean_token_accuracy": 0.7096063494682312, + "num_tokens": 35801318.0, + "step": 1387 + }, + { + "epoch": 0.15242697122776191, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.775477409362793, + "learning_rate": 5.076866764275256e-07, + "loss": 0.9902, + "mean_token_accuracy": 0.705070972442627, + "num_tokens": 35820852.0, + "step": 1388 + }, + { + "epoch": 0.1525367889303756, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.3143362998962402, + "learning_rate": 5.080527086383602e-07, + "loss": 1.0807, + "mean_token_accuracy": 0.6767557859420776, + "num_tokens": 35846653.0, + "step": 1389 + }, + { + "epoch": 0.15264660663298923, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.069141149520874, + "learning_rate": 5.084187408491947e-07, + "loss": 1.0867, + "mean_token_accuracy": 0.6756572723388672, + "num_tokens": 35877674.0, + "step": 1390 + }, + { + "epoch": 0.1527564243356029, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.083481550216675, + "learning_rate": 5.087847730600292e-07, + "loss": 1.08, + "mean_token_accuracy": 0.6804969310760498, + "num_tokens": 35908450.0, + "step": 1391 + }, + { + "epoch": 0.15286624203821655, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.246046304702759, + "learning_rate": 5.091508052708639e-07, + "loss": 1.0436, + "mean_token_accuracy": 0.6928403973579407, + "num_tokens": 35934858.0, + "step": 1392 + }, + { + "epoch": 0.15297605974083023, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.245915174484253, + "learning_rate": 5.095168374816984e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.6864023208618164, + "num_tokens": 35962276.0, + "step": 1393 + }, + { + "epoch": 0.15308587744344387, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.548100471496582, + "learning_rate": 5.09882869692533e-07, + "loss": 1.172, + "mean_token_accuracy": 0.6628578901290894, + "num_tokens": 35987873.0, + "step": 1394 + }, + { + "epoch": 0.15319569514605755, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.401832103729248, + "learning_rate": 5.102489019033675e-07, + "loss": 1.0454, + "mean_token_accuracy": 0.6862434148788452, + "num_tokens": 36013354.0, + "step": 1395 + }, + { + "epoch": 0.1533055128486712, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.226619005203247, + "learning_rate": 5.10614934114202e-07, + "loss": 1.0622, + "mean_token_accuracy": 0.682218074798584, + "num_tokens": 36040211.0, + "step": 1396 + }, + { + "epoch": 0.15341533055128487, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.244694709777832, + "learning_rate": 5.109809663250367e-07, + "loss": 1.0899, + "mean_token_accuracy": 0.6770281791687012, + "num_tokens": 36066374.0, + "step": 1397 + }, + { + "epoch": 0.15352514825389854, + "ewc_loss": 3.9674341678619385e-07, + "grad_norm": 2.2722277641296387, + "learning_rate": 5.113469985358712e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.7001955509185791, + "num_tokens": 36092114.0, + "step": 1398 + }, + { + "epoch": 0.1536349659565122, + "ewc_loss": 3.986060619354248e-07, + "grad_norm": 2.4249181747436523, + "learning_rate": 5.117130307467057e-07, + "loss": 1.0507, + "mean_token_accuracy": 0.689433217048645, + "num_tokens": 36117150.0, + "step": 1399 + }, + { + "epoch": 0.15374478365912586, + "ewc_loss": 3.986060619354248e-07, + "grad_norm": 2.113125801086426, + "learning_rate": 5.120790629575403e-07, + "loss": 1.1444, + "mean_token_accuracy": 0.6675103902816772, + "num_tokens": 36146676.0, + "step": 1400 + }, + { + "epoch": 0.1538546013617395, + "ewc_loss": 3.986060619354248e-07, + "grad_norm": 2.3464279174804688, + "learning_rate": 5.124450951683748e-07, + "loss": 1.0988, + "mean_token_accuracy": 0.67307049036026, + "num_tokens": 36173244.0, + "step": 1401 + }, + { + "epoch": 0.15396441906435318, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.1573054790496826, + "learning_rate": 5.128111273792094e-07, + "loss": 1.1008, + "mean_token_accuracy": 0.6777298450469971, + "num_tokens": 36203197.0, + "step": 1402 + }, + { + "epoch": 0.15407423676696683, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.2143101692199707, + "learning_rate": 5.13177159590044e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.673502504825592, + "num_tokens": 36231276.0, + "step": 1403 + }, + { + "epoch": 0.1541840544695805, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.067795753479004, + "learning_rate": 5.135431918008785e-07, + "loss": 1.0664, + "mean_token_accuracy": 0.6916633248329163, + "num_tokens": 36261923.0, + "step": 1404 + }, + { + "epoch": 0.15429387217219415, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.391885757446289, + "learning_rate": 5.13909224011713e-07, + "loss": 0.9809, + "mean_token_accuracy": 0.7106058597564697, + "num_tokens": 36285200.0, + "step": 1405 + }, + { + "epoch": 0.15440368987480782, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.264868974685669, + "learning_rate": 5.142752562225476e-07, + "loss": 1.0682, + "mean_token_accuracy": 0.6798052191734314, + "num_tokens": 36310732.0, + "step": 1406 + }, + { + "epoch": 0.1545135075774215, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.446103096008301, + "learning_rate": 5.14641288433382e-07, + "loss": 0.9954, + "mean_token_accuracy": 0.70147305727005, + "num_tokens": 36333833.0, + "step": 1407 + }, + { + "epoch": 0.15462332528003514, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.3469834327697754, + "learning_rate": 5.150073206442168e-07, + "loss": 1.0063, + "mean_token_accuracy": 0.7036315202713013, + "num_tokens": 36357548.0, + "step": 1408 + }, + { + "epoch": 0.1547331429826488, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.1215202808380127, + "learning_rate": 5.153733528550513e-07, + "loss": 0.964, + "mean_token_accuracy": 0.7070521712303162, + "num_tokens": 36386854.0, + "step": 1409 + }, + { + "epoch": 0.15484296068526246, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.011674165725708, + "learning_rate": 5.157393850658857e-07, + "loss": 1.1585, + "mean_token_accuracy": 0.6621065139770508, + "num_tokens": 36425311.0, + "step": 1410 + }, + { + "epoch": 0.15495277838787613, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.0742850303649902, + "learning_rate": 5.161054172767203e-07, + "loss": 0.9745, + "mean_token_accuracy": 0.7016199827194214, + "num_tokens": 36453543.0, + "step": 1411 + }, + { + "epoch": 0.15506259609048978, + "ewc_loss": 4.023313522338867e-07, + "grad_norm": 2.358114242553711, + "learning_rate": 5.164714494875548e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6879482269287109, + "num_tokens": 36479897.0, + "step": 1412 + }, + { + "epoch": 0.15517241379310345, + "ewc_loss": 4.041939973831177e-07, + "grad_norm": 2.317819118499756, + "learning_rate": 5.168374816983894e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.6820801496505737, + "num_tokens": 36508596.0, + "step": 1413 + }, + { + "epoch": 0.1552822314957171, + "ewc_loss": 4.041939973831177e-07, + "grad_norm": 2.4716484546661377, + "learning_rate": 5.17203513909224e-07, + "loss": 1.0727, + "mean_token_accuracy": 0.6786872148513794, + "num_tokens": 36532995.0, + "step": 1414 + }, + { + "epoch": 0.15539204919833077, + "ewc_loss": 4.041939973831177e-07, + "grad_norm": 2.1276473999023438, + "learning_rate": 5.175695461200585e-07, + "loss": 1.0825, + "mean_token_accuracy": 0.6786041259765625, + "num_tokens": 36565082.0, + "step": 1415 + }, + { + "epoch": 0.15550186690094442, + "ewc_loss": 4.0605664253234863e-07, + "grad_norm": 2.2295825481414795, + "learning_rate": 5.17935578330893e-07, + "loss": 1.1599, + "mean_token_accuracy": 0.6558036208152771, + "num_tokens": 36595531.0, + "step": 1416 + }, + { + "epoch": 0.1556116846035581, + "ewc_loss": 4.0605664253234863e-07, + "grad_norm": 2.413261890411377, + "learning_rate": 5.183016105417276e-07, + "loss": 1.0152, + "mean_token_accuracy": 0.7000076174736023, + "num_tokens": 36618337.0, + "step": 1417 + }, + { + "epoch": 0.15572150230617177, + "ewc_loss": 4.0605664253234863e-07, + "grad_norm": 2.5430643558502197, + "learning_rate": 5.186676427525622e-07, + "loss": 1.1123, + "mean_token_accuracy": 0.6701910495758057, + "num_tokens": 36641123.0, + "step": 1418 + }, + { + "epoch": 0.1558313200087854, + "ewc_loss": 4.0605664253234863e-07, + "grad_norm": 2.3288137912750244, + "learning_rate": 5.190336749633967e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6914616823196411, + "num_tokens": 36667870.0, + "step": 1419 + }, + { + "epoch": 0.15594113771139909, + "ewc_loss": 4.079192876815796e-07, + "grad_norm": 2.1815452575683594, + "learning_rate": 5.193997071742313e-07, + "loss": 1.0464, + "mean_token_accuracy": 0.6890373826026917, + "num_tokens": 36697458.0, + "step": 1420 + }, + { + "epoch": 0.15605095541401273, + "ewc_loss": 4.079192876815796e-07, + "grad_norm": 2.1332123279571533, + "learning_rate": 5.197657393850658e-07, + "loss": 1.1005, + "mean_token_accuracy": 0.6777047514915466, + "num_tokens": 36728999.0, + "step": 1421 + }, + { + "epoch": 0.1561607731166264, + "ewc_loss": 4.079192876815796e-07, + "grad_norm": 2.82550048828125, + "learning_rate": 5.201317715959003e-07, + "loss": 1.1375, + "mean_token_accuracy": 0.6674492955207825, + "num_tokens": 36751360.0, + "step": 1422 + }, + { + "epoch": 0.15627059081924005, + "ewc_loss": 4.079192876815796e-07, + "grad_norm": 2.2742574214935303, + "learning_rate": 5.204978038067349e-07, + "loss": 1.0946, + "mean_token_accuracy": 0.6809582114219666, + "num_tokens": 36778609.0, + "step": 1423 + }, + { + "epoch": 0.15638040852185373, + "ewc_loss": 4.0978193283081055e-07, + "grad_norm": 2.3463029861450195, + "learning_rate": 5.208638360175695e-07, + "loss": 1.0162, + "mean_token_accuracy": 0.6923478841781616, + "num_tokens": 36801124.0, + "step": 1424 + }, + { + "epoch": 0.15649022622446737, + "ewc_loss": 4.0978193283081055e-07, + "grad_norm": 2.3322715759277344, + "learning_rate": 5.212298682284041e-07, + "loss": 1.0967, + "mean_token_accuracy": 0.6731792688369751, + "num_tokens": 36826030.0, + "step": 1425 + }, + { + "epoch": 0.15660004392708105, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.209386110305786, + "learning_rate": 5.215959004392386e-07, + "loss": 1.1202, + "mean_token_accuracy": 0.6699249744415283, + "num_tokens": 36857373.0, + "step": 1426 + }, + { + "epoch": 0.15670986162969472, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.341738224029541, + "learning_rate": 5.219619326500731e-07, + "loss": 1.0214, + "mean_token_accuracy": 0.6969302892684937, + "num_tokens": 36882743.0, + "step": 1427 + }, + { + "epoch": 0.15681967933230836, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.2337422370910645, + "learning_rate": 5.223279648609077e-07, + "loss": 1.0356, + "mean_token_accuracy": 0.6902812123298645, + "num_tokens": 36908317.0, + "step": 1428 + }, + { + "epoch": 0.15692949703492204, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.372529983520508, + "learning_rate": 5.226939970717423e-07, + "loss": 1.0684, + "mean_token_accuracy": 0.6784020662307739, + "num_tokens": 36933934.0, + "step": 1429 + }, + { + "epoch": 0.15703931473753568, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.1968986988067627, + "learning_rate": 5.230600292825768e-07, + "loss": 1.006, + "mean_token_accuracy": 0.6969850063323975, + "num_tokens": 36961658.0, + "step": 1430 + }, + { + "epoch": 0.15714913244014936, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.2884068489074707, + "learning_rate": 5.234260614934114e-07, + "loss": 1.0354, + "mean_token_accuracy": 0.6854759454727173, + "num_tokens": 36989890.0, + "step": 1431 + }, + { + "epoch": 0.157258950142763, + "ewc_loss": 4.1350722312927246e-07, + "grad_norm": 2.5071940422058105, + "learning_rate": 5.237920937042459e-07, + "loss": 1.1269, + "mean_token_accuracy": 0.6712466478347778, + "num_tokens": 37014359.0, + "step": 1432 + }, + { + "epoch": 0.15736876784537668, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.44963002204895, + "learning_rate": 5.241581259150804e-07, + "loss": 1.0332, + "mean_token_accuracy": 0.6929571032524109, + "num_tokens": 37036466.0, + "step": 1433 + }, + { + "epoch": 0.15747858554799032, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.6143555641174316, + "learning_rate": 5.245241581259151e-07, + "loss": 0.9807, + "mean_token_accuracy": 0.7055132985115051, + "num_tokens": 37056557.0, + "step": 1434 + }, + { + "epoch": 0.157588403250604, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.3989529609680176, + "learning_rate": 5.248901903367496e-07, + "loss": 1.0949, + "mean_token_accuracy": 0.6745932102203369, + "num_tokens": 37079546.0, + "step": 1435 + }, + { + "epoch": 0.15769822095321767, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.527090311050415, + "learning_rate": 5.252562225475841e-07, + "loss": 1.1211, + "mean_token_accuracy": 0.668968141078949, + "num_tokens": 37103143.0, + "step": 1436 + }, + { + "epoch": 0.15780803865583132, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.3068692684173584, + "learning_rate": 5.256222547584187e-07, + "loss": 0.9865, + "mean_token_accuracy": 0.7207356691360474, + "num_tokens": 37128958.0, + "step": 1437 + }, + { + "epoch": 0.157917856358445, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.603882074356079, + "learning_rate": 5.259882869692532e-07, + "loss": 0.9805, + "mean_token_accuracy": 0.706449031829834, + "num_tokens": 37148862.0, + "step": 1438 + }, + { + "epoch": 0.15802767406105864, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.5776290893554688, + "learning_rate": 5.263543191800877e-07, + "loss": 0.9406, + "mean_token_accuracy": 0.7124354839324951, + "num_tokens": 37168500.0, + "step": 1439 + }, + { + "epoch": 0.1581374917636723, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.4955642223358154, + "learning_rate": 5.267203513909224e-07, + "loss": 1.0643, + "mean_token_accuracy": 0.6817752122879028, + "num_tokens": 37191202.0, + "step": 1440 + }, + { + "epoch": 0.15824730946628596, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.4530575275421143, + "learning_rate": 5.270863836017569e-07, + "loss": 1.1209, + "mean_token_accuracy": 0.6645269393920898, + "num_tokens": 37215939.0, + "step": 1441 + }, + { + "epoch": 0.15835712716889963, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.189472198486328, + "learning_rate": 5.274524158125915e-07, + "loss": 1.2379, + "mean_token_accuracy": 0.6605477333068848, + "num_tokens": 37246502.0, + "step": 1442 + }, + { + "epoch": 0.15846694487151328, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.474318027496338, + "learning_rate": 5.27818448023426e-07, + "loss": 0.943, + "mean_token_accuracy": 0.7150951623916626, + "num_tokens": 37269519.0, + "step": 1443 + }, + { + "epoch": 0.15857676257412695, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.169731616973877, + "learning_rate": 5.281844802342605e-07, + "loss": 1.0124, + "mean_token_accuracy": 0.6999486684799194, + "num_tokens": 37295472.0, + "step": 1444 + }, + { + "epoch": 0.15868658027674062, + "ewc_loss": 4.153698682785034e-07, + "grad_norm": 2.2408945560455322, + "learning_rate": 5.285505124450952e-07, + "loss": 1.0919, + "mean_token_accuracy": 0.6792910099029541, + "num_tokens": 37323089.0, + "step": 1445 + }, + { + "epoch": 0.15879639797935427, + "ewc_loss": 4.172325134277344e-07, + "grad_norm": 2.39764666557312, + "learning_rate": 5.289165446559297e-07, + "loss": 1.099, + "mean_token_accuracy": 0.6702117919921875, + "num_tokens": 37349083.0, + "step": 1446 + }, + { + "epoch": 0.15890621568196794, + "ewc_loss": 4.172325134277344e-07, + "grad_norm": 2.531017541885376, + "learning_rate": 5.292825768667642e-07, + "loss": 0.9143, + "mean_token_accuracy": 0.7186024785041809, + "num_tokens": 37370736.0, + "step": 1447 + }, + { + "epoch": 0.1590160333845816, + "ewc_loss": 4.172325134277344e-07, + "grad_norm": 2.3899762630462646, + "learning_rate": 5.296486090775988e-07, + "loss": 1.0756, + "mean_token_accuracy": 0.6822143197059631, + "num_tokens": 37395554.0, + "step": 1448 + }, + { + "epoch": 0.15912585108719526, + "ewc_loss": 4.209578037261963e-07, + "grad_norm": 2.380434989929199, + "learning_rate": 5.300146412884333e-07, + "loss": 1.0413, + "mean_token_accuracy": 0.688288688659668, + "num_tokens": 37418863.0, + "step": 1449 + }, + { + "epoch": 0.1592356687898089, + "ewc_loss": 4.209578037261963e-07, + "grad_norm": 2.927323818206787, + "learning_rate": 5.303806734992679e-07, + "loss": 1.0325, + "mean_token_accuracy": 0.6836248636245728, + "num_tokens": 37435803.0, + "step": 1450 + }, + { + "epoch": 0.15934548649242258, + "ewc_loss": 4.209578037261963e-07, + "grad_norm": 2.312802314758301, + "learning_rate": 5.307467057101025e-07, + "loss": 1.1496, + "mean_token_accuracy": 0.6581254005432129, + "num_tokens": 37465254.0, + "step": 1451 + }, + { + "epoch": 0.15945530419503623, + "ewc_loss": 4.1909515857696533e-07, + "grad_norm": 2.300065517425537, + "learning_rate": 5.31112737920937e-07, + "loss": 0.9882, + "mean_token_accuracy": 0.7079020738601685, + "num_tokens": 37490392.0, + "step": 1452 + }, + { + "epoch": 0.1595651218976499, + "ewc_loss": 4.209578037261963e-07, + "grad_norm": 2.3052611351013184, + "learning_rate": 5.314787701317715e-07, + "loss": 1.0447, + "mean_token_accuracy": 0.6874296069145203, + "num_tokens": 37515868.0, + "step": 1453 + }, + { + "epoch": 0.15967493960026355, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 2.3532485961914062, + "learning_rate": 5.318448023426061e-07, + "loss": 0.9894, + "mean_token_accuracy": 0.7051087617874146, + "num_tokens": 37539607.0, + "step": 1454 + }, + { + "epoch": 0.15978475730287722, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 2.861908197402954, + "learning_rate": 5.322108345534406e-07, + "loss": 0.9219, + "mean_token_accuracy": 0.715927004814148, + "num_tokens": 37557033.0, + "step": 1455 + }, + { + "epoch": 0.1598945750054909, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 2.421177387237549, + "learning_rate": 5.325768667642752e-07, + "loss": 0.9801, + "mean_token_accuracy": 0.7004128694534302, + "num_tokens": 37581118.0, + "step": 1456 + }, + { + "epoch": 0.16000439270810454, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 2.3399248123168945, + "learning_rate": 5.329428989751098e-07, + "loss": 1.131, + "mean_token_accuracy": 0.6679432988166809, + "num_tokens": 37609075.0, + "step": 1457 + }, + { + "epoch": 0.16011421041071822, + "ewc_loss": 4.2282044887542725e-07, + "grad_norm": 2.1996843814849854, + "learning_rate": 5.333089311859443e-07, + "loss": 1.0401, + "mean_token_accuracy": 0.6893833875656128, + "num_tokens": 37636878.0, + "step": 1458 + }, + { + "epoch": 0.16022402811333186, + "ewc_loss": 4.246830940246582e-07, + "grad_norm": 2.110522985458374, + "learning_rate": 5.336749633967789e-07, + "loss": 1.0989, + "mean_token_accuracy": 0.6678248047828674, + "num_tokens": 37667267.0, + "step": 1459 + }, + { + "epoch": 0.16033384581594554, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.64953875541687, + "learning_rate": 5.340409956076134e-07, + "loss": 1.0072, + "mean_token_accuracy": 0.6997532844543457, + "num_tokens": 37688240.0, + "step": 1460 + }, + { + "epoch": 0.16044366351855918, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.62261962890625, + "learning_rate": 5.34407027818448e-07, + "loss": 0.928, + "mean_token_accuracy": 0.7138038873672485, + "num_tokens": 37707416.0, + "step": 1461 + }, + { + "epoch": 0.16055348122117286, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.0310049057006836, + "learning_rate": 5.347730600292826e-07, + "loss": 1.0772, + "mean_token_accuracy": 0.6837600469589233, + "num_tokens": 37740039.0, + "step": 1462 + }, + { + "epoch": 0.1606632989237865, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.0679986476898193, + "learning_rate": 5.351390922401171e-07, + "loss": 1.1346, + "mean_token_accuracy": 0.6697371006011963, + "num_tokens": 37773476.0, + "step": 1463 + }, + { + "epoch": 0.16077311662640018, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.39149808883667, + "learning_rate": 5.355051244509516e-07, + "loss": 1.057, + "mean_token_accuracy": 0.6905989646911621, + "num_tokens": 37802625.0, + "step": 1464 + }, + { + "epoch": 0.16088293432901385, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.802654504776001, + "learning_rate": 5.358711566617862e-07, + "loss": 0.9617, + "mean_token_accuracy": 0.7090286612510681, + "num_tokens": 37822336.0, + "step": 1465 + }, + { + "epoch": 0.1609927520316275, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.2836978435516357, + "learning_rate": 5.362371888726208e-07, + "loss": 0.9974, + "mean_token_accuracy": 0.6980777978897095, + "num_tokens": 37845801.0, + "step": 1466 + }, + { + "epoch": 0.16110256973424117, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.3552958965301514, + "learning_rate": 5.366032210834553e-07, + "loss": 1.1076, + "mean_token_accuracy": 0.6772837042808533, + "num_tokens": 37872498.0, + "step": 1467 + }, + { + "epoch": 0.16121238743685481, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 1.956943154335022, + "learning_rate": 5.369692532942899e-07, + "loss": 1.0482, + "mean_token_accuracy": 0.6874778866767883, + "num_tokens": 37904152.0, + "step": 1468 + }, + { + "epoch": 0.1613222051394685, + "ewc_loss": 4.2654573917388916e-07, + "grad_norm": 2.1268317699432373, + "learning_rate": 5.373352855051244e-07, + "loss": 0.9743, + "mean_token_accuracy": 0.7111175060272217, + "num_tokens": 37931150.0, + "step": 1469 + }, + { + "epoch": 0.16143202284208213, + "ewc_loss": 4.284083843231201e-07, + "grad_norm": 2.3304691314697266, + "learning_rate": 5.377013177159589e-07, + "loss": 1.0473, + "mean_token_accuracy": 0.6845659613609314, + "num_tokens": 37955011.0, + "step": 1470 + }, + { + "epoch": 0.1615418405446958, + "ewc_loss": 4.284083843231201e-07, + "grad_norm": 2.38373064994812, + "learning_rate": 5.380673499267935e-07, + "loss": 1.1135, + "mean_token_accuracy": 0.6706697344779968, + "num_tokens": 37981518.0, + "step": 1471 + }, + { + "epoch": 0.16165165824730945, + "ewc_loss": 4.302710294723511e-07, + "grad_norm": 2.252142906188965, + "learning_rate": 5.384333821376281e-07, + "loss": 1.0808, + "mean_token_accuracy": 0.6816897392272949, + "num_tokens": 38008938.0, + "step": 1472 + }, + { + "epoch": 0.16176147594992313, + "ewc_loss": 4.33996319770813e-07, + "grad_norm": 2.277287244796753, + "learning_rate": 5.387994143484626e-07, + "loss": 1.0387, + "mean_token_accuracy": 0.6866186857223511, + "num_tokens": 38036626.0, + "step": 1473 + }, + { + "epoch": 0.1618712936525368, + "ewc_loss": 4.33996319770813e-07, + "grad_norm": 1.9780954122543335, + "learning_rate": 5.391654465592972e-07, + "loss": 1.0965, + "mean_token_accuracy": 0.6741881966590881, + "num_tokens": 38070141.0, + "step": 1474 + }, + { + "epoch": 0.16198111135515045, + "ewc_loss": 4.33996319770813e-07, + "grad_norm": 2.685669183731079, + "learning_rate": 5.395314787701317e-07, + "loss": 1.0144, + "mean_token_accuracy": 0.6937788128852844, + "num_tokens": 38092134.0, + "step": 1475 + }, + { + "epoch": 0.16209092905776412, + "ewc_loss": 4.33996319770813e-07, + "grad_norm": 2.6417431831359863, + "learning_rate": 5.398975109809663e-07, + "loss": 0.9781, + "mean_token_accuracy": 0.7005565762519836, + "num_tokens": 38111775.0, + "step": 1476 + }, + { + "epoch": 0.16220074676037777, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.5663411617279053, + "learning_rate": 5.402635431918009e-07, + "loss": 1.0879, + "mean_token_accuracy": 0.6761643886566162, + "num_tokens": 38134366.0, + "step": 1477 + }, + { + "epoch": 0.16231056446299144, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.299452543258667, + "learning_rate": 5.406295754026354e-07, + "loss": 1.0773, + "mean_token_accuracy": 0.6783204674720764, + "num_tokens": 38159731.0, + "step": 1478 + }, + { + "epoch": 0.1624203821656051, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.2231080532073975, + "learning_rate": 5.4099560761347e-07, + "loss": 1.0548, + "mean_token_accuracy": 0.6803619861602783, + "num_tokens": 38189414.0, + "step": 1479 + }, + { + "epoch": 0.16253019986821876, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.431046962738037, + "learning_rate": 5.413616398243045e-07, + "loss": 1.0225, + "mean_token_accuracy": 0.6909746527671814, + "num_tokens": 38213059.0, + "step": 1480 + }, + { + "epoch": 0.1626400175708324, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.165525436401367, + "learning_rate": 5.41727672035139e-07, + "loss": 1.1125, + "mean_token_accuracy": 0.6682407855987549, + "num_tokens": 38242367.0, + "step": 1481 + }, + { + "epoch": 0.16274983527344608, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.24509859085083, + "learning_rate": 5.420937042459737e-07, + "loss": 1.1282, + "mean_token_accuracy": 0.6618472337722778, + "num_tokens": 38269270.0, + "step": 1482 + }, + { + "epoch": 0.16285965297605975, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.2068145275115967, + "learning_rate": 5.424597364568082e-07, + "loss": 1.0912, + "mean_token_accuracy": 0.686382532119751, + "num_tokens": 38297171.0, + "step": 1483 + }, + { + "epoch": 0.1629694706786734, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.605560302734375, + "learning_rate": 5.428257686676427e-07, + "loss": 1.0914, + "mean_token_accuracy": 0.6745514273643494, + "num_tokens": 38319751.0, + "step": 1484 + }, + { + "epoch": 0.16307928838128707, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.376072406768799, + "learning_rate": 5.431918008784773e-07, + "loss": 1.0228, + "mean_token_accuracy": 0.6922314167022705, + "num_tokens": 38344590.0, + "step": 1485 + }, + { + "epoch": 0.16318910608390072, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.3244450092315674, + "learning_rate": 5.435578330893118e-07, + "loss": 1.0022, + "mean_token_accuracy": 0.7025253176689148, + "num_tokens": 38369984.0, + "step": 1486 + }, + { + "epoch": 0.1632989237865144, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.565553665161133, + "learning_rate": 5.439238653001463e-07, + "loss": 0.999, + "mean_token_accuracy": 0.7007066607475281, + "num_tokens": 38392266.0, + "step": 1487 + }, + { + "epoch": 0.16340874148912804, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.0023059844970703, + "learning_rate": 5.44289897510981e-07, + "loss": 1.1691, + "mean_token_accuracy": 0.6601068377494812, + "num_tokens": 38426150.0, + "step": 1488 + }, + { + "epoch": 0.1635185591917417, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.0041561126708984, + "learning_rate": 5.446559297218155e-07, + "loss": 1.0087, + "mean_token_accuracy": 0.7030335664749146, + "num_tokens": 38458153.0, + "step": 1489 + }, + { + "epoch": 0.16362837689435536, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.620889902114868, + "learning_rate": 5.4502196193265e-07, + "loss": 1.0422, + "mean_token_accuracy": 0.6949237585067749, + "num_tokens": 38478408.0, + "step": 1490 + }, + { + "epoch": 0.16373819459696903, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.182386875152588, + "learning_rate": 5.453879941434846e-07, + "loss": 0.9603, + "mean_token_accuracy": 0.7051841020584106, + "num_tokens": 38507099.0, + "step": 1491 + }, + { + "epoch": 0.16384801229958268, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.242269277572632, + "learning_rate": 5.457540263543191e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.6930509805679321, + "num_tokens": 38532923.0, + "step": 1492 + }, + { + "epoch": 0.16395783000219635, + "ewc_loss": 4.3585896492004395e-07, + "grad_norm": 2.3653037548065186, + "learning_rate": 5.461200585651538e-07, + "loss": 1.1054, + "mean_token_accuracy": 0.6739283800125122, + "num_tokens": 38558137.0, + "step": 1493 + }, + { + "epoch": 0.16406764770481003, + "ewc_loss": 4.377216100692749e-07, + "grad_norm": 2.206787347793579, + "learning_rate": 5.464860907759883e-07, + "loss": 1.0447, + "mean_token_accuracy": 0.6890017986297607, + "num_tokens": 38584890.0, + "step": 1494 + }, + { + "epoch": 0.16417746540742367, + "ewc_loss": 4.377216100692749e-07, + "grad_norm": 2.2258787155151367, + "learning_rate": 5.468521229868228e-07, + "loss": 1.0479, + "mean_token_accuracy": 0.6866344809532166, + "num_tokens": 38611310.0, + "step": 1495 + }, + { + "epoch": 0.16428728311003735, + "ewc_loss": 4.377216100692749e-07, + "grad_norm": 2.2468299865722656, + "learning_rate": 5.472181551976574e-07, + "loss": 1.0964, + "mean_token_accuracy": 0.6777801513671875, + "num_tokens": 38638118.0, + "step": 1496 + }, + { + "epoch": 0.164397100812651, + "ewc_loss": 4.377216100692749e-07, + "grad_norm": 2.4858100414276123, + "learning_rate": 5.475841874084919e-07, + "loss": 0.9591, + "mean_token_accuracy": 0.7091094851493835, + "num_tokens": 38660015.0, + "step": 1497 + }, + { + "epoch": 0.16450691851526467, + "ewc_loss": 4.377216100692749e-07, + "grad_norm": 2.1976943016052246, + "learning_rate": 5.479502196193265e-07, + "loss": 1.0696, + "mean_token_accuracy": 0.6817498207092285, + "num_tokens": 38688580.0, + "step": 1498 + }, + { + "epoch": 0.1646167362178783, + "ewc_loss": 4.3958425521850586e-07, + "grad_norm": 2.376209020614624, + "learning_rate": 5.483162518301611e-07, + "loss": 1.0803, + "mean_token_accuracy": 0.6798698306083679, + "num_tokens": 38714193.0, + "step": 1499 + }, + { + "epoch": 0.16472655392049199, + "ewc_loss": 4.3958425521850586e-07, + "grad_norm": 2.212318181991577, + "learning_rate": 5.486822840409956e-07, + "loss": 1.0704, + "mean_token_accuracy": 0.6797749996185303, + "num_tokens": 38742448.0, + "step": 1500 + }, + { + "epoch": 0.16483637162310563, + "ewc_loss": 4.3958425521850586e-07, + "grad_norm": 2.432159185409546, + "learning_rate": 5.490483162518301e-07, + "loss": 0.9773, + "mean_token_accuracy": 0.708882749080658, + "num_tokens": 38765620.0, + "step": 1501 + }, + { + "epoch": 0.1649461893257193, + "ewc_loss": 4.3958425521850586e-07, + "grad_norm": 2.404716968536377, + "learning_rate": 5.494143484626647e-07, + "loss": 0.9512, + "mean_token_accuracy": 0.7137305736541748, + "num_tokens": 38788077.0, + "step": 1502 + }, + { + "epoch": 0.16505600702833298, + "ewc_loss": 4.3958425521850586e-07, + "grad_norm": 2.3747453689575195, + "learning_rate": 5.497803806734992e-07, + "loss": 1.0138, + "mean_token_accuracy": 0.7004111409187317, + "num_tokens": 38811822.0, + "step": 1503 + }, + { + "epoch": 0.16516582473094663, + "ewc_loss": 4.414469003677368e-07, + "grad_norm": 2.3700168132781982, + "learning_rate": 5.501464128843338e-07, + "loss": 1.0327, + "mean_token_accuracy": 0.6901177167892456, + "num_tokens": 38835722.0, + "step": 1504 + }, + { + "epoch": 0.1652756424335603, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.2436742782592773, + "learning_rate": 5.505124450951684e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.701606035232544, + "num_tokens": 38859889.0, + "step": 1505 + }, + { + "epoch": 0.16538546013617395, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.063709020614624, + "learning_rate": 5.508784773060029e-07, + "loss": 0.9878, + "mean_token_accuracy": 0.71570885181427, + "num_tokens": 38890641.0, + "step": 1506 + }, + { + "epoch": 0.16549527783878762, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.364537239074707, + "learning_rate": 5.512445095168374e-07, + "loss": 1.0937, + "mean_token_accuracy": 0.6797841787338257, + "num_tokens": 38914684.0, + "step": 1507 + }, + { + "epoch": 0.16560509554140126, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.560711145401001, + "learning_rate": 5.51610541727672e-07, + "loss": 0.997, + "mean_token_accuracy": 0.7037476301193237, + "num_tokens": 38935499.0, + "step": 1508 + }, + { + "epoch": 0.16571491324401494, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.328373908996582, + "learning_rate": 5.519765739385066e-07, + "loss": 1.0357, + "mean_token_accuracy": 0.6875896453857422, + "num_tokens": 38960393.0, + "step": 1509 + }, + { + "epoch": 0.16582473094662858, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.4970526695251465, + "learning_rate": 5.523426061493412e-07, + "loss": 1.065, + "mean_token_accuracy": 0.690658688545227, + "num_tokens": 38983330.0, + "step": 1510 + }, + { + "epoch": 0.16593454864924226, + "ewc_loss": 4.4330954551696777e-07, + "grad_norm": 2.248551845550537, + "learning_rate": 5.527086383601757e-07, + "loss": 1.126, + "mean_token_accuracy": 0.6688722968101501, + "num_tokens": 39011686.0, + "step": 1511 + }, + { + "epoch": 0.16604436635185593, + "ewc_loss": 4.4517219066619873e-07, + "grad_norm": 2.341653823852539, + "learning_rate": 5.530746705710102e-07, + "loss": 1.0537, + "mean_token_accuracy": 0.6878066062927246, + "num_tokens": 39037527.0, + "step": 1512 + }, + { + "epoch": 0.16615418405446958, + "ewc_loss": 4.4517219066619873e-07, + "grad_norm": 2.4874565601348877, + "learning_rate": 5.534407027818448e-07, + "loss": 1.0977, + "mean_token_accuracy": 0.6733800172805786, + "num_tokens": 39061584.0, + "step": 1513 + }, + { + "epoch": 0.16626400175708325, + "ewc_loss": 4.470348358154297e-07, + "grad_norm": 2.368217706680298, + "learning_rate": 5.538067349926794e-07, + "loss": 1.0128, + "mean_token_accuracy": 0.6956267356872559, + "num_tokens": 39085730.0, + "step": 1514 + }, + { + "epoch": 0.1663738194596969, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.836392879486084, + "learning_rate": 5.541727672035139e-07, + "loss": 1.0241, + "mean_token_accuracy": 0.6986231803894043, + "num_tokens": 39103747.0, + "step": 1515 + }, + { + "epoch": 0.16648363716231057, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.5412259101867676, + "learning_rate": 5.545387994143485e-07, + "loss": 0.9796, + "mean_token_accuracy": 0.7006204724311829, + "num_tokens": 39124300.0, + "step": 1516 + }, + { + "epoch": 0.16659345486492422, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.2635552883148193, + "learning_rate": 5.54904831625183e-07, + "loss": 0.9462, + "mean_token_accuracy": 0.7094368934631348, + "num_tokens": 39148700.0, + "step": 1517 + }, + { + "epoch": 0.1667032725675379, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.62627911567688, + "learning_rate": 5.552708638360175e-07, + "loss": 1.1075, + "mean_token_accuracy": 0.669141411781311, + "num_tokens": 39172508.0, + "step": 1518 + }, + { + "epoch": 0.16681309027015154, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.17842698097229, + "learning_rate": 5.556368960468521e-07, + "loss": 0.9534, + "mean_token_accuracy": 0.7103272676467896, + "num_tokens": 39197782.0, + "step": 1519 + }, + { + "epoch": 0.1669229079727652, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.2511019706726074, + "learning_rate": 5.560029282576867e-07, + "loss": 1.0052, + "mean_token_accuracy": 0.6974204778671265, + "num_tokens": 39224616.0, + "step": 1520 + }, + { + "epoch": 0.16703272567537888, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.2776949405670166, + "learning_rate": 5.563689604685212e-07, + "loss": 0.9935, + "mean_token_accuracy": 0.6990765333175659, + "num_tokens": 39249771.0, + "step": 1521 + }, + { + "epoch": 0.16714254337799253, + "ewc_loss": 4.4889748096466064e-07, + "grad_norm": 2.4137980937957764, + "learning_rate": 5.567349926793558e-07, + "loss": 0.924, + "mean_token_accuracy": 0.7139554619789124, + "num_tokens": 39271524.0, + "step": 1522 + }, + { + "epoch": 0.1672523610806062, + "ewc_loss": 4.507601261138916e-07, + "grad_norm": 2.1932878494262695, + "learning_rate": 5.571010248901903e-07, + "loss": 0.9945, + "mean_token_accuracy": 0.7001978158950806, + "num_tokens": 39299086.0, + "step": 1523 + }, + { + "epoch": 0.16736217878321985, + "ewc_loss": 4.507601261138916e-07, + "grad_norm": 2.029993772506714, + "learning_rate": 5.574670571010248e-07, + "loss": 1.0062, + "mean_token_accuracy": 0.7040483951568604, + "num_tokens": 39331423.0, + "step": 1524 + }, + { + "epoch": 0.16747199648583352, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.5332517623901367, + "learning_rate": 5.578330893118595e-07, + "loss": 1.06, + "mean_token_accuracy": 0.6830785274505615, + "num_tokens": 39352747.0, + "step": 1525 + }, + { + "epoch": 0.16758181418844717, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.3846399784088135, + "learning_rate": 5.58199121522694e-07, + "loss": 1.0737, + "mean_token_accuracy": 0.6839577555656433, + "num_tokens": 39377552.0, + "step": 1526 + }, + { + "epoch": 0.16769163189106084, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.441253900527954, + "learning_rate": 5.585651537335286e-07, + "loss": 1.0491, + "mean_token_accuracy": 0.6856487393379211, + "num_tokens": 39401174.0, + "step": 1527 + }, + { + "epoch": 0.1678014495936745, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.19254469871521, + "learning_rate": 5.589311859443631e-07, + "loss": 0.9524, + "mean_token_accuracy": 0.711579442024231, + "num_tokens": 39427179.0, + "step": 1528 + }, + { + "epoch": 0.16791126729628816, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.5255300998687744, + "learning_rate": 5.592972181551976e-07, + "loss": 0.9455, + "mean_token_accuracy": 0.7088215947151184, + "num_tokens": 39446853.0, + "step": 1529 + }, + { + "epoch": 0.1680210849989018, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.513179063796997, + "learning_rate": 5.596632503660323e-07, + "loss": 0.9571, + "mean_token_accuracy": 0.7128690481185913, + "num_tokens": 39467826.0, + "step": 1530 + }, + { + "epoch": 0.16813090270151548, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.274428606033325, + "learning_rate": 5.600292825768668e-07, + "loss": 0.9145, + "mean_token_accuracy": 0.7189697027206421, + "num_tokens": 39492608.0, + "step": 1531 + }, + { + "epoch": 0.16824072040412916, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.5531020164489746, + "learning_rate": 5.603953147877013e-07, + "loss": 1.0577, + "mean_token_accuracy": 0.6892896890640259, + "num_tokens": 39513984.0, + "step": 1532 + }, + { + "epoch": 0.1683505381067428, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.1778252124786377, + "learning_rate": 5.607613469985359e-07, + "loss": 0.977, + "mean_token_accuracy": 0.7040027976036072, + "num_tokens": 39541191.0, + "step": 1533 + }, + { + "epoch": 0.16846035580935648, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.3389108180999756, + "learning_rate": 5.611273792093704e-07, + "loss": 0.9849, + "mean_token_accuracy": 0.7160824537277222, + "num_tokens": 39565937.0, + "step": 1534 + }, + { + "epoch": 0.16857017351197012, + "ewc_loss": 4.5262277126312256e-07, + "grad_norm": 2.6950862407684326, + "learning_rate": 5.614934114202049e-07, + "loss": 1.0267, + "mean_token_accuracy": 0.6965008974075317, + "num_tokens": 39585586.0, + "step": 1535 + }, + { + "epoch": 0.1686799912145838, + "ewc_loss": 4.5634806156158447e-07, + "grad_norm": 1.9925295114517212, + "learning_rate": 5.618594436310396e-07, + "loss": 1.1403, + "mean_token_accuracy": 0.6652028560638428, + "num_tokens": 39621116.0, + "step": 1536 + }, + { + "epoch": 0.16878980891719744, + "ewc_loss": 4.5634806156158447e-07, + "grad_norm": 2.4839437007904053, + "learning_rate": 5.622254758418741e-07, + "loss": 1.0732, + "mean_token_accuracy": 0.6836562752723694, + "num_tokens": 39645545.0, + "step": 1537 + }, + { + "epoch": 0.16889962661981112, + "ewc_loss": 4.5634806156158447e-07, + "grad_norm": 2.3034205436706543, + "learning_rate": 5.625915080527086e-07, + "loss": 1.0714, + "mean_token_accuracy": 0.6879509091377258, + "num_tokens": 39671881.0, + "step": 1538 + }, + { + "epoch": 0.16900944432242476, + "ewc_loss": 4.5634806156158447e-07, + "grad_norm": 2.4262194633483887, + "learning_rate": 5.629575402635432e-07, + "loss": 1.0155, + "mean_token_accuracy": 0.695371687412262, + "num_tokens": 39695290.0, + "step": 1539 + }, + { + "epoch": 0.16911926202503844, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.2080745697021484, + "learning_rate": 5.633235724743777e-07, + "loss": 1.0207, + "mean_token_accuracy": 0.7010234594345093, + "num_tokens": 39722135.0, + "step": 1540 + }, + { + "epoch": 0.1692290797276521, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.2720301151275635, + "learning_rate": 5.636896046852124e-07, + "loss": 1.15, + "mean_token_accuracy": 0.6610650420188904, + "num_tokens": 39752087.0, + "step": 1541 + }, + { + "epoch": 0.16933889743026576, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.091693639755249, + "learning_rate": 5.640556368960469e-07, + "loss": 1.0258, + "mean_token_accuracy": 0.6933150291442871, + "num_tokens": 39782108.0, + "step": 1542 + }, + { + "epoch": 0.16944871513287943, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.22568416595459, + "learning_rate": 5.644216691068814e-07, + "loss": 1.0726, + "mean_token_accuracy": 0.6750035285949707, + "num_tokens": 39810707.0, + "step": 1543 + }, + { + "epoch": 0.16955853283549308, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.28144907951355, + "learning_rate": 5.64787701317716e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.7014719247817993, + "num_tokens": 39836838.0, + "step": 1544 + }, + { + "epoch": 0.16966835053810675, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.4194061756134033, + "learning_rate": 5.651537335285505e-07, + "loss": 0.9681, + "mean_token_accuracy": 0.7106702327728271, + "num_tokens": 39860798.0, + "step": 1545 + }, + { + "epoch": 0.1697781682407204, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.34596586227417, + "learning_rate": 5.655197657393851e-07, + "loss": 1.1197, + "mean_token_accuracy": 0.6732560396194458, + "num_tokens": 39886558.0, + "step": 1546 + }, + { + "epoch": 0.16988798594333407, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.583641767501831, + "learning_rate": 5.658857979502197e-07, + "loss": 0.9475, + "mean_token_accuracy": 0.7071765065193176, + "num_tokens": 39907445.0, + "step": 1547 + }, + { + "epoch": 0.16999780364594771, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.144258499145508, + "learning_rate": 5.662518301610542e-07, + "loss": 0.9462, + "mean_token_accuracy": 0.7189558148384094, + "num_tokens": 39937347.0, + "step": 1548 + }, + { + "epoch": 0.1701076213485614, + "ewc_loss": 4.600733518600464e-07, + "grad_norm": 2.1537938117980957, + "learning_rate": 5.666178623718887e-07, + "loss": 0.9842, + "mean_token_accuracy": 0.7059580683708191, + "num_tokens": 39966512.0, + "step": 1549 + }, + { + "epoch": 0.17021743905117506, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.3738887310028076, + "learning_rate": 5.669838945827233e-07, + "loss": 1.0683, + "mean_token_accuracy": 0.6811109781265259, + "num_tokens": 39993941.0, + "step": 1550 + }, + { + "epoch": 0.1703272567537887, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.9554529190063477, + "learning_rate": 5.673499267935577e-07, + "loss": 0.8882, + "mean_token_accuracy": 0.731651782989502, + "num_tokens": 40012552.0, + "step": 1551 + }, + { + "epoch": 0.17043707445640238, + "ewc_loss": 4.637986421585083e-07, + "grad_norm": 2.164222240447998, + "learning_rate": 5.677159590043924e-07, + "loss": 1.0583, + "mean_token_accuracy": 0.6872156262397766, + "num_tokens": 40043398.0, + "step": 1552 + }, + { + "epoch": 0.17054689215901603, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.402651786804199, + "learning_rate": 5.68081991215227e-07, + "loss": 1.0925, + "mean_token_accuracy": 0.6783584356307983, + "num_tokens": 40070386.0, + "step": 1553 + }, + { + "epoch": 0.1706567098616297, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.3816516399383545, + "learning_rate": 5.684480234260615e-07, + "loss": 1.021, + "mean_token_accuracy": 0.6954878568649292, + "num_tokens": 40094339.0, + "step": 1554 + }, + { + "epoch": 0.17076652756424335, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.289524793624878, + "learning_rate": 5.688140556368959e-07, + "loss": 0.9347, + "mean_token_accuracy": 0.7216013669967651, + "num_tokens": 40118938.0, + "step": 1555 + }, + { + "epoch": 0.17087634526685702, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.196906328201294, + "learning_rate": 5.691800878477305e-07, + "loss": 1.1109, + "mean_token_accuracy": 0.6673885583877563, + "num_tokens": 40148530.0, + "step": 1556 + }, + { + "epoch": 0.17098616296947067, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.247025728225708, + "learning_rate": 5.695461200585652e-07, + "loss": 1.0662, + "mean_token_accuracy": 0.6876145601272583, + "num_tokens": 40175048.0, + "step": 1557 + }, + { + "epoch": 0.17109598067208434, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.432199239730835, + "learning_rate": 5.699121522693998e-07, + "loss": 1.0227, + "mean_token_accuracy": 0.6937360763549805, + "num_tokens": 40198815.0, + "step": 1558 + }, + { + "epoch": 0.17120579837469801, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.428457498550415, + "learning_rate": 5.702781844802342e-07, + "loss": 0.9951, + "mean_token_accuracy": 0.7069764137268066, + "num_tokens": 40221652.0, + "step": 1559 + }, + { + "epoch": 0.17131561607731166, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.282776117324829, + "learning_rate": 5.706442166910687e-07, + "loss": 1.0114, + "mean_token_accuracy": 0.6962447166442871, + "num_tokens": 40248573.0, + "step": 1560 + }, + { + "epoch": 0.17142543377992533, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.4415531158447266, + "learning_rate": 5.710102489019033e-07, + "loss": 1.0853, + "mean_token_accuracy": 0.6799036264419556, + "num_tokens": 40273242.0, + "step": 1561 + }, + { + "epoch": 0.17153525148253898, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.418911933898926, + "learning_rate": 5.71376281112738e-07, + "loss": 1.0099, + "mean_token_accuracy": 0.6966651678085327, + "num_tokens": 40295379.0, + "step": 1562 + }, + { + "epoch": 0.17164506918515265, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.8036234378814697, + "learning_rate": 5.717423133235724e-07, + "loss": 1.0148, + "mean_token_accuracy": 0.6941441297531128, + "num_tokens": 40313918.0, + "step": 1563 + }, + { + "epoch": 0.1717548868877663, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.3629491329193115, + "learning_rate": 5.72108345534407e-07, + "loss": 1.0322, + "mean_token_accuracy": 0.6919805407524109, + "num_tokens": 40338811.0, + "step": 1564 + }, + { + "epoch": 0.17186470459037997, + "ewc_loss": 4.6193599700927734e-07, + "grad_norm": 2.4235076904296875, + "learning_rate": 5.724743777452415e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6894504427909851, + "num_tokens": 40361823.0, + "step": 1565 + }, + { + "epoch": 0.17197452229299362, + "ewc_loss": 4.637986421585083e-07, + "grad_norm": 2.364290237426758, + "learning_rate": 5.72840409956076e-07, + "loss": 1.0723, + "mean_token_accuracy": 0.689789354801178, + "num_tokens": 40386526.0, + "step": 1566 + }, + { + "epoch": 0.1720843399956073, + "ewc_loss": 4.637986421585083e-07, + "grad_norm": 2.2664945125579834, + "learning_rate": 5.732064421669106e-07, + "loss": 0.9981, + "mean_token_accuracy": 0.7011144757270813, + "num_tokens": 40412670.0, + "step": 1567 + }, + { + "epoch": 0.17219415769822094, + "ewc_loss": 4.637986421585083e-07, + "grad_norm": 2.2856523990631104, + "learning_rate": 5.735724743777452e-07, + "loss": 1.0185, + "mean_token_accuracy": 0.6960831880569458, + "num_tokens": 40439359.0, + "step": 1568 + }, + { + "epoch": 0.1723039754008346, + "ewc_loss": 4.637986421585083e-07, + "grad_norm": 2.1928298473358154, + "learning_rate": 5.739385065885797e-07, + "loss": 1.0127, + "mean_token_accuracy": 0.6989273428916931, + "num_tokens": 40466494.0, + "step": 1569 + }, + { + "epoch": 0.1724137931034483, + "ewc_loss": 4.6566128730773926e-07, + "grad_norm": 2.538653612136841, + "learning_rate": 5.743045387994143e-07, + "loss": 0.9711, + "mean_token_accuracy": 0.7045537829399109, + "num_tokens": 40488412.0, + "step": 1570 + }, + { + "epoch": 0.17252361080606193, + "ewc_loss": 4.6566128730773926e-07, + "grad_norm": 2.353490114212036, + "learning_rate": 5.746705710102488e-07, + "loss": 1.0656, + "mean_token_accuracy": 0.6826671957969666, + "num_tokens": 40511460.0, + "step": 1571 + }, + { + "epoch": 0.1726334285086756, + "ewc_loss": 4.6938657760620117e-07, + "grad_norm": 2.2494962215423584, + "learning_rate": 5.750366032210833e-07, + "loss": 0.984, + "mean_token_accuracy": 0.706499457359314, + "num_tokens": 40537397.0, + "step": 1572 + }, + { + "epoch": 0.17274324621128925, + "ewc_loss": 4.7124922275543213e-07, + "grad_norm": 2.7351319789886475, + "learning_rate": 5.75402635431918e-07, + "loss": 0.9924, + "mean_token_accuracy": 0.7013146877288818, + "num_tokens": 40556838.0, + "step": 1573 + }, + { + "epoch": 0.17285306391390293, + "ewc_loss": 4.7124922275543213e-07, + "grad_norm": 2.7846500873565674, + "learning_rate": 5.757686676427525e-07, + "loss": 1.0026, + "mean_token_accuracy": 0.6944730281829834, + "num_tokens": 40575543.0, + "step": 1574 + }, + { + "epoch": 0.17296288161651657, + "ewc_loss": 4.7124922275543213e-07, + "grad_norm": 2.1599223613739014, + "learning_rate": 5.761346998535871e-07, + "loss": 1.056, + "mean_token_accuracy": 0.6878350973129272, + "num_tokens": 40604152.0, + "step": 1575 + }, + { + "epoch": 0.17307269931913025, + "ewc_loss": 4.7124922275543213e-07, + "grad_norm": 2.4137918949127197, + "learning_rate": 5.765007320644216e-07, + "loss": 1.0501, + "mean_token_accuracy": 0.6865657567977905, + "num_tokens": 40627574.0, + "step": 1576 + }, + { + "epoch": 0.1731825170217439, + "ewc_loss": 4.731118679046631e-07, + "grad_norm": 2.222836971282959, + "learning_rate": 5.768667642752561e-07, + "loss": 0.9902, + "mean_token_accuracy": 0.6999962329864502, + "num_tokens": 40651057.0, + "step": 1577 + }, + { + "epoch": 0.17329233472435757, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.226114273071289, + "learning_rate": 5.772327964860908e-07, + "loss": 1.0256, + "mean_token_accuracy": 0.6886547803878784, + "num_tokens": 40678662.0, + "step": 1578 + }, + { + "epoch": 0.17340215242697124, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.300564765930176, + "learning_rate": 5.775988286969253e-07, + "loss": 1.0892, + "mean_token_accuracy": 0.6837371587753296, + "num_tokens": 40704905.0, + "step": 1579 + }, + { + "epoch": 0.17351197012958489, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.571855068206787, + "learning_rate": 5.779648609077598e-07, + "loss": 0.9531, + "mean_token_accuracy": 0.7124948501586914, + "num_tokens": 40725916.0, + "step": 1580 + }, + { + "epoch": 0.17362178783219856, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.54884934425354, + "learning_rate": 5.783308931185944e-07, + "loss": 1.052, + "mean_token_accuracy": 0.6843062043190002, + "num_tokens": 40746748.0, + "step": 1581 + }, + { + "epoch": 0.1737316055348122, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.1548690795898438, + "learning_rate": 5.786969253294289e-07, + "loss": 1.1227, + "mean_token_accuracy": 0.6854970455169678, + "num_tokens": 40779091.0, + "step": 1582 + }, + { + "epoch": 0.17384142323742588, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.0745439529418945, + "learning_rate": 5.790629575402634e-07, + "loss": 1.1012, + "mean_token_accuracy": 0.6714450120925903, + "num_tokens": 40809761.0, + "step": 1583 + }, + { + "epoch": 0.17395124094003953, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.1170613765716553, + "learning_rate": 5.794289897510981e-07, + "loss": 1.0673, + "mean_token_accuracy": 0.6862552762031555, + "num_tokens": 40841486.0, + "step": 1584 + }, + { + "epoch": 0.1740610586426532, + "ewc_loss": 4.76837158203125e-07, + "grad_norm": 2.1109001636505127, + "learning_rate": 5.797950219619326e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.6838170886039734, + "num_tokens": 40870417.0, + "step": 1585 + }, + { + "epoch": 0.17417087634526685, + "ewc_loss": 4.805624485015869e-07, + "grad_norm": 2.308788776397705, + "learning_rate": 5.801610541727671e-07, + "loss": 1.0115, + "mean_token_accuracy": 0.7037114500999451, + "num_tokens": 40898422.0, + "step": 1586 + }, + { + "epoch": 0.17428069404788052, + "ewc_loss": 4.805624485015869e-07, + "grad_norm": 2.4821927547454834, + "learning_rate": 5.805270863836017e-07, + "loss": 1.1257, + "mean_token_accuracy": 0.6687886714935303, + "num_tokens": 40921246.0, + "step": 1587 + }, + { + "epoch": 0.1743905117504942, + "ewc_loss": 4.842877388000488e-07, + "grad_norm": 2.2205512523651123, + "learning_rate": 5.808931185944362e-07, + "loss": 1.0139, + "mean_token_accuracy": 0.6903723478317261, + "num_tokens": 40947697.0, + "step": 1588 + }, + { + "epoch": 0.17450032945310784, + "ewc_loss": 4.842877388000488e-07, + "grad_norm": 2.271677017211914, + "learning_rate": 5.812591508052709e-07, + "loss": 0.9397, + "mean_token_accuracy": 0.7165200710296631, + "num_tokens": 40971954.0, + "step": 1589 + }, + { + "epoch": 0.1746101471557215, + "ewc_loss": 4.842877388000488e-07, + "grad_norm": 2.117091655731201, + "learning_rate": 5.816251830161054e-07, + "loss": 1.0564, + "mean_token_accuracy": 0.6814695596694946, + "num_tokens": 41003026.0, + "step": 1590 + }, + { + "epoch": 0.17471996485833516, + "ewc_loss": 4.842877388000488e-07, + "grad_norm": 2.35204815864563, + "learning_rate": 5.819912152269399e-07, + "loss": 1.0637, + "mean_token_accuracy": 0.6795263290405273, + "num_tokens": 41027889.0, + "step": 1591 + }, + { + "epoch": 0.17482978256094883, + "ewc_loss": 4.842877388000488e-07, + "grad_norm": 2.669445276260376, + "learning_rate": 5.823572474377745e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.6897211670875549, + "num_tokens": 41048983.0, + "step": 1592 + }, + { + "epoch": 0.17493960026356248, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.331493616104126, + "learning_rate": 5.82723279648609e-07, + "loss": 1.0843, + "mean_token_accuracy": 0.6727827787399292, + "num_tokens": 41074025.0, + "step": 1593 + }, + { + "epoch": 0.17504941796617615, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.1738240718841553, + "learning_rate": 5.830893118594436e-07, + "loss": 1.0838, + "mean_token_accuracy": 0.6773414015769958, + "num_tokens": 41104142.0, + "step": 1594 + }, + { + "epoch": 0.1751592356687898, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.3425488471984863, + "learning_rate": 5.834553440702782e-07, + "loss": 1.0497, + "mean_token_accuracy": 0.6881778240203857, + "num_tokens": 41128610.0, + "step": 1595 + }, + { + "epoch": 0.17526905337140347, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.2021970748901367, + "learning_rate": 5.838213762811127e-07, + "loss": 0.9905, + "mean_token_accuracy": 0.7023000717163086, + "num_tokens": 41154215.0, + "step": 1596 + }, + { + "epoch": 0.17537887107401715, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.1350626945495605, + "learning_rate": 5.841874084919472e-07, + "loss": 1.0805, + "mean_token_accuracy": 0.6736564636230469, + "num_tokens": 41186683.0, + "step": 1597 + }, + { + "epoch": 0.1754886887766308, + "ewc_loss": 4.880130290985107e-07, + "grad_norm": 2.3052823543548584, + "learning_rate": 5.845534407027818e-07, + "loss": 0.9991, + "mean_token_accuracy": 0.6954741477966309, + "num_tokens": 41211138.0, + "step": 1598 + }, + { + "epoch": 0.17559850647924446, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.787879467010498, + "learning_rate": 5.849194729136163e-07, + "loss": 0.927, + "mean_token_accuracy": 0.7180095314979553, + "num_tokens": 41229133.0, + "step": 1599 + }, + { + "epoch": 0.1757083241818581, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.431694269180298, + "learning_rate": 5.852855051244509e-07, + "loss": 1.0599, + "mean_token_accuracy": 0.6884384155273438, + "num_tokens": 41254358.0, + "step": 1600 + }, + { + "epoch": 0.17581814188447178, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.2019264698028564, + "learning_rate": 5.856515373352855e-07, + "loss": 1.0874, + "mean_token_accuracy": 0.6845288872718811, + "num_tokens": 41282152.0, + "step": 1601 + }, + { + "epoch": 0.17592795958708543, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.3466360569000244, + "learning_rate": 5.8601756954612e-07, + "loss": 0.9627, + "mean_token_accuracy": 0.713539719581604, + "num_tokens": 41307760.0, + "step": 1602 + }, + { + "epoch": 0.1760377772896991, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.283881425857544, + "learning_rate": 5.863836017569545e-07, + "loss": 0.9816, + "mean_token_accuracy": 0.7030545473098755, + "num_tokens": 41333522.0, + "step": 1603 + }, + { + "epoch": 0.17614759499231275, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.3235597610473633, + "learning_rate": 5.867496339677891e-07, + "loss": 1.0175, + "mean_token_accuracy": 0.6970276832580566, + "num_tokens": 41357614.0, + "step": 1604 + }, + { + "epoch": 0.17625741269492642, + "ewc_loss": 4.917383193969727e-07, + "grad_norm": 2.231829881668091, + "learning_rate": 5.871156661786237e-07, + "loss": 0.9704, + "mean_token_accuracy": 0.7095940709114075, + "num_tokens": 41383591.0, + "step": 1605 + }, + { + "epoch": 0.17636723039754007, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.2448689937591553, + "learning_rate": 5.874816983894583e-07, + "loss": 1.0965, + "mean_token_accuracy": 0.6732184886932373, + "num_tokens": 41410821.0, + "step": 1606 + }, + { + "epoch": 0.17647704810015374, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.5431013107299805, + "learning_rate": 5.878477306002928e-07, + "loss": 1.0693, + "mean_token_accuracy": 0.6855653524398804, + "num_tokens": 41433212.0, + "step": 1607 + }, + { + "epoch": 0.17658686580276742, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.226187229156494, + "learning_rate": 5.882137628111273e-07, + "loss": 1.1441, + "mean_token_accuracy": 0.6599902510643005, + "num_tokens": 41461618.0, + "step": 1608 + }, + { + "epoch": 0.17669668350538106, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.1890337467193604, + "learning_rate": 5.885797950219619e-07, + "loss": 1.0556, + "mean_token_accuracy": 0.6817139983177185, + "num_tokens": 41490369.0, + "step": 1609 + }, + { + "epoch": 0.17680650120799474, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.283975601196289, + "learning_rate": 5.889458272327965e-07, + "loss": 1.2098, + "mean_token_accuracy": 0.6475147008895874, + "num_tokens": 41518493.0, + "step": 1610 + }, + { + "epoch": 0.17691631891060838, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.24771785736084, + "learning_rate": 5.89311859443631e-07, + "loss": 1.1191, + "mean_token_accuracy": 0.67461097240448, + "num_tokens": 41545826.0, + "step": 1611 + }, + { + "epoch": 0.17702613661322206, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.492363691329956, + "learning_rate": 5.896778916544656e-07, + "loss": 1.0137, + "mean_token_accuracy": 0.690750241279602, + "num_tokens": 41568877.0, + "step": 1612 + }, + { + "epoch": 0.1771359543158357, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.3553860187530518, + "learning_rate": 5.900439238653001e-07, + "loss": 0.9823, + "mean_token_accuracy": 0.7047868967056274, + "num_tokens": 41594228.0, + "step": 1613 + }, + { + "epoch": 0.17724577201844938, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.3776135444641113, + "learning_rate": 5.904099560761346e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6793109178543091, + "num_tokens": 41620861.0, + "step": 1614 + }, + { + "epoch": 0.17735558972106302, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.3186962604522705, + "learning_rate": 5.907759882869692e-07, + "loss": 1.0884, + "mean_token_accuracy": 0.6809999346733093, + "num_tokens": 41646859.0, + "step": 1615 + }, + { + "epoch": 0.1774654074236767, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.4220130443573, + "learning_rate": 5.911420204978038e-07, + "loss": 1.0502, + "mean_token_accuracy": 0.6884857416152954, + "num_tokens": 41670100.0, + "step": 1616 + }, + { + "epoch": 0.17757522512629037, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.5604095458984375, + "learning_rate": 5.915080527086383e-07, + "loss": 0.9899, + "mean_token_accuracy": 0.7075480222702026, + "num_tokens": 41690417.0, + "step": 1617 + }, + { + "epoch": 0.17768504282890402, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.007789373397827, + "learning_rate": 5.918740849194729e-07, + "loss": 0.9553, + "mean_token_accuracy": 0.7199646830558777, + "num_tokens": 41723702.0, + "step": 1618 + }, + { + "epoch": 0.1777948605315177, + "ewc_loss": 4.954636096954346e-07, + "grad_norm": 2.3095271587371826, + "learning_rate": 5.922401171303074e-07, + "loss": 0.9714, + "mean_token_accuracy": 0.7041807174682617, + "num_tokens": 41747292.0, + "step": 1619 + }, + { + "epoch": 0.17790467823413134, + "ewc_loss": 4.991888999938965e-07, + "grad_norm": 2.5659983158111572, + "learning_rate": 5.926061493411419e-07, + "loss": 1.0749, + "mean_token_accuracy": 0.6804576516151428, + "num_tokens": 41771664.0, + "step": 1620 + }, + { + "epoch": 0.178014495936745, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 2.1015143394470215, + "learning_rate": 5.929721815519766e-07, + "loss": 1.0839, + "mean_token_accuracy": 0.6781492233276367, + "num_tokens": 41802272.0, + "step": 1621 + }, + { + "epoch": 0.17812431363935866, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 2.263484239578247, + "learning_rate": 5.933382137628111e-07, + "loss": 1.0253, + "mean_token_accuracy": 0.6897919774055481, + "num_tokens": 41826323.0, + "step": 1622 + }, + { + "epoch": 0.17823413134197233, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 2.0391833782196045, + "learning_rate": 5.937042459736457e-07, + "loss": 1.1345, + "mean_token_accuracy": 0.6649907231330872, + "num_tokens": 41859043.0, + "step": 1623 + }, + { + "epoch": 0.17834394904458598, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 2.4955220222473145, + "learning_rate": 5.940702781844802e-07, + "loss": 0.999, + "mean_token_accuracy": 0.7002605199813843, + "num_tokens": 41880055.0, + "step": 1624 + }, + { + "epoch": 0.17845376674719965, + "ewc_loss": 5.029141902923584e-07, + "grad_norm": 2.2956888675689697, + "learning_rate": 5.944363103953147e-07, + "loss": 1.0648, + "mean_token_accuracy": 0.6774371862411499, + "num_tokens": 41906987.0, + "step": 1625 + }, + { + "epoch": 0.17856358444981332, + "ewc_loss": 5.066394805908203e-07, + "grad_norm": 2.387073278427124, + "learning_rate": 5.948023426061494e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.7006371021270752, + "num_tokens": 41929196.0, + "step": 1626 + }, + { + "epoch": 0.17867340215242697, + "ewc_loss": 5.066394805908203e-07, + "grad_norm": 2.315016269683838, + "learning_rate": 5.951683748169839e-07, + "loss": 1.1469, + "mean_token_accuracy": 0.6668835282325745, + "num_tokens": 41955354.0, + "step": 1627 + }, + { + "epoch": 0.17878321985504064, + "ewc_loss": 5.066394805908203e-07, + "grad_norm": 2.2529585361480713, + "learning_rate": 5.955344070278184e-07, + "loss": 1.146, + "mean_token_accuracy": 0.6583242416381836, + "num_tokens": 41982518.0, + "step": 1628 + }, + { + "epoch": 0.1788930375576543, + "ewc_loss": 5.103647708892822e-07, + "grad_norm": 2.331498146057129, + "learning_rate": 5.95900439238653e-07, + "loss": 0.9967, + "mean_token_accuracy": 0.6989402174949646, + "num_tokens": 42006184.0, + "step": 1629 + }, + { + "epoch": 0.17900285526026796, + "ewc_loss": 5.103647708892822e-07, + "grad_norm": 2.7535502910614014, + "learning_rate": 5.962664714494875e-07, + "loss": 0.9919, + "mean_token_accuracy": 0.7014331817626953, + "num_tokens": 42025392.0, + "step": 1630 + }, + { + "epoch": 0.1791126729628816, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.155017614364624, + "learning_rate": 5.96632503660322e-07, + "loss": 1.174, + "mean_token_accuracy": 0.6611683368682861, + "num_tokens": 42055514.0, + "step": 1631 + }, + { + "epoch": 0.17922249066549528, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.4515411853790283, + "learning_rate": 5.969985358711567e-07, + "loss": 1.07, + "mean_token_accuracy": 0.6827492713928223, + "num_tokens": 42081033.0, + "step": 1632 + }, + { + "epoch": 0.17933230836810893, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3987045288085938, + "learning_rate": 5.973645680819912e-07, + "loss": 0.9758, + "mean_token_accuracy": 0.7112588286399841, + "num_tokens": 42104399.0, + "step": 1633 + }, + { + "epoch": 0.1794421260707226, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.339503049850464, + "learning_rate": 5.977306002928257e-07, + "loss": 1.054, + "mean_token_accuracy": 0.68462735414505, + "num_tokens": 42128912.0, + "step": 1634 + }, + { + "epoch": 0.17955194377333628, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.622363328933716, + "learning_rate": 5.980966325036603e-07, + "loss": 1.0898, + "mean_token_accuracy": 0.6749926209449768, + "num_tokens": 42150892.0, + "step": 1635 + }, + { + "epoch": 0.17966176147594992, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.49153733253479, + "learning_rate": 5.984626647144948e-07, + "loss": 0.9753, + "mean_token_accuracy": 0.7127807140350342, + "num_tokens": 42172510.0, + "step": 1636 + }, + { + "epoch": 0.1797715791785636, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.202242851257324, + "learning_rate": 5.988286969253294e-07, + "loss": 1.0506, + "mean_token_accuracy": 0.6836763620376587, + "num_tokens": 42199310.0, + "step": 1637 + }, + { + "epoch": 0.17988139688117724, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.1099603176116943, + "learning_rate": 5.99194729136164e-07, + "loss": 1.0862, + "mean_token_accuracy": 0.6744233965873718, + "num_tokens": 42229990.0, + "step": 1638 + }, + { + "epoch": 0.17999121458379091, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.570056676864624, + "learning_rate": 5.995607613469985e-07, + "loss": 0.9581, + "mean_token_accuracy": 0.7126553654670715, + "num_tokens": 42249858.0, + "step": 1639 + }, + { + "epoch": 0.18010103228640456, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.5596365928649902, + "learning_rate": 5.999267935578331e-07, + "loss": 0.9531, + "mean_token_accuracy": 0.708564281463623, + "num_tokens": 42271355.0, + "step": 1640 + }, + { + "epoch": 0.18021084998901823, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.2806806564331055, + "learning_rate": 6.002928257686676e-07, + "loss": 1.0221, + "mean_token_accuracy": 0.6989628672599792, + "num_tokens": 42297624.0, + "step": 1641 + }, + { + "epoch": 0.18032066769163188, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.285176992416382, + "learning_rate": 6.006588579795022e-07, + "loss": 0.9632, + "mean_token_accuracy": 0.7062237858772278, + "num_tokens": 42322395.0, + "step": 1642 + }, + { + "epoch": 0.18043048539424555, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.610074043273926, + "learning_rate": 6.010248901903368e-07, + "loss": 1.0161, + "mean_token_accuracy": 0.7086877226829529, + "num_tokens": 42344634.0, + "step": 1643 + }, + { + "epoch": 0.1805403030968592, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3644063472747803, + "learning_rate": 6.013909224011713e-07, + "loss": 1.0981, + "mean_token_accuracy": 0.6769929528236389, + "num_tokens": 42370501.0, + "step": 1644 + }, + { + "epoch": 0.18065012079947287, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.23724365234375, + "learning_rate": 6.017569546120058e-07, + "loss": 1.0461, + "mean_token_accuracy": 0.6916109323501587, + "num_tokens": 42396966.0, + "step": 1645 + }, + { + "epoch": 0.18075993850208655, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.291255235671997, + "learning_rate": 6.021229868228404e-07, + "loss": 1.0936, + "mean_token_accuracy": 0.6862958669662476, + "num_tokens": 42423033.0, + "step": 1646 + }, + { + "epoch": 0.1808697562047002, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 1.9695687294006348, + "learning_rate": 6.024890190336749e-07, + "loss": 1.0421, + "mean_token_accuracy": 0.6940698623657227, + "num_tokens": 42456449.0, + "step": 1647 + }, + { + "epoch": 0.18097957390731387, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.377075433731079, + "learning_rate": 6.028550512445095e-07, + "loss": 1.0067, + "mean_token_accuracy": 0.6997823715209961, + "num_tokens": 42480035.0, + "step": 1648 + }, + { + "epoch": 0.1810893916099275, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3983075618743896, + "learning_rate": 6.032210834553441e-07, + "loss": 1.0168, + "mean_token_accuracy": 0.7002097964286804, + "num_tokens": 42503308.0, + "step": 1649 + }, + { + "epoch": 0.1811992093125412, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.198943614959717, + "learning_rate": 6.035871156661786e-07, + "loss": 1.0824, + "mean_token_accuracy": 0.6786623597145081, + "num_tokens": 42532075.0, + "step": 1650 + }, + { + "epoch": 0.18130902701515483, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3333592414855957, + "learning_rate": 6.039531478770131e-07, + "loss": 0.9936, + "mean_token_accuracy": 0.7039831876754761, + "num_tokens": 42556184.0, + "step": 1651 + }, + { + "epoch": 0.1814188447177685, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.548895835876465, + "learning_rate": 6.043191800878477e-07, + "loss": 1.0942, + "mean_token_accuracy": 0.6797338128089905, + "num_tokens": 42578435.0, + "step": 1652 + }, + { + "epoch": 0.18152866242038215, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.391845703125, + "learning_rate": 6.046852122986823e-07, + "loss": 1.0486, + "mean_token_accuracy": 0.6865143179893494, + "num_tokens": 42604295.0, + "step": 1653 + }, + { + "epoch": 0.18163848012299583, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.2641923427581787, + "learning_rate": 6.050512445095168e-07, + "loss": 1.12, + "mean_token_accuracy": 0.6805433034896851, + "num_tokens": 42632476.0, + "step": 1654 + }, + { + "epoch": 0.1817482978256095, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.4238028526306152, + "learning_rate": 6.054172767203514e-07, + "loss": 1.0986, + "mean_token_accuracy": 0.6755067706108093, + "num_tokens": 42657921.0, + "step": 1655 + }, + { + "epoch": 0.18185811552822315, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.4561197757720947, + "learning_rate": 6.057833089311859e-07, + "loss": 1.1406, + "mean_token_accuracy": 0.6651204824447632, + "num_tokens": 42683798.0, + "step": 1656 + }, + { + "epoch": 0.18196793323083682, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3477070331573486, + "learning_rate": 6.061493411420205e-07, + "loss": 0.9935, + "mean_token_accuracy": 0.7021312713623047, + "num_tokens": 42707333.0, + "step": 1657 + }, + { + "epoch": 0.18207775093345047, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.2486793994903564, + "learning_rate": 6.065153733528551e-07, + "loss": 1.0453, + "mean_token_accuracy": 0.6944440603256226, + "num_tokens": 42733776.0, + "step": 1658 + }, + { + "epoch": 0.18218756863606414, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 1.9383103847503662, + "learning_rate": 6.068814055636896e-07, + "loss": 1.0868, + "mean_token_accuracy": 0.6780873537063599, + "num_tokens": 42769503.0, + "step": 1659 + }, + { + "epoch": 0.18229738633867779, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.551542282104492, + "learning_rate": 6.072474377745242e-07, + "loss": 1.004, + "mean_token_accuracy": 0.7046754956245422, + "num_tokens": 42791718.0, + "step": 1660 + }, + { + "epoch": 0.18240720404129146, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.338453769683838, + "learning_rate": 6.076134699853587e-07, + "loss": 1.0301, + "mean_token_accuracy": 0.6922087669372559, + "num_tokens": 42816722.0, + "step": 1661 + }, + { + "epoch": 0.1825170217439051, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.11456561088562, + "learning_rate": 6.079795021961932e-07, + "loss": 0.9812, + "mean_token_accuracy": 0.6995911002159119, + "num_tokens": 42844961.0, + "step": 1662 + }, + { + "epoch": 0.18262683944651878, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.1910483837127686, + "learning_rate": 6.083455344070278e-07, + "loss": 0.9392, + "mean_token_accuracy": 0.7109293937683105, + "num_tokens": 42870511.0, + "step": 1663 + }, + { + "epoch": 0.18273665714913245, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.173222780227661, + "learning_rate": 6.087115666178624e-07, + "loss": 0.9657, + "mean_token_accuracy": 0.7074779272079468, + "num_tokens": 42897649.0, + "step": 1664 + }, + { + "epoch": 0.1828464748517461, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.074831485748291, + "learning_rate": 6.090775988286969e-07, + "loss": 0.9761, + "mean_token_accuracy": 0.704648494720459, + "num_tokens": 42926306.0, + "step": 1665 + }, + { + "epoch": 0.18295629255435977, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.272789716720581, + "learning_rate": 6.094436310395315e-07, + "loss": 1.0154, + "mean_token_accuracy": 0.6939423680305481, + "num_tokens": 42951568.0, + "step": 1666 + }, + { + "epoch": 0.18306611025697342, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.751945972442627, + "learning_rate": 6.09809663250366e-07, + "loss": 1.0956, + "mean_token_accuracy": 0.6871529817581177, + "num_tokens": 42975783.0, + "step": 1667 + }, + { + "epoch": 0.1831759279595871, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.1628434658050537, + "learning_rate": 6.101756954612005e-07, + "loss": 1.1179, + "mean_token_accuracy": 0.6717609167098999, + "num_tokens": 43004209.0, + "step": 1668 + }, + { + "epoch": 0.18328574566220074, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.40207576751709, + "learning_rate": 6.105417276720352e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6870927810668945, + "num_tokens": 43028852.0, + "step": 1669 + }, + { + "epoch": 0.1833955633648144, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3878164291381836, + "learning_rate": 6.109077598828697e-07, + "loss": 1.0037, + "mean_token_accuracy": 0.6946159601211548, + "num_tokens": 43051679.0, + "step": 1670 + }, + { + "epoch": 0.18350538106742806, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.2352356910705566, + "learning_rate": 6.112737920937042e-07, + "loss": 1.0925, + "mean_token_accuracy": 0.6775729656219482, + "num_tokens": 43079437.0, + "step": 1671 + }, + { + "epoch": 0.18361519877004173, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.0088889598846436, + "learning_rate": 6.116398243045388e-07, + "loss": 1.0636, + "mean_token_accuracy": 0.6819295883178711, + "num_tokens": 43111882.0, + "step": 1672 + }, + { + "epoch": 0.1837250164726554, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.335367441177368, + "learning_rate": 6.120058565153733e-07, + "loss": 1.0687, + "mean_token_accuracy": 0.6920658349990845, + "num_tokens": 43138332.0, + "step": 1673 + }, + { + "epoch": 0.18383483417526905, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.3162198066711426, + "learning_rate": 6.12371888726208e-07, + "loss": 1.0655, + "mean_token_accuracy": 0.6864722371101379, + "num_tokens": 43163166.0, + "step": 1674 + }, + { + "epoch": 0.18394465187788273, + "ewc_loss": 5.140900611877441e-07, + "grad_norm": 2.0847392082214355, + "learning_rate": 6.127379209370425e-07, + "loss": 1.0183, + "mean_token_accuracy": 0.6984710693359375, + "num_tokens": 43192052.0, + "step": 1675 + }, + { + "epoch": 0.18405446958049637, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.192821979522705, + "learning_rate": 6.13103953147877e-07, + "loss": 0.9763, + "mean_token_accuracy": 0.7018531560897827, + "num_tokens": 43218645.0, + "step": 1676 + }, + { + "epoch": 0.18416428728311005, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.4484033584594727, + "learning_rate": 6.134699853587116e-07, + "loss": 1.0748, + "mean_token_accuracy": 0.6745939254760742, + "num_tokens": 43241509.0, + "step": 1677 + }, + { + "epoch": 0.1842741049857237, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.513072967529297, + "learning_rate": 6.138360175695461e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6952076554298401, + "num_tokens": 43264698.0, + "step": 1678 + }, + { + "epoch": 0.18438392268833736, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.4769132137298584, + "learning_rate": 6.142020497803806e-07, + "loss": 0.987, + "mean_token_accuracy": 0.7066450119018555, + "num_tokens": 43287337.0, + "step": 1679 + }, + { + "epoch": 0.184493740390951, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.4286301136016846, + "learning_rate": 6.145680819912153e-07, + "loss": 1.0729, + "mean_token_accuracy": 0.6844549775123596, + "num_tokens": 43313430.0, + "step": 1680 + }, + { + "epoch": 0.18460355809356468, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.340475082397461, + "learning_rate": 6.149341142020498e-07, + "loss": 1.0466, + "mean_token_accuracy": 0.6884039044380188, + "num_tokens": 43337787.0, + "step": 1681 + }, + { + "epoch": 0.18471337579617833, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 1.990899920463562, + "learning_rate": 6.153001464128843e-07, + "loss": 1.098, + "mean_token_accuracy": 0.6749856472015381, + "num_tokens": 43373378.0, + "step": 1682 + }, + { + "epoch": 0.184823193498792, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 1.9995598793029785, + "learning_rate": 6.156661786237189e-07, + "loss": 0.9561, + "mean_token_accuracy": 0.7098480463027954, + "num_tokens": 43404353.0, + "step": 1683 + }, + { + "epoch": 0.18493301120140568, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.132706880569458, + "learning_rate": 6.160322108345534e-07, + "loss": 1.0883, + "mean_token_accuracy": 0.6808222532272339, + "num_tokens": 43433934.0, + "step": 1684 + }, + { + "epoch": 0.18504282890401932, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.0984160900115967, + "learning_rate": 6.16398243045388e-07, + "loss": 0.9851, + "mean_token_accuracy": 0.6985746622085571, + "num_tokens": 43466980.0, + "step": 1685 + }, + { + "epoch": 0.185152646606633, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.2144389152526855, + "learning_rate": 6.167642752562226e-07, + "loss": 1.0157, + "mean_token_accuracy": 0.7038843631744385, + "num_tokens": 43494439.0, + "step": 1686 + }, + { + "epoch": 0.18526246430924664, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.2359862327575684, + "learning_rate": 6.171303074670571e-07, + "loss": 1.1314, + "mean_token_accuracy": 0.6717974543571472, + "num_tokens": 43521827.0, + "step": 1687 + }, + { + "epoch": 0.18537228201186032, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.4576475620269775, + "learning_rate": 6.174963396778916e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.6750383377075195, + "num_tokens": 43546844.0, + "step": 1688 + }, + { + "epoch": 0.18548209971447396, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.142108201980591, + "learning_rate": 6.178623718887262e-07, + "loss": 1.1066, + "mean_token_accuracy": 0.6680349111557007, + "num_tokens": 43579098.0, + "step": 1689 + }, + { + "epoch": 0.18559191741708764, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.605792760848999, + "learning_rate": 6.182284040995608e-07, + "loss": 0.905, + "mean_token_accuracy": 0.7247169613838196, + "num_tokens": 43599155.0, + "step": 1690 + }, + { + "epoch": 0.18570173511970128, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.431670665740967, + "learning_rate": 6.185944363103954e-07, + "loss": 1.0091, + "mean_token_accuracy": 0.6981000900268555, + "num_tokens": 43622669.0, + "step": 1691 + }, + { + "epoch": 0.18581155282231496, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.360811710357666, + "learning_rate": 6.189604685212299e-07, + "loss": 0.9982, + "mean_token_accuracy": 0.6990798711776733, + "num_tokens": 43646422.0, + "step": 1692 + }, + { + "epoch": 0.18592137052492863, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.068819522857666, + "learning_rate": 6.193265007320644e-07, + "loss": 1.0006, + "mean_token_accuracy": 0.6997963786125183, + "num_tokens": 43676400.0, + "step": 1693 + }, + { + "epoch": 0.18603118822754228, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.626195192337036, + "learning_rate": 6.19692532942899e-07, + "loss": 1.0074, + "mean_token_accuracy": 0.6986321210861206, + "num_tokens": 43697193.0, + "step": 1694 + }, + { + "epoch": 0.18614100593015595, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.46842885017395, + "learning_rate": 6.200585651537335e-07, + "loss": 1.047, + "mean_token_accuracy": 0.6894004344940186, + "num_tokens": 43721079.0, + "step": 1695 + }, + { + "epoch": 0.1862508236327696, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.33078932762146, + "learning_rate": 6.204245973645681e-07, + "loss": 1.0392, + "mean_token_accuracy": 0.6953245401382446, + "num_tokens": 43747631.0, + "step": 1696 + }, + { + "epoch": 0.18636064133538327, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.140864610671997, + "learning_rate": 6.207906295754027e-07, + "loss": 1.0098, + "mean_token_accuracy": 0.6963469982147217, + "num_tokens": 43778341.0, + "step": 1697 + }, + { + "epoch": 0.18647045903799692, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.415252208709717, + "learning_rate": 6.211566617862372e-07, + "loss": 1.007, + "mean_token_accuracy": 0.698569655418396, + "num_tokens": 43802076.0, + "step": 1698 + }, + { + "epoch": 0.1865802767406106, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.304410457611084, + "learning_rate": 6.215226939970716e-07, + "loss": 1.0826, + "mean_token_accuracy": 0.6781333684921265, + "num_tokens": 43829357.0, + "step": 1699 + }, + { + "epoch": 0.18669009444322424, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.4533486366271973, + "learning_rate": 6.218887262079063e-07, + "loss": 0.9575, + "mean_token_accuracy": 0.7045809030532837, + "num_tokens": 43852687.0, + "step": 1700 + }, + { + "epoch": 0.1867999121458379, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.3564019203186035, + "learning_rate": 6.222547584187409e-07, + "loss": 1.0536, + "mean_token_accuracy": 0.6918487548828125, + "num_tokens": 43877297.0, + "step": 1701 + }, + { + "epoch": 0.18690972984845158, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.458688974380493, + "learning_rate": 6.226207906295753e-07, + "loss": 0.9949, + "mean_token_accuracy": 0.6969630718231201, + "num_tokens": 43899048.0, + "step": 1702 + }, + { + "epoch": 0.18701954755106523, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.403898000717163, + "learning_rate": 6.2298682284041e-07, + "loss": 0.9271, + "mean_token_accuracy": 0.7160347104072571, + "num_tokens": 43921618.0, + "step": 1703 + }, + { + "epoch": 0.1871293652536789, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 1.9918298721313477, + "learning_rate": 6.233528550512444e-07, + "loss": 1.0803, + "mean_token_accuracy": 0.6891622543334961, + "num_tokens": 43952213.0, + "step": 1704 + }, + { + "epoch": 0.18723918295629255, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.1615591049194336, + "learning_rate": 6.237188872620789e-07, + "loss": 1.0083, + "mean_token_accuracy": 0.7016129493713379, + "num_tokens": 43979710.0, + "step": 1705 + }, + { + "epoch": 0.18734900065890622, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.253680944442749, + "learning_rate": 6.240849194729137e-07, + "loss": 1.0101, + "mean_token_accuracy": 0.6968215703964233, + "num_tokens": 44005736.0, + "step": 1706 + }, + { + "epoch": 0.18745881836151987, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.0014572143554688, + "learning_rate": 6.244509516837481e-07, + "loss": 1.1085, + "mean_token_accuracy": 0.679379403591156, + "num_tokens": 44036907.0, + "step": 1707 + }, + { + "epoch": 0.18756863606413354, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 1.996730923652649, + "learning_rate": 6.248169838945828e-07, + "loss": 1.0651, + "mean_token_accuracy": 0.681178867816925, + "num_tokens": 44067574.0, + "step": 1708 + }, + { + "epoch": 0.1876784537667472, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.2858901023864746, + "learning_rate": 6.251830161054172e-07, + "loss": 1.0979, + "mean_token_accuracy": 0.6746199131011963, + "num_tokens": 44092129.0, + "step": 1709 + }, + { + "epoch": 0.18778827146936086, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.054088592529297, + "learning_rate": 6.255490483162517e-07, + "loss": 1.1199, + "mean_token_accuracy": 0.672714352607727, + "num_tokens": 44124520.0, + "step": 1710 + }, + { + "epoch": 0.18789808917197454, + "ewc_loss": 5.178153514862061e-07, + "grad_norm": 2.320007085800171, + "learning_rate": 6.259150805270863e-07, + "loss": 1.0703, + "mean_token_accuracy": 0.6904482841491699, + "num_tokens": 44150189.0, + "step": 1711 + }, + { + "epoch": 0.18800790687458818, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.496978759765625, + "learning_rate": 6.262811127379209e-07, + "loss": 0.9826, + "mean_token_accuracy": 0.7058804035186768, + "num_tokens": 44171622.0, + "step": 1712 + }, + { + "epoch": 0.18811772457720186, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.319201946258545, + "learning_rate": 6.266471449487554e-07, + "loss": 1.0736, + "mean_token_accuracy": 0.6910052299499512, + "num_tokens": 44196722.0, + "step": 1713 + }, + { + "epoch": 0.1882275422798155, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.0927734375, + "learning_rate": 6.2701317715959e-07, + "loss": 1.1115, + "mean_token_accuracy": 0.6694464683532715, + "num_tokens": 44228788.0, + "step": 1714 + }, + { + "epoch": 0.18833735998242918, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.469974994659424, + "learning_rate": 6.273792093704245e-07, + "loss": 1.0581, + "mean_token_accuracy": 0.6844947338104248, + "num_tokens": 44253946.0, + "step": 1715 + }, + { + "epoch": 0.18844717768504282, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.4026682376861572, + "learning_rate": 6.27745241581259e-07, + "loss": 1.0628, + "mean_token_accuracy": 0.6821643710136414, + "num_tokens": 44277666.0, + "step": 1716 + }, + { + "epoch": 0.1885569953876565, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.0739338397979736, + "learning_rate": 6.281112737920937e-07, + "loss": 0.9945, + "mean_token_accuracy": 0.7001961469650269, + "num_tokens": 44306333.0, + "step": 1717 + }, + { + "epoch": 0.18866681309027014, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.1937952041625977, + "learning_rate": 6.284773060029282e-07, + "loss": 1.0778, + "mean_token_accuracy": 0.6765192151069641, + "num_tokens": 44333933.0, + "step": 1718 + }, + { + "epoch": 0.18877663079288381, + "ewc_loss": 5.21540641784668e-07, + "grad_norm": 2.3441145420074463, + "learning_rate": 6.288433382137627e-07, + "loss": 0.9913, + "mean_token_accuracy": 0.6985836625099182, + "num_tokens": 44358642.0, + "step": 1719 + }, + { + "epoch": 0.18888644849549746, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.4782936573028564, + "learning_rate": 6.292093704245973e-07, + "loss": 1.0239, + "mean_token_accuracy": 0.6884479522705078, + "num_tokens": 44381083.0, + "step": 1720 + }, + { + "epoch": 0.18899626619811113, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.2824208736419678, + "learning_rate": 6.295754026354318e-07, + "loss": 1.0758, + "mean_token_accuracy": 0.6853042840957642, + "num_tokens": 44406847.0, + "step": 1721 + }, + { + "epoch": 0.1891060839007248, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.2155370712280273, + "learning_rate": 6.299414348462665e-07, + "loss": 1.0621, + "mean_token_accuracy": 0.6770586967468262, + "num_tokens": 44433809.0, + "step": 1722 + }, + { + "epoch": 0.18921590160333845, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.417977809906006, + "learning_rate": 6.30307467057101e-07, + "loss": 1.0529, + "mean_token_accuracy": 0.6832685470581055, + "num_tokens": 44458042.0, + "step": 1723 + }, + { + "epoch": 0.18932571930595213, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.681828260421753, + "learning_rate": 6.306734992679355e-07, + "loss": 1.0052, + "mean_token_accuracy": 0.6971968412399292, + "num_tokens": 44478035.0, + "step": 1724 + }, + { + "epoch": 0.18943553700856577, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 1.9545317888259888, + "learning_rate": 6.310395314787701e-07, + "loss": 1.1098, + "mean_token_accuracy": 0.6677041053771973, + "num_tokens": 44513865.0, + "step": 1725 + }, + { + "epoch": 0.18954535471117945, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.7498466968536377, + "learning_rate": 6.314055636896046e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6977046132087708, + "num_tokens": 44534733.0, + "step": 1726 + }, + { + "epoch": 0.1896551724137931, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.235507011413574, + "learning_rate": 6.317715959004391e-07, + "loss": 0.9805, + "mean_token_accuracy": 0.7017946839332581, + "num_tokens": 44560374.0, + "step": 1727 + }, + { + "epoch": 0.18976499011640677, + "ewc_loss": 5.252659320831299e-07, + "grad_norm": 2.7038049697875977, + "learning_rate": 6.321376281112738e-07, + "loss": 1.0198, + "mean_token_accuracy": 0.710088849067688, + "num_tokens": 44579550.0, + "step": 1728 + }, + { + "epoch": 0.1898748078190204, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.1646220684051514, + "learning_rate": 6.325036603221083e-07, + "loss": 1.056, + "mean_token_accuracy": 0.6816519498825073, + "num_tokens": 44605935.0, + "step": 1729 + }, + { + "epoch": 0.1899846255216341, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.49359130859375, + "learning_rate": 6.328696925329428e-07, + "loss": 1.0232, + "mean_token_accuracy": 0.694817841053009, + "num_tokens": 44628919.0, + "step": 1730 + }, + { + "epoch": 0.19009444322424776, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.2318029403686523, + "learning_rate": 6.332357247437774e-07, + "loss": 1.052, + "mean_token_accuracy": 0.6817079186439514, + "num_tokens": 44656105.0, + "step": 1731 + }, + { + "epoch": 0.1902042609268614, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.1262705326080322, + "learning_rate": 6.336017569546119e-07, + "loss": 1.0601, + "mean_token_accuracy": 0.6794731020927429, + "num_tokens": 44686900.0, + "step": 1732 + }, + { + "epoch": 0.19031407862947508, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.540600299835205, + "learning_rate": 6.339677891654465e-07, + "loss": 1.0049, + "mean_token_accuracy": 0.6988232135772705, + "num_tokens": 44710054.0, + "step": 1733 + }, + { + "epoch": 0.19042389633208873, + "ewc_loss": 5.289912223815918e-07, + "grad_norm": 2.6424877643585205, + "learning_rate": 6.343338213762811e-07, + "loss": 0.9874, + "mean_token_accuracy": 0.7056383490562439, + "num_tokens": 44730033.0, + "step": 1734 + }, + { + "epoch": 0.1905337140347024, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.2710156440734863, + "learning_rate": 6.346998535871156e-07, + "loss": 0.9536, + "mean_token_accuracy": 0.7075956463813782, + "num_tokens": 44756831.0, + "step": 1735 + }, + { + "epoch": 0.19064353173731605, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.300058126449585, + "learning_rate": 6.350658857979501e-07, + "loss": 1.0325, + "mean_token_accuracy": 0.6883590221405029, + "num_tokens": 44782982.0, + "step": 1736 + }, + { + "epoch": 0.19075334943992972, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.2407174110412598, + "learning_rate": 6.354319180087847e-07, + "loss": 1.1554, + "mean_token_accuracy": 0.6663603782653809, + "num_tokens": 44811212.0, + "step": 1737 + }, + { + "epoch": 0.19086316714254337, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.494108200073242, + "learning_rate": 6.357979502196193e-07, + "loss": 1.0082, + "mean_token_accuracy": 0.6991807818412781, + "num_tokens": 44833233.0, + "step": 1738 + }, + { + "epoch": 0.19097298484515704, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.0675039291381836, + "learning_rate": 6.361639824304539e-07, + "loss": 1.0553, + "mean_token_accuracy": 0.6886705160140991, + "num_tokens": 44861753.0, + "step": 1739 + }, + { + "epoch": 0.1910828025477707, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.206411838531494, + "learning_rate": 6.365300146412884e-07, + "loss": 1.0383, + "mean_token_accuracy": 0.6984043121337891, + "num_tokens": 44887824.0, + "step": 1740 + }, + { + "epoch": 0.19119262025038436, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.256047010421753, + "learning_rate": 6.368960468521229e-07, + "loss": 1.0236, + "mean_token_accuracy": 0.6866997480392456, + "num_tokens": 44914371.0, + "step": 1741 + }, + { + "epoch": 0.19130243795299803, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.078761100769043, + "learning_rate": 6.372620790629575e-07, + "loss": 1.091, + "mean_token_accuracy": 0.6753016710281372, + "num_tokens": 44943379.0, + "step": 1742 + }, + { + "epoch": 0.19141225565561168, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.4369213581085205, + "learning_rate": 6.37628111273792e-07, + "loss": 1.1079, + "mean_token_accuracy": 0.6800948977470398, + "num_tokens": 44967204.0, + "step": 1743 + }, + { + "epoch": 0.19152207335822535, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.184662103652954, + "learning_rate": 6.379941434846266e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.6990865468978882, + "num_tokens": 44993218.0, + "step": 1744 + }, + { + "epoch": 0.191631891060839, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.4075536727905273, + "learning_rate": 6.383601756954612e-07, + "loss": 1.0148, + "mean_token_accuracy": 0.6933621764183044, + "num_tokens": 45016771.0, + "step": 1745 + }, + { + "epoch": 0.19174170876345267, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.4768335819244385, + "learning_rate": 6.387262079062957e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6860824823379517, + "num_tokens": 45041696.0, + "step": 1746 + }, + { + "epoch": 0.19185152646606632, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.265629768371582, + "learning_rate": 6.390922401171302e-07, + "loss": 1.0685, + "mean_token_accuracy": 0.6841386556625366, + "num_tokens": 45067615.0, + "step": 1747 + }, + { + "epoch": 0.19196134416868, + "ewc_loss": 5.327165126800537e-07, + "grad_norm": 2.419712543487549, + "learning_rate": 6.394582723279648e-07, + "loss": 1.0677, + "mean_token_accuracy": 0.691484808921814, + "num_tokens": 45092065.0, + "step": 1748 + }, + { + "epoch": 0.19207116187129367, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.3977785110473633, + "learning_rate": 6.398243045387994e-07, + "loss": 1.0852, + "mean_token_accuracy": 0.6818916201591492, + "num_tokens": 45116555.0, + "step": 1749 + }, + { + "epoch": 0.1921809795739073, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.2443485260009766, + "learning_rate": 6.401903367496339e-07, + "loss": 1.0939, + "mean_token_accuracy": 0.6792292594909668, + "num_tokens": 45142968.0, + "step": 1750 + }, + { + "epoch": 0.19229079727652099, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.607665538787842, + "learning_rate": 6.405563689604685e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7002283334732056, + "num_tokens": 45162421.0, + "step": 1751 + }, + { + "epoch": 0.19240061497913463, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.242281675338745, + "learning_rate": 6.40922401171303e-07, + "loss": 0.9506, + "mean_token_accuracy": 0.7106649279594421, + "num_tokens": 45188853.0, + "step": 1752 + }, + { + "epoch": 0.1925104326817483, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.352003812789917, + "learning_rate": 6.412884333821375e-07, + "loss": 1.0902, + "mean_token_accuracy": 0.6963208913803101, + "num_tokens": 45211908.0, + "step": 1753 + }, + { + "epoch": 0.19262025038436195, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.254190683364868, + "learning_rate": 6.416544655929722e-07, + "loss": 0.9509, + "mean_token_accuracy": 0.711993932723999, + "num_tokens": 45235897.0, + "step": 1754 + }, + { + "epoch": 0.19273006808697563, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.163469076156616, + "learning_rate": 6.420204978038067e-07, + "loss": 1.1879, + "mean_token_accuracy": 0.6540935039520264, + "num_tokens": 45266668.0, + "step": 1755 + }, + { + "epoch": 0.19283988578958927, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.3179306983947754, + "learning_rate": 6.423865300146413e-07, + "loss": 1.0848, + "mean_token_accuracy": 0.6739643812179565, + "num_tokens": 45291961.0, + "step": 1756 + }, + { + "epoch": 0.19294970349220295, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.1143102645874023, + "learning_rate": 6.427525622254758e-07, + "loss": 1.0497, + "mean_token_accuracy": 0.6889088749885559, + "num_tokens": 45321224.0, + "step": 1757 + }, + { + "epoch": 0.1930595211948166, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.2101922035217285, + "learning_rate": 6.431185944363103e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6934995651245117, + "num_tokens": 45349120.0, + "step": 1758 + }, + { + "epoch": 0.19316933889743026, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.504025936126709, + "learning_rate": 6.434846266471449e-07, + "loss": 0.9593, + "mean_token_accuracy": 0.70952969789505, + "num_tokens": 45370609.0, + "step": 1759 + }, + { + "epoch": 0.19327915660004394, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.0376639366149902, + "learning_rate": 6.438506588579795e-07, + "loss": 1.021, + "mean_token_accuracy": 0.6955846548080444, + "num_tokens": 45402004.0, + "step": 1760 + }, + { + "epoch": 0.19338897430265758, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.5588817596435547, + "learning_rate": 6.44216691068814e-07, + "loss": 0.9696, + "mean_token_accuracy": 0.7079539895057678, + "num_tokens": 45423421.0, + "step": 1761 + }, + { + "epoch": 0.19349879200527126, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.1387534141540527, + "learning_rate": 6.445827232796486e-07, + "loss": 0.9749, + "mean_token_accuracy": 0.7094268798828125, + "num_tokens": 45449405.0, + "step": 1762 + }, + { + "epoch": 0.1936086097078849, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.492577314376831, + "learning_rate": 6.449487554904831e-07, + "loss": 0.9844, + "mean_token_accuracy": 0.694514274597168, + "num_tokens": 45470264.0, + "step": 1763 + }, + { + "epoch": 0.19371842741049858, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.036320447921753, + "learning_rate": 6.453147877013176e-07, + "loss": 0.9831, + "mean_token_accuracy": 0.7099088430404663, + "num_tokens": 45501935.0, + "step": 1764 + }, + { + "epoch": 0.19382824511311222, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.2585532665252686, + "learning_rate": 6.456808199121523e-07, + "loss": 1.079, + "mean_token_accuracy": 0.6811540126800537, + "num_tokens": 45528933.0, + "step": 1765 + }, + { + "epoch": 0.1939380628157259, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.395641326904297, + "learning_rate": 6.460468521229868e-07, + "loss": 0.9391, + "mean_token_accuracy": 0.7133556604385376, + "num_tokens": 45551417.0, + "step": 1766 + }, + { + "epoch": 0.19404788051833954, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.0248024463653564, + "learning_rate": 6.464128843338213e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6932620406150818, + "num_tokens": 45582175.0, + "step": 1767 + }, + { + "epoch": 0.19415769822095322, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.410290479660034, + "learning_rate": 6.467789165446559e-07, + "loss": 1.0995, + "mean_token_accuracy": 0.6726081371307373, + "num_tokens": 45606661.0, + "step": 1768 + }, + { + "epoch": 0.1942675159235669, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.3708536624908447, + "learning_rate": 6.471449487554904e-07, + "loss": 1.0597, + "mean_token_accuracy": 0.6815528273582458, + "num_tokens": 45632245.0, + "step": 1769 + }, + { + "epoch": 0.19437733362618054, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 1.9431648254394531, + "learning_rate": 6.475109809663251e-07, + "loss": 1.0769, + "mean_token_accuracy": 0.6836221218109131, + "num_tokens": 45666175.0, + "step": 1770 + }, + { + "epoch": 0.1944871513287942, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.333839178085327, + "learning_rate": 6.478770131771596e-07, + "loss": 1.0409, + "mean_token_accuracy": 0.6985114812850952, + "num_tokens": 45692819.0, + "step": 1771 + }, + { + "epoch": 0.19459696903140786, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.091235399246216, + "learning_rate": 6.482430453879941e-07, + "loss": 1.0389, + "mean_token_accuracy": 0.6945105195045471, + "num_tokens": 45721537.0, + "step": 1772 + }, + { + "epoch": 0.19470678673402153, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.6093590259552, + "learning_rate": 6.486090775988287e-07, + "loss": 1.0791, + "mean_token_accuracy": 0.6868964433670044, + "num_tokens": 45745690.0, + "step": 1773 + }, + { + "epoch": 0.19481660443663518, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.375331163406372, + "learning_rate": 6.489751098096632e-07, + "loss": 1.1063, + "mean_token_accuracy": 0.6743952035903931, + "num_tokens": 45771981.0, + "step": 1774 + }, + { + "epoch": 0.19492642213924885, + "ewc_loss": 5.364418029785156e-07, + "grad_norm": 2.442711353302002, + "learning_rate": 6.493411420204977e-07, + "loss": 1.036, + "mean_token_accuracy": 0.6943848133087158, + "num_tokens": 45796644.0, + "step": 1775 + }, + { + "epoch": 0.1950362398418625, + "ewc_loss": 5.438923835754395e-07, + "grad_norm": 2.7975332736968994, + "learning_rate": 6.497071742313324e-07, + "loss": 1.0749, + "mean_token_accuracy": 0.6783679127693176, + "num_tokens": 45815886.0, + "step": 1776 + }, + { + "epoch": 0.19514605754447617, + "ewc_loss": 5.438923835754395e-07, + "grad_norm": 2.215841054916382, + "learning_rate": 6.500732064421669e-07, + "loss": 1.0485, + "mean_token_accuracy": 0.688444972038269, + "num_tokens": 45843574.0, + "step": 1777 + }, + { + "epoch": 0.19525587524708984, + "ewc_loss": 5.438923835754395e-07, + "grad_norm": 2.20106840133667, + "learning_rate": 6.504392386530014e-07, + "loss": 1.0159, + "mean_token_accuracy": 0.6985558867454529, + "num_tokens": 45870615.0, + "step": 1778 + }, + { + "epoch": 0.1953656929497035, + "ewc_loss": 5.438923835754395e-07, + "grad_norm": 2.479685068130493, + "learning_rate": 6.50805270863836e-07, + "loss": 1.0237, + "mean_token_accuracy": 0.6967913508415222, + "num_tokens": 45892574.0, + "step": 1779 + }, + { + "epoch": 0.19547551065231716, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.371727228164673, + "learning_rate": 6.511713030746705e-07, + "loss": 1.093, + "mean_token_accuracy": 0.6768169403076172, + "num_tokens": 45917428.0, + "step": 1780 + }, + { + "epoch": 0.1955853283549308, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.0536623001098633, + "learning_rate": 6.515373352855051e-07, + "loss": 1.0182, + "mean_token_accuracy": 0.6907603740692139, + "num_tokens": 45947191.0, + "step": 1781 + }, + { + "epoch": 0.19569514605754448, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.2375051975250244, + "learning_rate": 6.519033674963397e-07, + "loss": 0.9957, + "mean_token_accuracy": 0.6996767520904541, + "num_tokens": 45973562.0, + "step": 1782 + }, + { + "epoch": 0.19580496376015813, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3120336532592773, + "learning_rate": 6.522693997071742e-07, + "loss": 1.0019, + "mean_token_accuracy": 0.7025423645973206, + "num_tokens": 46000779.0, + "step": 1783 + }, + { + "epoch": 0.1959147814627718, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.276975631713867, + "learning_rate": 6.526354319180087e-07, + "loss": 1.0346, + "mean_token_accuracy": 0.6932718753814697, + "num_tokens": 46025924.0, + "step": 1784 + }, + { + "epoch": 0.19602459916538545, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.304128646850586, + "learning_rate": 6.530014641288433e-07, + "loss": 1.0456, + "mean_token_accuracy": 0.6860873103141785, + "num_tokens": 46052335.0, + "step": 1785 + }, + { + "epoch": 0.19613441686799912, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3609843254089355, + "learning_rate": 6.533674963396779e-07, + "loss": 0.9709, + "mean_token_accuracy": 0.7079428434371948, + "num_tokens": 46075495.0, + "step": 1786 + }, + { + "epoch": 0.1962442345706128, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.663933277130127, + "learning_rate": 6.537335285505125e-07, + "loss": 0.9914, + "mean_token_accuracy": 0.6975936889648438, + "num_tokens": 46094772.0, + "step": 1787 + }, + { + "epoch": 0.19635405227322644, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3119938373565674, + "learning_rate": 6.54099560761347e-07, + "loss": 1.0982, + "mean_token_accuracy": 0.6729925870895386, + "num_tokens": 46123044.0, + "step": 1788 + }, + { + "epoch": 0.19646386997584012, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3306901454925537, + "learning_rate": 6.544655929721815e-07, + "loss": 1.0331, + "mean_token_accuracy": 0.691886305809021, + "num_tokens": 46146887.0, + "step": 1789 + }, + { + "epoch": 0.19657368767845376, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.486380100250244, + "learning_rate": 6.548316251830161e-07, + "loss": 1.0255, + "mean_token_accuracy": 0.6998759508132935, + "num_tokens": 46168260.0, + "step": 1790 + }, + { + "epoch": 0.19668350538106744, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.4880316257476807, + "learning_rate": 6.551976573938506e-07, + "loss": 1.0834, + "mean_token_accuracy": 0.6857962608337402, + "num_tokens": 46190266.0, + "step": 1791 + }, + { + "epoch": 0.19679332308368108, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3510165214538574, + "learning_rate": 6.555636896046852e-07, + "loss": 1.051, + "mean_token_accuracy": 0.6864098310470581, + "num_tokens": 46214139.0, + "step": 1792 + }, + { + "epoch": 0.19690314078629476, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.637827157974243, + "learning_rate": 6.559297218155198e-07, + "loss": 1.0538, + "mean_token_accuracy": 0.6917741298675537, + "num_tokens": 46235396.0, + "step": 1793 + }, + { + "epoch": 0.1970129584889084, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.214090585708618, + "learning_rate": 6.562957540263543e-07, + "loss": 1.0153, + "mean_token_accuracy": 0.6941612958908081, + "num_tokens": 46261539.0, + "step": 1794 + }, + { + "epoch": 0.19712277619152208, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.2850074768066406, + "learning_rate": 6.566617862371888e-07, + "loss": 1.0194, + "mean_token_accuracy": 0.6959185600280762, + "num_tokens": 46286833.0, + "step": 1795 + }, + { + "epoch": 0.19723259389413572, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.234736919403076, + "learning_rate": 6.570278184480234e-07, + "loss": 1.0646, + "mean_token_accuracy": 0.6824132204055786, + "num_tokens": 46313733.0, + "step": 1796 + }, + { + "epoch": 0.1973424115967494, + "ewc_loss": 5.476176738739014e-07, + "grad_norm": 2.3256125450134277, + "learning_rate": 6.57393850658858e-07, + "loss": 1.0665, + "mean_token_accuracy": 0.6839491128921509, + "num_tokens": 46337087.0, + "step": 1797 + }, + { + "epoch": 0.19745222929936307, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.7499196529388428, + "learning_rate": 6.577598828696925e-07, + "loss": 0.9642, + "mean_token_accuracy": 0.7027812004089355, + "num_tokens": 46356648.0, + "step": 1798 + }, + { + "epoch": 0.19756204700197671, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.34763765335083, + "learning_rate": 6.581259150805271e-07, + "loss": 0.9345, + "mean_token_accuracy": 0.7137355804443359, + "num_tokens": 46378939.0, + "step": 1799 + }, + { + "epoch": 0.1976718647045904, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.1230340003967285, + "learning_rate": 6.584919472913616e-07, + "loss": 1.0367, + "mean_token_accuracy": 0.6862177848815918, + "num_tokens": 46408171.0, + "step": 1800 + }, + { + "epoch": 0.19778168240720403, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.420435667037964, + "learning_rate": 6.588579795021961e-07, + "loss": 0.9938, + "mean_token_accuracy": 0.702739953994751, + "num_tokens": 46431563.0, + "step": 1801 + }, + { + "epoch": 0.1978915001098177, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.4191837310791016, + "learning_rate": 6.592240117130308e-07, + "loss": 1.0243, + "mean_token_accuracy": 0.7003772258758545, + "num_tokens": 46455897.0, + "step": 1802 + }, + { + "epoch": 0.19800131781243135, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.125570058822632, + "learning_rate": 6.595900439238653e-07, + "loss": 1.0823, + "mean_token_accuracy": 0.6751330494880676, + "num_tokens": 46485421.0, + "step": 1803 + }, + { + "epoch": 0.19811113551504503, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.746060609817505, + "learning_rate": 6.599560761346999e-07, + "loss": 0.9976, + "mean_token_accuracy": 0.6988588571548462, + "num_tokens": 46504392.0, + "step": 1804 + }, + { + "epoch": 0.19822095321765867, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.444326162338257, + "learning_rate": 6.603221083455344e-07, + "loss": 0.9462, + "mean_token_accuracy": 0.7149683237075806, + "num_tokens": 46526987.0, + "step": 1805 + }, + { + "epoch": 0.19833077092027235, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.328035831451416, + "learning_rate": 6.606881405563689e-07, + "loss": 1.0104, + "mean_token_accuracy": 0.6976950168609619, + "num_tokens": 46552167.0, + "step": 1806 + }, + { + "epoch": 0.19844058862288602, + "ewc_loss": 5.550682544708252e-07, + "grad_norm": 2.339167356491089, + "learning_rate": 6.610541727672035e-07, + "loss": 1.031, + "mean_token_accuracy": 0.6862456798553467, + "num_tokens": 46578431.0, + "step": 1807 + }, + { + "epoch": 0.19855040632549967, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.4543662071228027, + "learning_rate": 6.614202049780381e-07, + "loss": 0.9136, + "mean_token_accuracy": 0.7199428081512451, + "num_tokens": 46597882.0, + "step": 1808 + }, + { + "epoch": 0.19866022402811334, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.1521079540252686, + "learning_rate": 6.617862371888726e-07, + "loss": 1.0879, + "mean_token_accuracy": 0.6917779445648193, + "num_tokens": 46627173.0, + "step": 1809 + }, + { + "epoch": 0.198770041730727, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.3586251735687256, + "learning_rate": 6.621522693997072e-07, + "loss": 1.0425, + "mean_token_accuracy": 0.6945071220397949, + "num_tokens": 46652241.0, + "step": 1810 + }, + { + "epoch": 0.19887985943334066, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.2119719982147217, + "learning_rate": 6.625183016105417e-07, + "loss": 1.085, + "mean_token_accuracy": 0.6750094890594482, + "num_tokens": 46681172.0, + "step": 1811 + }, + { + "epoch": 0.1989896771359543, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.0704996585845947, + "learning_rate": 6.628843338213762e-07, + "loss": 1.0671, + "mean_token_accuracy": 0.6828509569168091, + "num_tokens": 46711892.0, + "step": 1812 + }, + { + "epoch": 0.19909949483856798, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 1.9977296590805054, + "learning_rate": 6.632503660322109e-07, + "loss": 1.0687, + "mean_token_accuracy": 0.6801787614822388, + "num_tokens": 46744839.0, + "step": 1813 + }, + { + "epoch": 0.19920931254118163, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.320178508758545, + "learning_rate": 6.636163982430454e-07, + "loss": 0.9339, + "mean_token_accuracy": 0.7054101228713989, + "num_tokens": 46768706.0, + "step": 1814 + }, + { + "epoch": 0.1993191302437953, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.665846347808838, + "learning_rate": 6.639824304538799e-07, + "loss": 0.9881, + "mean_token_accuracy": 0.7014845013618469, + "num_tokens": 46789029.0, + "step": 1815 + }, + { + "epoch": 0.19942894794640897, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.1461727619171143, + "learning_rate": 6.643484626647145e-07, + "loss": 1.0696, + "mean_token_accuracy": 0.6820883750915527, + "num_tokens": 46819385.0, + "step": 1816 + }, + { + "epoch": 0.19953876564902262, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.2884116172790527, + "learning_rate": 6.64714494875549e-07, + "loss": 1.051, + "mean_token_accuracy": 0.6846237182617188, + "num_tokens": 46845617.0, + "step": 1817 + }, + { + "epoch": 0.1996485833516363, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.4460513591766357, + "learning_rate": 6.650805270863836e-07, + "loss": 0.9772, + "mean_token_accuracy": 0.703492283821106, + "num_tokens": 46867168.0, + "step": 1818 + }, + { + "epoch": 0.19975840105424994, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.362349510192871, + "learning_rate": 6.654465592972182e-07, + "loss": 0.9782, + "mean_token_accuracy": 0.7050442695617676, + "num_tokens": 46893199.0, + "step": 1819 + }, + { + "epoch": 0.1998682187568636, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.6868062019348145, + "learning_rate": 6.658125915080527e-07, + "loss": 1.0174, + "mean_token_accuracy": 0.6956318020820618, + "num_tokens": 46912484.0, + "step": 1820 + }, + { + "epoch": 0.19997803645947726, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 1.9471032619476318, + "learning_rate": 6.661786237188873e-07, + "loss": 1.0998, + "mean_token_accuracy": 0.6842852830886841, + "num_tokens": 46949071.0, + "step": 1821 + }, + { + "epoch": 0.20008785416209093, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.1902215480804443, + "learning_rate": 6.665446559297218e-07, + "loss": 0.9892, + "mean_token_accuracy": 0.6963264346122742, + "num_tokens": 46976660.0, + "step": 1822 + }, + { + "epoch": 0.20019767186470458, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.221085548400879, + "learning_rate": 6.669106881405563e-07, + "loss": 0.9123, + "mean_token_accuracy": 0.7260616421699524, + "num_tokens": 47003264.0, + "step": 1823 + }, + { + "epoch": 0.20030748956731825, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 1.9750112295150757, + "learning_rate": 6.67276720351391e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.6871023774147034, + "num_tokens": 47036317.0, + "step": 1824 + }, + { + "epoch": 0.2004173072699319, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.458155870437622, + "learning_rate": 6.676427525622255e-07, + "loss": 1.0601, + "mean_token_accuracy": 0.6855038404464722, + "num_tokens": 47059959.0, + "step": 1825 + }, + { + "epoch": 0.20052712497254557, + "ewc_loss": 5.62518835067749e-07, + "grad_norm": 2.4019274711608887, + "learning_rate": 6.6800878477306e-07, + "loss": 1.0087, + "mean_token_accuracy": 0.6958675384521484, + "num_tokens": 47083996.0, + "step": 1826 + }, + { + "epoch": 0.20063694267515925, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.145258665084839, + "learning_rate": 6.683748169838946e-07, + "loss": 1.0125, + "mean_token_accuracy": 0.7175220251083374, + "num_tokens": 47111746.0, + "step": 1827 + }, + { + "epoch": 0.2007467603777729, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.2670836448669434, + "learning_rate": 6.687408491947291e-07, + "loss": 1.088, + "mean_token_accuracy": 0.6848232746124268, + "num_tokens": 47137621.0, + "step": 1828 + }, + { + "epoch": 0.20085657808038657, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.3141589164733887, + "learning_rate": 6.691068814055637e-07, + "loss": 0.9873, + "mean_token_accuracy": 0.6996139287948608, + "num_tokens": 47162400.0, + "step": 1829 + }, + { + "epoch": 0.2009663957830002, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.5826761722564697, + "learning_rate": 6.694729136163983e-07, + "loss": 0.9845, + "mean_token_accuracy": 0.7024260759353638, + "num_tokens": 47186479.0, + "step": 1830 + }, + { + "epoch": 0.20107621348561389, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.0256221294403076, + "learning_rate": 6.698389458272328e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6888349056243896, + "num_tokens": 47217322.0, + "step": 1831 + }, + { + "epoch": 0.20118603118822753, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.041508913040161, + "learning_rate": 6.702049780380673e-07, + "loss": 1.0206, + "mean_token_accuracy": 0.6888980865478516, + "num_tokens": 47247815.0, + "step": 1832 + }, + { + "epoch": 0.2012958488908412, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.278397798538208, + "learning_rate": 6.705710102489019e-07, + "loss": 1.0132, + "mean_token_accuracy": 0.6968945264816284, + "num_tokens": 47274328.0, + "step": 1833 + }, + { + "epoch": 0.20140566659345485, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.073148250579834, + "learning_rate": 6.709370424597365e-07, + "loss": 1.0299, + "mean_token_accuracy": 0.6865630149841309, + "num_tokens": 47303590.0, + "step": 1834 + }, + { + "epoch": 0.20151548429606853, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.1410343647003174, + "learning_rate": 6.71303074670571e-07, + "loss": 0.8967, + "mean_token_accuracy": 0.7296860218048096, + "num_tokens": 47331291.0, + "step": 1835 + }, + { + "epoch": 0.2016253019986822, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.274385452270508, + "learning_rate": 6.716691068814056e-07, + "loss": 1.0077, + "mean_token_accuracy": 0.6969131827354431, + "num_tokens": 47356358.0, + "step": 1836 + }, + { + "epoch": 0.20173511970129585, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.1928508281707764, + "learning_rate": 6.720351390922401e-07, + "loss": 0.9963, + "mean_token_accuracy": 0.7005110383033752, + "num_tokens": 47383864.0, + "step": 1837 + }, + { + "epoch": 0.20184493740390952, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.324761390686035, + "learning_rate": 6.724011713030747e-07, + "loss": 1.0272, + "mean_token_accuracy": 0.6921390295028687, + "num_tokens": 47408405.0, + "step": 1838 + }, + { + "epoch": 0.20195475510652316, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.3886241912841797, + "learning_rate": 6.727672035139092e-07, + "loss": 0.9985, + "mean_token_accuracy": 0.703193724155426, + "num_tokens": 47432899.0, + "step": 1839 + }, + { + "epoch": 0.20206457280913684, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.3093113899230957, + "learning_rate": 6.731332357247438e-07, + "loss": 1.1237, + "mean_token_accuracy": 0.6711214780807495, + "num_tokens": 47459200.0, + "step": 1840 + }, + { + "epoch": 0.20217439051175048, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.1377081871032715, + "learning_rate": 6.734992679355784e-07, + "loss": 1.1152, + "mean_token_accuracy": 0.6730726957321167, + "num_tokens": 47489755.0, + "step": 1841 + }, + { + "epoch": 0.20228420821436416, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 1.957831859588623, + "learning_rate": 6.738653001464129e-07, + "loss": 1.0378, + "mean_token_accuracy": 0.688765823841095, + "num_tokens": 47521733.0, + "step": 1842 + }, + { + "epoch": 0.2023940259169778, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.564394235610962, + "learning_rate": 6.742313323572474e-07, + "loss": 0.9532, + "mean_token_accuracy": 0.7135330438613892, + "num_tokens": 47541998.0, + "step": 1843 + }, + { + "epoch": 0.20250384361959148, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.1324422359466553, + "learning_rate": 6.74597364568082e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.688355028629303, + "num_tokens": 47569865.0, + "step": 1844 + }, + { + "epoch": 0.20261366132220515, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.2668676376342773, + "learning_rate": 6.749633967789166e-07, + "loss": 1.116, + "mean_token_accuracy": 0.675367534160614, + "num_tokens": 47598795.0, + "step": 1845 + }, + { + "epoch": 0.2027234790248188, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.1716604232788086, + "learning_rate": 6.753294289897511e-07, + "loss": 1.1154, + "mean_token_accuracy": 0.6714271903038025, + "num_tokens": 47627078.0, + "step": 1846 + }, + { + "epoch": 0.20283329672743247, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 1.978462815284729, + "learning_rate": 6.756954612005857e-07, + "loss": 1.0227, + "mean_token_accuracy": 0.6966361403465271, + "num_tokens": 47660542.0, + "step": 1847 + }, + { + "epoch": 0.20294311443004612, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.302252769470215, + "learning_rate": 6.760614934114202e-07, + "loss": 1.0803, + "mean_token_accuracy": 0.6796773672103882, + "num_tokens": 47687250.0, + "step": 1848 + }, + { + "epoch": 0.2030529321326598, + "ewc_loss": 5.662441253662109e-07, + "grad_norm": 2.5732908248901367, + "learning_rate": 6.764275256222546e-07, + "loss": 0.9771, + "mean_token_accuracy": 0.6956726908683777, + "num_tokens": 47708672.0, + "step": 1849 + }, + { + "epoch": 0.20316274983527344, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.5699918270111084, + "learning_rate": 6.767935578330894e-07, + "loss": 0.9582, + "mean_token_accuracy": 0.7129244804382324, + "num_tokens": 47730052.0, + "step": 1850 + }, + { + "epoch": 0.2032725675378871, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.329662561416626, + "learning_rate": 6.771595900439239e-07, + "loss": 1.0948, + "mean_token_accuracy": 0.6731921434402466, + "num_tokens": 47757372.0, + "step": 1851 + }, + { + "epoch": 0.20338238524050076, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.2962050437927246, + "learning_rate": 6.775256222547583e-07, + "loss": 1.0531, + "mean_token_accuracy": 0.6786946654319763, + "num_tokens": 47784543.0, + "step": 1852 + }, + { + "epoch": 0.20349220294311443, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.125247001647949, + "learning_rate": 6.77891654465593e-07, + "loss": 1.0458, + "mean_token_accuracy": 0.6858481168746948, + "num_tokens": 47813373.0, + "step": 1853 + }, + { + "epoch": 0.2036020206457281, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.298245668411255, + "learning_rate": 6.782576866764274e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.6880471706390381, + "num_tokens": 47837686.0, + "step": 1854 + }, + { + "epoch": 0.20371183834834175, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.0872628688812256, + "learning_rate": 6.78623718887262e-07, + "loss": 1.0683, + "mean_token_accuracy": 0.6896799802780151, + "num_tokens": 47870281.0, + "step": 1855 + }, + { + "epoch": 0.20382165605095542, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.1536762714385986, + "learning_rate": 6.789897510980966e-07, + "loss": 1.0868, + "mean_token_accuracy": 0.6744853258132935, + "num_tokens": 47899080.0, + "step": 1856 + }, + { + "epoch": 0.20393147375356907, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.265085458755493, + "learning_rate": 6.793557833089311e-07, + "loss": 1.0636, + "mean_token_accuracy": 0.6862195730209351, + "num_tokens": 47923965.0, + "step": 1857 + }, + { + "epoch": 0.20404129145618274, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.506613254547119, + "learning_rate": 6.797218155197657e-07, + "loss": 0.9923, + "mean_token_accuracy": 0.7008299827575684, + "num_tokens": 47945835.0, + "step": 1858 + }, + { + "epoch": 0.2041511091587964, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.3761303424835205, + "learning_rate": 6.800878477306002e-07, + "loss": 1.0316, + "mean_token_accuracy": 0.6893604397773743, + "num_tokens": 47971101.0, + "step": 1859 + }, + { + "epoch": 0.20426092686141006, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.436009168624878, + "learning_rate": 6.804538799414347e-07, + "loss": 1.0883, + "mean_token_accuracy": 0.6923087239265442, + "num_tokens": 47995297.0, + "step": 1860 + }, + { + "epoch": 0.2043707445640237, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.0293500423431396, + "learning_rate": 6.808199121522694e-07, + "loss": 1.0752, + "mean_token_accuracy": 0.6783560514450073, + "num_tokens": 48028033.0, + "step": 1861 + }, + { + "epoch": 0.20448056226663738, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.408912181854248, + "learning_rate": 6.811859443631039e-07, + "loss": 1.042, + "mean_token_accuracy": 0.6858952045440674, + "num_tokens": 48050471.0, + "step": 1862 + }, + { + "epoch": 0.20459037996925103, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.4393956661224365, + "learning_rate": 6.815519765739384e-07, + "loss": 1.0345, + "mean_token_accuracy": 0.6904147267341614, + "num_tokens": 48073222.0, + "step": 1863 + }, + { + "epoch": 0.2047001976718647, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.3465182781219482, + "learning_rate": 6.81918008784773e-07, + "loss": 1.0245, + "mean_token_accuracy": 0.6922323107719421, + "num_tokens": 48097610.0, + "step": 1864 + }, + { + "epoch": 0.20481001537447838, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.5690996646881104, + "learning_rate": 6.822840409956075e-07, + "loss": 0.9988, + "mean_token_accuracy": 0.7007539868354797, + "num_tokens": 48119076.0, + "step": 1865 + }, + { + "epoch": 0.20491983307709202, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.2155494689941406, + "learning_rate": 6.826500732064421e-07, + "loss": 0.988, + "mean_token_accuracy": 0.7009710669517517, + "num_tokens": 48145067.0, + "step": 1866 + }, + { + "epoch": 0.2050296507797057, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.3307645320892334, + "learning_rate": 6.830161054172767e-07, + "loss": 1.0525, + "mean_token_accuracy": 0.6778779625892639, + "num_tokens": 48169280.0, + "step": 1867 + }, + { + "epoch": 0.20513946848231934, + "ewc_loss": 5.699694156646729e-07, + "grad_norm": 2.0250909328460693, + "learning_rate": 6.833821376281112e-07, + "loss": 0.9638, + "mean_token_accuracy": 0.704438328742981, + "num_tokens": 48201385.0, + "step": 1868 + }, + { + "epoch": 0.20524928618493302, + "ewc_loss": 5.736947059631348e-07, + "grad_norm": 2.3939223289489746, + "learning_rate": 6.837481698389457e-07, + "loss": 0.9981, + "mean_token_accuracy": 0.6963509321212769, + "num_tokens": 48227893.0, + "step": 1869 + }, + { + "epoch": 0.20535910388754666, + "ewc_loss": 5.736947059631348e-07, + "grad_norm": 2.2723734378814697, + "learning_rate": 6.841142020497803e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6861289739608765, + "num_tokens": 48255573.0, + "step": 1870 + }, + { + "epoch": 0.20546892159016034, + "ewc_loss": 5.736947059631348e-07, + "grad_norm": 2.440389633178711, + "learning_rate": 6.844802342606148e-07, + "loss": 0.9399, + "mean_token_accuracy": 0.7208921909332275, + "num_tokens": 48276990.0, + "step": 1871 + }, + { + "epoch": 0.20557873929277398, + "ewc_loss": 5.736947059631348e-07, + "grad_norm": 2.4822354316711426, + "learning_rate": 6.848462664714495e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.6949270963668823, + "num_tokens": 48298559.0, + "step": 1872 + }, + { + "epoch": 0.20568855699538766, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.1348049640655518, + "learning_rate": 6.85212298682284e-07, + "loss": 1.0743, + "mean_token_accuracy": 0.6740399599075317, + "num_tokens": 48325901.0, + "step": 1873 + }, + { + "epoch": 0.20579837469800133, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.535555839538574, + "learning_rate": 6.855783308931185e-07, + "loss": 1.0836, + "mean_token_accuracy": 0.6850264072418213, + "num_tokens": 48347441.0, + "step": 1874 + }, + { + "epoch": 0.20590819240061498, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 3.1740052700042725, + "learning_rate": 6.859443631039531e-07, + "loss": 1.0297, + "mean_token_accuracy": 0.6950536966323853, + "num_tokens": 48365276.0, + "step": 1875 + }, + { + "epoch": 0.20601801010322865, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.6316120624542236, + "learning_rate": 6.863103953147876e-07, + "loss": 0.9368, + "mean_token_accuracy": 0.7148165702819824, + "num_tokens": 48385953.0, + "step": 1876 + }, + { + "epoch": 0.2061278278058423, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.3795809745788574, + "learning_rate": 6.866764275256222e-07, + "loss": 0.987, + "mean_token_accuracy": 0.7014436721801758, + "num_tokens": 48410901.0, + "step": 1877 + }, + { + "epoch": 0.20623764550845597, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.289370059967041, + "learning_rate": 6.870424597364568e-07, + "loss": 1.0808, + "mean_token_accuracy": 0.6816250085830688, + "num_tokens": 48438408.0, + "step": 1878 + }, + { + "epoch": 0.20634746321106961, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.6953327655792236, + "learning_rate": 6.874084919472913e-07, + "loss": 0.9981, + "mean_token_accuracy": 0.7021846771240234, + "num_tokens": 48457518.0, + "step": 1879 + }, + { + "epoch": 0.2064572809136833, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.1690778732299805, + "learning_rate": 6.877745241581258e-07, + "loss": 0.9848, + "mean_token_accuracy": 0.7089779376983643, + "num_tokens": 48486392.0, + "step": 1880 + }, + { + "epoch": 0.20656709861629693, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.1051154136657715, + "learning_rate": 6.881405563689604e-07, + "loss": 1.121, + "mean_token_accuracy": 0.6779626607894897, + "num_tokens": 48516058.0, + "step": 1881 + }, + { + "epoch": 0.2066769163189106, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.1170263290405273, + "learning_rate": 6.88506588579795e-07, + "loss": 1.0728, + "mean_token_accuracy": 0.6789088845252991, + "num_tokens": 48546658.0, + "step": 1882 + }, + { + "epoch": 0.20678673402152428, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.478931427001953, + "learning_rate": 6.888726207906295e-07, + "loss": 1.0567, + "mean_token_accuracy": 0.6815874576568604, + "num_tokens": 48569941.0, + "step": 1883 + }, + { + "epoch": 0.20689655172413793, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.669548988342285, + "learning_rate": 6.892386530014641e-07, + "loss": 0.9979, + "mean_token_accuracy": 0.6987749934196472, + "num_tokens": 48590774.0, + "step": 1884 + }, + { + "epoch": 0.2070063694267516, + "ewc_loss": 5.774199962615967e-07, + "grad_norm": 2.13659405708313, + "learning_rate": 6.896046852122986e-07, + "loss": 1.0589, + "mean_token_accuracy": 0.6808047294616699, + "num_tokens": 48617652.0, + "step": 1885 + }, + { + "epoch": 0.20711618712936525, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 2.2581684589385986, + "learning_rate": 6.899707174231331e-07, + "loss": 1.0053, + "mean_token_accuracy": 0.6908067464828491, + "num_tokens": 48642381.0, + "step": 1886 + }, + { + "epoch": 0.20722600483197892, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 2.097445487976074, + "learning_rate": 6.903367496339677e-07, + "loss": 0.9997, + "mean_token_accuracy": 0.7024556398391724, + "num_tokens": 48671585.0, + "step": 1887 + }, + { + "epoch": 0.20733582253459257, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 2.059688091278076, + "learning_rate": 6.907027818448023e-07, + "loss": 1.0675, + "mean_token_accuracy": 0.6799101233482361, + "num_tokens": 48703218.0, + "step": 1888 + }, + { + "epoch": 0.20744564023720624, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 2.1376821994781494, + "learning_rate": 6.910688140556369e-07, + "loss": 1.0691, + "mean_token_accuracy": 0.6797165274620056, + "num_tokens": 48733166.0, + "step": 1889 + }, + { + "epoch": 0.2075554579398199, + "ewc_loss": 5.811452865600586e-07, + "grad_norm": 2.380692720413208, + "learning_rate": 6.914348462664714e-07, + "loss": 1.0252, + "mean_token_accuracy": 0.6956621408462524, + "num_tokens": 48756462.0, + "step": 1890 + }, + { + "epoch": 0.20766527564243356, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.1521079540252686, + "learning_rate": 6.918008784773059e-07, + "loss": 1.0801, + "mean_token_accuracy": 0.690247118473053, + "num_tokens": 48785002.0, + "step": 1891 + }, + { + "epoch": 0.20777509334504723, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.2120401859283447, + "learning_rate": 6.921669106881405e-07, + "loss": 0.9716, + "mean_token_accuracy": 0.7091766595840454, + "num_tokens": 48810763.0, + "step": 1892 + }, + { + "epoch": 0.20788491104766088, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.3623316287994385, + "learning_rate": 6.925329428989751e-07, + "loss": 1.0004, + "mean_token_accuracy": 0.6970384120941162, + "num_tokens": 48835009.0, + "step": 1893 + }, + { + "epoch": 0.20799472875027455, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.4054954051971436, + "learning_rate": 6.928989751098096e-07, + "loss": 0.9974, + "mean_token_accuracy": 0.6985824108123779, + "num_tokens": 48858156.0, + "step": 1894 + }, + { + "epoch": 0.2081045464528882, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.482112407684326, + "learning_rate": 6.932650073206442e-07, + "loss": 1.0634, + "mean_token_accuracy": 0.6823287606239319, + "num_tokens": 48882263.0, + "step": 1895 + }, + { + "epoch": 0.20821436415550187, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.5224530696868896, + "learning_rate": 6.936310395314787e-07, + "loss": 1.0559, + "mean_token_accuracy": 0.6868070960044861, + "num_tokens": 48905461.0, + "step": 1896 + }, + { + "epoch": 0.20832418185811552, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.202742338180542, + "learning_rate": 6.939970717423132e-07, + "loss": 1.0281, + "mean_token_accuracy": 0.6911323666572571, + "num_tokens": 48931619.0, + "step": 1897 + }, + { + "epoch": 0.2084339995607292, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.304544448852539, + "learning_rate": 6.943631039531479e-07, + "loss": 1.0654, + "mean_token_accuracy": 0.6830876469612122, + "num_tokens": 48957444.0, + "step": 1898 + }, + { + "epoch": 0.20854381726334284, + "ewc_loss": 5.848705768585205e-07, + "grad_norm": 2.1497998237609863, + "learning_rate": 6.947291361639824e-07, + "loss": 0.9666, + "mean_token_accuracy": 0.703726053237915, + "num_tokens": 48986415.0, + "step": 1899 + }, + { + "epoch": 0.2086536349659565, + "ewc_loss": 5.923211574554443e-07, + "grad_norm": 2.404876232147217, + "learning_rate": 6.950951683748169e-07, + "loss": 0.9962, + "mean_token_accuracy": 0.7006197571754456, + "num_tokens": 49010061.0, + "step": 1900 + }, + { + "epoch": 0.20876345266857016, + "ewc_loss": 5.923211574554443e-07, + "grad_norm": 2.500014066696167, + "learning_rate": 6.954612005856515e-07, + "loss": 0.9535, + "mean_token_accuracy": 0.7136905193328857, + "num_tokens": 49033376.0, + "step": 1901 + }, + { + "epoch": 0.20887327037118383, + "ewc_loss": 5.923211574554443e-07, + "grad_norm": 2.366835117340088, + "learning_rate": 6.95827232796486e-07, + "loss": 1.019, + "mean_token_accuracy": 0.695049524307251, + "num_tokens": 49057563.0, + "step": 1902 + }, + { + "epoch": 0.2089830880737975, + "ewc_loss": 5.923211574554443e-07, + "grad_norm": 2.2513961791992188, + "learning_rate": 6.961932650073206e-07, + "loss": 1.0993, + "mean_token_accuracy": 0.6698797941207886, + "num_tokens": 49085030.0, + "step": 1903 + }, + { + "epoch": 0.20909290577641115, + "ewc_loss": 5.960464477539062e-07, + "grad_norm": 2.5069987773895264, + "learning_rate": 6.965592972181552e-07, + "loss": 0.9362, + "mean_token_accuracy": 0.7103137969970703, + "num_tokens": 49104353.0, + "step": 1904 + }, + { + "epoch": 0.20920272347902483, + "ewc_loss": 5.960464477539062e-07, + "grad_norm": 2.2049615383148193, + "learning_rate": 6.969253294289897e-07, + "loss": 1.0409, + "mean_token_accuracy": 0.6958218812942505, + "num_tokens": 49134413.0, + "step": 1905 + }, + { + "epoch": 0.20931254118163847, + "ewc_loss": 5.960464477539062e-07, + "grad_norm": 2.1908860206604004, + "learning_rate": 6.972913616398243e-07, + "loss": 1.0396, + "mean_token_accuracy": 0.692293643951416, + "num_tokens": 49164137.0, + "step": 1906 + }, + { + "epoch": 0.20942235888425215, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.1161587238311768, + "learning_rate": 6.976573938506588e-07, + "loss": 1.0233, + "mean_token_accuracy": 0.687333881855011, + "num_tokens": 49192401.0, + "step": 1907 + }, + { + "epoch": 0.2095321765868658, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.2871479988098145, + "learning_rate": 6.980234260614933e-07, + "loss": 1.0246, + "mean_token_accuracy": 0.6914472579956055, + "num_tokens": 49218341.0, + "step": 1908 + }, + { + "epoch": 0.20964199428947947, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.5866098403930664, + "learning_rate": 6.98389458272328e-07, + "loss": 1.0101, + "mean_token_accuracy": 0.6983204483985901, + "num_tokens": 49239163.0, + "step": 1909 + }, + { + "epoch": 0.2097518119920931, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.1347055435180664, + "learning_rate": 6.987554904831625e-07, + "loss": 1.0334, + "mean_token_accuracy": 0.6922913789749146, + "num_tokens": 49268705.0, + "step": 1910 + }, + { + "epoch": 0.20986162969470679, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.347529411315918, + "learning_rate": 6.99121522693997e-07, + "loss": 0.9607, + "mean_token_accuracy": 0.7095894813537598, + "num_tokens": 49293437.0, + "step": 1911 + }, + { + "epoch": 0.20997144739732046, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.3498308658599854, + "learning_rate": 6.994875549048316e-07, + "loss": 1.0947, + "mean_token_accuracy": 0.6693801879882812, + "num_tokens": 49317640.0, + "step": 1912 + }, + { + "epoch": 0.2100812650999341, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.071821928024292, + "learning_rate": 6.998535871156661e-07, + "loss": 1.0363, + "mean_token_accuracy": 0.6913620233535767, + "num_tokens": 49348204.0, + "step": 1913 + }, + { + "epoch": 0.21019108280254778, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.2398934364318848, + "learning_rate": 7.002196193265007e-07, + "loss": 1.0217, + "mean_token_accuracy": 0.699631929397583, + "num_tokens": 49376892.0, + "step": 1914 + }, + { + "epoch": 0.21030090050516143, + "ewc_loss": 5.997717380523682e-07, + "grad_norm": 2.1272850036621094, + "learning_rate": 7.005856515373353e-07, + "loss": 0.9944, + "mean_token_accuracy": 0.691286027431488, + "num_tokens": 49405788.0, + "step": 1915 + }, + { + "epoch": 0.2104107182077751, + "ewc_loss": 6.034970283508301e-07, + "grad_norm": 2.4020283222198486, + "learning_rate": 7.009516837481698e-07, + "loss": 1.0524, + "mean_token_accuracy": 0.6856772899627686, + "num_tokens": 49428328.0, + "step": 1916 + }, + { + "epoch": 0.21052053591038875, + "ewc_loss": 6.034970283508301e-07, + "grad_norm": 2.1344661712646484, + "learning_rate": 7.013177159590043e-07, + "loss": 1.0719, + "mean_token_accuracy": 0.6913021802902222, + "num_tokens": 49456756.0, + "step": 1917 + }, + { + "epoch": 0.21063035361300242, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.375584363937378, + "learning_rate": 7.016837481698389e-07, + "loss": 1.0239, + "mean_token_accuracy": 0.6944637298583984, + "num_tokens": 49481855.0, + "step": 1918 + }, + { + "epoch": 0.21074017131561606, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.3438973426818848, + "learning_rate": 7.020497803806734e-07, + "loss": 1.0535, + "mean_token_accuracy": 0.6857942342758179, + "num_tokens": 49507527.0, + "step": 1919 + }, + { + "epoch": 0.21084998901822974, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.1972227096557617, + "learning_rate": 7.024158125915081e-07, + "loss": 1.0614, + "mean_token_accuracy": 0.6789485216140747, + "num_tokens": 49536120.0, + "step": 1920 + }, + { + "epoch": 0.2109598067208434, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.5200037956237793, + "learning_rate": 7.027818448023426e-07, + "loss": 0.9942, + "mean_token_accuracy": 0.6998884677886963, + "num_tokens": 49557633.0, + "step": 1921 + }, + { + "epoch": 0.21106962442345706, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.375657796859741, + "learning_rate": 7.031478770131771e-07, + "loss": 0.9894, + "mean_token_accuracy": 0.7035784721374512, + "num_tokens": 49581184.0, + "step": 1922 + }, + { + "epoch": 0.21117944212607073, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.178527593612671, + "learning_rate": 7.035139092240117e-07, + "loss": 1.054, + "mean_token_accuracy": 0.686231255531311, + "num_tokens": 49612774.0, + "step": 1923 + }, + { + "epoch": 0.21128925982868438, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.727598190307617, + "learning_rate": 7.038799414348462e-07, + "loss": 0.9896, + "mean_token_accuracy": 0.6991550922393799, + "num_tokens": 49632537.0, + "step": 1924 + }, + { + "epoch": 0.21139907753129805, + "ewc_loss": 6.07222318649292e-07, + "grad_norm": 2.1049726009368896, + "learning_rate": 7.042459736456808e-07, + "loss": 0.9419, + "mean_token_accuracy": 0.7214992046356201, + "num_tokens": 49661533.0, + "step": 1925 + }, + { + "epoch": 0.2115088952339117, + "ewc_loss": 6.109476089477539e-07, + "grad_norm": 2.079573631286621, + "learning_rate": 7.046120058565154e-07, + "loss": 0.9914, + "mean_token_accuracy": 0.6987837553024292, + "num_tokens": 49691168.0, + "step": 1926 + }, + { + "epoch": 0.21161871293652537, + "ewc_loss": 6.109476089477539e-07, + "grad_norm": 2.2003700733184814, + "learning_rate": 7.049780380673499e-07, + "loss": 1.1253, + "mean_token_accuracy": 0.6714727282524109, + "num_tokens": 49718846.0, + "step": 1927 + }, + { + "epoch": 0.21172853063913902, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.6519901752471924, + "learning_rate": 7.053440702781844e-07, + "loss": 1.0217, + "mean_token_accuracy": 0.696784496307373, + "num_tokens": 49738425.0, + "step": 1928 + }, + { + "epoch": 0.2118383483417527, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.2267019748687744, + "learning_rate": 7.05710102489019e-07, + "loss": 0.9685, + "mean_token_accuracy": 0.7071729302406311, + "num_tokens": 49762843.0, + "step": 1929 + }, + { + "epoch": 0.21194816604436637, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.232034683227539, + "learning_rate": 7.060761346998536e-07, + "loss": 0.9167, + "mean_token_accuracy": 0.7264955043792725, + "num_tokens": 49787535.0, + "step": 1930 + }, + { + "epoch": 0.21205798374698, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.4973325729370117, + "learning_rate": 7.064421669106881e-07, + "loss": 0.9493, + "mean_token_accuracy": 0.7176687121391296, + "num_tokens": 49808954.0, + "step": 1931 + }, + { + "epoch": 0.21216780144959368, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.2670228481292725, + "learning_rate": 7.068081991215227e-07, + "loss": 0.959, + "mean_token_accuracy": 0.713511049747467, + "num_tokens": 49833932.0, + "step": 1932 + }, + { + "epoch": 0.21227761915220733, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.2474610805511475, + "learning_rate": 7.071742313323572e-07, + "loss": 1.0455, + "mean_token_accuracy": 0.6893892884254456, + "num_tokens": 49861910.0, + "step": 1933 + }, + { + "epoch": 0.212387436854821, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.4244954586029053, + "learning_rate": 7.075402635431917e-07, + "loss": 1.0505, + "mean_token_accuracy": 0.6887348294258118, + "num_tokens": 49885140.0, + "step": 1934 + }, + { + "epoch": 0.21249725455743465, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.1919572353363037, + "learning_rate": 7.079062957540263e-07, + "loss": 0.9451, + "mean_token_accuracy": 0.7130303382873535, + "num_tokens": 49909653.0, + "step": 1935 + }, + { + "epoch": 0.21260707226004832, + "ewc_loss": 6.146728992462158e-07, + "grad_norm": 2.268815279006958, + "learning_rate": 7.082723279648609e-07, + "loss": 1.0562, + "mean_token_accuracy": 0.6871442198753357, + "num_tokens": 49936584.0, + "step": 1936 + }, + { + "epoch": 0.21271688996266197, + "ewc_loss": 6.183981895446777e-07, + "grad_norm": 2.593046188354492, + "learning_rate": 7.086383601756955e-07, + "loss": 1.0206, + "mean_token_accuracy": 0.6898239254951477, + "num_tokens": 49956702.0, + "step": 1937 + }, + { + "epoch": 0.21282670766527564, + "ewc_loss": 6.183981895446777e-07, + "grad_norm": 2.4847021102905273, + "learning_rate": 7.0900439238653e-07, + "loss": 0.9723, + "mean_token_accuracy": 0.6959961652755737, + "num_tokens": 49977896.0, + "step": 1938 + }, + { + "epoch": 0.2129365253678893, + "ewc_loss": 6.183981895446777e-07, + "grad_norm": 2.229855537414551, + "learning_rate": 7.093704245973645e-07, + "loss": 0.9603, + "mean_token_accuracy": 0.7108920216560364, + "num_tokens": 50003208.0, + "step": 1939 + }, + { + "epoch": 0.21304634307050296, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.479989767074585, + "learning_rate": 7.097364568081991e-07, + "loss": 1.0252, + "mean_token_accuracy": 0.6930962800979614, + "num_tokens": 50024940.0, + "step": 1940 + }, + { + "epoch": 0.21315616077311664, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.088320016860962, + "learning_rate": 7.101024890190337e-07, + "loss": 0.9143, + "mean_token_accuracy": 0.7229870557785034, + "num_tokens": 50051034.0, + "step": 1941 + }, + { + "epoch": 0.21326597847573028, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.2728655338287354, + "learning_rate": 7.104685212298682e-07, + "loss": 1.0465, + "mean_token_accuracy": 0.6863420009613037, + "num_tokens": 50077175.0, + "step": 1942 + }, + { + "epoch": 0.21337579617834396, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 1.9131008386611938, + "learning_rate": 7.108345534407028e-07, + "loss": 1.1602, + "mean_token_accuracy": 0.6534422636032104, + "num_tokens": 50115357.0, + "step": 1943 + }, + { + "epoch": 0.2134856138809576, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.1057846546173096, + "learning_rate": 7.112005856515373e-07, + "loss": 0.9505, + "mean_token_accuracy": 0.712334394454956, + "num_tokens": 50143631.0, + "step": 1944 + }, + { + "epoch": 0.21359543158357128, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.3924221992492676, + "learning_rate": 7.115666178623718e-07, + "loss": 0.9845, + "mean_token_accuracy": 0.7070652842521667, + "num_tokens": 50167265.0, + "step": 1945 + }, + { + "epoch": 0.21370524928618492, + "ewc_loss": 6.221234798431396e-07, + "grad_norm": 2.4939563274383545, + "learning_rate": 7.119326500732065e-07, + "loss": 1.1158, + "mean_token_accuracy": 0.668609619140625, + "num_tokens": 50191169.0, + "step": 1946 + }, + { + "epoch": 0.2138150669887986, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.203537702560425, + "learning_rate": 7.12298682284041e-07, + "loss": 1.0336, + "mean_token_accuracy": 0.6889874935150146, + "num_tokens": 50218920.0, + "step": 1947 + }, + { + "epoch": 0.21392488469141224, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.3951478004455566, + "learning_rate": 7.126647144948755e-07, + "loss": 1.1289, + "mean_token_accuracy": 0.6654617190361023, + "num_tokens": 50244790.0, + "step": 1948 + }, + { + "epoch": 0.21403470239402592, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.2349934577941895, + "learning_rate": 7.130307467057101e-07, + "loss": 0.9944, + "mean_token_accuracy": 0.7020395994186401, + "num_tokens": 50270316.0, + "step": 1949 + }, + { + "epoch": 0.2141445200966396, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.2244913578033447, + "learning_rate": 7.133967789165446e-07, + "loss": 1.0073, + "mean_token_accuracy": 0.6984720826148987, + "num_tokens": 50296368.0, + "step": 1950 + }, + { + "epoch": 0.21425433779925324, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.5475564002990723, + "learning_rate": 7.137628111273791e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.6952183246612549, + "num_tokens": 50317194.0, + "step": 1951 + }, + { + "epoch": 0.2143641555018669, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.393441915512085, + "learning_rate": 7.141288433382138e-07, + "loss": 0.9549, + "mean_token_accuracy": 0.7107990384101868, + "num_tokens": 50338475.0, + "step": 1952 + }, + { + "epoch": 0.21447397320448056, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.3307278156280518, + "learning_rate": 7.144948755490483e-07, + "loss": 1.0678, + "mean_token_accuracy": 0.6765064001083374, + "num_tokens": 50363753.0, + "step": 1953 + }, + { + "epoch": 0.21458379090709423, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.3420052528381348, + "learning_rate": 7.148609077598829e-07, + "loss": 0.967, + "mean_token_accuracy": 0.7046284675598145, + "num_tokens": 50387871.0, + "step": 1954 + }, + { + "epoch": 0.21469360860970788, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.493579149246216, + "learning_rate": 7.152269399707174e-07, + "loss": 1.0107, + "mean_token_accuracy": 0.6963728666305542, + "num_tokens": 50408307.0, + "step": 1955 + }, + { + "epoch": 0.21480342631232155, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.2972564697265625, + "learning_rate": 7.155929721815519e-07, + "loss": 1.0824, + "mean_token_accuracy": 0.6900067925453186, + "num_tokens": 50434392.0, + "step": 1956 + }, + { + "epoch": 0.2149132440149352, + "ewc_loss": 6.258487701416016e-07, + "grad_norm": 2.0803143978118896, + "learning_rate": 7.159590043923866e-07, + "loss": 1.036, + "mean_token_accuracy": 0.6872872114181519, + "num_tokens": 50464382.0, + "step": 1957 + }, + { + "epoch": 0.21502306171754887, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.629495859146118, + "learning_rate": 7.163250366032211e-07, + "loss": 1.1376, + "mean_token_accuracy": 0.657119870185852, + "num_tokens": 50486233.0, + "step": 1958 + }, + { + "epoch": 0.21513287942016254, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.1558570861816406, + "learning_rate": 7.166910688140556e-07, + "loss": 1.0802, + "mean_token_accuracy": 0.6713361740112305, + "num_tokens": 50516822.0, + "step": 1959 + }, + { + "epoch": 0.2152426971227762, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.443256378173828, + "learning_rate": 7.170571010248902e-07, + "loss": 1.108, + "mean_token_accuracy": 0.6743561625480652, + "num_tokens": 50541270.0, + "step": 1960 + }, + { + "epoch": 0.21535251482538986, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.2257630825042725, + "learning_rate": 7.174231332357247e-07, + "loss": 0.9793, + "mean_token_accuracy": 0.701767086982727, + "num_tokens": 50566736.0, + "step": 1961 + }, + { + "epoch": 0.2154623325280035, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.2529613971710205, + "learning_rate": 7.177891654465593e-07, + "loss": 1.0177, + "mean_token_accuracy": 0.6927697658538818, + "num_tokens": 50591212.0, + "step": 1962 + }, + { + "epoch": 0.21557215023061718, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.5827510356903076, + "learning_rate": 7.181551976573939e-07, + "loss": 0.9424, + "mean_token_accuracy": 0.7123467922210693, + "num_tokens": 50613334.0, + "step": 1963 + }, + { + "epoch": 0.21568196793323083, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.024623155593872, + "learning_rate": 7.185212298682284e-07, + "loss": 0.9493, + "mean_token_accuracy": 0.7110533714294434, + "num_tokens": 50642793.0, + "step": 1964 + }, + { + "epoch": 0.2157917856358445, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.254960060119629, + "learning_rate": 7.188872620790629e-07, + "loss": 1.1179, + "mean_token_accuracy": 0.6793215274810791, + "num_tokens": 50670184.0, + "step": 1965 + }, + { + "epoch": 0.21590160333845815, + "ewc_loss": 6.332993507385254e-07, + "grad_norm": 2.8324732780456543, + "learning_rate": 7.192532942898975e-07, + "loss": 0.9421, + "mean_token_accuracy": 0.7067493200302124, + "num_tokens": 50688140.0, + "step": 1966 + }, + { + "epoch": 0.21601142104107182, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.330122470855713, + "learning_rate": 7.19619326500732e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6844772100448608, + "num_tokens": 50712549.0, + "step": 1967 + }, + { + "epoch": 0.2161212387436855, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.409881353378296, + "learning_rate": 7.199853587115667e-07, + "loss": 1.1311, + "mean_token_accuracy": 0.6649882793426514, + "num_tokens": 50738520.0, + "step": 1968 + }, + { + "epoch": 0.21623105644629914, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.270113706588745, + "learning_rate": 7.203513909224012e-07, + "loss": 0.9686, + "mean_token_accuracy": 0.715347170829773, + "num_tokens": 50763757.0, + "step": 1969 + }, + { + "epoch": 0.21634087414891282, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.019913673400879, + "learning_rate": 7.207174231332357e-07, + "loss": 1.0266, + "mean_token_accuracy": 0.7030934691429138, + "num_tokens": 50793832.0, + "step": 1970 + }, + { + "epoch": 0.21645069185152646, + "ewc_loss": 6.444752216339111e-07, + "grad_norm": 2.582186460494995, + "learning_rate": 7.210834553440703e-07, + "loss": 0.9389, + "mean_token_accuracy": 0.7103941440582275, + "num_tokens": 50814416.0, + "step": 1971 + }, + { + "epoch": 0.21656050955414013, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.6982784271240234, + "learning_rate": 7.214494875549048e-07, + "loss": 0.9434, + "mean_token_accuracy": 0.70694899559021, + "num_tokens": 50833389.0, + "step": 1972 + }, + { + "epoch": 0.21667032725675378, + "ewc_loss": 6.444752216339111e-07, + "grad_norm": 2.165907144546509, + "learning_rate": 7.218155197657394e-07, + "loss": 1.1335, + "mean_token_accuracy": 0.6717070937156677, + "num_tokens": 50863979.0, + "step": 1973 + }, + { + "epoch": 0.21678014495936745, + "ewc_loss": 6.444752216339111e-07, + "grad_norm": 2.4460673332214355, + "learning_rate": 7.22181551976574e-07, + "loss": 0.9576, + "mean_token_accuracy": 0.7051109075546265, + "num_tokens": 50886208.0, + "step": 1974 + }, + { + "epoch": 0.2168899626619811, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.472478151321411, + "learning_rate": 7.225475841874085e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.6859462261199951, + "num_tokens": 50908463.0, + "step": 1975 + }, + { + "epoch": 0.21699978036459477, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.469829797744751, + "learning_rate": 7.22913616398243e-07, + "loss": 0.9831, + "mean_token_accuracy": 0.7108577489852905, + "num_tokens": 50931728.0, + "step": 1976 + }, + { + "epoch": 0.21710959806720842, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.2148852348327637, + "learning_rate": 7.232796486090776e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6826114058494568, + "num_tokens": 50956836.0, + "step": 1977 + }, + { + "epoch": 0.2172194157698221, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.1239168643951416, + "learning_rate": 7.236456808199122e-07, + "loss": 1.1398, + "mean_token_accuracy": 0.6605716943740845, + "num_tokens": 50988874.0, + "step": 1978 + }, + { + "epoch": 0.21732923347243577, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.244614362716675, + "learning_rate": 7.240117130307467e-07, + "loss": 1.1104, + "mean_token_accuracy": 0.6836626529693604, + "num_tokens": 51017825.0, + "step": 1979 + }, + { + "epoch": 0.2174390511750494, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.1270289421081543, + "learning_rate": 7.243777452415813e-07, + "loss": 0.993, + "mean_token_accuracy": 0.7006430625915527, + "num_tokens": 51045955.0, + "step": 1980 + }, + { + "epoch": 0.2175488688776631, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.4303715229034424, + "learning_rate": 7.247437774524158e-07, + "loss": 1.0379, + "mean_token_accuracy": 0.6803948879241943, + "num_tokens": 51069330.0, + "step": 1981 + }, + { + "epoch": 0.21765868658027673, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.5775861740112305, + "learning_rate": 7.251098096632503e-07, + "loss": 1.0687, + "mean_token_accuracy": 0.6934622526168823, + "num_tokens": 51092178.0, + "step": 1982 + }, + { + "epoch": 0.2177685042828904, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.3366973400115967, + "learning_rate": 7.254758418740849e-07, + "loss": 0.9921, + "mean_token_accuracy": 0.6938226819038391, + "num_tokens": 51115554.0, + "step": 1983 + }, + { + "epoch": 0.21787832198550405, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.445744037628174, + "learning_rate": 7.258418740849195e-07, + "loss": 0.9792, + "mean_token_accuracy": 0.7030560374259949, + "num_tokens": 51135508.0, + "step": 1984 + }, + { + "epoch": 0.21798813968811773, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.222715377807617, + "learning_rate": 7.262079062957541e-07, + "loss": 0.9788, + "mean_token_accuracy": 0.7063740491867065, + "num_tokens": 51159938.0, + "step": 1985 + }, + { + "epoch": 0.21809795739073137, + "ewc_loss": 6.407499313354492e-07, + "grad_norm": 2.1245298385620117, + "learning_rate": 7.265739385065886e-07, + "loss": 1.1239, + "mean_token_accuracy": 0.6684843301773071, + "num_tokens": 51189162.0, + "step": 1986 + }, + { + "epoch": 0.21820777509334505, + "ewc_loss": 6.444752216339111e-07, + "grad_norm": 2.3526527881622314, + "learning_rate": 7.269399707174231e-07, + "loss": 1.0145, + "mean_token_accuracy": 0.6919564008712769, + "num_tokens": 51213141.0, + "step": 1987 + }, + { + "epoch": 0.21831759279595872, + "ewc_loss": 6.444752216339111e-07, + "grad_norm": 2.12548565864563, + "learning_rate": 7.273060029282577e-07, + "loss": 1.0366, + "mean_token_accuracy": 0.6885750889778137, + "num_tokens": 51243740.0, + "step": 1988 + }, + { + "epoch": 0.21842741049857237, + "ewc_loss": 6.48200511932373e-07, + "grad_norm": 2.3992698192596436, + "learning_rate": 7.276720351390923e-07, + "loss": 1.0282, + "mean_token_accuracy": 0.6993045806884766, + "num_tokens": 51267656.0, + "step": 1989 + }, + { + "epoch": 0.21853722820118604, + "ewc_loss": 6.48200511932373e-07, + "grad_norm": 2.310408115386963, + "learning_rate": 7.280380673499268e-07, + "loss": 0.9433, + "mean_token_accuracy": 0.7135748267173767, + "num_tokens": 51291884.0, + "step": 1990 + }, + { + "epoch": 0.21864704590379969, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.4678380489349365, + "learning_rate": 7.284040995607614e-07, + "loss": 1.0478, + "mean_token_accuracy": 0.6851907968521118, + "num_tokens": 51313957.0, + "step": 1991 + }, + { + "epoch": 0.21875686360641336, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.690791606903076, + "learning_rate": 7.287701317715959e-07, + "loss": 0.9864, + "mean_token_accuracy": 0.7011317610740662, + "num_tokens": 51333146.0, + "step": 1992 + }, + { + "epoch": 0.218866681309027, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.5552570819854736, + "learning_rate": 7.291361639824304e-07, + "loss": 1.0721, + "mean_token_accuracy": 0.681782066822052, + "num_tokens": 51354676.0, + "step": 1993 + }, + { + "epoch": 0.21897649901164068, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.3738458156585693, + "learning_rate": 7.295021961932651e-07, + "loss": 0.9987, + "mean_token_accuracy": 0.7001981735229492, + "num_tokens": 51377314.0, + "step": 1994 + }, + { + "epoch": 0.21908631671425433, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.3508005142211914, + "learning_rate": 7.298682284040996e-07, + "loss": 1.0531, + "mean_token_accuracy": 0.6864467263221741, + "num_tokens": 51404483.0, + "step": 1995 + }, + { + "epoch": 0.219196134416868, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.3834006786346436, + "learning_rate": 7.30234260614934e-07, + "loss": 1.0033, + "mean_token_accuracy": 0.6969408988952637, + "num_tokens": 51427270.0, + "step": 1996 + }, + { + "epoch": 0.21930595211948167, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.4959402084350586, + "learning_rate": 7.306002928257687e-07, + "loss": 1.0679, + "mean_token_accuracy": 0.6867492198944092, + "num_tokens": 51451033.0, + "step": 1997 + }, + { + "epoch": 0.21941576982209532, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.2815027236938477, + "learning_rate": 7.309663250366031e-07, + "loss": 0.9915, + "mean_token_accuracy": 0.7026482820510864, + "num_tokens": 51477625.0, + "step": 1998 + }, + { + "epoch": 0.219525587524709, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.5717971324920654, + "learning_rate": 7.313323572474376e-07, + "loss": 1.0433, + "mean_token_accuracy": 0.6882876753807068, + "num_tokens": 51499313.0, + "step": 1999 + }, + { + "epoch": 0.21963540522732264, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.4305615425109863, + "learning_rate": 7.316983894582724e-07, + "loss": 1.007, + "mean_token_accuracy": 0.7044018507003784, + "num_tokens": 51523584.0, + "step": 2000 + }, + { + "epoch": 0.2197452229299363, + "ewc_loss": 6.556510925292969e-07, + "grad_norm": 2.538261890411377, + "learning_rate": 7.320644216691068e-07, + "loss": 1.0676, + "mean_token_accuracy": 0.6813294887542725, + "num_tokens": 51544804.0, + "step": 2001 + }, + { + "epoch": 0.21985504063254996, + "ewc_loss": 6.593763828277588e-07, + "grad_norm": 2.3532445430755615, + "learning_rate": 7.324304538799415e-07, + "loss": 0.9749, + "mean_token_accuracy": 0.7029596567153931, + "num_tokens": 51568871.0, + "step": 2002 + }, + { + "epoch": 0.21996485833516363, + "ewc_loss": 6.593763828277588e-07, + "grad_norm": 2.219332456588745, + "learning_rate": 7.32796486090776e-07, + "loss": 1.0165, + "mean_token_accuracy": 0.7049084901809692, + "num_tokens": 51596345.0, + "step": 2003 + }, + { + "epoch": 0.22007467603777728, + "ewc_loss": 6.593763828277588e-07, + "grad_norm": 2.100615978240967, + "learning_rate": 7.331625183016104e-07, + "loss": 1.0616, + "mean_token_accuracy": 0.6864186525344849, + "num_tokens": 51624644.0, + "step": 2004 + }, + { + "epoch": 0.22018449374039095, + "ewc_loss": 6.593763828277588e-07, + "grad_norm": 2.3250133991241455, + "learning_rate": 7.335285505124452e-07, + "loss": 0.9296, + "mean_token_accuracy": 0.7163770794868469, + "num_tokens": 51647338.0, + "step": 2005 + }, + { + "epoch": 0.22029431144300463, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.440725088119507, + "learning_rate": 7.338945827232796e-07, + "loss": 1.0393, + "mean_token_accuracy": 0.6905108690261841, + "num_tokens": 51670588.0, + "step": 2006 + }, + { + "epoch": 0.22040412914561827, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.2501494884490967, + "learning_rate": 7.342606149341141e-07, + "loss": 0.9995, + "mean_token_accuracy": 0.7099248766899109, + "num_tokens": 51694915.0, + "step": 2007 + }, + { + "epoch": 0.22051394684823195, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.162712812423706, + "learning_rate": 7.346266471449487e-07, + "loss": 1.1016, + "mean_token_accuracy": 0.6701804399490356, + "num_tokens": 51722461.0, + "step": 2008 + }, + { + "epoch": 0.2206237645508456, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.217799663543701, + "learning_rate": 7.349926793557832e-07, + "loss": 0.9919, + "mean_token_accuracy": 0.6983795166015625, + "num_tokens": 51750114.0, + "step": 2009 + }, + { + "epoch": 0.22073358225345927, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.5059750080108643, + "learning_rate": 7.353587115666178e-07, + "loss": 0.9719, + "mean_token_accuracy": 0.7051356434822083, + "num_tokens": 51769519.0, + "step": 2010 + }, + { + "epoch": 0.2208433999560729, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.4988715648651123, + "learning_rate": 7.357247437774524e-07, + "loss": 0.9554, + "mean_token_accuracy": 0.7141982913017273, + "num_tokens": 51790952.0, + "step": 2011 + }, + { + "epoch": 0.22095321765868658, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.048137664794922, + "learning_rate": 7.360907759882869e-07, + "loss": 1.0142, + "mean_token_accuracy": 0.7006543874740601, + "num_tokens": 51820766.0, + "step": 2012 + }, + { + "epoch": 0.22106303536130023, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.192528247833252, + "learning_rate": 7.364568081991214e-07, + "loss": 1.078, + "mean_token_accuracy": 0.6816781759262085, + "num_tokens": 51848779.0, + "step": 2013 + }, + { + "epoch": 0.2211728530639139, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.4805243015289307, + "learning_rate": 7.36822840409956e-07, + "loss": 1.0509, + "mean_token_accuracy": 0.6877478957176208, + "num_tokens": 51872683.0, + "step": 2014 + }, + { + "epoch": 0.22128267076652755, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.5492374897003174, + "learning_rate": 7.371888726207905e-07, + "loss": 0.9815, + "mean_token_accuracy": 0.7032919526100159, + "num_tokens": 51892998.0, + "step": 2015 + }, + { + "epoch": 0.22139248846914122, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.3985419273376465, + "learning_rate": 7.375549048316251e-07, + "loss": 0.9959, + "mean_token_accuracy": 0.6964727640151978, + "num_tokens": 51916800.0, + "step": 2016 + }, + { + "epoch": 0.2215023061717549, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.4732284545898438, + "learning_rate": 7.379209370424597e-07, + "loss": 0.9716, + "mean_token_accuracy": 0.7074753046035767, + "num_tokens": 51938004.0, + "step": 2017 + }, + { + "epoch": 0.22161212387436854, + "ewc_loss": 6.631016731262207e-07, + "grad_norm": 2.5119190216064453, + "learning_rate": 7.382869692532942e-07, + "loss": 0.9874, + "mean_token_accuracy": 0.6980758905410767, + "num_tokens": 51957847.0, + "step": 2018 + }, + { + "epoch": 0.22172194157698222, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.429246664047241, + "learning_rate": 7.386530014641288e-07, + "loss": 0.9771, + "mean_token_accuracy": 0.7042545676231384, + "num_tokens": 51980724.0, + "step": 2019 + }, + { + "epoch": 0.22183175927959586, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.255251169204712, + "learning_rate": 7.390190336749633e-07, + "loss": 0.9546, + "mean_token_accuracy": 0.7044240236282349, + "num_tokens": 52004838.0, + "step": 2020 + }, + { + "epoch": 0.22194157698220954, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.1788978576660156, + "learning_rate": 7.393850658857979e-07, + "loss": 1.0906, + "mean_token_accuracy": 0.687958836555481, + "num_tokens": 52031951.0, + "step": 2021 + }, + { + "epoch": 0.22205139468482318, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.5171401500701904, + "learning_rate": 7.397510980966325e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6850983500480652, + "num_tokens": 52055321.0, + "step": 2022 + }, + { + "epoch": 0.22216121238743686, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.2787277698516846, + "learning_rate": 7.40117130307467e-07, + "loss": 1.09, + "mean_token_accuracy": 0.673316478729248, + "num_tokens": 52081626.0, + "step": 2023 + }, + { + "epoch": 0.2222710300900505, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 1.9384326934814453, + "learning_rate": 7.404831625183015e-07, + "loss": 1.0603, + "mean_token_accuracy": 0.6849178075790405, + "num_tokens": 52116812.0, + "step": 2024 + }, + { + "epoch": 0.22238084779266418, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.214083671569824, + "learning_rate": 7.408491947291361e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.693297803401947, + "num_tokens": 52146142.0, + "step": 2025 + }, + { + "epoch": 0.22249066549527785, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.2951626777648926, + "learning_rate": 7.412152269399707e-07, + "loss": 1.0484, + "mean_token_accuracy": 0.6865289211273193, + "num_tokens": 52171070.0, + "step": 2026 + }, + { + "epoch": 0.2226004831978915, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.178964853286743, + "learning_rate": 7.415812591508052e-07, + "loss": 0.9423, + "mean_token_accuracy": 0.7090268135070801, + "num_tokens": 52198273.0, + "step": 2027 + }, + { + "epoch": 0.22271030090050517, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.4676411151885986, + "learning_rate": 7.419472913616398e-07, + "loss": 1.0171, + "mean_token_accuracy": 0.696051836013794, + "num_tokens": 52220792.0, + "step": 2028 + }, + { + "epoch": 0.22282011860311882, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.219816207885742, + "learning_rate": 7.423133235724743e-07, + "loss": 1.0188, + "mean_token_accuracy": 0.7064456343650818, + "num_tokens": 52248973.0, + "step": 2029 + }, + { + "epoch": 0.2229299363057325, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.4537086486816406, + "learning_rate": 7.426793557833088e-07, + "loss": 1.0891, + "mean_token_accuracy": 0.6785963773727417, + "num_tokens": 52271488.0, + "step": 2030 + }, + { + "epoch": 0.22303975400834614, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.220693349838257, + "learning_rate": 7.430453879941434e-07, + "loss": 1.1303, + "mean_token_accuracy": 0.663001298904419, + "num_tokens": 52299754.0, + "step": 2031 + }, + { + "epoch": 0.2231495717109598, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.396688222885132, + "learning_rate": 7.43411420204978e-07, + "loss": 1.0904, + "mean_token_accuracy": 0.6785857677459717, + "num_tokens": 52324906.0, + "step": 2032 + }, + { + "epoch": 0.22325938941357346, + "ewc_loss": 6.668269634246826e-07, + "grad_norm": 2.3566646575927734, + "learning_rate": 7.437774524158125e-07, + "loss": 0.9797, + "mean_token_accuracy": 0.7019416093826294, + "num_tokens": 52348354.0, + "step": 2033 + }, + { + "epoch": 0.22336920711618713, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.5967819690704346, + "learning_rate": 7.441434846266471e-07, + "loss": 1.0286, + "mean_token_accuracy": 0.689714252948761, + "num_tokens": 52369396.0, + "step": 2034 + }, + { + "epoch": 0.2234790248188008, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.4522697925567627, + "learning_rate": 7.445095168374816e-07, + "loss": 0.8776, + "mean_token_accuracy": 0.7216777205467224, + "num_tokens": 52391080.0, + "step": 2035 + }, + { + "epoch": 0.22358884252141445, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.6191213130950928, + "learning_rate": 7.448755490483162e-07, + "loss": 0.9616, + "mean_token_accuracy": 0.7084084749221802, + "num_tokens": 52410968.0, + "step": 2036 + }, + { + "epoch": 0.22369866022402812, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.024977207183838, + "learning_rate": 7.452415812591508e-07, + "loss": 1.0981, + "mean_token_accuracy": 0.6697908043861389, + "num_tokens": 52441502.0, + "step": 2037 + }, + { + "epoch": 0.22380847792664177, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.4258899688720703, + "learning_rate": 7.456076134699853e-07, + "loss": 1.0378, + "mean_token_accuracy": 0.6886254549026489, + "num_tokens": 52464356.0, + "step": 2038 + }, + { + "epoch": 0.22391829562925544, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.211570978164673, + "learning_rate": 7.459736456808199e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.6903018951416016, + "num_tokens": 52491196.0, + "step": 2039 + }, + { + "epoch": 0.2240281133318691, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.70186448097229, + "learning_rate": 7.463396778916544e-07, + "loss": 1.024, + "mean_token_accuracy": 0.6926898956298828, + "num_tokens": 52510079.0, + "step": 2040 + }, + { + "epoch": 0.22413793103448276, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.27390193939209, + "learning_rate": 7.467057101024889e-07, + "loss": 1.1054, + "mean_token_accuracy": 0.6776164770126343, + "num_tokens": 52536899.0, + "step": 2041 + }, + { + "epoch": 0.2242477487370964, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.2594974040985107, + "learning_rate": 7.470717423133236e-07, + "loss": 1.1122, + "mean_token_accuracy": 0.6739320755004883, + "num_tokens": 52565125.0, + "step": 2042 + }, + { + "epoch": 0.22435756643971008, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.267228126525879, + "learning_rate": 7.474377745241581e-07, + "loss": 0.9959, + "mean_token_accuracy": 0.7011277079582214, + "num_tokens": 52589724.0, + "step": 2043 + }, + { + "epoch": 0.22446738414232376, + "ewc_loss": 6.705522537231445e-07, + "grad_norm": 2.1470985412597656, + "learning_rate": 7.478038067349926e-07, + "loss": 1.127, + "mean_token_accuracy": 0.6665428876876831, + "num_tokens": 52620905.0, + "step": 2044 + }, + { + "epoch": 0.2245772018449374, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.454991579055786, + "learning_rate": 7.481698389458272e-07, + "loss": 0.9891, + "mean_token_accuracy": 0.7045999765396118, + "num_tokens": 52647382.0, + "step": 2045 + }, + { + "epoch": 0.22468701954755108, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.418382406234741, + "learning_rate": 7.485358711566617e-07, + "loss": 1.0325, + "mean_token_accuracy": 0.6873356103897095, + "num_tokens": 52670035.0, + "step": 2046 + }, + { + "epoch": 0.22479683725016472, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 1.9592264890670776, + "learning_rate": 7.489019033674962e-07, + "loss": 1.0071, + "mean_token_accuracy": 0.6929365992546082, + "num_tokens": 52702091.0, + "step": 2047 + }, + { + "epoch": 0.2249066549527784, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.3854410648345947, + "learning_rate": 7.492679355783309e-07, + "loss": 0.9777, + "mean_token_accuracy": 0.70247483253479, + "num_tokens": 52723476.0, + "step": 2048 + }, + { + "epoch": 0.22501647265539204, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.125800132751465, + "learning_rate": 7.496339677891654e-07, + "loss": 0.9282, + "mean_token_accuracy": 0.7159138917922974, + "num_tokens": 52750250.0, + "step": 2049 + }, + { + "epoch": 0.22512629035800572, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.289501667022705, + "learning_rate": 7.5e-07, + "loss": 1.0068, + "mean_token_accuracy": 0.6943444013595581, + "num_tokens": 52774913.0, + "step": 2050 + }, + { + "epoch": 0.22523610806061936, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.3088438510894775, + "learning_rate": 7.503660322108345e-07, + "loss": 1.0749, + "mean_token_accuracy": 0.6820449233055115, + "num_tokens": 52799061.0, + "step": 2051 + }, + { + "epoch": 0.22534592576323303, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.147425889968872, + "learning_rate": 7.50732064421669e-07, + "loss": 1.031, + "mean_token_accuracy": 0.6884940266609192, + "num_tokens": 52827682.0, + "step": 2052 + }, + { + "epoch": 0.22545574346584668, + "ewc_loss": 6.742775440216064e-07, + "grad_norm": 2.2369465827941895, + "learning_rate": 7.510980966325037e-07, + "loss": 1.0784, + "mean_token_accuracy": 0.6771905422210693, + "num_tokens": 52854004.0, + "step": 2053 + }, + { + "epoch": 0.22556556116846035, + "ewc_loss": 6.780028343200684e-07, + "grad_norm": 2.6061277389526367, + "learning_rate": 7.514641288433382e-07, + "loss": 1.0488, + "mean_token_accuracy": 0.696411669254303, + "num_tokens": 52875334.0, + "step": 2054 + }, + { + "epoch": 0.22567537887107403, + "ewc_loss": 6.780028343200684e-07, + "grad_norm": 2.2882044315338135, + "learning_rate": 7.518301610541727e-07, + "loss": 1.0387, + "mean_token_accuracy": 0.6934179067611694, + "num_tokens": 52901072.0, + "step": 2055 + }, + { + "epoch": 0.22578519657368767, + "ewc_loss": 6.780028343200684e-07, + "grad_norm": 2.169013023376465, + "learning_rate": 7.521961932650073e-07, + "loss": 1.038, + "mean_token_accuracy": 0.6925950050354004, + "num_tokens": 52929140.0, + "step": 2056 + }, + { + "epoch": 0.22589501427630135, + "ewc_loss": 6.780028343200684e-07, + "grad_norm": 2.366581439971924, + "learning_rate": 7.525622254758418e-07, + "loss": 0.9773, + "mean_token_accuracy": 0.7031658887863159, + "num_tokens": 52953279.0, + "step": 2057 + }, + { + "epoch": 0.226004831978915, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 2.0425052642822266, + "learning_rate": 7.529282576866763e-07, + "loss": 1.0719, + "mean_token_accuracy": 0.6766951084136963, + "num_tokens": 52985489.0, + "step": 2058 + }, + { + "epoch": 0.22611464968152867, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 2.596311569213867, + "learning_rate": 7.53294289897511e-07, + "loss": 0.994, + "mean_token_accuracy": 0.696973979473114, + "num_tokens": 53008432.0, + "step": 2059 + }, + { + "epoch": 0.2262244673841423, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 2.5627384185791016, + "learning_rate": 7.536603221083455e-07, + "loss": 1.0051, + "mean_token_accuracy": 0.7073812484741211, + "num_tokens": 53030196.0, + "step": 2060 + }, + { + "epoch": 0.226334285086756, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 2.2271156311035156, + "learning_rate": 7.5402635431918e-07, + "loss": 1.0035, + "mean_token_accuracy": 0.7031721472740173, + "num_tokens": 53059442.0, + "step": 2061 + }, + { + "epoch": 0.22644410278936963, + "ewc_loss": 6.817281246185303e-07, + "grad_norm": 2.5695574283599854, + "learning_rate": 7.543923865300146e-07, + "loss": 0.9892, + "mean_token_accuracy": 0.7016521692276001, + "num_tokens": 53081527.0, + "step": 2062 + }, + { + "epoch": 0.2265539204919833, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.1499037742614746, + "learning_rate": 7.547584187408491e-07, + "loss": 1.0428, + "mean_token_accuracy": 0.689662754535675, + "num_tokens": 53109281.0, + "step": 2063 + }, + { + "epoch": 0.22666373819459698, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.225102663040161, + "learning_rate": 7.551244509516837e-07, + "loss": 1.0, + "mean_token_accuracy": 0.7054787874221802, + "num_tokens": 53134575.0, + "step": 2064 + }, + { + "epoch": 0.22677355589721063, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.206437110900879, + "learning_rate": 7.554904831625183e-07, + "loss": 1.0005, + "mean_token_accuracy": 0.6990825533866882, + "num_tokens": 53163685.0, + "step": 2065 + }, + { + "epoch": 0.2268833735998243, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.0839314460754395, + "learning_rate": 7.558565153733528e-07, + "loss": 1.0318, + "mean_token_accuracy": 0.6940550804138184, + "num_tokens": 53193254.0, + "step": 2066 + }, + { + "epoch": 0.22699319130243795, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.443721055984497, + "learning_rate": 7.562225475841874e-07, + "loss": 1.0241, + "mean_token_accuracy": 0.6929562091827393, + "num_tokens": 53216268.0, + "step": 2067 + }, + { + "epoch": 0.22710300900505162, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.355457305908203, + "learning_rate": 7.565885797950219e-07, + "loss": 0.9272, + "mean_token_accuracy": 0.7147189378738403, + "num_tokens": 53239331.0, + "step": 2068 + }, + { + "epoch": 0.22721282670766527, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.1629412174224854, + "learning_rate": 7.569546120058565e-07, + "loss": 1.0613, + "mean_token_accuracy": 0.690603494644165, + "num_tokens": 53269366.0, + "step": 2069 + }, + { + "epoch": 0.22732264441027894, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.4086198806762695, + "learning_rate": 7.573206442166911e-07, + "loss": 1.0185, + "mean_token_accuracy": 0.6937807202339172, + "num_tokens": 53293551.0, + "step": 2070 + }, + { + "epoch": 0.22743246211289259, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.3513097763061523, + "learning_rate": 7.576866764275256e-07, + "loss": 0.9671, + "mean_token_accuracy": 0.7014544010162354, + "num_tokens": 53317407.0, + "step": 2071 + }, + { + "epoch": 0.22754227981550626, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.0852835178375244, + "learning_rate": 7.580527086383601e-07, + "loss": 1.1031, + "mean_token_accuracy": 0.6823588013648987, + "num_tokens": 53349129.0, + "step": 2072 + }, + { + "epoch": 0.22765209751811993, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.4038608074188232, + "learning_rate": 7.584187408491947e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.6980442404747009, + "num_tokens": 53373807.0, + "step": 2073 + }, + { + "epoch": 0.22776191522073358, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.4144554138183594, + "learning_rate": 7.587847730600292e-07, + "loss": 1.1018, + "mean_token_accuracy": 0.6733400821685791, + "num_tokens": 53398993.0, + "step": 2074 + }, + { + "epoch": 0.22787173292334725, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.09798264503479, + "learning_rate": 7.591508052708638e-07, + "loss": 1.1134, + "mean_token_accuracy": 0.6736618876457214, + "num_tokens": 53430064.0, + "step": 2075 + }, + { + "epoch": 0.2279815506259609, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.1467444896698, + "learning_rate": 7.595168374816984e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.6943974494934082, + "num_tokens": 53459080.0, + "step": 2076 + }, + { + "epoch": 0.22809136832857457, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.1919736862182617, + "learning_rate": 7.598828696925329e-07, + "loss": 1.1084, + "mean_token_accuracy": 0.6734632253646851, + "num_tokens": 53488249.0, + "step": 2077 + }, + { + "epoch": 0.22820118603118822, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.2549023628234863, + "learning_rate": 7.602489019033674e-07, + "loss": 0.8922, + "mean_token_accuracy": 0.7258670330047607, + "num_tokens": 53512224.0, + "step": 2078 + }, + { + "epoch": 0.2283110037338019, + "ewc_loss": 6.854534149169922e-07, + "grad_norm": 2.258207321166992, + "learning_rate": 7.60614934114202e-07, + "loss": 1.0439, + "mean_token_accuracy": 0.6851093173027039, + "num_tokens": 53539216.0, + "step": 2079 + }, + { + "epoch": 0.22842082143641554, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.6186306476593018, + "learning_rate": 7.609809663250366e-07, + "loss": 0.9557, + "mean_token_accuracy": 0.7080814838409424, + "num_tokens": 53558834.0, + "step": 2080 + }, + { + "epoch": 0.2285306391390292, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.1262893676757812, + "learning_rate": 7.613469985358711e-07, + "loss": 1.0275, + "mean_token_accuracy": 0.6914466023445129, + "num_tokens": 53587249.0, + "step": 2081 + }, + { + "epoch": 0.2286404568416429, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.463347911834717, + "learning_rate": 7.617130307467057e-07, + "loss": 0.9877, + "mean_token_accuracy": 0.7016960382461548, + "num_tokens": 53610048.0, + "step": 2082 + }, + { + "epoch": 0.22875027454425653, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.1454782485961914, + "learning_rate": 7.620790629575402e-07, + "loss": 1.01, + "mean_token_accuracy": 0.6987793445587158, + "num_tokens": 53639322.0, + "step": 2083 + }, + { + "epoch": 0.2288600922468702, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.4223105907440186, + "learning_rate": 7.624450951683748e-07, + "loss": 0.959, + "mean_token_accuracy": 0.7094864249229431, + "num_tokens": 53661912.0, + "step": 2084 + }, + { + "epoch": 0.22896990994948385, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.182314395904541, + "learning_rate": 7.628111273792094e-07, + "loss": 1.0006, + "mean_token_accuracy": 0.7025343775749207, + "num_tokens": 53692100.0, + "step": 2085 + }, + { + "epoch": 0.22907972765209753, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.075594425201416, + "learning_rate": 7.631771595900439e-07, + "loss": 1.062, + "mean_token_accuracy": 0.6828126907348633, + "num_tokens": 53721317.0, + "step": 2086 + }, + { + "epoch": 0.22918954535471117, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.4595117568969727, + "learning_rate": 7.635431918008785e-07, + "loss": 1.0695, + "mean_token_accuracy": 0.6891855001449585, + "num_tokens": 53745152.0, + "step": 2087 + }, + { + "epoch": 0.22929936305732485, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.6316640377044678, + "learning_rate": 7.63909224011713e-07, + "loss": 0.9706, + "mean_token_accuracy": 0.7115713357925415, + "num_tokens": 53766720.0, + "step": 2088 + }, + { + "epoch": 0.2294091807599385, + "ewc_loss": 6.92903995513916e-07, + "grad_norm": 2.145355701446533, + "learning_rate": 7.642752562225475e-07, + "loss": 0.992, + "mean_token_accuracy": 0.695417582988739, + "num_tokens": 53794504.0, + "step": 2089 + }, + { + "epoch": 0.22951899846255217, + "ewc_loss": 6.966292858123779e-07, + "grad_norm": 2.354388475418091, + "learning_rate": 7.646412884333821e-07, + "loss": 1.0227, + "mean_token_accuracy": 0.6966847777366638, + "num_tokens": 53820018.0, + "step": 2090 + }, + { + "epoch": 0.2296288161651658, + "ewc_loss": 7.003545761108398e-07, + "grad_norm": 2.0371017456054688, + "learning_rate": 7.650073206442167e-07, + "loss": 1.0512, + "mean_token_accuracy": 0.686445951461792, + "num_tokens": 53851435.0, + "step": 2091 + }, + { + "epoch": 0.22973863386777948, + "ewc_loss": 7.003545761108398e-07, + "grad_norm": 2.527630567550659, + "learning_rate": 7.653733528550512e-07, + "loss": 0.9811, + "mean_token_accuracy": 0.7030225992202759, + "num_tokens": 53873241.0, + "step": 2092 + }, + { + "epoch": 0.22984845157039316, + "ewc_loss": 7.003545761108398e-07, + "grad_norm": 2.1227433681488037, + "learning_rate": 7.657393850658858e-07, + "loss": 1.0528, + "mean_token_accuracy": 0.688561737537384, + "num_tokens": 53903030.0, + "step": 2093 + }, + { + "epoch": 0.2299582692730068, + "ewc_loss": 7.003545761108398e-07, + "grad_norm": 2.536992311477661, + "learning_rate": 7.661054172767203e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.6947961449623108, + "num_tokens": 53925218.0, + "step": 2094 + }, + { + "epoch": 0.23006808697562048, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.41267991065979, + "learning_rate": 7.664714494875548e-07, + "loss": 1.0771, + "mean_token_accuracy": 0.6835402250289917, + "num_tokens": 53948843.0, + "step": 2095 + }, + { + "epoch": 0.23017790467823412, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.0573925971984863, + "learning_rate": 7.668374816983895e-07, + "loss": 0.9747, + "mean_token_accuracy": 0.7031269073486328, + "num_tokens": 53977112.0, + "step": 2096 + }, + { + "epoch": 0.2302877223808478, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.335757255554199, + "learning_rate": 7.67203513909224e-07, + "loss": 1.0051, + "mean_token_accuracy": 0.7042970657348633, + "num_tokens": 54002810.0, + "step": 2097 + }, + { + "epoch": 0.23039754008346144, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.2713489532470703, + "learning_rate": 7.675695461200585e-07, + "loss": 1.0023, + "mean_token_accuracy": 0.7058870792388916, + "num_tokens": 54029780.0, + "step": 2098 + }, + { + "epoch": 0.23050735778607512, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.2495415210723877, + "learning_rate": 7.679355783308931e-07, + "loss": 1.0768, + "mean_token_accuracy": 0.6853649616241455, + "num_tokens": 54054899.0, + "step": 2099 + }, + { + "epoch": 0.23061717548868876, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.1612942218780518, + "learning_rate": 7.683016105417276e-07, + "loss": 1.1003, + "mean_token_accuracy": 0.682387113571167, + "num_tokens": 54084137.0, + "step": 2100 + }, + { + "epoch": 0.23072699319130244, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.338146924972534, + "learning_rate": 7.686676427525623e-07, + "loss": 0.959, + "mean_token_accuracy": 0.7164239883422852, + "num_tokens": 54108823.0, + "step": 2101 + }, + { + "epoch": 0.2308368108939161, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.40804123878479, + "learning_rate": 7.690336749633968e-07, + "loss": 0.9793, + "mean_token_accuracy": 0.7028541564941406, + "num_tokens": 54131262.0, + "step": 2102 + }, + { + "epoch": 0.23094662859652976, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.3674521446228027, + "learning_rate": 7.693997071742313e-07, + "loss": 1.0128, + "mean_token_accuracy": 0.6923926472663879, + "num_tokens": 54154640.0, + "step": 2103 + }, + { + "epoch": 0.23105644629914343, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.013824701309204, + "learning_rate": 7.697657393850659e-07, + "loss": 0.9929, + "mean_token_accuracy": 0.7001086473464966, + "num_tokens": 54185330.0, + "step": 2104 + }, + { + "epoch": 0.23116626400175708, + "ewc_loss": 7.040798664093018e-07, + "grad_norm": 2.540423631668091, + "learning_rate": 7.701317715959004e-07, + "loss": 0.9629, + "mean_token_accuracy": 0.7076435089111328, + "num_tokens": 54206266.0, + "step": 2105 + }, + { + "epoch": 0.23127608170437075, + "ewc_loss": 7.078051567077637e-07, + "grad_norm": 2.1622984409332275, + "learning_rate": 7.704978038067349e-07, + "loss": 1.0736, + "mean_token_accuracy": 0.6767585277557373, + "num_tokens": 54234830.0, + "step": 2106 + }, + { + "epoch": 0.2313858994069844, + "ewc_loss": 7.078051567077637e-07, + "grad_norm": 2.3259363174438477, + "learning_rate": 7.708638360175696e-07, + "loss": 0.8863, + "mean_token_accuracy": 0.7247002720832825, + "num_tokens": 54258518.0, + "step": 2107 + }, + { + "epoch": 0.23149571710959807, + "ewc_loss": 7.078051567077637e-07, + "grad_norm": 2.1797075271606445, + "learning_rate": 7.712298682284041e-07, + "loss": 1.0781, + "mean_token_accuracy": 0.6800583600997925, + "num_tokens": 54285139.0, + "step": 2108 + }, + { + "epoch": 0.23160553481221172, + "ewc_loss": 7.078051567077637e-07, + "grad_norm": 2.381295919418335, + "learning_rate": 7.715959004392386e-07, + "loss": 0.9835, + "mean_token_accuracy": 0.7022603750228882, + "num_tokens": 54308341.0, + "step": 2109 + }, + { + "epoch": 0.2317153525148254, + "ewc_loss": 7.115304470062256e-07, + "grad_norm": 2.1865687370300293, + "learning_rate": 7.719619326500732e-07, + "loss": 1.1337, + "mean_token_accuracy": 0.6650663614273071, + "num_tokens": 54338009.0, + "step": 2110 + }, + { + "epoch": 0.23182517021743906, + "ewc_loss": 7.115304470062256e-07, + "grad_norm": 2.4573354721069336, + "learning_rate": 7.723279648609077e-07, + "loss": 1.0826, + "mean_token_accuracy": 0.6814857125282288, + "num_tokens": 54361156.0, + "step": 2111 + }, + { + "epoch": 0.2319349879200527, + "ewc_loss": 7.115304470062256e-07, + "grad_norm": 1.94013512134552, + "learning_rate": 7.726939970717423e-07, + "loss": 0.8927, + "mean_token_accuracy": 0.7265783548355103, + "num_tokens": 54390645.0, + "step": 2112 + }, + { + "epoch": 0.23204480562266638, + "ewc_loss": 7.115304470062256e-07, + "grad_norm": 2.478229522705078, + "learning_rate": 7.730600292825769e-07, + "loss": 0.9938, + "mean_token_accuracy": 0.6943752765655518, + "num_tokens": 54412841.0, + "step": 2113 + }, + { + "epoch": 0.23215462332528003, + "ewc_loss": 7.152557373046875e-07, + "grad_norm": 2.1249940395355225, + "learning_rate": 7.734260614934114e-07, + "loss": 0.9693, + "mean_token_accuracy": 0.7043212056159973, + "num_tokens": 54440731.0, + "step": 2114 + }, + { + "epoch": 0.2322644410278937, + "ewc_loss": 7.152557373046875e-07, + "grad_norm": 2.7181479930877686, + "learning_rate": 7.737920937042459e-07, + "loss": 1.011, + "mean_token_accuracy": 0.6940478086471558, + "num_tokens": 54460803.0, + "step": 2115 + }, + { + "epoch": 0.23237425873050735, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.471736431121826, + "learning_rate": 7.741581259150805e-07, + "loss": 0.9861, + "mean_token_accuracy": 0.7079150676727295, + "num_tokens": 54483265.0, + "step": 2116 + }, + { + "epoch": 0.23248407643312102, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.1720387935638428, + "learning_rate": 7.745241581259151e-07, + "loss": 0.9866, + "mean_token_accuracy": 0.7014706134796143, + "num_tokens": 54509995.0, + "step": 2117 + }, + { + "epoch": 0.23259389413573467, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.4622230529785156, + "learning_rate": 7.748901903367497e-07, + "loss": 0.983, + "mean_token_accuracy": 0.7001964449882507, + "num_tokens": 54533911.0, + "step": 2118 + }, + { + "epoch": 0.23270371183834834, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.5217642784118652, + "learning_rate": 7.752562225475842e-07, + "loss": 0.9896, + "mean_token_accuracy": 0.6998158097267151, + "num_tokens": 54555628.0, + "step": 2119 + }, + { + "epoch": 0.23281352954096202, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.1244752407073975, + "learning_rate": 7.756222547584187e-07, + "loss": 1.0421, + "mean_token_accuracy": 0.6923226118087769, + "num_tokens": 54584663.0, + "step": 2120 + }, + { + "epoch": 0.23292334724357566, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.706854820251465, + "learning_rate": 7.759882869692533e-07, + "loss": 1.0186, + "mean_token_accuracy": 0.7017096281051636, + "num_tokens": 54606332.0, + "step": 2121 + }, + { + "epoch": 0.23303316494618934, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.046887159347534, + "learning_rate": 7.763543191800878e-07, + "loss": 1.1003, + "mean_token_accuracy": 0.6761085987091064, + "num_tokens": 54639008.0, + "step": 2122 + }, + { + "epoch": 0.23314298264880298, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.0942888259887695, + "learning_rate": 7.767203513909224e-07, + "loss": 1.0415, + "mean_token_accuracy": 0.6855410933494568, + "num_tokens": 54668941.0, + "step": 2123 + }, + { + "epoch": 0.23325280035141666, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.505439043045044, + "learning_rate": 7.77086383601757e-07, + "loss": 0.9902, + "mean_token_accuracy": 0.6997741460800171, + "num_tokens": 54691348.0, + "step": 2124 + }, + { + "epoch": 0.2333626180540303, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.265299081802368, + "learning_rate": 7.774524158125915e-07, + "loss": 0.9813, + "mean_token_accuracy": 0.7058357000350952, + "num_tokens": 54715840.0, + "step": 2125 + }, + { + "epoch": 0.23347243575664398, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 2.2750415802001953, + "learning_rate": 7.77818448023426e-07, + "loss": 1.1086, + "mean_token_accuracy": 0.667043924331665, + "num_tokens": 54742261.0, + "step": 2126 + }, + { + "epoch": 0.23358225345925762, + "ewc_loss": 7.227063179016113e-07, + "grad_norm": 1.9627997875213623, + "learning_rate": 7.781844802342606e-07, + "loss": 1.0915, + "mean_token_accuracy": 0.6735649108886719, + "num_tokens": 54775001.0, + "step": 2127 + }, + { + "epoch": 0.2336920711618713, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.289632797241211, + "learning_rate": 7.785505124450952e-07, + "loss": 1.0661, + "mean_token_accuracy": 0.682169497013092, + "num_tokens": 54800287.0, + "step": 2128 + }, + { + "epoch": 0.23380188886448494, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.1954967975616455, + "learning_rate": 7.789165446559297e-07, + "loss": 1.0867, + "mean_token_accuracy": 0.6821330189704895, + "num_tokens": 54827212.0, + "step": 2129 + }, + { + "epoch": 0.23391170656709862, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.2296090126037598, + "learning_rate": 7.792825768667643e-07, + "loss": 1.1596, + "mean_token_accuracy": 0.6587291955947876, + "num_tokens": 54855265.0, + "step": 2130 + }, + { + "epoch": 0.2340215242697123, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.4497578144073486, + "learning_rate": 7.796486090775988e-07, + "loss": 1.0712, + "mean_token_accuracy": 0.676094651222229, + "num_tokens": 54880558.0, + "step": 2131 + }, + { + "epoch": 0.23413134197232593, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.2171242237091064, + "learning_rate": 7.800146412884333e-07, + "loss": 1.1037, + "mean_token_accuracy": 0.6748670339584351, + "num_tokens": 54908760.0, + "step": 2132 + }, + { + "epoch": 0.2342411596749396, + "ewc_loss": 7.301568984985352e-07, + "grad_norm": 2.429786205291748, + "learning_rate": 7.80380673499268e-07, + "loss": 0.9196, + "mean_token_accuracy": 0.7144325375556946, + "num_tokens": 54927989.0, + "step": 2133 + }, + { + "epoch": 0.23435097737755325, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.193485736846924, + "learning_rate": 7.807467057101025e-07, + "loss": 1.0473, + "mean_token_accuracy": 0.6869644522666931, + "num_tokens": 54955289.0, + "step": 2134 + }, + { + "epoch": 0.23446079508016693, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.2257018089294434, + "learning_rate": 7.811127379209371e-07, + "loss": 0.9779, + "mean_token_accuracy": 0.6993560791015625, + "num_tokens": 54981124.0, + "step": 2135 + }, + { + "epoch": 0.23457061278278057, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.1034514904022217, + "learning_rate": 7.814787701317716e-07, + "loss": 1.0692, + "mean_token_accuracy": 0.6832391023635864, + "num_tokens": 55009205.0, + "step": 2136 + }, + { + "epoch": 0.23468043048539425, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.2423877716064453, + "learning_rate": 7.818448023426061e-07, + "loss": 1.1212, + "mean_token_accuracy": 0.6686295866966248, + "num_tokens": 55036118.0, + "step": 2137 + }, + { + "epoch": 0.2347902481880079, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.265578269958496, + "learning_rate": 7.822108345534407e-07, + "loss": 0.9565, + "mean_token_accuracy": 0.7043547630310059, + "num_tokens": 55060467.0, + "step": 2138 + }, + { + "epoch": 0.23490006589062157, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.241112232208252, + "learning_rate": 7.825768667642753e-07, + "loss": 1.0968, + "mean_token_accuracy": 0.6739183664321899, + "num_tokens": 55087375.0, + "step": 2139 + }, + { + "epoch": 0.23500988359323524, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.2632288932800293, + "learning_rate": 7.829428989751098e-07, + "loss": 1.0294, + "mean_token_accuracy": 0.697332501411438, + "num_tokens": 55112810.0, + "step": 2140 + }, + { + "epoch": 0.2351197012958489, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.4865589141845703, + "learning_rate": 7.833089311859444e-07, + "loss": 1.0214, + "mean_token_accuracy": 0.698387622833252, + "num_tokens": 55134221.0, + "step": 2141 + }, + { + "epoch": 0.23522951899846256, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.1704983711242676, + "learning_rate": 7.836749633967789e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.6949067115783691, + "num_tokens": 55161282.0, + "step": 2142 + }, + { + "epoch": 0.2353393367010762, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.408538341522217, + "learning_rate": 7.840409956076133e-07, + "loss": 1.0512, + "mean_token_accuracy": 0.6889861822128296, + "num_tokens": 55185586.0, + "step": 2143 + }, + { + "epoch": 0.23544915440368988, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.099564790725708, + "learning_rate": 7.844070278184481e-07, + "loss": 1.1548, + "mean_token_accuracy": 0.6635032892227173, + "num_tokens": 55214146.0, + "step": 2144 + }, + { + "epoch": 0.23555897210630353, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.504340887069702, + "learning_rate": 7.847730600292826e-07, + "loss": 1.0017, + "mean_token_accuracy": 0.693455696105957, + "num_tokens": 55234676.0, + "step": 2145 + }, + { + "epoch": 0.2356687898089172, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.338927984237671, + "learning_rate": 7.85139092240117e-07, + "loss": 0.9358, + "mean_token_accuracy": 0.7106335163116455, + "num_tokens": 55257852.0, + "step": 2146 + }, + { + "epoch": 0.23577860751153085, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.4368345737457275, + "learning_rate": 7.855051244509517e-07, + "loss": 1.0164, + "mean_token_accuracy": 0.6933399438858032, + "num_tokens": 55280103.0, + "step": 2147 + }, + { + "epoch": 0.23588842521414452, + "ewc_loss": 7.338821887969971e-07, + "grad_norm": 2.224120855331421, + "learning_rate": 7.858711566617861e-07, + "loss": 1.0526, + "mean_token_accuracy": 0.6788272261619568, + "num_tokens": 55306689.0, + "step": 2148 + }, + { + "epoch": 0.2359982429167582, + "ewc_loss": 7.37607479095459e-07, + "grad_norm": 2.1840240955352783, + "learning_rate": 7.862371888726209e-07, + "loss": 1.0881, + "mean_token_accuracy": 0.675077497959137, + "num_tokens": 55335557.0, + "step": 2149 + }, + { + "epoch": 0.23610806061937184, + "ewc_loss": 7.37607479095459e-07, + "grad_norm": 2.3841304779052734, + "learning_rate": 7.866032210834554e-07, + "loss": 1.0385, + "mean_token_accuracy": 0.687624454498291, + "num_tokens": 55361347.0, + "step": 2150 + }, + { + "epoch": 0.2362178783219855, + "ewc_loss": 7.37607479095459e-07, + "grad_norm": 2.301534414291382, + "learning_rate": 7.869692532942898e-07, + "loss": 1.0293, + "mean_token_accuracy": 0.6886767745018005, + "num_tokens": 55386252.0, + "step": 2151 + }, + { + "epoch": 0.23632769602459916, + "ewc_loss": 7.37607479095459e-07, + "grad_norm": 2.4039273262023926, + "learning_rate": 7.873352855051244e-07, + "loss": 1.0568, + "mean_token_accuracy": 0.6936601400375366, + "num_tokens": 55407899.0, + "step": 2152 + }, + { + "epoch": 0.23643751372721283, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.3046505451202393, + "learning_rate": 7.877013177159589e-07, + "loss": 0.8689, + "mean_token_accuracy": 0.7302806377410889, + "num_tokens": 55430617.0, + "step": 2153 + }, + { + "epoch": 0.23654733142982648, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.5952084064483643, + "learning_rate": 7.880673499267934e-07, + "loss": 0.8572, + "mean_token_accuracy": 0.7316926121711731, + "num_tokens": 55449753.0, + "step": 2154 + }, + { + "epoch": 0.23665714913244015, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.1997785568237305, + "learning_rate": 7.884333821376281e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6876391172409058, + "num_tokens": 55477552.0, + "step": 2155 + }, + { + "epoch": 0.2367669668350538, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.346431255340576, + "learning_rate": 7.887994143484626e-07, + "loss": 0.9919, + "mean_token_accuracy": 0.7023364305496216, + "num_tokens": 55500986.0, + "step": 2156 + }, + { + "epoch": 0.23687678453766747, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.0732264518737793, + "learning_rate": 7.891654465592971e-07, + "loss": 1.056, + "mean_token_accuracy": 0.6866101026535034, + "num_tokens": 55533557.0, + "step": 2157 + }, + { + "epoch": 0.23698660224028115, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.183729887008667, + "learning_rate": 7.895314787701317e-07, + "loss": 1.0502, + "mean_token_accuracy": 0.6816227436065674, + "num_tokens": 55562766.0, + "step": 2158 + }, + { + "epoch": 0.2370964199428948, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.2649595737457275, + "learning_rate": 7.898975109809662e-07, + "loss": 1.1047, + "mean_token_accuracy": 0.670096218585968, + "num_tokens": 55590781.0, + "step": 2159 + }, + { + "epoch": 0.23720623764550847, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.392021656036377, + "learning_rate": 7.902635431918008e-07, + "loss": 1.0011, + "mean_token_accuracy": 0.6993933916091919, + "num_tokens": 55615755.0, + "step": 2160 + }, + { + "epoch": 0.2373160553481221, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.3249025344848633, + "learning_rate": 7.906295754026354e-07, + "loss": 0.9895, + "mean_token_accuracy": 0.701496958732605, + "num_tokens": 55639958.0, + "step": 2161 + }, + { + "epoch": 0.2374258730507358, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.519862651824951, + "learning_rate": 7.909956076134699e-07, + "loss": 1.0646, + "mean_token_accuracy": 0.6821128129959106, + "num_tokens": 55661257.0, + "step": 2162 + }, + { + "epoch": 0.23753569075334943, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.187288522720337, + "learning_rate": 7.913616398243044e-07, + "loss": 0.9418, + "mean_token_accuracy": 0.7206969857215881, + "num_tokens": 55687147.0, + "step": 2163 + }, + { + "epoch": 0.2376455084559631, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.028041124343872, + "learning_rate": 7.91727672035139e-07, + "loss": 1.0544, + "mean_token_accuracy": 0.6847788095474243, + "num_tokens": 55718756.0, + "step": 2164 + }, + { + "epoch": 0.23775532615857675, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.2959048748016357, + "learning_rate": 7.920937042459736e-07, + "loss": 1.1005, + "mean_token_accuracy": 0.6741660833358765, + "num_tokens": 55745692.0, + "step": 2165 + }, + { + "epoch": 0.23786514386119043, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.2109568119049072, + "learning_rate": 7.924597364568082e-07, + "loss": 0.9953, + "mean_token_accuracy": 0.7060422301292419, + "num_tokens": 55771595.0, + "step": 2166 + }, + { + "epoch": 0.23797496156380407, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.268162727355957, + "learning_rate": 7.928257686676427e-07, + "loss": 1.0269, + "mean_token_accuracy": 0.6885089874267578, + "num_tokens": 55795574.0, + "step": 2167 + }, + { + "epoch": 0.23808477926641775, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.1910746097564697, + "learning_rate": 7.931918008784772e-07, + "loss": 1.062, + "mean_token_accuracy": 0.684564471244812, + "num_tokens": 55823116.0, + "step": 2168 + }, + { + "epoch": 0.23819459696903142, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.1827635765075684, + "learning_rate": 7.935578330893118e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.686618447303772, + "num_tokens": 55849908.0, + "step": 2169 + }, + { + "epoch": 0.23830441467164507, + "ewc_loss": 7.413327693939209e-07, + "grad_norm": 2.0806124210357666, + "learning_rate": 7.939238653001463e-07, + "loss": 1.0587, + "mean_token_accuracy": 0.684181809425354, + "num_tokens": 55879307.0, + "step": 2170 + }, + { + "epoch": 0.23841423237425874, + "ewc_loss": 7.487833499908447e-07, + "grad_norm": 2.345564603805542, + "learning_rate": 7.942898975109809e-07, + "loss": 1.098, + "mean_token_accuracy": 0.6756201982498169, + "num_tokens": 55907879.0, + "step": 2171 + }, + { + "epoch": 0.23852405007687238, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.477658271789551, + "learning_rate": 7.946559297218155e-07, + "loss": 1.1096, + "mean_token_accuracy": 0.677456259727478, + "num_tokens": 55930065.0, + "step": 2172 + }, + { + "epoch": 0.23863386777948606, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.4329042434692383, + "learning_rate": 7.9502196193265e-07, + "loss": 1.0095, + "mean_token_accuracy": 0.6959583759307861, + "num_tokens": 55954500.0, + "step": 2173 + }, + { + "epoch": 0.2387436854820997, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.2222187519073486, + "learning_rate": 7.953879941434845e-07, + "loss": 0.9363, + "mean_token_accuracy": 0.7125355005264282, + "num_tokens": 55980294.0, + "step": 2174 + }, + { + "epoch": 0.23885350318471338, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 1.9904783964157104, + "learning_rate": 7.957540263543191e-07, + "loss": 1.022, + "mean_token_accuracy": 0.6963059902191162, + "num_tokens": 56012046.0, + "step": 2175 + }, + { + "epoch": 0.23896332088732702, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.3351824283599854, + "learning_rate": 7.961200585651537e-07, + "loss": 1.0084, + "mean_token_accuracy": 0.6971991062164307, + "num_tokens": 56039172.0, + "step": 2176 + }, + { + "epoch": 0.2390731385899407, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.494513511657715, + "learning_rate": 7.964860907759882e-07, + "loss": 1.0008, + "mean_token_accuracy": 0.7018762826919556, + "num_tokens": 56061746.0, + "step": 2177 + }, + { + "epoch": 0.23918295629255437, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.334972381591797, + "learning_rate": 7.968521229868228e-07, + "loss": 1.0433, + "mean_token_accuracy": 0.6824252605438232, + "num_tokens": 56086553.0, + "step": 2178 + }, + { + "epoch": 0.23929277399516802, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.0977416038513184, + "learning_rate": 7.972181551976573e-07, + "loss": 1.1002, + "mean_token_accuracy": 0.6740286350250244, + "num_tokens": 56118829.0, + "step": 2179 + }, + { + "epoch": 0.2394025916977817, + "ewc_loss": 7.525086402893066e-07, + "grad_norm": 2.2044119834899902, + "learning_rate": 7.975841874084918e-07, + "loss": 1.0474, + "mean_token_accuracy": 0.6870430707931519, + "num_tokens": 56145714.0, + "step": 2180 + }, + { + "epoch": 0.23951240940039534, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.3833534717559814, + "learning_rate": 7.979502196193265e-07, + "loss": 1.0465, + "mean_token_accuracy": 0.6835265755653381, + "num_tokens": 56170868.0, + "step": 2181 + }, + { + "epoch": 0.239622227103009, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.358966588973999, + "learning_rate": 7.98316251830161e-07, + "loss": 1.0203, + "mean_token_accuracy": 0.6946984529495239, + "num_tokens": 56195851.0, + "step": 2182 + }, + { + "epoch": 0.23973204480562266, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.0377678871154785, + "learning_rate": 7.986822840409956e-07, + "loss": 1.0769, + "mean_token_accuracy": 0.6871508359909058, + "num_tokens": 56227161.0, + "step": 2183 + }, + { + "epoch": 0.23984186250823633, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.4489023685455322, + "learning_rate": 7.990483162518301e-07, + "loss": 1.0875, + "mean_token_accuracy": 0.6780064105987549, + "num_tokens": 56249842.0, + "step": 2184 + }, + { + "epoch": 0.23995168021084998, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.227240562438965, + "learning_rate": 7.994143484626646e-07, + "loss": 1.0423, + "mean_token_accuracy": 0.6887598633766174, + "num_tokens": 56276810.0, + "step": 2185 + }, + { + "epoch": 0.24006149791346365, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.2768290042877197, + "learning_rate": 7.997803806734992e-07, + "loss": 0.9925, + "mean_token_accuracy": 0.6975774168968201, + "num_tokens": 56301375.0, + "step": 2186 + }, + { + "epoch": 0.24017131561607732, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.1889407634735107, + "learning_rate": 8.001464128843338e-07, + "loss": 1.051, + "mean_token_accuracy": 0.6839318871498108, + "num_tokens": 56327061.0, + "step": 2187 + }, + { + "epoch": 0.24028113331869097, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.0895044803619385, + "learning_rate": 8.005124450951683e-07, + "loss": 1.0188, + "mean_token_accuracy": 0.6946792602539062, + "num_tokens": 56357185.0, + "step": 2188 + }, + { + "epoch": 0.24039095102130464, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.3678932189941406, + "learning_rate": 8.008784773060029e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6859965324401855, + "num_tokens": 56380411.0, + "step": 2189 + }, + { + "epoch": 0.2405007687239183, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.5175838470458984, + "learning_rate": 8.012445095168374e-07, + "loss": 1.0015, + "mean_token_accuracy": 0.7032307982444763, + "num_tokens": 56403539.0, + "step": 2190 + }, + { + "epoch": 0.24061058642653196, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.2357280254364014, + "learning_rate": 8.016105417276719e-07, + "loss": 1.0091, + "mean_token_accuracy": 0.7004963159561157, + "num_tokens": 56430207.0, + "step": 2191 + }, + { + "epoch": 0.2407204041291456, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.0564169883728027, + "learning_rate": 8.019765739385066e-07, + "loss": 1.0832, + "mean_token_accuracy": 0.674830973148346, + "num_tokens": 56460386.0, + "step": 2192 + }, + { + "epoch": 0.24083022183175928, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.1685173511505127, + "learning_rate": 8.023426061493411e-07, + "loss": 0.9325, + "mean_token_accuracy": 0.717305600643158, + "num_tokens": 56487140.0, + "step": 2193 + }, + { + "epoch": 0.24094003953437293, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.573408603668213, + "learning_rate": 8.027086383601756e-07, + "loss": 1.1057, + "mean_token_accuracy": 0.6702275276184082, + "num_tokens": 56508781.0, + "step": 2194 + }, + { + "epoch": 0.2410498572369866, + "ewc_loss": 7.562339305877686e-07, + "grad_norm": 2.3805646896362305, + "learning_rate": 8.030746705710102e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6867263317108154, + "num_tokens": 56532472.0, + "step": 2195 + }, + { + "epoch": 0.24115967493960028, + "ewc_loss": 7.636845111846924e-07, + "grad_norm": 1.9415044784545898, + "learning_rate": 8.034407027818447e-07, + "loss": 1.0747, + "mean_token_accuracy": 0.6826355457305908, + "num_tokens": 56567185.0, + "step": 2196 + }, + { + "epoch": 0.24126949264221392, + "ewc_loss": 7.636845111846924e-07, + "grad_norm": 2.0125784873962402, + "learning_rate": 8.038067349926793e-07, + "loss": 0.9833, + "mean_token_accuracy": 0.7065773606300354, + "num_tokens": 56596186.0, + "step": 2197 + }, + { + "epoch": 0.2413793103448276, + "ewc_loss": 7.636845111846924e-07, + "grad_norm": 2.216244697570801, + "learning_rate": 8.041727672035139e-07, + "loss": 0.9672, + "mean_token_accuracy": 0.7013516426086426, + "num_tokens": 56621026.0, + "step": 2198 + }, + { + "epoch": 0.24148912804744124, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 2.5167222023010254, + "learning_rate": 8.045387994143484e-07, + "loss": 0.9939, + "mean_token_accuracy": 0.7007800340652466, + "num_tokens": 56641149.0, + "step": 2199 + }, + { + "epoch": 0.24159894575005492, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 2.360529661178589, + "learning_rate": 8.04904831625183e-07, + "loss": 0.9815, + "mean_token_accuracy": 0.7012075185775757, + "num_tokens": 56663743.0, + "step": 2200 + }, + { + "epoch": 0.24170876345266856, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 2.1039445400238037, + "learning_rate": 8.052708638360175e-07, + "loss": 0.9652, + "mean_token_accuracy": 0.7152014970779419, + "num_tokens": 56690487.0, + "step": 2201 + }, + { + "epoch": 0.24181858115528224, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 2.3649470806121826, + "learning_rate": 8.05636896046852e-07, + "loss": 0.9849, + "mean_token_accuracy": 0.6954383254051208, + "num_tokens": 56714174.0, + "step": 2202 + }, + { + "epoch": 0.24192839885789588, + "ewc_loss": 7.711350917816162e-07, + "grad_norm": 2.204493761062622, + "learning_rate": 8.060029282576867e-07, + "loss": 1.0901, + "mean_token_accuracy": 0.6773586273193359, + "num_tokens": 56743225.0, + "step": 2203 + }, + { + "epoch": 0.24203821656050956, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.313694953918457, + "learning_rate": 8.063689604685212e-07, + "loss": 1.0519, + "mean_token_accuracy": 0.6812951564788818, + "num_tokens": 56769588.0, + "step": 2204 + }, + { + "epoch": 0.2421480342631232, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.3195106983184814, + "learning_rate": 8.067349926793557e-07, + "loss": 1.0161, + "mean_token_accuracy": 0.6940523982048035, + "num_tokens": 56793483.0, + "step": 2205 + }, + { + "epoch": 0.24225785196573688, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.077000141143799, + "learning_rate": 8.071010248901903e-07, + "loss": 1.0645, + "mean_token_accuracy": 0.6899275183677673, + "num_tokens": 56821789.0, + "step": 2206 + }, + { + "epoch": 0.24236766966835055, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.246633768081665, + "learning_rate": 8.074670571010248e-07, + "loss": 0.957, + "mean_token_accuracy": 0.7069027423858643, + "num_tokens": 56847788.0, + "step": 2207 + }, + { + "epoch": 0.2424774873709642, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.243356943130493, + "learning_rate": 8.078330893118594e-07, + "loss": 0.9673, + "mean_token_accuracy": 0.7084348797798157, + "num_tokens": 56874038.0, + "step": 2208 + }, + { + "epoch": 0.24258730507357787, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.5208888053894043, + "learning_rate": 8.08199121522694e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.6898074746131897, + "num_tokens": 56895943.0, + "step": 2209 + }, + { + "epoch": 0.24269712277619152, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.3192427158355713, + "learning_rate": 8.085651537335285e-07, + "loss": 0.9834, + "mean_token_accuracy": 0.7015976905822754, + "num_tokens": 56920459.0, + "step": 2210 + }, + { + "epoch": 0.2428069404788052, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.2532167434692383, + "learning_rate": 8.08931185944363e-07, + "loss": 1.1118, + "mean_token_accuracy": 0.6722769737243652, + "num_tokens": 56947307.0, + "step": 2211 + }, + { + "epoch": 0.24291675818141883, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.19537353515625, + "learning_rate": 8.092972181551976e-07, + "loss": 1.0386, + "mean_token_accuracy": 0.6891639232635498, + "num_tokens": 56974344.0, + "step": 2212 + }, + { + "epoch": 0.2430265758840325, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.0746774673461914, + "learning_rate": 8.096632503660322e-07, + "loss": 1.1602, + "mean_token_accuracy": 0.6685266494750977, + "num_tokens": 57005194.0, + "step": 2213 + }, + { + "epoch": 0.24313639358664615, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.46140456199646, + "learning_rate": 8.100292825768668e-07, + "loss": 0.9577, + "mean_token_accuracy": 0.7028598785400391, + "num_tokens": 57027359.0, + "step": 2214 + }, + { + "epoch": 0.24324621128925983, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.427110433578491, + "learning_rate": 8.103953147877013e-07, + "loss": 0.9921, + "mean_token_accuracy": 0.6994390487670898, + "num_tokens": 57050898.0, + "step": 2215 + }, + { + "epoch": 0.2433560289918735, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.399646282196045, + "learning_rate": 8.107613469985358e-07, + "loss": 0.9632, + "mean_token_accuracy": 0.7026087045669556, + "num_tokens": 57075058.0, + "step": 2216 + }, + { + "epoch": 0.24346584669448715, + "ewc_loss": 7.748603820800781e-07, + "grad_norm": 2.267608880996704, + "learning_rate": 8.111273792093704e-07, + "loss": 0.9594, + "mean_token_accuracy": 0.7088035345077515, + "num_tokens": 57100386.0, + "step": 2217 + }, + { + "epoch": 0.24357566439710082, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.1770386695861816, + "learning_rate": 8.114934114202049e-07, + "loss": 0.9976, + "mean_token_accuracy": 0.6974363327026367, + "num_tokens": 57126772.0, + "step": 2218 + }, + { + "epoch": 0.24368548209971447, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.2884557247161865, + "learning_rate": 8.118594436310395e-07, + "loss": 1.0437, + "mean_token_accuracy": 0.6885262727737427, + "num_tokens": 57152044.0, + "step": 2219 + }, + { + "epoch": 0.24379529980232814, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.258120536804199, + "learning_rate": 8.122254758418741e-07, + "loss": 1.0452, + "mean_token_accuracy": 0.692441463470459, + "num_tokens": 57176202.0, + "step": 2220 + }, + { + "epoch": 0.2439051175049418, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.4312613010406494, + "learning_rate": 8.125915080527086e-07, + "loss": 0.9746, + "mean_token_accuracy": 0.7035017013549805, + "num_tokens": 57198532.0, + "step": 2221 + }, + { + "epoch": 0.24401493520755546, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.144096851348877, + "learning_rate": 8.129575402635431e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6909068822860718, + "num_tokens": 57228410.0, + "step": 2222 + }, + { + "epoch": 0.2441247529101691, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.036912679672241, + "learning_rate": 8.133235724743777e-07, + "loss": 0.9788, + "mean_token_accuracy": 0.7042288780212402, + "num_tokens": 57257091.0, + "step": 2223 + }, + { + "epoch": 0.24423457061278278, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.1551859378814697, + "learning_rate": 8.136896046852123e-07, + "loss": 0.9903, + "mean_token_accuracy": 0.6975374221801758, + "num_tokens": 57282667.0, + "step": 2224 + }, + { + "epoch": 0.24434438831539645, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.5004312992095947, + "learning_rate": 8.140556368960468e-07, + "loss": 0.94, + "mean_token_accuracy": 0.7113661766052246, + "num_tokens": 57301885.0, + "step": 2225 + }, + { + "epoch": 0.2444542060180101, + "ewc_loss": 7.7858567237854e-07, + "grad_norm": 2.5652315616607666, + "learning_rate": 8.144216691068814e-07, + "loss": 1.0265, + "mean_token_accuracy": 0.7021713256835938, + "num_tokens": 57322637.0, + "step": 2226 + }, + { + "epoch": 0.24456402372062377, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.477993965148926, + "learning_rate": 8.147877013177159e-07, + "loss": 1.0653, + "mean_token_accuracy": 0.6881588697433472, + "num_tokens": 57346366.0, + "step": 2227 + }, + { + "epoch": 0.24467384142323742, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.2368111610412598, + "learning_rate": 8.151537335285504e-07, + "loss": 1.0484, + "mean_token_accuracy": 0.6951442956924438, + "num_tokens": 57373168.0, + "step": 2228 + }, + { + "epoch": 0.2447836591258511, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.188389778137207, + "learning_rate": 8.155197657393851e-07, + "loss": 0.9945, + "mean_token_accuracy": 0.6960116624832153, + "num_tokens": 57402621.0, + "step": 2229 + }, + { + "epoch": 0.24489347682846474, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.4128503799438477, + "learning_rate": 8.158857979502196e-07, + "loss": 1.0459, + "mean_token_accuracy": 0.6957787275314331, + "num_tokens": 57426230.0, + "step": 2230 + }, + { + "epoch": 0.2450032945310784, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.2308130264282227, + "learning_rate": 8.162518301610542e-07, + "loss": 0.9918, + "mean_token_accuracy": 0.7025278806686401, + "num_tokens": 57453725.0, + "step": 2231 + }, + { + "epoch": 0.24511311223369206, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.082031488418579, + "learning_rate": 8.166178623718887e-07, + "loss": 1.0541, + "mean_token_accuracy": 0.6845980882644653, + "num_tokens": 57484306.0, + "step": 2232 + }, + { + "epoch": 0.24522292993630573, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.26924991607666, + "learning_rate": 8.169838945827232e-07, + "loss": 1.0399, + "mean_token_accuracy": 0.6915467977523804, + "num_tokens": 57510966.0, + "step": 2233 + }, + { + "epoch": 0.2453327476389194, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.190493583679199, + "learning_rate": 8.173499267935578e-07, + "loss": 1.0223, + "mean_token_accuracy": 0.6911072731018066, + "num_tokens": 57538533.0, + "step": 2234 + }, + { + "epoch": 0.24544256534153305, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.312734603881836, + "learning_rate": 8.177159590043924e-07, + "loss": 1.012, + "mean_token_accuracy": 0.6921509504318237, + "num_tokens": 57564275.0, + "step": 2235 + }, + { + "epoch": 0.24555238304414673, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.0116519927978516, + "learning_rate": 8.180819912152269e-07, + "loss": 1.0016, + "mean_token_accuracy": 0.6993246674537659, + "num_tokens": 57594837.0, + "step": 2236 + }, + { + "epoch": 0.24566220074676037, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.1752328872680664, + "learning_rate": 8.184480234260615e-07, + "loss": 0.9777, + "mean_token_accuracy": 0.7133125066757202, + "num_tokens": 57620478.0, + "step": 2237 + }, + { + "epoch": 0.24577201844937405, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.22259259223938, + "learning_rate": 8.18814055636896e-07, + "loss": 1.0756, + "mean_token_accuracy": 0.6830554008483887, + "num_tokens": 57647752.0, + "step": 2238 + }, + { + "epoch": 0.2458818361519877, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.3081817626953125, + "learning_rate": 8.191800878477305e-07, + "loss": 0.9959, + "mean_token_accuracy": 0.6992312669754028, + "num_tokens": 57672591.0, + "step": 2239 + }, + { + "epoch": 0.24599165385460137, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.300724983215332, + "learning_rate": 8.195461200585652e-07, + "loss": 0.9834, + "mean_token_accuracy": 0.7016465663909912, + "num_tokens": 57696420.0, + "step": 2240 + }, + { + "epoch": 0.246101471557215, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.3606786727905273, + "learning_rate": 8.199121522693997e-07, + "loss": 1.0148, + "mean_token_accuracy": 0.6905494928359985, + "num_tokens": 57720386.0, + "step": 2241 + }, + { + "epoch": 0.2462112892598287, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.5220766067504883, + "learning_rate": 8.202781844802342e-07, + "loss": 1.0323, + "mean_token_accuracy": 0.6950512528419495, + "num_tokens": 57746173.0, + "step": 2242 + }, + { + "epoch": 0.24632110696244233, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.370365858078003, + "learning_rate": 8.206442166910688e-07, + "loss": 1.0284, + "mean_token_accuracy": 0.6902545690536499, + "num_tokens": 57770085.0, + "step": 2243 + }, + { + "epoch": 0.246430924665056, + "ewc_loss": 7.82310962677002e-07, + "grad_norm": 2.029823064804077, + "learning_rate": 8.210102489019033e-07, + "loss": 0.9406, + "mean_token_accuracy": 0.7112561464309692, + "num_tokens": 57799026.0, + "step": 2244 + }, + { + "epoch": 0.24654074236766968, + "ewc_loss": 7.860362529754639e-07, + "grad_norm": 2.3039636611938477, + "learning_rate": 8.213762811127379e-07, + "loss": 1.0164, + "mean_token_accuracy": 0.694900393486023, + "num_tokens": 57824626.0, + "step": 2245 + }, + { + "epoch": 0.24665056007028333, + "ewc_loss": 7.860362529754639e-07, + "grad_norm": 2.484905242919922, + "learning_rate": 8.217423133235725e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.6821101307868958, + "num_tokens": 57847648.0, + "step": 2246 + }, + { + "epoch": 0.246760377772897, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 2.3750710487365723, + "learning_rate": 8.22108345534407e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.6868919134140015, + "num_tokens": 57870239.0, + "step": 2247 + }, + { + "epoch": 0.24687019547551065, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 2.1043426990509033, + "learning_rate": 8.224743777452416e-07, + "loss": 1.0077, + "mean_token_accuracy": 0.6927772760391235, + "num_tokens": 57899702.0, + "step": 2248 + }, + { + "epoch": 0.24698001317812432, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 2.7743475437164307, + "learning_rate": 8.228404099560761e-07, + "loss": 0.9869, + "mean_token_accuracy": 0.7042101621627808, + "num_tokens": 57918400.0, + "step": 2249 + }, + { + "epoch": 0.24708983088073797, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 2.529362916946411, + "learning_rate": 8.232064421669106e-07, + "loss": 1.0148, + "mean_token_accuracy": 0.6933884024620056, + "num_tokens": 57940924.0, + "step": 2250 + }, + { + "epoch": 0.24719964858335164, + "ewc_loss": 7.897615432739258e-07, + "grad_norm": 2.6914010047912598, + "learning_rate": 8.235724743777453e-07, + "loss": 0.9383, + "mean_token_accuracy": 0.7210505604743958, + "num_tokens": 57959917.0, + "step": 2251 + }, + { + "epoch": 0.24730946628596528, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.0785582065582275, + "learning_rate": 8.239385065885798e-07, + "loss": 1.0618, + "mean_token_accuracy": 0.6889795064926147, + "num_tokens": 57989935.0, + "step": 2252 + }, + { + "epoch": 0.24741928398857896, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.5959839820861816, + "learning_rate": 8.243045387994143e-07, + "loss": 0.9701, + "mean_token_accuracy": 0.7128550410270691, + "num_tokens": 58009541.0, + "step": 2253 + }, + { + "epoch": 0.24752910169119263, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 1.9768136739730835, + "learning_rate": 8.246705710102489e-07, + "loss": 1.0098, + "mean_token_accuracy": 0.6914093494415283, + "num_tokens": 58042087.0, + "step": 2254 + }, + { + "epoch": 0.24763891939380628, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.1475679874420166, + "learning_rate": 8.250366032210834e-07, + "loss": 0.9606, + "mean_token_accuracy": 0.7140957117080688, + "num_tokens": 58070764.0, + "step": 2255 + }, + { + "epoch": 0.24774873709641995, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.679403781890869, + "learning_rate": 8.25402635431918e-07, + "loss": 0.9772, + "mean_token_accuracy": 0.7038108110427856, + "num_tokens": 58090387.0, + "step": 2256 + }, + { + "epoch": 0.2478585547990336, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.3456058502197266, + "learning_rate": 8.257686676427526e-07, + "loss": 1.0648, + "mean_token_accuracy": 0.6869672536849976, + "num_tokens": 58113755.0, + "step": 2257 + }, + { + "epoch": 0.24796837250164727, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.4982807636260986, + "learning_rate": 8.261346998535871e-07, + "loss": 0.9425, + "mean_token_accuracy": 0.7162103056907654, + "num_tokens": 58135562.0, + "step": 2258 + }, + { + "epoch": 0.24807819020426092, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.305555820465088, + "learning_rate": 8.265007320644216e-07, + "loss": 1.052, + "mean_token_accuracy": 0.6825451850891113, + "num_tokens": 58161079.0, + "step": 2259 + }, + { + "epoch": 0.2481880079068746, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 2.276738405227661, + "learning_rate": 8.268667642752562e-07, + "loss": 0.947, + "mean_token_accuracy": 0.7175992727279663, + "num_tokens": 58186104.0, + "step": 2260 + }, + { + "epoch": 0.24829782560948824, + "ewc_loss": 7.972121238708496e-07, + "grad_norm": 1.9633992910385132, + "learning_rate": 8.272327964860908e-07, + "loss": 1.0835, + "mean_token_accuracy": 0.6789299249649048, + "num_tokens": 58221477.0, + "step": 2261 + }, + { + "epoch": 0.2484076433121019, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.253297805786133, + "learning_rate": 8.275988286969253e-07, + "loss": 1.0678, + "mean_token_accuracy": 0.6839488744735718, + "num_tokens": 58247976.0, + "step": 2262 + }, + { + "epoch": 0.24851746101471558, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.584174156188965, + "learning_rate": 8.279648609077599e-07, + "loss": 0.9545, + "mean_token_accuracy": 0.7119195461273193, + "num_tokens": 58267331.0, + "step": 2263 + }, + { + "epoch": 0.24862727871732923, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.3311288356781006, + "learning_rate": 8.283308931185944e-07, + "loss": 1.0444, + "mean_token_accuracy": 0.6830295324325562, + "num_tokens": 58292582.0, + "step": 2264 + }, + { + "epoch": 0.2487370964199429, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.277219295501709, + "learning_rate": 8.28696925329429e-07, + "loss": 1.1317, + "mean_token_accuracy": 0.6630065441131592, + "num_tokens": 58318884.0, + "step": 2265 + }, + { + "epoch": 0.24884691412255655, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.2940361499786377, + "learning_rate": 8.290629575402635e-07, + "loss": 1.0176, + "mean_token_accuracy": 0.6918901205062866, + "num_tokens": 58344826.0, + "step": 2266 + }, + { + "epoch": 0.24895673182517022, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.29892635345459, + "learning_rate": 8.294289897510981e-07, + "loss": 1.0245, + "mean_token_accuracy": 0.6902559399604797, + "num_tokens": 58368859.0, + "step": 2267 + }, + { + "epoch": 0.24906654952778387, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.2213282585144043, + "learning_rate": 8.297950219619327e-07, + "loss": 0.9945, + "mean_token_accuracy": 0.7035078406333923, + "num_tokens": 58397377.0, + "step": 2268 + }, + { + "epoch": 0.24917636723039754, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.3410961627960205, + "learning_rate": 8.301610541727672e-07, + "loss": 0.8878, + "mean_token_accuracy": 0.7313247323036194, + "num_tokens": 58419411.0, + "step": 2269 + }, + { + "epoch": 0.2492861849330112, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.1189143657684326, + "learning_rate": 8.305270863836017e-07, + "loss": 0.974, + "mean_token_accuracy": 0.7017532587051392, + "num_tokens": 58447935.0, + "step": 2270 + }, + { + "epoch": 0.24939600263562486, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.119518756866455, + "learning_rate": 8.308931185944363e-07, + "loss": 0.9949, + "mean_token_accuracy": 0.6998024582862854, + "num_tokens": 58474769.0, + "step": 2271 + }, + { + "epoch": 0.24950582033823854, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.2337474822998047, + "learning_rate": 8.312591508052709e-07, + "loss": 1.1365, + "mean_token_accuracy": 0.6627380847930908, + "num_tokens": 58504789.0, + "step": 2272 + }, + { + "epoch": 0.24961563804085218, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.581651449203491, + "learning_rate": 8.316251830161054e-07, + "loss": 0.9784, + "mean_token_accuracy": 0.700608491897583, + "num_tokens": 58526700.0, + "step": 2273 + }, + { + "epoch": 0.24972545574346586, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.104372262954712, + "learning_rate": 8.3199121522694e-07, + "loss": 1.0281, + "mean_token_accuracy": 0.6866216659545898, + "num_tokens": 58555555.0, + "step": 2274 + }, + { + "epoch": 0.2498352734460795, + "ewc_loss": 8.046627044677734e-07, + "grad_norm": 2.3447704315185547, + "learning_rate": 8.323572474377745e-07, + "loss": 0.8656, + "mean_token_accuracy": 0.7357082366943359, + "num_tokens": 58577667.0, + "step": 2275 + }, + { + "epoch": 0.24994509114869318, + "ewc_loss": 8.083879947662354e-07, + "grad_norm": 2.1119728088378906, + "learning_rate": 8.32723279648609e-07, + "loss": 1.0371, + "mean_token_accuracy": 0.6928800344467163, + "num_tokens": 58606031.0, + "step": 2276 + }, + { + "epoch": 0.2500549088513068, + "ewc_loss": 8.083879947662354e-07, + "grad_norm": 2.350693464279175, + "learning_rate": 8.330893118594437e-07, + "loss": 0.9759, + "mean_token_accuracy": 0.7050377130508423, + "num_tokens": 58631967.0, + "step": 2277 + }, + { + "epoch": 0.25016472655392047, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.1099514961242676, + "learning_rate": 8.334553440702782e-07, + "loss": 1.044, + "mean_token_accuracy": 0.6854469180107117, + "num_tokens": 58662429.0, + "step": 2278 + }, + { + "epoch": 0.25027454425653417, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.242572069168091, + "learning_rate": 8.338213762811127e-07, + "loss": 1.072, + "mean_token_accuracy": 0.673525333404541, + "num_tokens": 58689786.0, + "step": 2279 + }, + { + "epoch": 0.2503843619591478, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.395101308822632, + "learning_rate": 8.341874084919473e-07, + "loss": 0.9761, + "mean_token_accuracy": 0.7039283514022827, + "num_tokens": 58711364.0, + "step": 2280 + }, + { + "epoch": 0.25049417966176146, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.1509242057800293, + "learning_rate": 8.345534407027818e-07, + "loss": 1.0424, + "mean_token_accuracy": 0.6881393194198608, + "num_tokens": 58737840.0, + "step": 2281 + }, + { + "epoch": 0.25060399736437516, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.1637356281280518, + "learning_rate": 8.349194729136164e-07, + "loss": 1.095, + "mean_token_accuracy": 0.6781297922134399, + "num_tokens": 58766062.0, + "step": 2282 + }, + { + "epoch": 0.2507138150669888, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.2739267349243164, + "learning_rate": 8.35285505124451e-07, + "loss": 1.0348, + "mean_token_accuracy": 0.6849196553230286, + "num_tokens": 58792625.0, + "step": 2283 + }, + { + "epoch": 0.25082363276960246, + "ewc_loss": 8.121132850646973e-07, + "grad_norm": 2.395364999771118, + "learning_rate": 8.356515373352855e-07, + "loss": 1.1554, + "mean_token_accuracy": 0.6621044874191284, + "num_tokens": 58817547.0, + "step": 2284 + }, + { + "epoch": 0.2509334504722161, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.087430953979492, + "learning_rate": 8.360175695461201e-07, + "loss": 1.0337, + "mean_token_accuracy": 0.6881876587867737, + "num_tokens": 58845967.0, + "step": 2285 + }, + { + "epoch": 0.2510432681748298, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.06792950630188, + "learning_rate": 8.363836017569546e-07, + "loss": 1.021, + "mean_token_accuracy": 0.6959413886070251, + "num_tokens": 58876861.0, + "step": 2286 + }, + { + "epoch": 0.25115308587744345, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.3614156246185303, + "learning_rate": 8.36749633967789e-07, + "loss": 1.0422, + "mean_token_accuracy": 0.6884152889251709, + "num_tokens": 58900848.0, + "step": 2287 + }, + { + "epoch": 0.2512629035800571, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.310098648071289, + "learning_rate": 8.371156661786238e-07, + "loss": 0.9782, + "mean_token_accuracy": 0.7039041519165039, + "num_tokens": 58922927.0, + "step": 2288 + }, + { + "epoch": 0.25137272128267074, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.149210214614868, + "learning_rate": 8.374816983894583e-07, + "loss": 1.019, + "mean_token_accuracy": 0.6881542205810547, + "num_tokens": 58951123.0, + "step": 2289 + }, + { + "epoch": 0.25148253898528444, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.57000470161438, + "learning_rate": 8.378477306002928e-07, + "loss": 1.049, + "mean_token_accuracy": 0.6846756935119629, + "num_tokens": 58970496.0, + "step": 2290 + }, + { + "epoch": 0.2515923566878981, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.241467237472534, + "learning_rate": 8.382137628111274e-07, + "loss": 1.0429, + "mean_token_accuracy": 0.6905703544616699, + "num_tokens": 58994989.0, + "step": 2291 + }, + { + "epoch": 0.25170217439051173, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.5502822399139404, + "learning_rate": 8.385797950219619e-07, + "loss": 0.9669, + "mean_token_accuracy": 0.7054662704467773, + "num_tokens": 59015235.0, + "step": 2292 + }, + { + "epoch": 0.25181199209312544, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.3193612098693848, + "learning_rate": 8.389458272327965e-07, + "loss": 0.9743, + "mean_token_accuracy": 0.6992229223251343, + "num_tokens": 59038421.0, + "step": 2293 + }, + { + "epoch": 0.2519218097957391, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.2676899433135986, + "learning_rate": 8.393118594436311e-07, + "loss": 1.0652, + "mean_token_accuracy": 0.6762639284133911, + "num_tokens": 59064410.0, + "step": 2294 + }, + { + "epoch": 0.25203162749835273, + "ewc_loss": 8.158385753631592e-07, + "grad_norm": 2.394340991973877, + "learning_rate": 8.396778916544656e-07, + "loss": 1.0375, + "mean_token_accuracy": 0.6895688772201538, + "num_tokens": 59087245.0, + "step": 2295 + }, + { + "epoch": 0.2521414452009664, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.5116050243377686, + "learning_rate": 8.400439238653e-07, + "loss": 1.0255, + "mean_token_accuracy": 0.6904302835464478, + "num_tokens": 59111988.0, + "step": 2296 + }, + { + "epoch": 0.2522512629035801, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.48549222946167, + "learning_rate": 8.404099560761346e-07, + "loss": 1.0239, + "mean_token_accuracy": 0.699267566204071, + "num_tokens": 59137371.0, + "step": 2297 + }, + { + "epoch": 0.2523610806061937, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.323254108428955, + "learning_rate": 8.407759882869691e-07, + "loss": 1.0192, + "mean_token_accuracy": 0.6966637372970581, + "num_tokens": 59163108.0, + "step": 2298 + }, + { + "epoch": 0.25247089830880737, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.2108006477355957, + "learning_rate": 8.411420204978039e-07, + "loss": 1.0757, + "mean_token_accuracy": 0.6761759519577026, + "num_tokens": 59191057.0, + "step": 2299 + }, + { + "epoch": 0.252580716011421, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.1156065464019775, + "learning_rate": 8.415080527086383e-07, + "loss": 1.083, + "mean_token_accuracy": 0.6787328720092773, + "num_tokens": 59220690.0, + "step": 2300 + }, + { + "epoch": 0.2526905337140347, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.1514925956726074, + "learning_rate": 8.418740849194728e-07, + "loss": 0.9864, + "mean_token_accuracy": 0.7056663036346436, + "num_tokens": 59247310.0, + "step": 2301 + }, + { + "epoch": 0.25280035141664836, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.3013458251953125, + "learning_rate": 8.422401171303074e-07, + "loss": 1.0034, + "mean_token_accuracy": 0.7013258337974548, + "num_tokens": 59271797.0, + "step": 2302 + }, + { + "epoch": 0.252910169119262, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.2498157024383545, + "learning_rate": 8.426061493411419e-07, + "loss": 0.9935, + "mean_token_accuracy": 0.7011861205101013, + "num_tokens": 59299996.0, + "step": 2303 + }, + { + "epoch": 0.2530199868218757, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.2237446308135986, + "learning_rate": 8.429721815519765e-07, + "loss": 1.0738, + "mean_token_accuracy": 0.680635929107666, + "num_tokens": 59326639.0, + "step": 2304 + }, + { + "epoch": 0.25312980452448935, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.309659719467163, + "learning_rate": 8.433382137628111e-07, + "loss": 1.0075, + "mean_token_accuracy": 0.704028308391571, + "num_tokens": 59353176.0, + "step": 2305 + }, + { + "epoch": 0.253239622227103, + "ewc_loss": 8.195638656616211e-07, + "grad_norm": 2.102404832839966, + "learning_rate": 8.437042459736456e-07, + "loss": 1.0135, + "mean_token_accuracy": 0.6986289024353027, + "num_tokens": 59381590.0, + "step": 2306 + }, + { + "epoch": 0.25334943992971665, + "ewc_loss": 8.23289155960083e-07, + "grad_norm": 1.976577877998352, + "learning_rate": 8.440702781844801e-07, + "loss": 0.9764, + "mean_token_accuracy": 0.7052708864212036, + "num_tokens": 59411696.0, + "step": 2307 + }, + { + "epoch": 0.25345925763233035, + "ewc_loss": 8.307397365570068e-07, + "grad_norm": 2.204167127609253, + "learning_rate": 8.444363103953147e-07, + "loss": 1.0492, + "mean_token_accuracy": 0.680863618850708, + "num_tokens": 59438691.0, + "step": 2308 + }, + { + "epoch": 0.253569075334944, + "ewc_loss": 8.270144462585449e-07, + "grad_norm": 2.0892860889434814, + "learning_rate": 8.448023426061493e-07, + "loss": 1.0056, + "mean_token_accuracy": 0.6968332529067993, + "num_tokens": 59467636.0, + "step": 2309 + }, + { + "epoch": 0.25367889303755764, + "ewc_loss": 8.344650268554688e-07, + "grad_norm": 2.0956287384033203, + "learning_rate": 8.451683748169838e-07, + "loss": 1.0417, + "mean_token_accuracy": 0.6905860304832458, + "num_tokens": 59496548.0, + "step": 2310 + }, + { + "epoch": 0.25378871074017134, + "ewc_loss": 8.307397365570068e-07, + "grad_norm": 2.447556972503662, + "learning_rate": 8.455344070278184e-07, + "loss": 1.0385, + "mean_token_accuracy": 0.6904148459434509, + "num_tokens": 59518194.0, + "step": 2311 + }, + { + "epoch": 0.253898528442785, + "ewc_loss": 8.270144462585449e-07, + "grad_norm": 2.4609971046447754, + "learning_rate": 8.459004392386529e-07, + "loss": 0.9084, + "mean_token_accuracy": 0.721746563911438, + "num_tokens": 59540317.0, + "step": 2312 + }, + { + "epoch": 0.25400834614539863, + "ewc_loss": 8.344650268554688e-07, + "grad_norm": 2.2403981685638428, + "learning_rate": 8.462664714494874e-07, + "loss": 1.0747, + "mean_token_accuracy": 0.6745030283927917, + "num_tokens": 59569101.0, + "step": 2313 + }, + { + "epoch": 0.2541181638480123, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.2211103439331055, + "learning_rate": 8.46632503660322e-07, + "loss": 1.064, + "mean_token_accuracy": 0.6805329322814941, + "num_tokens": 59597709.0, + "step": 2314 + }, + { + "epoch": 0.254227981550626, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.0929629802703857, + "learning_rate": 8.469985358711566e-07, + "loss": 1.0145, + "mean_token_accuracy": 0.6941653490066528, + "num_tokens": 59626571.0, + "step": 2315 + }, + { + "epoch": 0.2543377992532396, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.27048921585083, + "learning_rate": 8.473645680819912e-07, + "loss": 1.0187, + "mean_token_accuracy": 0.6865277290344238, + "num_tokens": 59653507.0, + "step": 2316 + }, + { + "epoch": 0.2544476169558533, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.2957394123077393, + "learning_rate": 8.477306002928257e-07, + "loss": 0.9823, + "mean_token_accuracy": 0.7105328440666199, + "num_tokens": 59682791.0, + "step": 2317 + }, + { + "epoch": 0.2545574346584669, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.3880679607391357, + "learning_rate": 8.480966325036602e-07, + "loss": 1.115, + "mean_token_accuracy": 0.6690894365310669, + "num_tokens": 59708535.0, + "step": 2318 + }, + { + "epoch": 0.2546672523610806, + "ewc_loss": 8.419156074523926e-07, + "grad_norm": 2.754415988922119, + "learning_rate": 8.484626647144948e-07, + "loss": 0.9824, + "mean_token_accuracy": 0.703912615776062, + "num_tokens": 59728949.0, + "step": 2319 + }, + { + "epoch": 0.25477707006369427, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.415449619293213, + "learning_rate": 8.488286969253294e-07, + "loss": 0.9753, + "mean_token_accuracy": 0.7099558711051941, + "num_tokens": 59750730.0, + "step": 2320 + }, + { + "epoch": 0.2548868877663079, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.356590986251831, + "learning_rate": 8.491947291361639e-07, + "loss": 1.0072, + "mean_token_accuracy": 0.6988837718963623, + "num_tokens": 59773449.0, + "step": 2321 + }, + { + "epoch": 0.2549967054689216, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.304842472076416, + "learning_rate": 8.495607613469985e-07, + "loss": 1.0477, + "mean_token_accuracy": 0.6833726763725281, + "num_tokens": 59797987.0, + "step": 2322 + }, + { + "epoch": 0.25510652317153526, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.474630832672119, + "learning_rate": 8.49926793557833e-07, + "loss": 0.9675, + "mean_token_accuracy": 0.7002127766609192, + "num_tokens": 59819138.0, + "step": 2323 + }, + { + "epoch": 0.2552163408741489, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.4168171882629395, + "learning_rate": 8.502928257686675e-07, + "loss": 1.0627, + "mean_token_accuracy": 0.6908302307128906, + "num_tokens": 59842604.0, + "step": 2324 + }, + { + "epoch": 0.25532615857676255, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.1886861324310303, + "learning_rate": 8.506588579795022e-07, + "loss": 1.0282, + "mean_token_accuracy": 0.6914516687393188, + "num_tokens": 59869413.0, + "step": 2325 + }, + { + "epoch": 0.25543597627937625, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.2473440170288086, + "learning_rate": 8.510248901903367e-07, + "loss": 1.0412, + "mean_token_accuracy": 0.6857048273086548, + "num_tokens": 59896911.0, + "step": 2326 + }, + { + "epoch": 0.2555457939819899, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.328911781311035, + "learning_rate": 8.513909224011712e-07, + "loss": 0.9694, + "mean_token_accuracy": 0.7056839466094971, + "num_tokens": 59920209.0, + "step": 2327 + }, + { + "epoch": 0.25565561168460355, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.2788608074188232, + "learning_rate": 8.517569546120058e-07, + "loss": 1.0592, + "mean_token_accuracy": 0.6827645301818848, + "num_tokens": 59946626.0, + "step": 2328 + }, + { + "epoch": 0.25576542938721725, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.138448476791382, + "learning_rate": 8.521229868228403e-07, + "loss": 1.0555, + "mean_token_accuracy": 0.6893564462661743, + "num_tokens": 59975887.0, + "step": 2329 + }, + { + "epoch": 0.2558752470898309, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.2858879566192627, + "learning_rate": 8.524890190336748e-07, + "loss": 0.9413, + "mean_token_accuracy": 0.713649332523346, + "num_tokens": 60000691.0, + "step": 2330 + }, + { + "epoch": 0.25598506479244454, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.286694049835205, + "learning_rate": 8.528550512445095e-07, + "loss": 1.0177, + "mean_token_accuracy": 0.6973402500152588, + "num_tokens": 60025315.0, + "step": 2331 + }, + { + "epoch": 0.2560948824950582, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.144456148147583, + "learning_rate": 8.53221083455344e-07, + "loss": 1.0889, + "mean_token_accuracy": 0.6875259876251221, + "num_tokens": 60057120.0, + "step": 2332 + }, + { + "epoch": 0.2562047001976719, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.1515865325927734, + "learning_rate": 8.535871156661786e-07, + "loss": 1.0223, + "mean_token_accuracy": 0.6993914246559143, + "num_tokens": 60084156.0, + "step": 2333 + }, + { + "epoch": 0.25631451790028553, + "ewc_loss": 8.493661880493164e-07, + "grad_norm": 2.3585002422332764, + "learning_rate": 8.539531478770131e-07, + "loss": 1.0782, + "mean_token_accuracy": 0.6807236671447754, + "num_tokens": 60109948.0, + "step": 2334 + }, + { + "epoch": 0.2564243356028992, + "ewc_loss": 8.456408977508545e-07, + "grad_norm": 2.6379051208496094, + "learning_rate": 8.543191800878476e-07, + "loss": 0.9356, + "mean_token_accuracy": 0.7099153995513916, + "num_tokens": 60129571.0, + "step": 2335 + }, + { + "epoch": 0.2565341533055128, + "ewc_loss": 8.456408977508545e-07, + "grad_norm": 2.196425676345825, + "learning_rate": 8.546852122986823e-07, + "loss": 1.0558, + "mean_token_accuracy": 0.6833006143569946, + "num_tokens": 60157568.0, + "step": 2336 + }, + { + "epoch": 0.2566439710081265, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 2.257822275161743, + "learning_rate": 8.550512445095168e-07, + "loss": 0.9993, + "mean_token_accuracy": 0.6969596147537231, + "num_tokens": 60185034.0, + "step": 2337 + }, + { + "epoch": 0.25675378871074017, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 1.9164763689041138, + "learning_rate": 8.554172767203513e-07, + "loss": 1.0282, + "mean_token_accuracy": 0.6935347318649292, + "num_tokens": 60218705.0, + "step": 2338 + }, + { + "epoch": 0.2568636064133538, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 2.257467269897461, + "learning_rate": 8.557833089311859e-07, + "loss": 0.9391, + "mean_token_accuracy": 0.7166616320610046, + "num_tokens": 60243522.0, + "step": 2339 + }, + { + "epoch": 0.2569734241159675, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 2.204711675643921, + "learning_rate": 8.561493411420204e-07, + "loss": 1.0087, + "mean_token_accuracy": 0.6990578174591064, + "num_tokens": 60271606.0, + "step": 2340 + }, + { + "epoch": 0.25708324181858117, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 2.110626459121704, + "learning_rate": 8.56515373352855e-07, + "loss": 1.0342, + "mean_token_accuracy": 0.6935571432113647, + "num_tokens": 60300332.0, + "step": 2341 + }, + { + "epoch": 0.2571930595211948, + "ewc_loss": 8.530914783477783e-07, + "grad_norm": 2.117602825164795, + "learning_rate": 8.568814055636896e-07, + "loss": 0.8968, + "mean_token_accuracy": 0.7281796932220459, + "num_tokens": 60328449.0, + "step": 2342 + }, + { + "epoch": 0.25730287722380846, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.178589344024658, + "learning_rate": 8.572474377745241e-07, + "loss": 1.0114, + "mean_token_accuracy": 0.6988641619682312, + "num_tokens": 60354910.0, + "step": 2343 + }, + { + "epoch": 0.25741269492642216, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.207472085952759, + "learning_rate": 8.576134699853586e-07, + "loss": 0.9908, + "mean_token_accuracy": 0.7053408622741699, + "num_tokens": 60380063.0, + "step": 2344 + }, + { + "epoch": 0.2575225126290358, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.3860087394714355, + "learning_rate": 8.579795021961932e-07, + "loss": 1.0469, + "mean_token_accuracy": 0.688835084438324, + "num_tokens": 60404706.0, + "step": 2345 + }, + { + "epoch": 0.25763233033164945, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.1587250232696533, + "learning_rate": 8.583455344070277e-07, + "loss": 0.9964, + "mean_token_accuracy": 0.7015678882598877, + "num_tokens": 60433576.0, + "step": 2346 + }, + { + "epoch": 0.2577421480342631, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.1756012439727783, + "learning_rate": 8.587115666178624e-07, + "loss": 1.0018, + "mean_token_accuracy": 0.6969996094703674, + "num_tokens": 60458975.0, + "step": 2347 + }, + { + "epoch": 0.2578519657368768, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.096583843231201, + "learning_rate": 8.590775988286969e-07, + "loss": 1.0163, + "mean_token_accuracy": 0.7050628066062927, + "num_tokens": 60490043.0, + "step": 2348 + }, + { + "epoch": 0.25796178343949044, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.374633550643921, + "learning_rate": 8.594436310395314e-07, + "loss": 1.0354, + "mean_token_accuracy": 0.6884514689445496, + "num_tokens": 60512267.0, + "step": 2349 + }, + { + "epoch": 0.2580716011421041, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.256713628768921, + "learning_rate": 8.59809663250366e-07, + "loss": 1.046, + "mean_token_accuracy": 0.6904391646385193, + "num_tokens": 60537869.0, + "step": 2350 + }, + { + "epoch": 0.2581814188447178, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.2099666595458984, + "learning_rate": 8.601756954612005e-07, + "loss": 1.0064, + "mean_token_accuracy": 0.6987895965576172, + "num_tokens": 60563603.0, + "step": 2351 + }, + { + "epoch": 0.25829123654733144, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.435405731201172, + "learning_rate": 8.605417276720351e-07, + "loss": 1.0129, + "mean_token_accuracy": 0.6930129528045654, + "num_tokens": 60587192.0, + "step": 2352 + }, + { + "epoch": 0.2584010542499451, + "ewc_loss": 8.605420589447021e-07, + "grad_norm": 2.3558619022369385, + "learning_rate": 8.609077598828697e-07, + "loss": 0.9629, + "mean_token_accuracy": 0.7023799419403076, + "num_tokens": 60610638.0, + "step": 2353 + }, + { + "epoch": 0.25851087195255873, + "ewc_loss": 8.642673492431641e-07, + "grad_norm": 2.3478620052337646, + "learning_rate": 8.612737920937042e-07, + "loss": 1.069, + "mean_token_accuracy": 0.688702404499054, + "num_tokens": 60636598.0, + "step": 2354 + }, + { + "epoch": 0.25862068965517243, + "ewc_loss": 8.642673492431641e-07, + "grad_norm": 2.0769264698028564, + "learning_rate": 8.616398243045387e-07, + "loss": 1.0598, + "mean_token_accuracy": 0.6868211627006531, + "num_tokens": 60666496.0, + "step": 2355 + }, + { + "epoch": 0.2587305073577861, + "ewc_loss": 8.642673492431641e-07, + "grad_norm": 2.010664701461792, + "learning_rate": 8.620058565153733e-07, + "loss": 1.0663, + "mean_token_accuracy": 0.682790994644165, + "num_tokens": 60696995.0, + "step": 2356 + }, + { + "epoch": 0.2588403250603997, + "ewc_loss": 8.754432201385498e-07, + "grad_norm": 2.552832841873169, + "learning_rate": 8.623718887262079e-07, + "loss": 0.9606, + "mean_token_accuracy": 0.7123295664787292, + "num_tokens": 60717345.0, + "step": 2357 + }, + { + "epoch": 0.2589501427630134, + "ewc_loss": 8.754432201385498e-07, + "grad_norm": 2.302194833755493, + "learning_rate": 8.627379209370424e-07, + "loss": 1.0558, + "mean_token_accuracy": 0.6933486461639404, + "num_tokens": 60744639.0, + "step": 2358 + }, + { + "epoch": 0.25905996046562707, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.0997064113616943, + "learning_rate": 8.63103953147877e-07, + "loss": 1.0869, + "mean_token_accuracy": 0.6817349195480347, + "num_tokens": 60773675.0, + "step": 2359 + }, + { + "epoch": 0.2591697781682407, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.186988115310669, + "learning_rate": 8.634699853587115e-07, + "loss": 1.0105, + "mean_token_accuracy": 0.6978214979171753, + "num_tokens": 60799408.0, + "step": 2360 + }, + { + "epoch": 0.25927959587085436, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.401716947555542, + "learning_rate": 8.63836017569546e-07, + "loss": 1.0616, + "mean_token_accuracy": 0.6837636232376099, + "num_tokens": 60823031.0, + "step": 2361 + }, + { + "epoch": 0.25938941357346806, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.155545711517334, + "learning_rate": 8.642020497803806e-07, + "loss": 1.0378, + "mean_token_accuracy": 0.6925840973854065, + "num_tokens": 60851807.0, + "step": 2362 + }, + { + "epoch": 0.2594992312760817, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.892040729522705, + "learning_rate": 8.645680819912152e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.6866194009780884, + "num_tokens": 60871805.0, + "step": 2363 + }, + { + "epoch": 0.25960904897869536, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 1.9429476261138916, + "learning_rate": 8.649341142020498e-07, + "loss": 1.1601, + "mean_token_accuracy": 0.6573328971862793, + "num_tokens": 60909599.0, + "step": 2364 + }, + { + "epoch": 0.259718866681309, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.21916127204895, + "learning_rate": 8.653001464128843e-07, + "loss": 1.0774, + "mean_token_accuracy": 0.6833018064498901, + "num_tokens": 60936870.0, + "step": 2365 + }, + { + "epoch": 0.2598286843839227, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.4080123901367188, + "learning_rate": 8.656661786237188e-07, + "loss": 1.0085, + "mean_token_accuracy": 0.6967458724975586, + "num_tokens": 60958742.0, + "step": 2366 + }, + { + "epoch": 0.25993850208653635, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.086545705795288, + "learning_rate": 8.660322108345534e-07, + "loss": 1.0296, + "mean_token_accuracy": 0.6961685419082642, + "num_tokens": 60987581.0, + "step": 2367 + }, + { + "epoch": 0.26004831978915, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.166337013244629, + "learning_rate": 8.66398243045388e-07, + "loss": 1.0692, + "mean_token_accuracy": 0.6815318465232849, + "num_tokens": 61015388.0, + "step": 2368 + }, + { + "epoch": 0.2601581374917637, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.025332450866699, + "learning_rate": 8.667642752562225e-07, + "loss": 1.0578, + "mean_token_accuracy": 0.6856439113616943, + "num_tokens": 61046490.0, + "step": 2369 + }, + { + "epoch": 0.26026795519437734, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.138963460922241, + "learning_rate": 8.671303074670571e-07, + "loss": 1.1067, + "mean_token_accuracy": 0.6728129386901855, + "num_tokens": 61076806.0, + "step": 2370 + }, + { + "epoch": 0.260377772896991, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 1.9887815713882446, + "learning_rate": 8.674963396778916e-07, + "loss": 1.1094, + "mean_token_accuracy": 0.678874135017395, + "num_tokens": 61109820.0, + "step": 2371 + }, + { + "epoch": 0.26048759059960463, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.362428903579712, + "learning_rate": 8.678623718887261e-07, + "loss": 0.9912, + "mean_token_accuracy": 0.6980836391448975, + "num_tokens": 61131765.0, + "step": 2372 + }, + { + "epoch": 0.26059740830221834, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.099519729614258, + "learning_rate": 8.682284040995608e-07, + "loss": 1.0387, + "mean_token_accuracy": 0.6995838284492493, + "num_tokens": 61160748.0, + "step": 2373 + }, + { + "epoch": 0.260707226004832, + "ewc_loss": 8.791685104370117e-07, + "grad_norm": 2.1155099868774414, + "learning_rate": 8.685944363103953e-07, + "loss": 0.9978, + "mean_token_accuracy": 0.7029142379760742, + "num_tokens": 61189525.0, + "step": 2374 + }, + { + "epoch": 0.26081704370744563, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.2285208702087402, + "learning_rate": 8.689604685212298e-07, + "loss": 1.0345, + "mean_token_accuracy": 0.6955642700195312, + "num_tokens": 61217544.0, + "step": 2375 + }, + { + "epoch": 0.2609268614100593, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.2502260208129883, + "learning_rate": 8.693265007320644e-07, + "loss": 1.0137, + "mean_token_accuracy": 0.6944160461425781, + "num_tokens": 61244972.0, + "step": 2376 + }, + { + "epoch": 0.261036679112673, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.2356154918670654, + "learning_rate": 8.696925329428989e-07, + "loss": 1.0831, + "mean_token_accuracy": 0.6722246408462524, + "num_tokens": 61272586.0, + "step": 2377 + }, + { + "epoch": 0.2611464968152866, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.439039945602417, + "learning_rate": 8.700585651537334e-07, + "loss": 0.8913, + "mean_token_accuracy": 0.7244110107421875, + "num_tokens": 61294173.0, + "step": 2378 + }, + { + "epoch": 0.26125631451790027, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.1609554290771484, + "learning_rate": 8.704245973645681e-07, + "loss": 1.0421, + "mean_token_accuracy": 0.6893982887268066, + "num_tokens": 61321813.0, + "step": 2379 + }, + { + "epoch": 0.26136613222051397, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.1346559524536133, + "learning_rate": 8.707906295754026e-07, + "loss": 1.1582, + "mean_token_accuracy": 0.6589488387107849, + "num_tokens": 61351003.0, + "step": 2380 + }, + { + "epoch": 0.2614759499231276, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.3008291721343994, + "learning_rate": 8.711566617862372e-07, + "loss": 1.0798, + "mean_token_accuracy": 0.6830000281333923, + "num_tokens": 61376842.0, + "step": 2381 + }, + { + "epoch": 0.26158576762574126, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 1.9978957176208496, + "learning_rate": 8.715226939970717e-07, + "loss": 1.0188, + "mean_token_accuracy": 0.6951146125793457, + "num_tokens": 61409980.0, + "step": 2382 + }, + { + "epoch": 0.2616955853283549, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.0127670764923096, + "learning_rate": 8.718887262079062e-07, + "loss": 1.0623, + "mean_token_accuracy": 0.6884126663208008, + "num_tokens": 61442720.0, + "step": 2383 + }, + { + "epoch": 0.2618054030309686, + "ewc_loss": 8.828938007354736e-07, + "grad_norm": 2.5565361976623535, + "learning_rate": 8.722547584187409e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.7022514939308167, + "num_tokens": 61468322.0, + "step": 2384 + }, + { + "epoch": 0.26191522073358225, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.4872546195983887, + "learning_rate": 8.726207906295754e-07, + "loss": 1.0242, + "mean_token_accuracy": 0.685827374458313, + "num_tokens": 61490210.0, + "step": 2385 + }, + { + "epoch": 0.2620250384361959, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.390836477279663, + "learning_rate": 8.729868228404099e-07, + "loss": 1.0235, + "mean_token_accuracy": 0.6902260184288025, + "num_tokens": 61513905.0, + "step": 2386 + }, + { + "epoch": 0.2621348561388096, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.523177146911621, + "learning_rate": 8.733528550512445e-07, + "loss": 1.0113, + "mean_token_accuracy": 0.6918133497238159, + "num_tokens": 61533984.0, + "step": 2387 + }, + { + "epoch": 0.26224467384142325, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.1697731018066406, + "learning_rate": 8.73718887262079e-07, + "loss": 0.9776, + "mean_token_accuracy": 0.7071133255958557, + "num_tokens": 61559133.0, + "step": 2388 + }, + { + "epoch": 0.2623544915440369, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.266057014465332, + "learning_rate": 8.740849194729136e-07, + "loss": 0.9973, + "mean_token_accuracy": 0.7008439302444458, + "num_tokens": 61583791.0, + "step": 2389 + }, + { + "epoch": 0.26246430924665054, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.5882561206817627, + "learning_rate": 8.744509516837482e-07, + "loss": 0.9763, + "mean_token_accuracy": 0.7076836824417114, + "num_tokens": 61604249.0, + "step": 2390 + }, + { + "epoch": 0.26257412694926424, + "ewc_loss": 8.903443813323975e-07, + "grad_norm": 2.446124315261841, + "learning_rate": 8.748169838945827e-07, + "loss": 0.9844, + "mean_token_accuracy": 0.706334114074707, + "num_tokens": 61627052.0, + "step": 2391 + }, + { + "epoch": 0.2626839446518779, + "ewc_loss": 8.940696716308594e-07, + "grad_norm": 2.1446685791015625, + "learning_rate": 8.751830161054172e-07, + "loss": 1.0854, + "mean_token_accuracy": 0.6831361055374146, + "num_tokens": 61655979.0, + "step": 2392 + }, + { + "epoch": 0.26279376235449153, + "ewc_loss": 8.940696716308594e-07, + "grad_norm": 2.2106239795684814, + "learning_rate": 8.755490483162518e-07, + "loss": 1.103, + "mean_token_accuracy": 0.6697862148284912, + "num_tokens": 61684326.0, + "step": 2393 + }, + { + "epoch": 0.2629035800571052, + "ewc_loss": 8.940696716308594e-07, + "grad_norm": 2.28749942779541, + "learning_rate": 8.759150805270863e-07, + "loss": 1.0776, + "mean_token_accuracy": 0.6799296736717224, + "num_tokens": 61712239.0, + "step": 2394 + }, + { + "epoch": 0.2630133977597189, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.156508445739746, + "learning_rate": 8.76281112737921e-07, + "loss": 1.0419, + "mean_token_accuracy": 0.6899658441543579, + "num_tokens": 61740619.0, + "step": 2395 + }, + { + "epoch": 0.2631232154623325, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.2723329067230225, + "learning_rate": 8.766471449487555e-07, + "loss": 1.0619, + "mean_token_accuracy": 0.6795587539672852, + "num_tokens": 61766457.0, + "step": 2396 + }, + { + "epoch": 0.2632330331649462, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.2957074642181396, + "learning_rate": 8.7701317715959e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.696161150932312, + "num_tokens": 61792109.0, + "step": 2397 + }, + { + "epoch": 0.2633428508675599, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.0605156421661377, + "learning_rate": 8.773792093704246e-07, + "loss": 1.0047, + "mean_token_accuracy": 0.695796012878418, + "num_tokens": 61821867.0, + "step": 2398 + }, + { + "epoch": 0.2634526685701735, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.034054756164551, + "learning_rate": 8.777452415812591e-07, + "loss": 0.9806, + "mean_token_accuracy": 0.7042685151100159, + "num_tokens": 61851076.0, + "step": 2399 + }, + { + "epoch": 0.26356248627278717, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.0759294033050537, + "learning_rate": 8.781112737920937e-07, + "loss": 0.9798, + "mean_token_accuracy": 0.6991807222366333, + "num_tokens": 61883955.0, + "step": 2400 + }, + { + "epoch": 0.2636723039754008, + "ewc_loss": 8.977949619293213e-07, + "grad_norm": 2.1596925258636475, + "learning_rate": 8.784773060029283e-07, + "loss": 0.9238, + "mean_token_accuracy": 0.7164724469184875, + "num_tokens": 61909733.0, + "step": 2401 + }, + { + "epoch": 0.2637821216780145, + "ewc_loss": 9.052455425262451e-07, + "grad_norm": 2.1890270709991455, + "learning_rate": 8.788433382137628e-07, + "loss": 1.0272, + "mean_token_accuracy": 0.6891052722930908, + "num_tokens": 61937771.0, + "step": 2402 + }, + { + "epoch": 0.26389193938062816, + "ewc_loss": 9.052455425262451e-07, + "grad_norm": 2.636765241622925, + "learning_rate": 8.792093704245973e-07, + "loss": 0.9995, + "mean_token_accuracy": 0.699215292930603, + "num_tokens": 61958207.0, + "step": 2403 + }, + { + "epoch": 0.2640017570832418, + "ewc_loss": 9.08970832824707e-07, + "grad_norm": 2.1441359519958496, + "learning_rate": 8.795754026354319e-07, + "loss": 1.0504, + "mean_token_accuracy": 0.6802067756652832, + "num_tokens": 61987806.0, + "step": 2404 + }, + { + "epoch": 0.2641115747858555, + "ewc_loss": 9.08970832824707e-07, + "grad_norm": 2.0948843955993652, + "learning_rate": 8.799414348462665e-07, + "loss": 1.1131, + "mean_token_accuracy": 0.6695915460586548, + "num_tokens": 62019461.0, + "step": 2405 + }, + { + "epoch": 0.26422139248846915, + "ewc_loss": 9.08970832824707e-07, + "grad_norm": 2.02112078666687, + "learning_rate": 8.80307467057101e-07, + "loss": 1.1091, + "mean_token_accuracy": 0.6727863550186157, + "num_tokens": 62052648.0, + "step": 2406 + }, + { + "epoch": 0.2643312101910828, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.3152194023132324, + "learning_rate": 8.806734992679356e-07, + "loss": 1.0196, + "mean_token_accuracy": 0.7021013498306274, + "num_tokens": 62077698.0, + "step": 2407 + }, + { + "epoch": 0.26444102789369645, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.228609800338745, + "learning_rate": 8.810395314787701e-07, + "loss": 1.0471, + "mean_token_accuracy": 0.69594806432724, + "num_tokens": 62105774.0, + "step": 2408 + }, + { + "epoch": 0.26455084559631015, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.4334380626678467, + "learning_rate": 8.814055636896046e-07, + "loss": 1.0257, + "mean_token_accuracy": 0.6922910213470459, + "num_tokens": 62128899.0, + "step": 2409 + }, + { + "epoch": 0.2646606632989238, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.3563852310180664, + "learning_rate": 8.817715959004392e-07, + "loss": 1.0678, + "mean_token_accuracy": 0.6775149703025818, + "num_tokens": 62152471.0, + "step": 2410 + }, + { + "epoch": 0.26477048100153744, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.5649290084838867, + "learning_rate": 8.821376281112738e-07, + "loss": 1.0395, + "mean_token_accuracy": 0.6917634010314941, + "num_tokens": 62174310.0, + "step": 2411 + }, + { + "epoch": 0.2648802987041511, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.6698973178863525, + "learning_rate": 8.825036603221084e-07, + "loss": 1.0208, + "mean_token_accuracy": 0.7069604992866516, + "num_tokens": 62193956.0, + "step": 2412 + }, + { + "epoch": 0.2649901164067648, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.183248996734619, + "learning_rate": 8.828696925329429e-07, + "loss": 1.0516, + "mean_token_accuracy": 0.6795995831489563, + "num_tokens": 62220723.0, + "step": 2413 + }, + { + "epoch": 0.26509993410937843, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.2630908489227295, + "learning_rate": 8.832357247437774e-07, + "loss": 1.0445, + "mean_token_accuracy": 0.6894117593765259, + "num_tokens": 62246768.0, + "step": 2414 + }, + { + "epoch": 0.2652097518119921, + "ewc_loss": 9.164214134216309e-07, + "grad_norm": 2.1179983615875244, + "learning_rate": 8.83601756954612e-07, + "loss": 0.9779, + "mean_token_accuracy": 0.708207905292511, + "num_tokens": 62274476.0, + "step": 2415 + }, + { + "epoch": 0.2653195695146058, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.2176947593688965, + "learning_rate": 8.839677891654466e-07, + "loss": 0.9728, + "mean_token_accuracy": 0.7065861225128174, + "num_tokens": 62300250.0, + "step": 2416 + }, + { + "epoch": 0.2654293872172194, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.153306484222412, + "learning_rate": 8.843338213762811e-07, + "loss": 0.975, + "mean_token_accuracy": 0.7097492218017578, + "num_tokens": 62325482.0, + "step": 2417 + }, + { + "epoch": 0.26553920491983307, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.4347283840179443, + "learning_rate": 8.846998535871157e-07, + "loss": 1.0335, + "mean_token_accuracy": 0.6947623491287231, + "num_tokens": 62348385.0, + "step": 2418 + }, + { + "epoch": 0.2656490226224467, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.3346097469329834, + "learning_rate": 8.850658857979502e-07, + "loss": 0.9942, + "mean_token_accuracy": 0.6999633312225342, + "num_tokens": 62371942.0, + "step": 2419 + }, + { + "epoch": 0.2657588403250604, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.2857372760772705, + "learning_rate": 8.854319180087847e-07, + "loss": 0.9866, + "mean_token_accuracy": 0.7057726383209229, + "num_tokens": 62398163.0, + "step": 2420 + }, + { + "epoch": 0.26586865802767407, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.5722153186798096, + "learning_rate": 8.857979502196194e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6853552460670471, + "num_tokens": 62419047.0, + "step": 2421 + }, + { + "epoch": 0.2659784757302877, + "ewc_loss": 9.201467037200928e-07, + "grad_norm": 2.321805953979492, + "learning_rate": 8.861639824304539e-07, + "loss": 1.1139, + "mean_token_accuracy": 0.6813833713531494, + "num_tokens": 62446956.0, + "step": 2422 + }, + { + "epoch": 0.26608829343290136, + "ewc_loss": 9.238719940185547e-07, + "grad_norm": 2.5054383277893066, + "learning_rate": 8.865300146412884e-07, + "loss": 1.0473, + "mean_token_accuracy": 0.6902263760566711, + "num_tokens": 62468806.0, + "step": 2423 + }, + { + "epoch": 0.26619811113551506, + "ewc_loss": 9.238719940185547e-07, + "grad_norm": 2.3486242294311523, + "learning_rate": 8.86896046852123e-07, + "loss": 1.0549, + "mean_token_accuracy": 0.6872930526733398, + "num_tokens": 62493696.0, + "step": 2424 + }, + { + "epoch": 0.2663079288381287, + "ewc_loss": 9.238719940185547e-07, + "grad_norm": 2.83315110206604, + "learning_rate": 8.872620790629575e-07, + "loss": 0.9242, + "mean_token_accuracy": 0.714908242225647, + "num_tokens": 62510545.0, + "step": 2425 + }, + { + "epoch": 0.26641774654074235, + "ewc_loss": 9.238719940185547e-07, + "grad_norm": 2.0956196784973145, + "learning_rate": 8.87628111273792e-07, + "loss": 1.0106, + "mean_token_accuracy": 0.7034958004951477, + "num_tokens": 62538602.0, + "step": 2426 + }, + { + "epoch": 0.26652756424335605, + "ewc_loss": 9.238719940185547e-07, + "grad_norm": 2.8106849193573, + "learning_rate": 8.879941434846267e-07, + "loss": 0.924, + "mean_token_accuracy": 0.7162008285522461, + "num_tokens": 62556858.0, + "step": 2427 + }, + { + "epoch": 0.2666373819459697, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.3180174827575684, + "learning_rate": 8.883601756954612e-07, + "loss": 1.0037, + "mean_token_accuracy": 0.7013291120529175, + "num_tokens": 62580368.0, + "step": 2428 + }, + { + "epoch": 0.26674719964858334, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.221654176712036, + "learning_rate": 8.887262079062958e-07, + "loss": 1.0144, + "mean_token_accuracy": 0.6954865455627441, + "num_tokens": 62605952.0, + "step": 2429 + }, + { + "epoch": 0.266857017351197, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.088409662246704, + "learning_rate": 8.890922401171303e-07, + "loss": 1.0351, + "mean_token_accuracy": 0.6877374649047852, + "num_tokens": 62635903.0, + "step": 2430 + }, + { + "epoch": 0.2669668350538107, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 1.8788186311721802, + "learning_rate": 8.894582723279648e-07, + "loss": 1.107, + "mean_token_accuracy": 0.6687682867050171, + "num_tokens": 62675543.0, + "step": 2431 + }, + { + "epoch": 0.26707665275642434, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.0972957611083984, + "learning_rate": 8.898243045387995e-07, + "loss": 1.0774, + "mean_token_accuracy": 0.6757497191429138, + "num_tokens": 62705144.0, + "step": 2432 + }, + { + "epoch": 0.267186470459038, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.3486878871917725, + "learning_rate": 8.90190336749634e-07, + "loss": 1.0069, + "mean_token_accuracy": 0.6934388875961304, + "num_tokens": 62730740.0, + "step": 2433 + }, + { + "epoch": 0.2672962881616517, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.2832906246185303, + "learning_rate": 8.905563689604685e-07, + "loss": 1.0619, + "mean_token_accuracy": 0.684400737285614, + "num_tokens": 62755879.0, + "step": 2434 + }, + { + "epoch": 0.26740610586426533, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.057945966720581, + "learning_rate": 8.909224011713031e-07, + "loss": 0.9948, + "mean_token_accuracy": 0.7015471458435059, + "num_tokens": 62786905.0, + "step": 2435 + }, + { + "epoch": 0.267515923566879, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.0763137340545654, + "learning_rate": 8.912884333821376e-07, + "loss": 1.0407, + "mean_token_accuracy": 0.6921597719192505, + "num_tokens": 62815251.0, + "step": 2436 + }, + { + "epoch": 0.2676257412694926, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.2370309829711914, + "learning_rate": 8.916544655929722e-07, + "loss": 0.9785, + "mean_token_accuracy": 0.7127030491828918, + "num_tokens": 62840071.0, + "step": 2437 + }, + { + "epoch": 0.2677355589721063, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.3403823375701904, + "learning_rate": 8.920204978038068e-07, + "loss": 0.9428, + "mean_token_accuracy": 0.7118615508079529, + "num_tokens": 62863599.0, + "step": 2438 + }, + { + "epoch": 0.26784537667471997, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.4008703231811523, + "learning_rate": 8.923865300146413e-07, + "loss": 0.9879, + "mean_token_accuracy": 0.699706494808197, + "num_tokens": 62886063.0, + "step": 2439 + }, + { + "epoch": 0.2679551943773336, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.29353404045105, + "learning_rate": 8.927525622254757e-07, + "loss": 1.0132, + "mean_token_accuracy": 0.7043805122375488, + "num_tokens": 62911701.0, + "step": 2440 + }, + { + "epoch": 0.26806501207994726, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.6659622192382812, + "learning_rate": 8.931185944363104e-07, + "loss": 0.9628, + "mean_token_accuracy": 0.7064584493637085, + "num_tokens": 62931389.0, + "step": 2441 + }, + { + "epoch": 0.26817482978256096, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.181292772293091, + "learning_rate": 8.934846266471448e-07, + "loss": 1.0912, + "mean_token_accuracy": 0.6744272112846375, + "num_tokens": 62959298.0, + "step": 2442 + }, + { + "epoch": 0.2682846474851746, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.128387928009033, + "learning_rate": 8.938506588579794e-07, + "loss": 0.9921, + "mean_token_accuracy": 0.7069084644317627, + "num_tokens": 62987228.0, + "step": 2443 + }, + { + "epoch": 0.26839446518778826, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.0538389682769775, + "learning_rate": 8.94216691068814e-07, + "loss": 0.9986, + "mean_token_accuracy": 0.6960195302963257, + "num_tokens": 63016848.0, + "step": 2444 + }, + { + "epoch": 0.26850428289040196, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.147291898727417, + "learning_rate": 8.945827232796485e-07, + "loss": 1.1022, + "mean_token_accuracy": 0.6751184463500977, + "num_tokens": 63047314.0, + "step": 2445 + }, + { + "epoch": 0.2686141005930156, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 1.9944604635238647, + "learning_rate": 8.949487554904832e-07, + "loss": 1.1057, + "mean_token_accuracy": 0.6719263792037964, + "num_tokens": 63081054.0, + "step": 2446 + }, + { + "epoch": 0.26872391829562925, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.5273585319519043, + "learning_rate": 8.953147877013176e-07, + "loss": 1.0074, + "mean_token_accuracy": 0.6970745325088501, + "num_tokens": 63102763.0, + "step": 2447 + }, + { + "epoch": 0.2688337359982429, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.5699269771575928, + "learning_rate": 8.956808199121522e-07, + "loss": 1.0142, + "mean_token_accuracy": 0.6937205791473389, + "num_tokens": 63124068.0, + "step": 2448 + }, + { + "epoch": 0.2689435537008566, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.341902256011963, + "learning_rate": 8.960468521229869e-07, + "loss": 0.9632, + "mean_token_accuracy": 0.7081130743026733, + "num_tokens": 63145050.0, + "step": 2449 + }, + { + "epoch": 0.26905337140347024, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.3964898586273193, + "learning_rate": 8.964128843338213e-07, + "loss": 0.9742, + "mean_token_accuracy": 0.7127792239189148, + "num_tokens": 63167810.0, + "step": 2450 + }, + { + "epoch": 0.2691631891060839, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.0974035263061523, + "learning_rate": 8.967789165446558e-07, + "loss": 0.9838, + "mean_token_accuracy": 0.7058429718017578, + "num_tokens": 63197274.0, + "step": 2451 + }, + { + "epoch": 0.26927300680869753, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.2299692630767822, + "learning_rate": 8.971449487554904e-07, + "loss": 1.0368, + "mean_token_accuracy": 0.6877426505088806, + "num_tokens": 63221700.0, + "step": 2452 + }, + { + "epoch": 0.26938282451131124, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.315755605697632, + "learning_rate": 8.97510980966325e-07, + "loss": 1.0601, + "mean_token_accuracy": 0.6812056303024292, + "num_tokens": 63248117.0, + "step": 2453 + }, + { + "epoch": 0.2694926422139249, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.3702898025512695, + "learning_rate": 8.978770131771595e-07, + "loss": 0.9248, + "mean_token_accuracy": 0.7176526188850403, + "num_tokens": 63271257.0, + "step": 2454 + }, + { + "epoch": 0.26960245991653853, + "ewc_loss": 9.275972843170166e-07, + "grad_norm": 2.4765422344207764, + "learning_rate": 8.982430453879941e-07, + "loss": 1.0219, + "mean_token_accuracy": 0.6880168914794922, + "num_tokens": 63293129.0, + "step": 2455 + }, + { + "epoch": 0.26971227761915223, + "ewc_loss": 9.313225746154785e-07, + "grad_norm": 2.518122673034668, + "learning_rate": 8.986090775988286e-07, + "loss": 1.0055, + "mean_token_accuracy": 0.698010265827179, + "num_tokens": 63315767.0, + "step": 2456 + }, + { + "epoch": 0.2698220953217659, + "ewc_loss": 9.313225746154785e-07, + "grad_norm": 2.127065658569336, + "learning_rate": 8.989751098096631e-07, + "loss": 1.0274, + "mean_token_accuracy": 0.6962459087371826, + "num_tokens": 63345045.0, + "step": 2457 + }, + { + "epoch": 0.2699319130243795, + "ewc_loss": 9.387731552124023e-07, + "grad_norm": 2.348280668258667, + "learning_rate": 8.993411420204977e-07, + "loss": 0.9952, + "mean_token_accuracy": 0.6971726417541504, + "num_tokens": 63370935.0, + "step": 2458 + }, + { + "epoch": 0.27004173072699317, + "ewc_loss": 9.387731552124023e-07, + "grad_norm": 2.3392527103424072, + "learning_rate": 8.997071742313323e-07, + "loss": 1.0197, + "mean_token_accuracy": 0.6882064938545227, + "num_tokens": 63395023.0, + "step": 2459 + }, + { + "epoch": 0.27015154842960687, + "ewc_loss": 9.387731552124023e-07, + "grad_norm": 1.9765123128890991, + "learning_rate": 9.000732064421668e-07, + "loss": 1.0996, + "mean_token_accuracy": 0.6745239496231079, + "num_tokens": 63429774.0, + "step": 2460 + }, + { + "epoch": 0.2702613661322205, + "ewc_loss": 9.424984455108643e-07, + "grad_norm": 2.200136184692383, + "learning_rate": 9.004392386530014e-07, + "loss": 0.9564, + "mean_token_accuracy": 0.7068440318107605, + "num_tokens": 63454711.0, + "step": 2461 + }, + { + "epoch": 0.27037118383483416, + "ewc_loss": 9.424984455108643e-07, + "grad_norm": 2.2406580448150635, + "learning_rate": 9.008052708638359e-07, + "loss": 0.9864, + "mean_token_accuracy": 0.6990323662757874, + "num_tokens": 63481256.0, + "step": 2462 + }, + { + "epoch": 0.27048100153744786, + "ewc_loss": 9.424984455108643e-07, + "grad_norm": 2.382497787475586, + "learning_rate": 9.011713030746705e-07, + "loss": 1.0558, + "mean_token_accuracy": 0.6845338344573975, + "num_tokens": 63507253.0, + "step": 2463 + }, + { + "epoch": 0.2705908192400615, + "ewc_loss": 9.424984455108643e-07, + "grad_norm": 2.314277410507202, + "learning_rate": 9.015373352855051e-07, + "loss": 1.0018, + "mean_token_accuracy": 0.689064621925354, + "num_tokens": 63532003.0, + "step": 2464 + }, + { + "epoch": 0.27070063694267515, + "ewc_loss": 9.424984455108643e-07, + "grad_norm": 2.461638927459717, + "learning_rate": 9.019033674963396e-07, + "loss": 0.9901, + "mean_token_accuracy": 0.6978940963745117, + "num_tokens": 63554003.0, + "step": 2465 + }, + { + "epoch": 0.2708104546452888, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 2.5258305072784424, + "learning_rate": 9.022693997071742e-07, + "loss": 0.8856, + "mean_token_accuracy": 0.7269500494003296, + "num_tokens": 63574058.0, + "step": 2466 + }, + { + "epoch": 0.2709202723479025, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 2.4645915031433105, + "learning_rate": 9.026354319180087e-07, + "loss": 0.962, + "mean_token_accuracy": 0.7046645879745483, + "num_tokens": 63596234.0, + "step": 2467 + }, + { + "epoch": 0.27103009005051615, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 2.0230801105499268, + "learning_rate": 9.030014641288432e-07, + "loss": 1.0566, + "mean_token_accuracy": 0.6850759983062744, + "num_tokens": 63627837.0, + "step": 2468 + }, + { + "epoch": 0.2711399077531298, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 2.0620839595794678, + "learning_rate": 9.033674963396779e-07, + "loss": 1.1236, + "mean_token_accuracy": 0.6672687530517578, + "num_tokens": 63661751.0, + "step": 2469 + }, + { + "epoch": 0.27124972545574344, + "ewc_loss": 9.462237358093262e-07, + "grad_norm": 2.3497676849365234, + "learning_rate": 9.037335285505124e-07, + "loss": 1.0436, + "mean_token_accuracy": 0.6897874474525452, + "num_tokens": 63686230.0, + "step": 2470 + }, + { + "epoch": 0.27135954315835714, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.382075786590576, + "learning_rate": 9.040995607613469e-07, + "loss": 1.0935, + "mean_token_accuracy": 0.6725575923919678, + "num_tokens": 63710974.0, + "step": 2471 + }, + { + "epoch": 0.2714693608609708, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.079828977584839, + "learning_rate": 9.044655929721815e-07, + "loss": 1.0844, + "mean_token_accuracy": 0.6738629937171936, + "num_tokens": 63741057.0, + "step": 2472 + }, + { + "epoch": 0.27157917856358443, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.2969348430633545, + "learning_rate": 9.04831625183016e-07, + "loss": 1.086, + "mean_token_accuracy": 0.6774975061416626, + "num_tokens": 63766625.0, + "step": 2473 + }, + { + "epoch": 0.27168899626619814, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.154664993286133, + "learning_rate": 9.051976573938505e-07, + "loss": 1.0755, + "mean_token_accuracy": 0.682532548904419, + "num_tokens": 63792679.0, + "step": 2474 + }, + { + "epoch": 0.2717988139688118, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.3215203285217285, + "learning_rate": 9.055636896046852e-07, + "loss": 1.0351, + "mean_token_accuracy": 0.6904095411300659, + "num_tokens": 63818259.0, + "step": 2475 + }, + { + "epoch": 0.2719086316714254, + "ewc_loss": 9.499490261077881e-07, + "grad_norm": 2.397031784057617, + "learning_rate": 9.059297218155197e-07, + "loss": 0.9706, + "mean_token_accuracy": 0.7028827667236328, + "num_tokens": 63839630.0, + "step": 2476 + }, + { + "epoch": 0.2720184493740391, + "ewc_loss": 9.5367431640625e-07, + "grad_norm": 2.0692148208618164, + "learning_rate": 9.062957540263542e-07, + "loss": 0.9806, + "mean_token_accuracy": 0.7030467987060547, + "num_tokens": 63868019.0, + "step": 2477 + }, + { + "epoch": 0.2721282670766528, + "ewc_loss": 9.5367431640625e-07, + "grad_norm": 2.0308499336242676, + "learning_rate": 9.066617862371888e-07, + "loss": 1.016, + "mean_token_accuracy": 0.7021428942680359, + "num_tokens": 63896304.0, + "step": 2478 + }, + { + "epoch": 0.2722380847792664, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.0583767890930176, + "learning_rate": 9.070278184480233e-07, + "loss": 0.9489, + "mean_token_accuracy": 0.7084751725196838, + "num_tokens": 63925436.0, + "step": 2479 + }, + { + "epoch": 0.27234790248188007, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.287651300430298, + "learning_rate": 9.07393850658858e-07, + "loss": 0.998, + "mean_token_accuracy": 0.6999735832214355, + "num_tokens": 63949487.0, + "step": 2480 + }, + { + "epoch": 0.27245772018449377, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.260589361190796, + "learning_rate": 9.077598828696925e-07, + "loss": 0.9266, + "mean_token_accuracy": 0.7218275666236877, + "num_tokens": 63972779.0, + "step": 2481 + }, + { + "epoch": 0.2725675378871074, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.888141393661499, + "learning_rate": 9.08125915080527e-07, + "loss": 0.9918, + "mean_token_accuracy": 0.7007204294204712, + "num_tokens": 63990229.0, + "step": 2482 + }, + { + "epoch": 0.27267735558972106, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.4176485538482666, + "learning_rate": 9.084919472913616e-07, + "loss": 1.0857, + "mean_token_accuracy": 0.6742740869522095, + "num_tokens": 64014964.0, + "step": 2483 + }, + { + "epoch": 0.2727871732923347, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.1934096813201904, + "learning_rate": 9.088579795021961e-07, + "loss": 0.9939, + "mean_token_accuracy": 0.6982271075248718, + "num_tokens": 64041383.0, + "step": 2484 + }, + { + "epoch": 0.2728969909949484, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.243675470352173, + "learning_rate": 9.092240117130307e-07, + "loss": 1.0586, + "mean_token_accuracy": 0.6804572343826294, + "num_tokens": 64070705.0, + "step": 2485 + }, + { + "epoch": 0.27300680869756205, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 2.111191511154175, + "learning_rate": 9.095900439238653e-07, + "loss": 0.9205, + "mean_token_accuracy": 0.714751660823822, + "num_tokens": 64098527.0, + "step": 2486 + }, + { + "epoch": 0.2731166264001757, + "ewc_loss": 9.611248970031738e-07, + "grad_norm": 1.9754902124404907, + "learning_rate": 9.099560761346998e-07, + "loss": 0.9739, + "mean_token_accuracy": 0.7016558051109314, + "num_tokens": 64129915.0, + "step": 2487 + }, + { + "epoch": 0.27322644410278935, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.385342597961426, + "learning_rate": 9.103221083455343e-07, + "loss": 1.03, + "mean_token_accuracy": 0.6915899515151978, + "num_tokens": 64154983.0, + "step": 2488 + }, + { + "epoch": 0.27333626180540305, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.054940700531006, + "learning_rate": 9.106881405563689e-07, + "loss": 1.1402, + "mean_token_accuracy": 0.6735585927963257, + "num_tokens": 64186278.0, + "step": 2489 + }, + { + "epoch": 0.2734460795080167, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.3479905128479004, + "learning_rate": 9.110541727672034e-07, + "loss": 1.0401, + "mean_token_accuracy": 0.6889550089836121, + "num_tokens": 64211406.0, + "step": 2490 + }, + { + "epoch": 0.27355589721063034, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.526282548904419, + "learning_rate": 9.11420204978038e-07, + "loss": 0.9514, + "mean_token_accuracy": 0.7103928923606873, + "num_tokens": 64236208.0, + "step": 2491 + }, + { + "epoch": 0.27366571491324404, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 1.9997447729110718, + "learning_rate": 9.117862371888726e-07, + "loss": 1.1191, + "mean_token_accuracy": 0.6681703329086304, + "num_tokens": 64269029.0, + "step": 2492 + }, + { + "epoch": 0.2737755326158577, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.2505626678466797, + "learning_rate": 9.121522693997071e-07, + "loss": 1.0826, + "mean_token_accuracy": 0.6707780957221985, + "num_tokens": 64295333.0, + "step": 2493 + }, + { + "epoch": 0.27388535031847133, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.1395654678344727, + "learning_rate": 9.125183016105416e-07, + "loss": 1.0418, + "mean_token_accuracy": 0.6858164072036743, + "num_tokens": 64325395.0, + "step": 2494 + }, + { + "epoch": 0.273995168021085, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.1774189472198486, + "learning_rate": 9.128843338213762e-07, + "loss": 1.031, + "mean_token_accuracy": 0.6855742335319519, + "num_tokens": 64353858.0, + "step": 2495 + }, + { + "epoch": 0.2741049857236987, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.3352150917053223, + "learning_rate": 9.132503660322108e-07, + "loss": 1.0546, + "mean_token_accuracy": 0.6813549399375916, + "num_tokens": 64378320.0, + "step": 2496 + }, + { + "epoch": 0.2742148034263123, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.357971668243408, + "learning_rate": 9.136163982430454e-07, + "loss": 0.9569, + "mean_token_accuracy": 0.7099453210830688, + "num_tokens": 64402520.0, + "step": 2497 + }, + { + "epoch": 0.27432462112892597, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.3043735027313232, + "learning_rate": 9.139824304538799e-07, + "loss": 0.9075, + "mean_token_accuracy": 0.7220866680145264, + "num_tokens": 64424227.0, + "step": 2498 + }, + { + "epoch": 0.2744344388315396, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.320030927658081, + "learning_rate": 9.143484626647144e-07, + "loss": 1.1136, + "mean_token_accuracy": 0.6746435165405273, + "num_tokens": 64449016.0, + "step": 2499 + }, + { + "epoch": 0.2745442565341533, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.27927565574646, + "learning_rate": 9.14714494875549e-07, + "loss": 0.8578, + "mean_token_accuracy": 0.7320678234100342, + "num_tokens": 64471600.0, + "step": 2500 + }, + { + "epoch": 0.27465407423676697, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.129209518432617, + "learning_rate": 9.150805270863836e-07, + "loss": 1.0935, + "mean_token_accuracy": 0.6796953678131104, + "num_tokens": 64501456.0, + "step": 2501 + }, + { + "epoch": 0.2747638919393806, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.362264394760132, + "learning_rate": 9.154465592972181e-07, + "loss": 1.0078, + "mean_token_accuracy": 0.6995470523834229, + "num_tokens": 64525896.0, + "step": 2502 + }, + { + "epoch": 0.2748737096419943, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.1185388565063477, + "learning_rate": 9.158125915080527e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6872254610061646, + "num_tokens": 64556903.0, + "step": 2503 + }, + { + "epoch": 0.27498352734460796, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.061845064163208, + "learning_rate": 9.161786237188872e-07, + "loss": 1.0569, + "mean_token_accuracy": 0.6806642413139343, + "num_tokens": 64586040.0, + "step": 2504 + }, + { + "epoch": 0.2750933450472216, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.202988862991333, + "learning_rate": 9.165446559297217e-07, + "loss": 1.0761, + "mean_token_accuracy": 0.6833640933036804, + "num_tokens": 64611675.0, + "step": 2505 + }, + { + "epoch": 0.27520316274983525, + "ewc_loss": 9.685754776000977e-07, + "grad_norm": 2.1530048847198486, + "learning_rate": 9.169106881405563e-07, + "loss": 0.9444, + "mean_token_accuracy": 0.7160211801528931, + "num_tokens": 64637574.0, + "step": 2506 + }, + { + "epoch": 0.27531298045244895, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.4047958850860596, + "learning_rate": 9.172767203513909e-07, + "loss": 0.8862, + "mean_token_accuracy": 0.7258065938949585, + "num_tokens": 64659812.0, + "step": 2507 + }, + { + "epoch": 0.2754227981550626, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.486893653869629, + "learning_rate": 9.176427525622254e-07, + "loss": 1.0892, + "mean_token_accuracy": 0.6784379482269287, + "num_tokens": 64683019.0, + "step": 2508 + }, + { + "epoch": 0.27553261585767624, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.3416202068328857, + "learning_rate": 9.1800878477306e-07, + "loss": 1.0298, + "mean_token_accuracy": 0.6960766315460205, + "num_tokens": 64708468.0, + "step": 2509 + }, + { + "epoch": 0.27564243356028995, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.0086474418640137, + "learning_rate": 9.183748169838945e-07, + "loss": 1.0493, + "mean_token_accuracy": 0.6797482967376709, + "num_tokens": 64740161.0, + "step": 2510 + }, + { + "epoch": 0.2757522512629036, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.5715975761413574, + "learning_rate": 9.18740849194729e-07, + "loss": 0.966, + "mean_token_accuracy": 0.7067722082138062, + "num_tokens": 64759273.0, + "step": 2511 + }, + { + "epoch": 0.27586206896551724, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.0090396404266357, + "learning_rate": 9.191068814055637e-07, + "loss": 0.9925, + "mean_token_accuracy": 0.6987178921699524, + "num_tokens": 64788926.0, + "step": 2512 + }, + { + "epoch": 0.2759718866681309, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.2100744247436523, + "learning_rate": 9.194729136163982e-07, + "loss": 1.0487, + "mean_token_accuracy": 0.6885888576507568, + "num_tokens": 64817940.0, + "step": 2513 + }, + { + "epoch": 0.2760817043707446, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 1.9142385721206665, + "learning_rate": 9.198389458272328e-07, + "loss": 1.1318, + "mean_token_accuracy": 0.679951012134552, + "num_tokens": 64855373.0, + "step": 2514 + }, + { + "epoch": 0.27619152207335823, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.130811929702759, + "learning_rate": 9.202049780380673e-07, + "loss": 1.0039, + "mean_token_accuracy": 0.6931898593902588, + "num_tokens": 64881957.0, + "step": 2515 + }, + { + "epoch": 0.2763013397759719, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.1524975299835205, + "learning_rate": 9.205710102489018e-07, + "loss": 0.9527, + "mean_token_accuracy": 0.7077026963233948, + "num_tokens": 64908585.0, + "step": 2516 + }, + { + "epoch": 0.2764111574785855, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.5577995777130127, + "learning_rate": 9.209370424597365e-07, + "loss": 0.9634, + "mean_token_accuracy": 0.7061382532119751, + "num_tokens": 64930242.0, + "step": 2517 + }, + { + "epoch": 0.2765209751811992, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.032625198364258, + "learning_rate": 9.21303074670571e-07, + "loss": 0.9798, + "mean_token_accuracy": 0.7017345428466797, + "num_tokens": 64959196.0, + "step": 2518 + }, + { + "epoch": 0.27663079288381287, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.1573967933654785, + "learning_rate": 9.216691068814055e-07, + "loss": 1.0115, + "mean_token_accuracy": 0.6909914612770081, + "num_tokens": 64987422.0, + "step": 2519 + }, + { + "epoch": 0.2767406105864265, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.161560297012329, + "learning_rate": 9.220351390922401e-07, + "loss": 1.09, + "mean_token_accuracy": 0.682878315448761, + "num_tokens": 65015218.0, + "step": 2520 + }, + { + "epoch": 0.2768504282890402, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.2122225761413574, + "learning_rate": 9.224011713030746e-07, + "loss": 1.0152, + "mean_token_accuracy": 0.6982520818710327, + "num_tokens": 65042379.0, + "step": 2521 + }, + { + "epoch": 0.27696024599165386, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.202345132827759, + "learning_rate": 9.227672035139091e-07, + "loss": 1.0405, + "mean_token_accuracy": 0.6865382194519043, + "num_tokens": 65069290.0, + "step": 2522 + }, + { + "epoch": 0.2770700636942675, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.2097904682159424, + "learning_rate": 9.231332357247438e-07, + "loss": 0.9789, + "mean_token_accuracy": 0.7000836133956909, + "num_tokens": 65094108.0, + "step": 2523 + }, + { + "epoch": 0.27717988139688116, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.1039865016937256, + "learning_rate": 9.234992679355783e-07, + "loss": 0.9712, + "mean_token_accuracy": 0.7072117328643799, + "num_tokens": 65122212.0, + "step": 2524 + }, + { + "epoch": 0.27728969909949486, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.0709621906280518, + "learning_rate": 9.238653001464128e-07, + "loss": 1.068, + "mean_token_accuracy": 0.684525728225708, + "num_tokens": 65153331.0, + "step": 2525 + }, + { + "epoch": 0.2773995168021085, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.265510320663452, + "learning_rate": 9.242313323572474e-07, + "loss": 0.9835, + "mean_token_accuracy": 0.706411600112915, + "num_tokens": 65178441.0, + "step": 2526 + }, + { + "epoch": 0.27750933450472215, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.057466506958008, + "learning_rate": 9.245973645680819e-07, + "loss": 0.9484, + "mean_token_accuracy": 0.708430826663971, + "num_tokens": 65207757.0, + "step": 2527 + }, + { + "epoch": 0.2776191522073358, + "ewc_loss": 9.760260581970215e-07, + "grad_norm": 2.246253728866577, + "learning_rate": 9.249633967789166e-07, + "loss": 1.0632, + "mean_token_accuracy": 0.688258171081543, + "num_tokens": 65236719.0, + "step": 2528 + }, + { + "epoch": 0.2777289699099495, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.209807872772217, + "learning_rate": 9.253294289897511e-07, + "loss": 0.9304, + "mean_token_accuracy": 0.7152511477470398, + "num_tokens": 65260911.0, + "step": 2529 + }, + { + "epoch": 0.27783878761256314, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.5903193950653076, + "learning_rate": 9.256954612005856e-07, + "loss": 0.9501, + "mean_token_accuracy": 0.7084629535675049, + "num_tokens": 65281523.0, + "step": 2530 + }, + { + "epoch": 0.2779486053151768, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.5469181537628174, + "learning_rate": 9.260614934114202e-07, + "loss": 1.0186, + "mean_token_accuracy": 0.6897922158241272, + "num_tokens": 65304936.0, + "step": 2531 + }, + { + "epoch": 0.2780584230177905, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.1805129051208496, + "learning_rate": 9.264275256222547e-07, + "loss": 0.8909, + "mean_token_accuracy": 0.7238150835037231, + "num_tokens": 65329771.0, + "step": 2532 + }, + { + "epoch": 0.27816824072040414, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.4388744831085205, + "learning_rate": 9.267935578330893e-07, + "loss": 1.1031, + "mean_token_accuracy": 0.6757985353469849, + "num_tokens": 65352739.0, + "step": 2533 + }, + { + "epoch": 0.2782780584230178, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.1098086833953857, + "learning_rate": 9.271595900439239e-07, + "loss": 1.0487, + "mean_token_accuracy": 0.6981881260871887, + "num_tokens": 65381469.0, + "step": 2534 + }, + { + "epoch": 0.27838787612563143, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.4882588386535645, + "learning_rate": 9.275256222547584e-07, + "loss": 0.9499, + "mean_token_accuracy": 0.7109100818634033, + "num_tokens": 65402894.0, + "step": 2535 + }, + { + "epoch": 0.27849769382824513, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.121279716491699, + "learning_rate": 9.278916544655929e-07, + "loss": 1.0885, + "mean_token_accuracy": 0.6838575005531311, + "num_tokens": 65433634.0, + "step": 2536 + }, + { + "epoch": 0.2786075115308588, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.7894394397735596, + "learning_rate": 9.282576866764275e-07, + "loss": 0.8893, + "mean_token_accuracy": 0.7234354019165039, + "num_tokens": 65452164.0, + "step": 2537 + }, + { + "epoch": 0.2787173292334724, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.199744701385498, + "learning_rate": 9.28623718887262e-07, + "loss": 1.0731, + "mean_token_accuracy": 0.6880776286125183, + "num_tokens": 65478967.0, + "step": 2538 + }, + { + "epoch": 0.2788271469360861, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.623513698577881, + "learning_rate": 9.289897510980966e-07, + "loss": 0.9457, + "mean_token_accuracy": 0.7122887969017029, + "num_tokens": 65499590.0, + "step": 2539 + }, + { + "epoch": 0.27893696463869977, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.349973440170288, + "learning_rate": 9.293557833089312e-07, + "loss": 0.9594, + "mean_token_accuracy": 0.7067301273345947, + "num_tokens": 65525619.0, + "step": 2540 + }, + { + "epoch": 0.2790467823413134, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.460005521774292, + "learning_rate": 9.297218155197657e-07, + "loss": 0.9713, + "mean_token_accuracy": 0.70538330078125, + "num_tokens": 65547328.0, + "step": 2541 + }, + { + "epoch": 0.27915660004392706, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.222975254058838, + "learning_rate": 9.300878477306002e-07, + "loss": 1.0789, + "mean_token_accuracy": 0.6787455081939697, + "num_tokens": 65574613.0, + "step": 2542 + }, + { + "epoch": 0.27926641774654076, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.118726968765259, + "learning_rate": 9.304538799414348e-07, + "loss": 0.9607, + "mean_token_accuracy": 0.7032681107521057, + "num_tokens": 65603136.0, + "step": 2543 + }, + { + "epoch": 0.2793762354491544, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.3274624347686768, + "learning_rate": 9.308199121522694e-07, + "loss": 1.0, + "mean_token_accuracy": 0.6975988149642944, + "num_tokens": 65628228.0, + "step": 2544 + }, + { + "epoch": 0.27948605315176805, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.299837827682495, + "learning_rate": 9.31185944363104e-07, + "loss": 1.0292, + "mean_token_accuracy": 0.6882340312004089, + "num_tokens": 65655181.0, + "step": 2545 + }, + { + "epoch": 0.2795958708543817, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.0584824085235596, + "learning_rate": 9.315519765739385e-07, + "loss": 1.1155, + "mean_token_accuracy": 0.6734782457351685, + "num_tokens": 65684935.0, + "step": 2546 + }, + { + "epoch": 0.2797056885569954, + "ewc_loss": 9.834766387939453e-07, + "grad_norm": 2.427539110183716, + "learning_rate": 9.31918008784773e-07, + "loss": 0.9993, + "mean_token_accuracy": 0.6944599151611328, + "num_tokens": 65708943.0, + "step": 2547 + }, + { + "epoch": 0.27981550625960905, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.147676944732666, + "learning_rate": 9.322840409956076e-07, + "loss": 0.9864, + "mean_token_accuracy": 0.7094884514808655, + "num_tokens": 65737123.0, + "step": 2548 + }, + { + "epoch": 0.2799253239622227, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.4097530841827393, + "learning_rate": 9.326500732064422e-07, + "loss": 1.0342, + "mean_token_accuracy": 0.697170615196228, + "num_tokens": 65760927.0, + "step": 2549 + }, + { + "epoch": 0.2800351416648364, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.1682002544403076, + "learning_rate": 9.330161054172767e-07, + "loss": 1.0291, + "mean_token_accuracy": 0.6898674964904785, + "num_tokens": 65788313.0, + "step": 2550 + }, + { + "epoch": 0.28014495936745004, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.221975803375244, + "learning_rate": 9.333821376281113e-07, + "loss": 1.0174, + "mean_token_accuracy": 0.6920077800750732, + "num_tokens": 65812857.0, + "step": 2551 + }, + { + "epoch": 0.2802547770700637, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.301445484161377, + "learning_rate": 9.337481698389458e-07, + "loss": 0.9769, + "mean_token_accuracy": 0.7080467343330383, + "num_tokens": 65836955.0, + "step": 2552 + }, + { + "epoch": 0.28036459477267733, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.170053720474243, + "learning_rate": 9.341142020497803e-07, + "loss": 1.007, + "mean_token_accuracy": 0.6972931027412415, + "num_tokens": 65863603.0, + "step": 2553 + }, + { + "epoch": 0.28047441247529104, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.307508707046509, + "learning_rate": 9.344802342606149e-07, + "loss": 1.0374, + "mean_token_accuracy": 0.6932936310768127, + "num_tokens": 65890120.0, + "step": 2554 + }, + { + "epoch": 0.2805842301779047, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 1.9375488758087158, + "learning_rate": 9.348462664714495e-07, + "loss": 1.0326, + "mean_token_accuracy": 0.6961002945899963, + "num_tokens": 65926236.0, + "step": 2555 + }, + { + "epoch": 0.2806940478805183, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.464931011199951, + "learning_rate": 9.35212298682284e-07, + "loss": 0.95, + "mean_token_accuracy": 0.7102112174034119, + "num_tokens": 65947643.0, + "step": 2556 + }, + { + "epoch": 0.28080386558313203, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.456763982772827, + "learning_rate": 9.355783308931186e-07, + "loss": 0.9753, + "mean_token_accuracy": 0.7118079662322998, + "num_tokens": 65970667.0, + "step": 2557 + }, + { + "epoch": 0.2809136832857457, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.277244806289673, + "learning_rate": 9.359443631039531e-07, + "loss": 1.0523, + "mean_token_accuracy": 0.6865025758743286, + "num_tokens": 65996991.0, + "step": 2558 + }, + { + "epoch": 0.2810235009883593, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.0966176986694336, + "learning_rate": 9.363103953147876e-07, + "loss": 1.0171, + "mean_token_accuracy": 0.6925055980682373, + "num_tokens": 66027766.0, + "step": 2559 + }, + { + "epoch": 0.28113331869097297, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.647862672805786, + "learning_rate": 9.366764275256223e-07, + "loss": 0.9187, + "mean_token_accuracy": 0.714387059211731, + "num_tokens": 66046417.0, + "step": 2560 + }, + { + "epoch": 0.28124313639358667, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.1410257816314697, + "learning_rate": 9.370424597364568e-07, + "loss": 0.9643, + "mean_token_accuracy": 0.7066755294799805, + "num_tokens": 66073856.0, + "step": 2561 + }, + { + "epoch": 0.2813529540962003, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.2410242557525635, + "learning_rate": 9.374084919472914e-07, + "loss": 1.0053, + "mean_token_accuracy": 0.6988779306411743, + "num_tokens": 66101429.0, + "step": 2562 + }, + { + "epoch": 0.28146277179881396, + "ewc_loss": 9.909272193908691e-07, + "grad_norm": 2.210664987564087, + "learning_rate": 9.377745241581259e-07, + "loss": 1.0979, + "mean_token_accuracy": 0.6726076602935791, + "num_tokens": 66129613.0, + "step": 2563 + }, + { + "epoch": 0.2815725895014276, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.136277675628662, + "learning_rate": 9.381405563689604e-07, + "loss": 1.023, + "mean_token_accuracy": 0.6901999711990356, + "num_tokens": 66159666.0, + "step": 2564 + }, + { + "epoch": 0.2816824072040413, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.664320230484009, + "learning_rate": 9.385065885797951e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.7009594440460205, + "num_tokens": 66181171.0, + "step": 2565 + }, + { + "epoch": 0.28179222490665495, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.246600866317749, + "learning_rate": 9.388726207906296e-07, + "loss": 0.9317, + "mean_token_accuracy": 0.7204675078392029, + "num_tokens": 66205013.0, + "step": 2566 + }, + { + "epoch": 0.2819020426092686, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.2059848308563232, + "learning_rate": 9.392386530014641e-07, + "loss": 1.0009, + "mean_token_accuracy": 0.70255446434021, + "num_tokens": 66230664.0, + "step": 2567 + }, + { + "epoch": 0.2820118603118823, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.031280517578125, + "learning_rate": 9.396046852122987e-07, + "loss": 0.9996, + "mean_token_accuracy": 0.6995728015899658, + "num_tokens": 66260812.0, + "step": 2568 + }, + { + "epoch": 0.28212167801449595, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.4277071952819824, + "learning_rate": 9.399707174231332e-07, + "loss": 1.0274, + "mean_token_accuracy": 0.6862865090370178, + "num_tokens": 66282638.0, + "step": 2569 + }, + { + "epoch": 0.2822314957171096, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 1.941545009613037, + "learning_rate": 9.403367496339677e-07, + "loss": 1.0551, + "mean_token_accuracy": 0.6875970363616943, + "num_tokens": 66315689.0, + "step": 2570 + }, + { + "epoch": 0.28234131341972324, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.0596437454223633, + "learning_rate": 9.407027818448024e-07, + "loss": 0.9669, + "mean_token_accuracy": 0.7075703144073486, + "num_tokens": 66343772.0, + "step": 2571 + }, + { + "epoch": 0.28245113112233694, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.260118007659912, + "learning_rate": 9.410688140556369e-07, + "loss": 0.8995, + "mean_token_accuracy": 0.7200985550880432, + "num_tokens": 66366792.0, + "step": 2572 + }, + { + "epoch": 0.2825609488249506, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.3846099376678467, + "learning_rate": 9.414348462664714e-07, + "loss": 0.9675, + "mean_token_accuracy": 0.7037813067436218, + "num_tokens": 66389908.0, + "step": 2573 + }, + { + "epoch": 0.28267076652756423, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 1.972403883934021, + "learning_rate": 9.41800878477306e-07, + "loss": 1.0527, + "mean_token_accuracy": 0.6878178119659424, + "num_tokens": 66423245.0, + "step": 2574 + }, + { + "epoch": 0.2827805842301779, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.2638463973999023, + "learning_rate": 9.421669106881405e-07, + "loss": 1.0815, + "mean_token_accuracy": 0.6745361089706421, + "num_tokens": 66453024.0, + "step": 2575 + }, + { + "epoch": 0.2828904019327916, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.1977155208587646, + "learning_rate": 9.425329428989752e-07, + "loss": 1.0158, + "mean_token_accuracy": 0.6961653232574463, + "num_tokens": 66479736.0, + "step": 2576 + }, + { + "epoch": 0.2830002196354052, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.161379814147949, + "learning_rate": 9.428989751098097e-07, + "loss": 1.0591, + "mean_token_accuracy": 0.6798284649848938, + "num_tokens": 66509498.0, + "step": 2577 + }, + { + "epoch": 0.28311003733801887, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.3268203735351562, + "learning_rate": 9.432650073206442e-07, + "loss": 1.0121, + "mean_token_accuracy": 0.6968196630477905, + "num_tokens": 66534943.0, + "step": 2578 + }, + { + "epoch": 0.2832198550406326, + "ewc_loss": 9.98377799987793e-07, + "grad_norm": 2.139615058898926, + "learning_rate": 9.436310395314788e-07, + "loss": 0.9961, + "mean_token_accuracy": 0.6979909539222717, + "num_tokens": 66561088.0, + "step": 2579 + }, + { + "epoch": 0.2833296727432462, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 2.3226211071014404, + "learning_rate": 9.439970717423133e-07, + "loss": 0.8794, + "mean_token_accuracy": 0.7306951284408569, + "num_tokens": 66584705.0, + "step": 2580 + }, + { + "epoch": 0.28343949044585987, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 2.0125415325164795, + "learning_rate": 9.443631039531479e-07, + "loss": 1.0605, + "mean_token_accuracy": 0.681972861289978, + "num_tokens": 66617156.0, + "step": 2581 + }, + { + "epoch": 0.2835493081484735, + "ewc_loss": 1.0058283805847168e-06, + "grad_norm": 2.3153798580169678, + "learning_rate": 9.447291361639825e-07, + "loss": 1.0803, + "mean_token_accuracy": 0.6730157136917114, + "num_tokens": 66644407.0, + "step": 2582 + }, + { + "epoch": 0.2836591258510872, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.113649368286133, + "learning_rate": 9.45095168374817e-07, + "loss": 1.1095, + "mean_token_accuracy": 0.6769934892654419, + "num_tokens": 66680089.0, + "step": 2583 + }, + { + "epoch": 0.28376894355370086, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 1.853410005569458, + "learning_rate": 9.454612005856515e-07, + "loss": 1.0216, + "mean_token_accuracy": 0.6918900012969971, + "num_tokens": 66716109.0, + "step": 2584 + }, + { + "epoch": 0.2838787612563145, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.4148850440979004, + "learning_rate": 9.458272327964861e-07, + "loss": 1.0456, + "mean_token_accuracy": 0.6922370195388794, + "num_tokens": 66740383.0, + "step": 2585 + }, + { + "epoch": 0.2839885789589282, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 1.9425861835479736, + "learning_rate": 9.461932650073206e-07, + "loss": 1.0432, + "mean_token_accuracy": 0.6945605874061584, + "num_tokens": 66774475.0, + "step": 2586 + }, + { + "epoch": 0.28409839666154185, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.0579404830932617, + "learning_rate": 9.465592972181552e-07, + "loss": 0.9334, + "mean_token_accuracy": 0.7169609665870667, + "num_tokens": 66802774.0, + "step": 2587 + }, + { + "epoch": 0.2842082143641555, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.6452090740203857, + "learning_rate": 9.469253294289898e-07, + "loss": 0.9772, + "mean_token_accuracy": 0.6977560520172119, + "num_tokens": 66822965.0, + "step": 2588 + }, + { + "epoch": 0.28431803206676914, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.1371049880981445, + "learning_rate": 9.472913616398243e-07, + "loss": 1.0466, + "mean_token_accuracy": 0.690489649772644, + "num_tokens": 66852734.0, + "step": 2589 + }, + { + "epoch": 0.28442784976938285, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.287914752960205, + "learning_rate": 9.476573938506587e-07, + "loss": 0.9323, + "mean_token_accuracy": 0.7187459468841553, + "num_tokens": 66877006.0, + "step": 2590 + }, + { + "epoch": 0.2845376674719965, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.3290672302246094, + "learning_rate": 9.480234260614933e-07, + "loss": 1.0557, + "mean_token_accuracy": 0.6888517141342163, + "num_tokens": 66902300.0, + "step": 2591 + }, + { + "epoch": 0.28464748517461014, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.114229202270508, + "learning_rate": 9.48389458272328e-07, + "loss": 0.9276, + "mean_token_accuracy": 0.7194996476173401, + "num_tokens": 66930991.0, + "step": 2592 + }, + { + "epoch": 0.2847573028772238, + "ewc_loss": 1.0132789611816406e-06, + "grad_norm": 2.251889705657959, + "learning_rate": 9.487554904831626e-07, + "loss": 1.0195, + "mean_token_accuracy": 0.697418212890625, + "num_tokens": 66957004.0, + "step": 2593 + }, + { + "epoch": 0.2848671205798375, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.512632131576538, + "learning_rate": 9.49121522693997e-07, + "loss": 1.0261, + "mean_token_accuracy": 0.7024351358413696, + "num_tokens": 66980931.0, + "step": 2594 + }, + { + "epoch": 0.28497693828245113, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.346083879470825, + "learning_rate": 9.494875549048315e-07, + "loss": 0.9832, + "mean_token_accuracy": 0.708209753036499, + "num_tokens": 67005148.0, + "step": 2595 + }, + { + "epoch": 0.2850867559850648, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.6041042804718018, + "learning_rate": 9.498535871156661e-07, + "loss": 1.0135, + "mean_token_accuracy": 0.6829590797424316, + "num_tokens": 67024780.0, + "step": 2596 + }, + { + "epoch": 0.2851965736876785, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.238016366958618, + "learning_rate": 9.502196193265007e-07, + "loss": 1.0349, + "mean_token_accuracy": 0.6936049461364746, + "num_tokens": 67051088.0, + "step": 2597 + }, + { + "epoch": 0.2853063913902921, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.557631731033325, + "learning_rate": 9.505856515373352e-07, + "loss": 1.0029, + "mean_token_accuracy": 0.7019079923629761, + "num_tokens": 67073148.0, + "step": 2598 + }, + { + "epoch": 0.28541620909290577, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.1630778312683105, + "learning_rate": 9.509516837481698e-07, + "loss": 1.0762, + "mean_token_accuracy": 0.6859018206596375, + "num_tokens": 67103289.0, + "step": 2599 + }, + { + "epoch": 0.2855260267955194, + "ewc_loss": 1.0207295417785645e-06, + "grad_norm": 2.1439924240112305, + "learning_rate": 9.513177159590043e-07, + "loss": 1.1288, + "mean_token_accuracy": 0.664668619632721, + "num_tokens": 67134940.0, + "step": 2600 + }, + { + "epoch": 0.2856358444981331, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.149962902069092, + "learning_rate": 9.516837481698388e-07, + "loss": 1.035, + "mean_token_accuracy": 0.6911154985427856, + "num_tokens": 67162861.0, + "step": 2601 + }, + { + "epoch": 0.28574566220074676, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.5111920833587646, + "learning_rate": 9.520497803806734e-07, + "loss": 0.9813, + "mean_token_accuracy": 0.7014998197555542, + "num_tokens": 67183291.0, + "step": 2602 + }, + { + "epoch": 0.2858554799033604, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.320974349975586, + "learning_rate": 9.52415812591508e-07, + "loss": 1.0173, + "mean_token_accuracy": 0.6941670775413513, + "num_tokens": 67208977.0, + "step": 2603 + }, + { + "epoch": 0.28596529760597406, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.304386854171753, + "learning_rate": 9.527818448023425e-07, + "loss": 0.9884, + "mean_token_accuracy": 0.7025833129882812, + "num_tokens": 67233401.0, + "step": 2604 + }, + { + "epoch": 0.28607511530858776, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.240940809249878, + "learning_rate": 9.531478770131771e-07, + "loss": 1.0357, + "mean_token_accuracy": 0.6983247399330139, + "num_tokens": 67258898.0, + "step": 2605 + }, + { + "epoch": 0.2861849330112014, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.0121054649353027, + "learning_rate": 9.535139092240116e-07, + "loss": 0.9927, + "mean_token_accuracy": 0.7022718787193298, + "num_tokens": 67288247.0, + "step": 2606 + }, + { + "epoch": 0.28629475071381505, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.267521858215332, + "learning_rate": 9.538799414348462e-07, + "loss": 1.0078, + "mean_token_accuracy": 0.696111798286438, + "num_tokens": 67312209.0, + "step": 2607 + }, + { + "epoch": 0.28640456841642875, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.233549118041992, + "learning_rate": 9.542459736456808e-07, + "loss": 1.0055, + "mean_token_accuracy": 0.7012632489204407, + "num_tokens": 67336959.0, + "step": 2608 + }, + { + "epoch": 0.2865143861190424, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.074258804321289, + "learning_rate": 9.546120058565154e-07, + "loss": 0.999, + "mean_token_accuracy": 0.6955048441886902, + "num_tokens": 67365192.0, + "step": 2609 + }, + { + "epoch": 0.28662420382165604, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.2402966022491455, + "learning_rate": 9.549780380673498e-07, + "loss": 0.9999, + "mean_token_accuracy": 0.6947858333587646, + "num_tokens": 67392262.0, + "step": 2610 + }, + { + "epoch": 0.2867340215242697, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.127249240875244, + "learning_rate": 9.553440702781844e-07, + "loss": 1.0043, + "mean_token_accuracy": 0.7019510269165039, + "num_tokens": 67419572.0, + "step": 2611 + }, + { + "epoch": 0.2868438392268834, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.4750845432281494, + "learning_rate": 9.55710102489019e-07, + "loss": 0.9456, + "mean_token_accuracy": 0.7178497314453125, + "num_tokens": 67440259.0, + "step": 2612 + }, + { + "epoch": 0.28695365692949704, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.431568145751953, + "learning_rate": 9.560761346998536e-07, + "loss": 0.9463, + "mean_token_accuracy": 0.7070683240890503, + "num_tokens": 67464103.0, + "step": 2613 + }, + { + "epoch": 0.2870634746321107, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 1.930794596672058, + "learning_rate": 9.56442166910688e-07, + "loss": 1.095, + "mean_token_accuracy": 0.6735770106315613, + "num_tokens": 67498304.0, + "step": 2614 + }, + { + "epoch": 0.2871732923347244, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.5264604091644287, + "learning_rate": 9.568081991215226e-07, + "loss": 1.0773, + "mean_token_accuracy": 0.6778865456581116, + "num_tokens": 67521821.0, + "step": 2615 + }, + { + "epoch": 0.28728311003733803, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.5928988456726074, + "learning_rate": 9.571742313323572e-07, + "loss": 0.9654, + "mean_token_accuracy": 0.708061933517456, + "num_tokens": 67541624.0, + "step": 2616 + }, + { + "epoch": 0.2873929277399517, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.389620542526245, + "learning_rate": 9.575402635431916e-07, + "loss": 0.9853, + "mean_token_accuracy": 0.7101089358329773, + "num_tokens": 67563863.0, + "step": 2617 + }, + { + "epoch": 0.2875027454425653, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.3514113426208496, + "learning_rate": 9.579062957540262e-07, + "loss": 1.0741, + "mean_token_accuracy": 0.675766110420227, + "num_tokens": 67588335.0, + "step": 2618 + }, + { + "epoch": 0.287612563145179, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.44118332862854, + "learning_rate": 9.582723279648608e-07, + "loss": 0.9332, + "mean_token_accuracy": 0.7118979096412659, + "num_tokens": 67610513.0, + "step": 2619 + }, + { + "epoch": 0.28772238084779267, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.07631778717041, + "learning_rate": 9.586383601756954e-07, + "loss": 1.1314, + "mean_token_accuracy": 0.6655588150024414, + "num_tokens": 67643857.0, + "step": 2620 + }, + { + "epoch": 0.2878321985504063, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.411073684692383, + "learning_rate": 9.5900439238653e-07, + "loss": 0.9807, + "mean_token_accuracy": 0.6983012557029724, + "num_tokens": 67665589.0, + "step": 2621 + }, + { + "epoch": 0.28794201625301996, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.029632568359375, + "learning_rate": 9.593704245973644e-07, + "loss": 1.017, + "mean_token_accuracy": 0.6967320442199707, + "num_tokens": 67695418.0, + "step": 2622 + }, + { + "epoch": 0.28805183395563366, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.446043014526367, + "learning_rate": 9.59736456808199e-07, + "loss": 0.9848, + "mean_token_accuracy": 0.6986489295959473, + "num_tokens": 67716503.0, + "step": 2623 + }, + { + "epoch": 0.2881616516582473, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.128350257873535, + "learning_rate": 9.601024890190336e-07, + "loss": 0.9928, + "mean_token_accuracy": 0.7043564915657043, + "num_tokens": 67743036.0, + "step": 2624 + }, + { + "epoch": 0.28827146936086095, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 1.9798619747161865, + "learning_rate": 9.604685212298682e-07, + "loss": 1.073, + "mean_token_accuracy": 0.680490255355835, + "num_tokens": 67775477.0, + "step": 2625 + }, + { + "epoch": 0.28838128706347466, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.594869613647461, + "learning_rate": 9.608345534407028e-07, + "loss": 1.0259, + "mean_token_accuracy": 0.6929917931556702, + "num_tokens": 67796246.0, + "step": 2626 + }, + { + "epoch": 0.2884911047660883, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.418630599975586, + "learning_rate": 9.612005856515372e-07, + "loss": 0.9831, + "mean_token_accuracy": 0.7089915871620178, + "num_tokens": 67820758.0, + "step": 2627 + }, + { + "epoch": 0.28860092246870195, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.311603307723999, + "learning_rate": 9.615666178623718e-07, + "loss": 1.0081, + "mean_token_accuracy": 0.6984657645225525, + "num_tokens": 67844105.0, + "step": 2628 + }, + { + "epoch": 0.2887107401713156, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.0268993377685547, + "learning_rate": 9.619326500732064e-07, + "loss": 1.0825, + "mean_token_accuracy": 0.6724879741668701, + "num_tokens": 67876536.0, + "step": 2629 + }, + { + "epoch": 0.2888205578739293, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.08427357673645, + "learning_rate": 9.62298682284041e-07, + "loss": 0.9765, + "mean_token_accuracy": 0.7088978290557861, + "num_tokens": 67906726.0, + "step": 2630 + }, + { + "epoch": 0.28893037557654294, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.1867988109588623, + "learning_rate": 9.626647144948754e-07, + "loss": 1.0275, + "mean_token_accuracy": 0.6819146275520325, + "num_tokens": 67934853.0, + "step": 2631 + }, + { + "epoch": 0.2890401932791566, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.2353696823120117, + "learning_rate": 9.6303074670571e-07, + "loss": 1.0169, + "mean_token_accuracy": 0.7035391330718994, + "num_tokens": 67960487.0, + "step": 2632 + }, + { + "epoch": 0.2891500109817703, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.232062339782715, + "learning_rate": 9.633967789165446e-07, + "loss": 0.9963, + "mean_token_accuracy": 0.6972788572311401, + "num_tokens": 67985099.0, + "step": 2633 + }, + { + "epoch": 0.28925982868438394, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.6495301723480225, + "learning_rate": 9.63762811127379e-07, + "loss": 1.1004, + "mean_token_accuracy": 0.6742867231369019, + "num_tokens": 68005907.0, + "step": 2634 + }, + { + "epoch": 0.2893696463869976, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.539590358734131, + "learning_rate": 9.641288433382138e-07, + "loss": 0.9633, + "mean_token_accuracy": 0.7115916609764099, + "num_tokens": 68025604.0, + "step": 2635 + }, + { + "epoch": 0.2894794640896112, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.1006741523742676, + "learning_rate": 9.644948755490482e-07, + "loss": 0.8925, + "mean_token_accuracy": 0.7262659072875977, + "num_tokens": 68053834.0, + "step": 2636 + }, + { + "epoch": 0.28958928179222493, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.4539875984191895, + "learning_rate": 9.648609077598828e-07, + "loss": 1.0183, + "mean_token_accuracy": 0.6911430358886719, + "num_tokens": 68075654.0, + "step": 2637 + }, + { + "epoch": 0.2896990994948386, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.295363187789917, + "learning_rate": 9.652269399707174e-07, + "loss": 0.9963, + "mean_token_accuracy": 0.7018066644668579, + "num_tokens": 68100679.0, + "step": 2638 + }, + { + "epoch": 0.2898089171974522, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.287890911102295, + "learning_rate": 9.655929721815518e-07, + "loss": 1.0437, + "mean_token_accuracy": 0.6935617327690125, + "num_tokens": 68126152.0, + "step": 2639 + }, + { + "epoch": 0.28991873490006587, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 1.864757776260376, + "learning_rate": 9.659590043923866e-07, + "loss": 1.0286, + "mean_token_accuracy": 0.6942620277404785, + "num_tokens": 68160904.0, + "step": 2640 + }, + { + "epoch": 0.29002855260267957, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.0355265140533447, + "learning_rate": 9.66325036603221e-07, + "loss": 1.1081, + "mean_token_accuracy": 0.6687662601470947, + "num_tokens": 68194352.0, + "step": 2641 + }, + { + "epoch": 0.2901383703052932, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.302302837371826, + "learning_rate": 9.666910688140556e-07, + "loss": 1.0267, + "mean_token_accuracy": 0.6961597800254822, + "num_tokens": 68220095.0, + "step": 2642 + }, + { + "epoch": 0.29024818800790686, + "ewc_loss": 1.0281801223754883e-06, + "grad_norm": 2.271454095840454, + "learning_rate": 9.670571010248902e-07, + "loss": 0.9872, + "mean_token_accuracy": 0.7030976414680481, + "num_tokens": 68245461.0, + "step": 2643 + }, + { + "epoch": 0.29035800571052056, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.2701597213745117, + "learning_rate": 9.674231332357246e-07, + "loss": 0.9958, + "mean_token_accuracy": 0.6936053037643433, + "num_tokens": 68270879.0, + "step": 2644 + }, + { + "epoch": 0.2904678234131342, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.4295897483825684, + "learning_rate": 9.677891654465592e-07, + "loss": 0.9275, + "mean_token_accuracy": 0.7116703391075134, + "num_tokens": 68292893.0, + "step": 2645 + }, + { + "epoch": 0.29057764111574785, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1352689266204834, + "learning_rate": 9.681551976573938e-07, + "loss": 1.0377, + "mean_token_accuracy": 0.6828693151473999, + "num_tokens": 68319075.0, + "step": 2646 + }, + { + "epoch": 0.2906874588183615, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.485948324203491, + "learning_rate": 9.685212298682284e-07, + "loss": 0.9454, + "mean_token_accuracy": 0.710414469242096, + "num_tokens": 68342098.0, + "step": 2647 + }, + { + "epoch": 0.2907972765209752, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1838114261627197, + "learning_rate": 9.688872620790628e-07, + "loss": 1.0877, + "mean_token_accuracy": 0.6711087822914124, + "num_tokens": 68371225.0, + "step": 2648 + }, + { + "epoch": 0.29090709422358885, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.3099443912506104, + "learning_rate": 9.692532942898974e-07, + "loss": 1.0454, + "mean_token_accuracy": 0.6859422922134399, + "num_tokens": 68398188.0, + "step": 2649 + }, + { + "epoch": 0.2910169119262025, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.19543194770813, + "learning_rate": 9.69619326500732e-07, + "loss": 0.9627, + "mean_token_accuracy": 0.7082858085632324, + "num_tokens": 68425520.0, + "step": 2650 + }, + { + "epoch": 0.29112672962881614, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 1.8355211019515991, + "learning_rate": 9.699853587115666e-07, + "loss": 1.0755, + "mean_token_accuracy": 0.6781904101371765, + "num_tokens": 68462456.0, + "step": 2651 + }, + { + "epoch": 0.29123654733142984, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.233691453933716, + "learning_rate": 9.703513909224012e-07, + "loss": 1.0857, + "mean_token_accuracy": 0.6775012612342834, + "num_tokens": 68489331.0, + "step": 2652 + }, + { + "epoch": 0.2913463650340435, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1977193355560303, + "learning_rate": 9.707174231332356e-07, + "loss": 1.0314, + "mean_token_accuracy": 0.688163161277771, + "num_tokens": 68516051.0, + "step": 2653 + }, + { + "epoch": 0.29145618273665713, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.298001527786255, + "learning_rate": 9.710834553440702e-07, + "loss": 1.0351, + "mean_token_accuracy": 0.6940558552742004, + "num_tokens": 68539805.0, + "step": 2654 + }, + { + "epoch": 0.29156600043927083, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.274324893951416, + "learning_rate": 9.714494875549048e-07, + "loss": 0.9997, + "mean_token_accuracy": 0.6982856392860413, + "num_tokens": 68565221.0, + "step": 2655 + }, + { + "epoch": 0.2916758181418845, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1749467849731445, + "learning_rate": 9.718155197657394e-07, + "loss": 0.9888, + "mean_token_accuracy": 0.7023922204971313, + "num_tokens": 68593703.0, + "step": 2656 + }, + { + "epoch": 0.2917856358444981, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.328507900238037, + "learning_rate": 9.72181551976574e-07, + "loss": 0.9858, + "mean_token_accuracy": 0.6999664306640625, + "num_tokens": 68615507.0, + "step": 2657 + }, + { + "epoch": 0.29189545354711177, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1173243522644043, + "learning_rate": 9.725475841874084e-07, + "loss": 1.061, + "mean_token_accuracy": 0.6985388398170471, + "num_tokens": 68644423.0, + "step": 2658 + }, + { + "epoch": 0.2920052712497255, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1830554008483887, + "learning_rate": 9.72913616398243e-07, + "loss": 1.0228, + "mean_token_accuracy": 0.6951591968536377, + "num_tokens": 68670329.0, + "step": 2659 + }, + { + "epoch": 0.2921150889523391, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.580571174621582, + "learning_rate": 9.732796486090776e-07, + "loss": 1.0361, + "mean_token_accuracy": 0.6977691650390625, + "num_tokens": 68691542.0, + "step": 2660 + }, + { + "epoch": 0.29222490665495277, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.4361393451690674, + "learning_rate": 9.736456808199122e-07, + "loss": 0.9833, + "mean_token_accuracy": 0.702410101890564, + "num_tokens": 68713333.0, + "step": 2661 + }, + { + "epoch": 0.29233472435756647, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.3626980781555176, + "learning_rate": 9.740117130307466e-07, + "loss": 0.9861, + "mean_token_accuracy": 0.7069014310836792, + "num_tokens": 68735818.0, + "step": 2662 + }, + { + "epoch": 0.2924445420601801, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.1744496822357178, + "learning_rate": 9.743777452415812e-07, + "loss": 1.0157, + "mean_token_accuracy": 0.6955697536468506, + "num_tokens": 68762899.0, + "step": 2663 + }, + { + "epoch": 0.29255435976279376, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.4680869579315186, + "learning_rate": 9.747437774524158e-07, + "loss": 1.0307, + "mean_token_accuracy": 0.6983917951583862, + "num_tokens": 68786750.0, + "step": 2664 + }, + { + "epoch": 0.2926641774654074, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.3901500701904297, + "learning_rate": 9.751098096632502e-07, + "loss": 0.9807, + "mean_token_accuracy": 0.7031207084655762, + "num_tokens": 68810131.0, + "step": 2665 + }, + { + "epoch": 0.2927739951680211, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.2986700534820557, + "learning_rate": 9.754758418740848e-07, + "loss": 1.0909, + "mean_token_accuracy": 0.677964448928833, + "num_tokens": 68837302.0, + "step": 2666 + }, + { + "epoch": 0.29288381287063475, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.114147901535034, + "learning_rate": 9.758418740849194e-07, + "loss": 1.0124, + "mean_token_accuracy": 0.7019091844558716, + "num_tokens": 68866848.0, + "step": 2667 + }, + { + "epoch": 0.2929936305732484, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.3773274421691895, + "learning_rate": 9.76207906295754e-07, + "loss": 1.0692, + "mean_token_accuracy": 0.684257984161377, + "num_tokens": 68892221.0, + "step": 2668 + }, + { + "epoch": 0.29310344827586204, + "ewc_loss": 1.0356307029724121e-06, + "grad_norm": 2.0886857509613037, + "learning_rate": 9.765739385065886e-07, + "loss": 0.9775, + "mean_token_accuracy": 0.7032097578048706, + "num_tokens": 68920022.0, + "step": 2669 + }, + { + "epoch": 0.29321326597847575, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.519730806350708, + "learning_rate": 9.76939970717423e-07, + "loss": 0.9786, + "mean_token_accuracy": 0.7099440097808838, + "num_tokens": 68941852.0, + "step": 2670 + }, + { + "epoch": 0.2933230836810894, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.1045684814453125, + "learning_rate": 9.773060029282576e-07, + "loss": 1.0609, + "mean_token_accuracy": 0.6812189817428589, + "num_tokens": 68973292.0, + "step": 2671 + }, + { + "epoch": 0.29343290138370304, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.5397472381591797, + "learning_rate": 9.776720351390922e-07, + "loss": 1.0485, + "mean_token_accuracy": 0.6907070875167847, + "num_tokens": 68995369.0, + "step": 2672 + }, + { + "epoch": 0.29354271908631674, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2715766429901123, + "learning_rate": 9.780380673499268e-07, + "loss": 1.0184, + "mean_token_accuracy": 0.6864160299301147, + "num_tokens": 69018419.0, + "step": 2673 + }, + { + "epoch": 0.2936525367889304, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.6666946411132812, + "learning_rate": 9.784040995607614e-07, + "loss": 0.9709, + "mean_token_accuracy": 0.7111411690711975, + "num_tokens": 69039570.0, + "step": 2674 + }, + { + "epoch": 0.29376235449154403, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2502212524414062, + "learning_rate": 9.787701317715958e-07, + "loss": 0.9692, + "mean_token_accuracy": 0.7091584205627441, + "num_tokens": 69065295.0, + "step": 2675 + }, + { + "epoch": 0.2938721721941577, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.4286675453186035, + "learning_rate": 9.791361639824304e-07, + "loss": 0.9895, + "mean_token_accuracy": 0.7077261209487915, + "num_tokens": 69088292.0, + "step": 2676 + }, + { + "epoch": 0.2939819898967714, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.5047292709350586, + "learning_rate": 9.79502196193265e-07, + "loss": 0.9748, + "mean_token_accuracy": 0.6984269618988037, + "num_tokens": 69109009.0, + "step": 2677 + }, + { + "epoch": 0.294091807599385, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.129883289337158, + "learning_rate": 9.798682284040996e-07, + "loss": 0.9691, + "mean_token_accuracy": 0.7132229208946228, + "num_tokens": 69135296.0, + "step": 2678 + }, + { + "epoch": 0.29420162530199867, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.067427158355713, + "learning_rate": 9.80234260614934e-07, + "loss": 1.0638, + "mean_token_accuracy": 0.6787179112434387, + "num_tokens": 69167474.0, + "step": 2679 + }, + { + "epoch": 0.2943114430046123, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.011132001876831, + "learning_rate": 9.806002928257686e-07, + "loss": 1.0445, + "mean_token_accuracy": 0.6908774375915527, + "num_tokens": 69199833.0, + "step": 2680 + }, + { + "epoch": 0.294421260707226, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.1258726119995117, + "learning_rate": 9.809663250366032e-07, + "loss": 1.0459, + "mean_token_accuracy": 0.6863870024681091, + "num_tokens": 69228030.0, + "step": 2681 + }, + { + "epoch": 0.29453107840983966, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2543013095855713, + "learning_rate": 9.813323572474376e-07, + "loss": 1.039, + "mean_token_accuracy": 0.6801669597625732, + "num_tokens": 69253176.0, + "step": 2682 + }, + { + "epoch": 0.2946408961124533, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.213120698928833, + "learning_rate": 9.816983894582724e-07, + "loss": 0.9606, + "mean_token_accuracy": 0.7096306085586548, + "num_tokens": 69279049.0, + "step": 2683 + }, + { + "epoch": 0.294750713815067, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 1.990372896194458, + "learning_rate": 9.820644216691068e-07, + "loss": 1.049, + "mean_token_accuracy": 0.6874637007713318, + "num_tokens": 69311963.0, + "step": 2684 + }, + { + "epoch": 0.29486053151768066, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2056519985198975, + "learning_rate": 9.824304538799414e-07, + "loss": 0.9493, + "mean_token_accuracy": 0.7135461568832397, + "num_tokens": 69338490.0, + "step": 2685 + }, + { + "epoch": 0.2949703492202943, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.4681472778320312, + "learning_rate": 9.82796486090776e-07, + "loss": 0.9794, + "mean_token_accuracy": 0.6986817121505737, + "num_tokens": 69360211.0, + "step": 2686 + }, + { + "epoch": 0.29508016692290795, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2485079765319824, + "learning_rate": 9.831625183016104e-07, + "loss": 1.1455, + "mean_token_accuracy": 0.6726541519165039, + "num_tokens": 69389732.0, + "step": 2687 + }, + { + "epoch": 0.29518998462552165, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.4229795932769775, + "learning_rate": 9.835285505124452e-07, + "loss": 0.9146, + "mean_token_accuracy": 0.7169537544250488, + "num_tokens": 69412227.0, + "step": 2688 + }, + { + "epoch": 0.2952998023281353, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.3043947219848633, + "learning_rate": 9.838945827232796e-07, + "loss": 1.033, + "mean_token_accuracy": 0.6919360756874084, + "num_tokens": 69435342.0, + "step": 2689 + }, + { + "epoch": 0.29540962003074894, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 1.9726496934890747, + "learning_rate": 9.842606149341142e-07, + "loss": 1.0373, + "mean_token_accuracy": 0.6855436563491821, + "num_tokens": 69466875.0, + "step": 2690 + }, + { + "epoch": 0.29551943773336264, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.228691816329956, + "learning_rate": 9.846266471449488e-07, + "loss": 1.0353, + "mean_token_accuracy": 0.7005301117897034, + "num_tokens": 69493073.0, + "step": 2691 + }, + { + "epoch": 0.2956292554359763, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.154053211212158, + "learning_rate": 9.849926793557832e-07, + "loss": 0.9646, + "mean_token_accuracy": 0.7056024670600891, + "num_tokens": 69520144.0, + "step": 2692 + }, + { + "epoch": 0.29573907313858994, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.2630999088287354, + "learning_rate": 9.853587115666178e-07, + "loss": 0.9913, + "mean_token_accuracy": 0.698582112789154, + "num_tokens": 69545286.0, + "step": 2693 + }, + { + "epoch": 0.2958488908412036, + "ewc_loss": 1.043081283569336e-06, + "grad_norm": 2.128019332885742, + "learning_rate": 9.857247437774524e-07, + "loss": 1.1236, + "mean_token_accuracy": 0.6725872755050659, + "num_tokens": 69575666.0, + "step": 2694 + }, + { + "epoch": 0.2959587085438173, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.5802252292633057, + "learning_rate": 9.86090775988287e-07, + "loss": 0.9863, + "mean_token_accuracy": 0.6986172199249268, + "num_tokens": 69597333.0, + "step": 2695 + }, + { + "epoch": 0.29606852624643093, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.7786014080047607, + "learning_rate": 9.864568081991214e-07, + "loss": 0.9578, + "mean_token_accuracy": 0.7055777311325073, + "num_tokens": 69615995.0, + "step": 2696 + }, + { + "epoch": 0.2961783439490446, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.338252544403076, + "learning_rate": 9.86822840409956e-07, + "loss": 1.0996, + "mean_token_accuracy": 0.6770498752593994, + "num_tokens": 69640263.0, + "step": 2697 + }, + { + "epoch": 0.2962881616516582, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.3815081119537354, + "learning_rate": 9.871888726207906e-07, + "loss": 1.0065, + "mean_token_accuracy": 0.7051924467086792, + "num_tokens": 69662277.0, + "step": 2698 + }, + { + "epoch": 0.2963979793542719, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 1.980958104133606, + "learning_rate": 9.875549048316252e-07, + "loss": 1.0821, + "mean_token_accuracy": 0.6784302592277527, + "num_tokens": 69694745.0, + "step": 2699 + }, + { + "epoch": 0.29650779705688557, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.2900350093841553, + "learning_rate": 9.879209370424598e-07, + "loss": 1.0142, + "mean_token_accuracy": 0.695386528968811, + "num_tokens": 69720316.0, + "step": 2700 + }, + { + "epoch": 0.2966176147594992, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1291704177856445, + "learning_rate": 9.882869692532942e-07, + "loss": 1.0346, + "mean_token_accuracy": 0.68851637840271, + "num_tokens": 69750100.0, + "step": 2701 + }, + { + "epoch": 0.2967274324621129, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1801304817199707, + "learning_rate": 9.886530014641288e-07, + "loss": 1.0461, + "mean_token_accuracy": 0.6887949705123901, + "num_tokens": 69777792.0, + "step": 2702 + }, + { + "epoch": 0.29683725016472656, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.2482028007507324, + "learning_rate": 9.890190336749634e-07, + "loss": 1.0667, + "mean_token_accuracy": 0.6742995977401733, + "num_tokens": 69804220.0, + "step": 2703 + }, + { + "epoch": 0.2969470678673402, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.085965394973755, + "learning_rate": 9.89385065885798e-07, + "loss": 1.0674, + "mean_token_accuracy": 0.6752371788024902, + "num_tokens": 69837893.0, + "step": 2704 + }, + { + "epoch": 0.29705688556995385, + "ewc_loss": 1.0505318641662598e-06, + "grad_norm": 2.2984461784362793, + "learning_rate": 9.897510980966326e-07, + "loss": 0.9821, + "mean_token_accuracy": 0.7101184129714966, + "num_tokens": 69862269.0, + "step": 2705 + }, + { + "epoch": 0.29716670327256756, + "ewc_loss": 1.0505318641662598e-06, + "grad_norm": 2.064162254333496, + "learning_rate": 9.90117130307467e-07, + "loss": 1.0805, + "mean_token_accuracy": 0.6746286153793335, + "num_tokens": 69893396.0, + "step": 2706 + }, + { + "epoch": 0.2972765209751812, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.193486213684082, + "learning_rate": 9.904831625183016e-07, + "loss": 0.9885, + "mean_token_accuracy": 0.6978503465652466, + "num_tokens": 69920648.0, + "step": 2707 + }, + { + "epoch": 0.29738633867779485, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1409554481506348, + "learning_rate": 9.908491947291362e-07, + "loss": 0.9965, + "mean_token_accuracy": 0.6993497610092163, + "num_tokens": 69947077.0, + "step": 2708 + }, + { + "epoch": 0.29749615638040855, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1808040142059326, + "learning_rate": 9.912152269399708e-07, + "loss": 1.0757, + "mean_token_accuracy": 0.6837053298950195, + "num_tokens": 69977260.0, + "step": 2709 + }, + { + "epoch": 0.2976059740830222, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.03906512260437, + "learning_rate": 9.915812591508052e-07, + "loss": 0.997, + "mean_token_accuracy": 0.7026085257530212, + "num_tokens": 70006314.0, + "step": 2710 + }, + { + "epoch": 0.29771579178563584, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.171144962310791, + "learning_rate": 9.919472913616398e-07, + "loss": 0.9911, + "mean_token_accuracy": 0.6998376846313477, + "num_tokens": 70033098.0, + "step": 2711 + }, + { + "epoch": 0.2978256094882495, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1696410179138184, + "learning_rate": 9.923133235724744e-07, + "loss": 1.0079, + "mean_token_accuracy": 0.7039709091186523, + "num_tokens": 70062200.0, + "step": 2712 + }, + { + "epoch": 0.2979354271908632, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.154754877090454, + "learning_rate": 9.926793557833088e-07, + "loss": 1.0668, + "mean_token_accuracy": 0.6955215334892273, + "num_tokens": 70089711.0, + "step": 2713 + }, + { + "epoch": 0.29804524489347684, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.2258803844451904, + "learning_rate": 9.930453879941434e-07, + "loss": 0.9233, + "mean_token_accuracy": 0.7200579643249512, + "num_tokens": 70113978.0, + "step": 2714 + }, + { + "epoch": 0.2981550625960905, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.2488887310028076, + "learning_rate": 9.93411420204978e-07, + "loss": 0.9837, + "mean_token_accuracy": 0.6999276280403137, + "num_tokens": 70138238.0, + "step": 2715 + }, + { + "epoch": 0.2982648802987041, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.372081756591797, + "learning_rate": 9.937774524158126e-07, + "loss": 0.9637, + "mean_token_accuracy": 0.7104094624519348, + "num_tokens": 70161906.0, + "step": 2716 + }, + { + "epoch": 0.29837469800131783, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.0707597732543945, + "learning_rate": 9.941434846266472e-07, + "loss": 0.9597, + "mean_token_accuracy": 0.7069994807243347, + "num_tokens": 70192723.0, + "step": 2717 + }, + { + "epoch": 0.2984845157039315, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.1998908519744873, + "learning_rate": 9.945095168374816e-07, + "loss": 0.9891, + "mean_token_accuracy": 0.7016829252243042, + "num_tokens": 70221730.0, + "step": 2718 + }, + { + "epoch": 0.2985943334065451, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.324044704437256, + "learning_rate": 9.948755490483162e-07, + "loss": 0.9695, + "mean_token_accuracy": 0.7084840536117554, + "num_tokens": 70249072.0, + "step": 2719 + }, + { + "epoch": 0.2987041511091588, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.688732385635376, + "learning_rate": 9.952415812591508e-07, + "loss": 0.9674, + "mean_token_accuracy": 0.7040243148803711, + "num_tokens": 70270451.0, + "step": 2720 + }, + { + "epoch": 0.29881396881177247, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.5073764324188232, + "learning_rate": 9.956076134699854e-07, + "loss": 0.9143, + "mean_token_accuracy": 0.7241379022598267, + "num_tokens": 70290888.0, + "step": 2721 + }, + { + "epoch": 0.2989237865143861, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.427448272705078, + "learning_rate": 9.9597364568082e-07, + "loss": 1.0865, + "mean_token_accuracy": 0.6756315231323242, + "num_tokens": 70314777.0, + "step": 2722 + }, + { + "epoch": 0.29903360421699976, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.078317403793335, + "learning_rate": 9.963396778916544e-07, + "loss": 0.9771, + "mean_token_accuracy": 0.7055584192276001, + "num_tokens": 70343692.0, + "step": 2723 + }, + { + "epoch": 0.29914342191961346, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.217998743057251, + "learning_rate": 9.96705710102489e-07, + "loss": 0.9931, + "mean_token_accuracy": 0.6956079006195068, + "num_tokens": 70367561.0, + "step": 2724 + }, + { + "epoch": 0.2992532396222271, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.4235708713531494, + "learning_rate": 9.970717423133236e-07, + "loss": 0.9266, + "mean_token_accuracy": 0.7114998698234558, + "num_tokens": 70389900.0, + "step": 2725 + }, + { + "epoch": 0.29936305732484075, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.378307819366455, + "learning_rate": 9.974377745241582e-07, + "loss": 0.967, + "mean_token_accuracy": 0.7082104682922363, + "num_tokens": 70415989.0, + "step": 2726 + }, + { + "epoch": 0.2994728750274544, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 1.9550451040267944, + "learning_rate": 9.978038067349926e-07, + "loss": 1.0585, + "mean_token_accuracy": 0.6819769740104675, + "num_tokens": 70448872.0, + "step": 2727 + }, + { + "epoch": 0.2995826927300681, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.0507349967956543, + "learning_rate": 9.981698389458272e-07, + "loss": 1.0276, + "mean_token_accuracy": 0.6931897401809692, + "num_tokens": 70479507.0, + "step": 2728 + }, + { + "epoch": 0.29969251043268175, + "ewc_loss": 1.0579824447631836e-06, + "grad_norm": 2.001276731491089, + "learning_rate": 9.985358711566618e-07, + "loss": 0.9733, + "mean_token_accuracy": 0.7010918855667114, + "num_tokens": 70510493.0, + "step": 2729 + }, + { + "epoch": 0.2998023281352954, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.5402090549468994, + "learning_rate": 9.989019033674961e-07, + "loss": 0.9336, + "mean_token_accuracy": 0.711152970790863, + "num_tokens": 70532115.0, + "step": 2730 + }, + { + "epoch": 0.2999121458379091, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 1.967177152633667, + "learning_rate": 9.99267935578331e-07, + "loss": 0.9075, + "mean_token_accuracy": 0.7116950750350952, + "num_tokens": 70563634.0, + "step": 2731 + }, + { + "epoch": 0.30002196354052274, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.1816422939300537, + "learning_rate": 9.996339677891654e-07, + "loss": 0.9947, + "mean_token_accuracy": 0.7061022520065308, + "num_tokens": 70591530.0, + "step": 2732 + }, + { + "epoch": 0.3001317812431364, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.61489200592041, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6905859112739563, + "num_tokens": 70611898.0, + "step": 2733 + }, + { + "epoch": 0.30024159894575003, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.5696818828582764, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7194846272468567, + "num_tokens": 70631779.0, + "step": 2734 + }, + { + "epoch": 0.30035141664836373, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.0361487865448, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.69693922996521, + "num_tokens": 70660246.0, + "step": 2735 + }, + { + "epoch": 0.3004612343509774, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.0434913635253906, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6905797123908997, + "num_tokens": 70691353.0, + "step": 2736 + }, + { + "epoch": 0.300571052053591, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.5053229331970215, + "learning_rate": 1e-06, + "loss": 1.0886, + "mean_token_accuracy": 0.674191951751709, + "num_tokens": 70714143.0, + "step": 2737 + }, + { + "epoch": 0.3006808697562047, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.263296365737915, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6874715089797974, + "num_tokens": 70738976.0, + "step": 2738 + }, + { + "epoch": 0.3007906874588184, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.430022954940796, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7164384126663208, + "num_tokens": 70760448.0, + "step": 2739 + }, + { + "epoch": 0.300900505161432, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.1414761543273926, + "learning_rate": 1e-06, + "loss": 1.0909, + "mean_token_accuracy": 0.6700775623321533, + "num_tokens": 70786483.0, + "step": 2740 + }, + { + "epoch": 0.30101032286404567, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.260298490524292, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7097564339637756, + "num_tokens": 70809208.0, + "step": 2741 + }, + { + "epoch": 0.30112014056665937, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.3503336906433105, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7071586847305298, + "num_tokens": 70834608.0, + "step": 2742 + }, + { + "epoch": 0.301229958269273, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.2377755641937256, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7119158506393433, + "num_tokens": 70861261.0, + "step": 2743 + }, + { + "epoch": 0.30133977597188666, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.219322443008423, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6868689060211182, + "num_tokens": 70889551.0, + "step": 2744 + }, + { + "epoch": 0.3014495936745003, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.31268572807312, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6956155896186829, + "num_tokens": 70916162.0, + "step": 2745 + }, + { + "epoch": 0.301559411377114, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.1380817890167236, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7088285684585571, + "num_tokens": 70944163.0, + "step": 2746 + }, + { + "epoch": 0.30166922907972765, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 2.1537258625030518, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7057334184646606, + "num_tokens": 70970646.0, + "step": 2747 + }, + { + "epoch": 0.3017790467823413, + "ewc_loss": 1.0728836059570312e-06, + "grad_norm": 1.9534322023391724, + "learning_rate": 1e-06, + "loss": 1.0746, + "mean_token_accuracy": 0.679431676864624, + "num_tokens": 71003040.0, + "step": 2748 + }, + { + "epoch": 0.301888864484955, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 1.9848694801330566, + "learning_rate": 1e-06, + "loss": 1.067, + "mean_token_accuracy": 0.6808602213859558, + "num_tokens": 71034419.0, + "step": 2749 + }, + { + "epoch": 0.30199868218756865, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 2.2811737060546875, + "learning_rate": 1e-06, + "loss": 1.0285, + "mean_token_accuracy": 0.6950998306274414, + "num_tokens": 71060266.0, + "step": 2750 + }, + { + "epoch": 0.3021084998901823, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 2.2812700271606445, + "learning_rate": 1e-06, + "loss": 1.1111, + "mean_token_accuracy": 0.6656150817871094, + "num_tokens": 71085939.0, + "step": 2751 + }, + { + "epoch": 0.30221831759279594, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 2.24465274810791, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6932014226913452, + "num_tokens": 71110518.0, + "step": 2752 + }, + { + "epoch": 0.30232813529540964, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 2.3992981910705566, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7075571417808533, + "num_tokens": 71132356.0, + "step": 2753 + }, + { + "epoch": 0.3024379529980233, + "ewc_loss": 1.080334186553955e-06, + "grad_norm": 2.3253867626190186, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6930763721466064, + "num_tokens": 71155990.0, + "step": 2754 + }, + { + "epoch": 0.30254777070063693, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.461444139480591, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7184371948242188, + "num_tokens": 71176187.0, + "step": 2755 + }, + { + "epoch": 0.3026575884032506, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.2646291255950928, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.700129508972168, + "num_tokens": 71200764.0, + "step": 2756 + }, + { + "epoch": 0.3027674061058643, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.26147723197937, + "learning_rate": 1e-06, + "loss": 1.0103, + "mean_token_accuracy": 0.6933740377426147, + "num_tokens": 71228226.0, + "step": 2757 + }, + { + "epoch": 0.3028772238084779, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.256983518600464, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6931467652320862, + "num_tokens": 71252338.0, + "step": 2758 + }, + { + "epoch": 0.30298704151109157, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.4432523250579834, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6808759570121765, + "num_tokens": 71276029.0, + "step": 2759 + }, + { + "epoch": 0.30309685921370527, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.351665496826172, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.71588134765625, + "num_tokens": 71300798.0, + "step": 2760 + }, + { + "epoch": 0.3032066769163189, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 1.9680774211883545, + "learning_rate": 1e-06, + "loss": 1.0931, + "mean_token_accuracy": 0.6725212335586548, + "num_tokens": 71335051.0, + "step": 2761 + }, + { + "epoch": 0.30331649461893256, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.2779955863952637, + "learning_rate": 1e-06, + "loss": 1.0416, + "mean_token_accuracy": 0.6907195448875427, + "num_tokens": 71361661.0, + "step": 2762 + }, + { + "epoch": 0.3034263123215462, + "ewc_loss": 1.087784767150879e-06, + "grad_norm": 2.557467460632324, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7000718116760254, + "num_tokens": 71382572.0, + "step": 2763 + }, + { + "epoch": 0.3035361300241599, + "ewc_loss": 1.0952353477478027e-06, + "grad_norm": 2.4746525287628174, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7152527570724487, + "num_tokens": 71403425.0, + "step": 2764 + }, + { + "epoch": 0.30364594772677356, + "ewc_loss": 1.0952353477478027e-06, + "grad_norm": 2.2072863578796387, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6950305104255676, + "num_tokens": 71428908.0, + "step": 2765 + }, + { + "epoch": 0.3037557654293872, + "ewc_loss": 1.0952353477478027e-06, + "grad_norm": 2.189800977706909, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7171036601066589, + "num_tokens": 71454772.0, + "step": 2766 + }, + { + "epoch": 0.3038655831320009, + "ewc_loss": 1.0952353477478027e-06, + "grad_norm": 2.2802834510803223, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6849822998046875, + "num_tokens": 71482693.0, + "step": 2767 + }, + { + "epoch": 0.30397540083461455, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.4393484592437744, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7166132926940918, + "num_tokens": 71504932.0, + "step": 2768 + }, + { + "epoch": 0.3040852185372282, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.729771614074707, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.708778440952301, + "num_tokens": 71524792.0, + "step": 2769 + }, + { + "epoch": 0.30419503623984184, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.1252992153167725, + "learning_rate": 1e-06, + "loss": 1.128, + "mean_token_accuracy": 0.6633667945861816, + "num_tokens": 71555938.0, + "step": 2770 + }, + { + "epoch": 0.30430485394245554, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.3480641841888428, + "learning_rate": 1e-06, + "loss": 1.0488, + "mean_token_accuracy": 0.6854193806648254, + "num_tokens": 71582278.0, + "step": 2771 + }, + { + "epoch": 0.3044146716450692, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.2388498783111572, + "learning_rate": 1e-06, + "loss": 1.0512, + "mean_token_accuracy": 0.6948184967041016, + "num_tokens": 71611249.0, + "step": 2772 + }, + { + "epoch": 0.30452448934768284, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.1169090270996094, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.6965290307998657, + "num_tokens": 71640985.0, + "step": 2773 + }, + { + "epoch": 0.3046343070502965, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.4409596920013428, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.7060574889183044, + "num_tokens": 71663987.0, + "step": 2774 + }, + { + "epoch": 0.3047441247529102, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.42557430267334, + "learning_rate": 1e-06, + "loss": 1.0759, + "mean_token_accuracy": 0.6816337704658508, + "num_tokens": 71688019.0, + "step": 2775 + }, + { + "epoch": 0.30485394245552383, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.0927422046661377, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7152233719825745, + "num_tokens": 71715223.0, + "step": 2776 + }, + { + "epoch": 0.3049637601581375, + "ewc_loss": 1.1026859283447266e-06, + "grad_norm": 2.2230050563812256, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.678336501121521, + "num_tokens": 71742935.0, + "step": 2777 + }, + { + "epoch": 0.3050735778607512, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.1243927478790283, + "learning_rate": 1e-06, + "loss": 1.1261, + "mean_token_accuracy": 0.6685529947280884, + "num_tokens": 71774364.0, + "step": 2778 + }, + { + "epoch": 0.3051833955633648, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 1.9985755681991577, + "learning_rate": 1e-06, + "loss": 1.1277, + "mean_token_accuracy": 0.670073390007019, + "num_tokens": 71808027.0, + "step": 2779 + }, + { + "epoch": 0.30529321326597847, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.3761205673217773, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7095134854316711, + "num_tokens": 71830722.0, + "step": 2780 + }, + { + "epoch": 0.3054030309685921, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.4000422954559326, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.7013285160064697, + "num_tokens": 71852928.0, + "step": 2781 + }, + { + "epoch": 0.3055128486712058, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.0942139625549316, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6920656561851501, + "num_tokens": 71882433.0, + "step": 2782 + }, + { + "epoch": 0.30562266637381946, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 1.9522238969802856, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6868772506713867, + "num_tokens": 71914594.0, + "step": 2783 + }, + { + "epoch": 0.3057324840764331, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.454535484313965, + "learning_rate": 1e-06, + "loss": 1.0607, + "mean_token_accuracy": 0.6823248863220215, + "num_tokens": 71938171.0, + "step": 2784 + }, + { + "epoch": 0.30584230177904675, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.1101138591766357, + "learning_rate": 1e-06, + "loss": 1.0853, + "mean_token_accuracy": 0.6904388070106506, + "num_tokens": 71966619.0, + "step": 2785 + }, + { + "epoch": 0.30595211948166046, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.1946253776550293, + "learning_rate": 1e-06, + "loss": 1.1093, + "mean_token_accuracy": 0.6683461666107178, + "num_tokens": 71996236.0, + "step": 2786 + }, + { + "epoch": 0.3060619371842741, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.428626537322998, + "learning_rate": 1e-06, + "loss": 1.0883, + "mean_token_accuracy": 0.676435649394989, + "num_tokens": 72020186.0, + "step": 2787 + }, + { + "epoch": 0.30617175488688775, + "ewc_loss": 1.1101365089416504e-06, + "grad_norm": 2.1764144897460938, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7059277296066284, + "num_tokens": 72045722.0, + "step": 2788 + }, + { + "epoch": 0.30628157258950145, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.300523042678833, + "learning_rate": 1e-06, + "loss": 1.1026, + "mean_token_accuracy": 0.673578679561615, + "num_tokens": 72072531.0, + "step": 2789 + }, + { + "epoch": 0.3063913902921151, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.214475154876709, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.697301983833313, + "num_tokens": 72098928.0, + "step": 2790 + }, + { + "epoch": 0.30650120799472874, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.3805272579193115, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6873730421066284, + "num_tokens": 72123697.0, + "step": 2791 + }, + { + "epoch": 0.3066110256973424, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.6498191356658936, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7051077485084534, + "num_tokens": 72144223.0, + "step": 2792 + }, + { + "epoch": 0.3067208433999561, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.3411214351654053, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7077229619026184, + "num_tokens": 72167845.0, + "step": 2793 + }, + { + "epoch": 0.30683066110256974, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.320006847381592, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.699617862701416, + "num_tokens": 72190607.0, + "step": 2794 + }, + { + "epoch": 0.3069404788051834, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.231365442276001, + "learning_rate": 1e-06, + "loss": 1.0256, + "mean_token_accuracy": 0.691577672958374, + "num_tokens": 72216993.0, + "step": 2795 + }, + { + "epoch": 0.3070502965077971, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.3514153957366943, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7062007784843445, + "num_tokens": 72240619.0, + "step": 2796 + }, + { + "epoch": 0.30716011421041073, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.5522799491882324, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6897523403167725, + "num_tokens": 72261970.0, + "step": 2797 + }, + { + "epoch": 0.3072699319130244, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.4895989894866943, + "learning_rate": 1e-06, + "loss": 1.084, + "mean_token_accuracy": 0.6798473596572876, + "num_tokens": 72286977.0, + "step": 2798 + }, + { + "epoch": 0.307379749615638, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.1646461486816406, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6897017955780029, + "num_tokens": 72312612.0, + "step": 2799 + }, + { + "epoch": 0.3074895673182517, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.080198049545288, + "learning_rate": 1e-06, + "loss": 1.0713, + "mean_token_accuracy": 0.6849983930587769, + "num_tokens": 72342397.0, + "step": 2800 + }, + { + "epoch": 0.30759938502086537, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.0299479961395264, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6856579780578613, + "num_tokens": 72373431.0, + "step": 2801 + }, + { + "epoch": 0.307709202723479, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.245434522628784, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.6960315704345703, + "num_tokens": 72399898.0, + "step": 2802 + }, + { + "epoch": 0.30781902042609266, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.3115148544311523, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7523594498634338, + "num_tokens": 72421490.0, + "step": 2803 + }, + { + "epoch": 0.30792883812870636, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.5370726585388184, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7171077728271484, + "num_tokens": 72441399.0, + "step": 2804 + }, + { + "epoch": 0.30803865583132, + "ewc_loss": 1.1324882507324219e-06, + "grad_norm": 2.4975643157958984, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7210124731063843, + "num_tokens": 72465359.0, + "step": 2805 + }, + { + "epoch": 0.30814847353393365, + "ewc_loss": 1.1324882507324219e-06, + "grad_norm": 2.1309313774108887, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7079709768295288, + "num_tokens": 72492832.0, + "step": 2806 + }, + { + "epoch": 0.30825829123654735, + "ewc_loss": 1.1324882507324219e-06, + "grad_norm": 2.339423418045044, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6913891434669495, + "num_tokens": 72517740.0, + "step": 2807 + }, + { + "epoch": 0.308368108939161, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.045559883117676, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7189905047416687, + "num_tokens": 72546634.0, + "step": 2808 + }, + { + "epoch": 0.30847792664177465, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.193141460418701, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6978036165237427, + "num_tokens": 72574884.0, + "step": 2809 + }, + { + "epoch": 0.3085877443443883, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.3068442344665527, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7110157012939453, + "num_tokens": 72599568.0, + "step": 2810 + }, + { + "epoch": 0.308697562047002, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.3797874450683594, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7052391767501831, + "num_tokens": 72623881.0, + "step": 2811 + }, + { + "epoch": 0.30880737974961564, + "ewc_loss": 1.1175870895385742e-06, + "grad_norm": 2.0173757076263428, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.697046160697937, + "num_tokens": 72656182.0, + "step": 2812 + }, + { + "epoch": 0.3089171974522293, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.210144519805908, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7150498032569885, + "num_tokens": 72681467.0, + "step": 2813 + }, + { + "epoch": 0.309027015154843, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.4720358848571777, + "learning_rate": 1e-06, + "loss": 1.0844, + "mean_token_accuracy": 0.6848042011260986, + "num_tokens": 72704525.0, + "step": 2814 + }, + { + "epoch": 0.30913683285745663, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.2720184326171875, + "learning_rate": 1e-06, + "loss": 1.0388, + "mean_token_accuracy": 0.6864475607872009, + "num_tokens": 72729940.0, + "step": 2815 + }, + { + "epoch": 0.3092466505600703, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.331531047821045, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7099554538726807, + "num_tokens": 72754018.0, + "step": 2816 + }, + { + "epoch": 0.3093564682626839, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.2986409664154053, + "learning_rate": 1e-06, + "loss": 1.074, + "mean_token_accuracy": 0.6849439144134521, + "num_tokens": 72780723.0, + "step": 2817 + }, + { + "epoch": 0.3094662859652976, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.18961763381958, + "learning_rate": 1e-06, + "loss": 1.0601, + "mean_token_accuracy": 0.6807681322097778, + "num_tokens": 72809617.0, + "step": 2818 + }, + { + "epoch": 0.3095761036679113, + "ewc_loss": 1.125037670135498e-06, + "grad_norm": 2.5800864696502686, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7372397780418396, + "num_tokens": 72829097.0, + "step": 2819 + }, + { + "epoch": 0.3096859213705249, + "ewc_loss": 1.1324882507324219e-06, + "grad_norm": 2.129892349243164, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7058975696563721, + "num_tokens": 72856357.0, + "step": 2820 + }, + { + "epoch": 0.30979573907313857, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 2.275116443634033, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.708595871925354, + "num_tokens": 72882233.0, + "step": 2821 + }, + { + "epoch": 0.30990555677575227, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 2.3454670906066895, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7114328145980835, + "num_tokens": 72905153.0, + "step": 2822 + }, + { + "epoch": 0.3100153744783659, + "ewc_loss": 1.1399388313293457e-06, + "grad_norm": 2.5912280082702637, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7179993987083435, + "num_tokens": 72926296.0, + "step": 2823 + }, + { + "epoch": 0.31012519218097956, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.105092763900757, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6915416121482849, + "num_tokens": 72955950.0, + "step": 2824 + }, + { + "epoch": 0.31023500988359326, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.333714485168457, + "learning_rate": 1e-06, + "loss": 1.0463, + "mean_token_accuracy": 0.681157112121582, + "num_tokens": 72982794.0, + "step": 2825 + }, + { + "epoch": 0.3103448275862069, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.352501153945923, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7010297179222107, + "num_tokens": 73007331.0, + "step": 2826 + }, + { + "epoch": 0.31045464528882055, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.559349298477173, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.695578396320343, + "num_tokens": 73027103.0, + "step": 2827 + }, + { + "epoch": 0.3105644629914342, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.413172483444214, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6961712837219238, + "num_tokens": 73049063.0, + "step": 2828 + }, + { + "epoch": 0.3106742806940479, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.0209662914276123, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.694956362247467, + "num_tokens": 73078869.0, + "step": 2829 + }, + { + "epoch": 0.31078409839666155, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.2910242080688477, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.698676347732544, + "num_tokens": 73105191.0, + "step": 2830 + }, + { + "epoch": 0.3108939160992752, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.1916065216064453, + "learning_rate": 1e-06, + "loss": 1.0611, + "mean_token_accuracy": 0.6890265941619873, + "num_tokens": 73131768.0, + "step": 2831 + }, + { + "epoch": 0.31100373380188884, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.2537801265716553, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6983481049537659, + "num_tokens": 73158912.0, + "step": 2832 + }, + { + "epoch": 0.31111355150450254, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.3883256912231445, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7021013498306274, + "num_tokens": 73183322.0, + "step": 2833 + }, + { + "epoch": 0.3112233692071162, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.0497567653656006, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6993557214736938, + "num_tokens": 73212338.0, + "step": 2834 + }, + { + "epoch": 0.31133318690972983, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.236583948135376, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6894751787185669, + "num_tokens": 73241951.0, + "step": 2835 + }, + { + "epoch": 0.31144300461234353, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.1855411529541016, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7194643020629883, + "num_tokens": 73269042.0, + "step": 2836 + }, + { + "epoch": 0.3115528223149572, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.2797844409942627, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6999952793121338, + "num_tokens": 73295104.0, + "step": 2837 + }, + { + "epoch": 0.3116626400175708, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.1447367668151855, + "learning_rate": 1e-06, + "loss": 1.0645, + "mean_token_accuracy": 0.6837138533592224, + "num_tokens": 73322207.0, + "step": 2838 + }, + { + "epoch": 0.31177245772018447, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.4223105907440186, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6904106140136719, + "num_tokens": 73344542.0, + "step": 2839 + }, + { + "epoch": 0.31188227542279817, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.495274543762207, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7052809596061707, + "num_tokens": 73366839.0, + "step": 2840 + }, + { + "epoch": 0.3119920931254118, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.4070544242858887, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6967655420303345, + "num_tokens": 73390568.0, + "step": 2841 + }, + { + "epoch": 0.31210191082802546, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.4595274925231934, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7107052803039551, + "num_tokens": 73411547.0, + "step": 2842 + }, + { + "epoch": 0.31221172853063917, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.4723453521728516, + "learning_rate": 1e-06, + "loss": 1.0227, + "mean_token_accuracy": 0.6964868307113647, + "num_tokens": 73436034.0, + "step": 2843 + }, + { + "epoch": 0.3123215462332528, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.3231120109558105, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6949306726455688, + "num_tokens": 73461319.0, + "step": 2844 + }, + { + "epoch": 0.31243136393586646, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.419408082962036, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6957634687423706, + "num_tokens": 73483607.0, + "step": 2845 + }, + { + "epoch": 0.3125411816384801, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.196916103363037, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6817166209220886, + "num_tokens": 73510148.0, + "step": 2846 + }, + { + "epoch": 0.3126509993410938, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.8180198669433594, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.738418698310852, + "num_tokens": 73527606.0, + "step": 2847 + }, + { + "epoch": 0.31276081704370745, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.4468390941619873, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7015833258628845, + "num_tokens": 73548554.0, + "step": 2848 + }, + { + "epoch": 0.3128706347463211, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.0997672080993652, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7260187864303589, + "num_tokens": 73575796.0, + "step": 2849 + }, + { + "epoch": 0.31298045244893474, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.106748342514038, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.700107216835022, + "num_tokens": 73605313.0, + "step": 2850 + }, + { + "epoch": 0.31309027015154844, + "ewc_loss": 1.1473894119262695e-06, + "grad_norm": 2.3348982334136963, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6892622709274292, + "num_tokens": 73630136.0, + "step": 2851 + }, + { + "epoch": 0.3132000878541621, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.5576345920562744, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7130545377731323, + "num_tokens": 73650701.0, + "step": 2852 + }, + { + "epoch": 0.31330990555677574, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.4029643535614014, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.718564510345459, + "num_tokens": 73675451.0, + "step": 2853 + }, + { + "epoch": 0.31341972325938944, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.384722948074341, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6928967237472534, + "num_tokens": 73699631.0, + "step": 2854 + }, + { + "epoch": 0.3135295409620031, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.0383665561676025, + "learning_rate": 1e-06, + "loss": 1.1434, + "mean_token_accuracy": 0.669035792350769, + "num_tokens": 73734841.0, + "step": 2855 + }, + { + "epoch": 0.31363935866461673, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.105800151824951, + "learning_rate": 1e-06, + "loss": 1.0619, + "mean_token_accuracy": 0.6775798201560974, + "num_tokens": 73766927.0, + "step": 2856 + }, + { + "epoch": 0.3137491763672304, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.1937737464904785, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6981474757194519, + "num_tokens": 73794957.0, + "step": 2857 + }, + { + "epoch": 0.3138589940698441, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.613149404525757, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.701572835445404, + "num_tokens": 73814952.0, + "step": 2858 + }, + { + "epoch": 0.3139688117724577, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.130545139312744, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.6935024261474609, + "num_tokens": 73840972.0, + "step": 2859 + }, + { + "epoch": 0.31407862947507137, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.0317749977111816, + "learning_rate": 1e-06, + "loss": 1.044, + "mean_token_accuracy": 0.6815835237503052, + "num_tokens": 73871707.0, + "step": 2860 + }, + { + "epoch": 0.314188447177685, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.360640287399292, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7057831883430481, + "num_tokens": 73896564.0, + "step": 2861 + }, + { + "epoch": 0.3142982648802987, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.2109034061431885, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7038366198539734, + "num_tokens": 73923668.0, + "step": 2862 + }, + { + "epoch": 0.31440808258291236, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.186659097671509, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.688143253326416, + "num_tokens": 73949813.0, + "step": 2863 + }, + { + "epoch": 0.314517900285526, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.4996161460876465, + "learning_rate": 1e-06, + "loss": 1.0621, + "mean_token_accuracy": 0.6885272860527039, + "num_tokens": 73972394.0, + "step": 2864 + }, + { + "epoch": 0.3146277179881397, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.2534830570220947, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6960001587867737, + "num_tokens": 73996907.0, + "step": 2865 + }, + { + "epoch": 0.31473753569075336, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.1311402320861816, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7247806787490845, + "num_tokens": 74023279.0, + "step": 2866 + }, + { + "epoch": 0.314847353393367, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.064582586288452, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.699439287185669, + "num_tokens": 74055139.0, + "step": 2867 + }, + { + "epoch": 0.31495717109598065, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.495957374572754, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7036696076393127, + "num_tokens": 74077381.0, + "step": 2868 + }, + { + "epoch": 0.31506698879859435, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.367398262023926, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6974878311157227, + "num_tokens": 74102507.0, + "step": 2869 + }, + { + "epoch": 0.315176806501208, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.1698806285858154, + "learning_rate": 1e-06, + "loss": 1.0475, + "mean_token_accuracy": 0.6830735206604004, + "num_tokens": 74131402.0, + "step": 2870 + }, + { + "epoch": 0.31528662420382164, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.6825151443481445, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.716209888458252, + "num_tokens": 74150490.0, + "step": 2871 + }, + { + "epoch": 0.31539644190643534, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.2259185314178467, + "learning_rate": 1e-06, + "loss": 1.0646, + "mean_token_accuracy": 0.6801298260688782, + "num_tokens": 74177232.0, + "step": 2872 + }, + { + "epoch": 0.315506259609049, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.317154884338379, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.716282844543457, + "num_tokens": 74201663.0, + "step": 2873 + }, + { + "epoch": 0.31561607731166264, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.316943883895874, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7023614645004272, + "num_tokens": 74225721.0, + "step": 2874 + }, + { + "epoch": 0.3157258950142763, + "ewc_loss": 1.1548399925231934e-06, + "grad_norm": 2.0664007663726807, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7016449570655823, + "num_tokens": 74254833.0, + "step": 2875 + }, + { + "epoch": 0.31583571271689, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.260216474533081, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6980891823768616, + "num_tokens": 74280086.0, + "step": 2876 + }, + { + "epoch": 0.31594553041950363, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.232807159423828, + "learning_rate": 1e-06, + "loss": 1.1227, + "mean_token_accuracy": 0.6800645589828491, + "num_tokens": 74308202.0, + "step": 2877 + }, + { + "epoch": 0.3160553481221173, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 1.9517449140548706, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6966850161552429, + "num_tokens": 74340124.0, + "step": 2878 + }, + { + "epoch": 0.3161651658247309, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.448204278945923, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7042810320854187, + "num_tokens": 74363775.0, + "step": 2879 + }, + { + "epoch": 0.3162749835273446, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.385167121887207, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7154366970062256, + "num_tokens": 74385059.0, + "step": 2880 + }, + { + "epoch": 0.31638480122995827, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 1.9755498170852661, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6934359073638916, + "num_tokens": 74415274.0, + "step": 2881 + }, + { + "epoch": 0.3164946189325719, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.769946336746216, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6976061463356018, + "num_tokens": 74433509.0, + "step": 2882 + }, + { + "epoch": 0.3166044366351856, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.2350504398345947, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.7089056968688965, + "num_tokens": 74456649.0, + "step": 2883 + }, + { + "epoch": 0.31671425433779926, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.5342464447021484, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.6970925331115723, + "num_tokens": 74478535.0, + "step": 2884 + }, + { + "epoch": 0.3168240720404129, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.2755136489868164, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7120548486709595, + "num_tokens": 74501778.0, + "step": 2885 + }, + { + "epoch": 0.31693388974302655, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.144481897354126, + "learning_rate": 1e-06, + "loss": 1.0888, + "mean_token_accuracy": 0.6750342845916748, + "num_tokens": 74530204.0, + "step": 2886 + }, + { + "epoch": 0.31704370744564025, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.3535075187683105, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.736632764339447, + "num_tokens": 74552726.0, + "step": 2887 + }, + { + "epoch": 0.3171535251482539, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.401931047439575, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.7045223116874695, + "num_tokens": 74574343.0, + "step": 2888 + }, + { + "epoch": 0.31726334285086755, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.331334114074707, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7063310742378235, + "num_tokens": 74597489.0, + "step": 2889 + }, + { + "epoch": 0.31737316055348125, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 1.9100440740585327, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.699536919593811, + "num_tokens": 74630000.0, + "step": 2890 + }, + { + "epoch": 0.3174829782560949, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.2401583194732666, + "learning_rate": 1e-06, + "loss": 1.0956, + "mean_token_accuracy": 0.6781556606292725, + "num_tokens": 74656838.0, + "step": 2891 + }, + { + "epoch": 0.31759279595870854, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.2445061206817627, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6928693652153015, + "num_tokens": 74680459.0, + "step": 2892 + }, + { + "epoch": 0.3177026136613222, + "ewc_loss": 1.1622905731201172e-06, + "grad_norm": 2.0932657718658447, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7077671885490417, + "num_tokens": 74709266.0, + "step": 2893 + }, + { + "epoch": 0.3178124313639359, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.119235038757324, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7158539295196533, + "num_tokens": 74737699.0, + "step": 2894 + }, + { + "epoch": 0.31792224906654953, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.377788782119751, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7061030864715576, + "num_tokens": 74761079.0, + "step": 2895 + }, + { + "epoch": 0.3180320667691632, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 1.9941506385803223, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7012125849723816, + "num_tokens": 74790938.0, + "step": 2896 + }, + { + "epoch": 0.3181418844717768, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.0204262733459473, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7082377672195435, + "num_tokens": 74819771.0, + "step": 2897 + }, + { + "epoch": 0.3182517021743905, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.303299903869629, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6984621286392212, + "num_tokens": 74845767.0, + "step": 2898 + }, + { + "epoch": 0.3183615198770042, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.1517767906188965, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7058595418930054, + "num_tokens": 74873917.0, + "step": 2899 + }, + { + "epoch": 0.3184713375796178, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.2113754749298096, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7159238457679749, + "num_tokens": 74898683.0, + "step": 2900 + }, + { + "epoch": 0.3185811552822315, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.234954833984375, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6920828819274902, + "num_tokens": 74923942.0, + "step": 2901 + }, + { + "epoch": 0.31869097298484517, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.149538040161133, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7087187767028809, + "num_tokens": 74950771.0, + "step": 2902 + }, + { + "epoch": 0.3188007906874588, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.4815402030944824, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6951198577880859, + "num_tokens": 74973253.0, + "step": 2903 + }, + { + "epoch": 0.31891060839007246, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.2821850776672363, + "learning_rate": 1e-06, + "loss": 1.0785, + "mean_token_accuracy": 0.6794847846031189, + "num_tokens": 74999470.0, + "step": 2904 + }, + { + "epoch": 0.31902042609268616, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.637019157409668, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6876736879348755, + "num_tokens": 75019191.0, + "step": 2905 + }, + { + "epoch": 0.3191302437952998, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.136458396911621, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6861083507537842, + "num_tokens": 75047634.0, + "step": 2906 + }, + { + "epoch": 0.31924006149791345, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.06965970993042, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7054166197776794, + "num_tokens": 75077206.0, + "step": 2907 + }, + { + "epoch": 0.3193498792005271, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.336738109588623, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.7002944946289062, + "num_tokens": 75103531.0, + "step": 2908 + }, + { + "epoch": 0.3194596969031408, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.037144422531128, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7133669257164001, + "num_tokens": 75135218.0, + "step": 2909 + }, + { + "epoch": 0.31956951460575445, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.2554407119750977, + "learning_rate": 1e-06, + "loss": 1.0726, + "mean_token_accuracy": 0.6726482510566711, + "num_tokens": 75162189.0, + "step": 2910 + }, + { + "epoch": 0.3196793323083681, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.4616901874542236, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7268921136856079, + "num_tokens": 75183522.0, + "step": 2911 + }, + { + "epoch": 0.3197891500109818, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 2.0820541381835938, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.7042704820632935, + "num_tokens": 75211860.0, + "step": 2912 + }, + { + "epoch": 0.31989896771359544, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 1.7737488746643066, + "learning_rate": 1e-06, + "loss": 1.0499, + "mean_token_accuracy": 0.6841126680374146, + "num_tokens": 75250506.0, + "step": 2913 + }, + { + "epoch": 0.3200087854162091, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 1.9517902135849, + "learning_rate": 1e-06, + "loss": 1.1091, + "mean_token_accuracy": 0.6743314266204834, + "num_tokens": 75285783.0, + "step": 2914 + }, + { + "epoch": 0.32011860311882273, + "ewc_loss": 1.169741153717041e-06, + "grad_norm": 1.9922466278076172, + "learning_rate": 1e-06, + "loss": 1.1406, + "mean_token_accuracy": 0.6565972566604614, + "num_tokens": 75318059.0, + "step": 2915 + }, + { + "epoch": 0.32022842082143643, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.3251845836639404, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.6949347257614136, + "num_tokens": 75345039.0, + "step": 2916 + }, + { + "epoch": 0.3203382385240501, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.1223442554473877, + "learning_rate": 1e-06, + "loss": 1.0884, + "mean_token_accuracy": 0.6836878061294556, + "num_tokens": 75374543.0, + "step": 2917 + }, + { + "epoch": 0.3204480562266637, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.480969190597534, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.7024844884872437, + "num_tokens": 75397039.0, + "step": 2918 + }, + { + "epoch": 0.3205578739292774, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2619762420654297, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6906570196151733, + "num_tokens": 75422381.0, + "step": 2919 + }, + { + "epoch": 0.32066769163189107, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.3868982791900635, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7216507196426392, + "num_tokens": 75446255.0, + "step": 2920 + }, + { + "epoch": 0.3207775093345047, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.375962257385254, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7216372489929199, + "num_tokens": 75468639.0, + "step": 2921 + }, + { + "epoch": 0.32088732703711836, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.191216468811035, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7218745946884155, + "num_tokens": 75492994.0, + "step": 2922 + }, + { + "epoch": 0.32099714473973207, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.1504452228546143, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6911087036132812, + "num_tokens": 75520922.0, + "step": 2923 + }, + { + "epoch": 0.3211069624423457, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.1199464797973633, + "learning_rate": 1e-06, + "loss": 1.0678, + "mean_token_accuracy": 0.6897338032722473, + "num_tokens": 75548894.0, + "step": 2924 + }, + { + "epoch": 0.32121678014495936, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.4726510047912598, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.694390058517456, + "num_tokens": 75569737.0, + "step": 2925 + }, + { + "epoch": 0.321326597847573, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.1983649730682373, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7000510096549988, + "num_tokens": 75594651.0, + "step": 2926 + }, + { + "epoch": 0.3214364155501867, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2125234603881836, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6982830762863159, + "num_tokens": 75620646.0, + "step": 2927 + }, + { + "epoch": 0.32154623325280035, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.3457534313201904, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7142787575721741, + "num_tokens": 75642082.0, + "step": 2928 + }, + { + "epoch": 0.321656050955414, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.1039071083068848, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.6894119381904602, + "num_tokens": 75671471.0, + "step": 2929 + }, + { + "epoch": 0.3217658686580277, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.223022699356079, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6864365339279175, + "num_tokens": 75698969.0, + "step": 2930 + }, + { + "epoch": 0.32187568636064134, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.0929839611053467, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7260828018188477, + "num_tokens": 75726030.0, + "step": 2931 + }, + { + "epoch": 0.321985504063255, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.3712751865386963, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7104500532150269, + "num_tokens": 75749806.0, + "step": 2932 + }, + { + "epoch": 0.32209532176586864, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.4740803241729736, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.7025508880615234, + "num_tokens": 75774356.0, + "step": 2933 + }, + { + "epoch": 0.32220513946848234, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2149386405944824, + "learning_rate": 1e-06, + "loss": 1.0439, + "mean_token_accuracy": 0.6867733001708984, + "num_tokens": 75801477.0, + "step": 2934 + }, + { + "epoch": 0.322314957171096, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2355659008026123, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7109934687614441, + "num_tokens": 75826278.0, + "step": 2935 + }, + { + "epoch": 0.32242477487370963, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2032294273376465, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.7004952430725098, + "num_tokens": 75853384.0, + "step": 2936 + }, + { + "epoch": 0.3225345925763233, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.3715367317199707, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7160061001777649, + "num_tokens": 75875598.0, + "step": 2937 + }, + { + "epoch": 0.322644410278937, + "ewc_loss": 1.1771917343139648e-06, + "grad_norm": 2.2221779823303223, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6997705698013306, + "num_tokens": 75901014.0, + "step": 2938 + }, + { + "epoch": 0.3227542279815506, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 2.2192223072052, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7065728902816772, + "num_tokens": 75927081.0, + "step": 2939 + }, + { + "epoch": 0.32286404568416427, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 2.4901461601257324, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7376536726951599, + "num_tokens": 75947874.0, + "step": 2940 + }, + { + "epoch": 0.32297386338677797, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 2.2304649353027344, + "learning_rate": 1e-06, + "loss": 1.0863, + "mean_token_accuracy": 0.677490234375, + "num_tokens": 75974473.0, + "step": 2941 + }, + { + "epoch": 0.3230836810893916, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.4629247188568115, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7043586373329163, + "num_tokens": 75997171.0, + "step": 2942 + }, + { + "epoch": 0.32319349879200526, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.4615023136138916, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7059720754623413, + "num_tokens": 76020257.0, + "step": 2943 + }, + { + "epoch": 0.3233033164946189, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.3529677391052246, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7057569622993469, + "num_tokens": 76042908.0, + "step": 2944 + }, + { + "epoch": 0.3234131341972326, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.112067937850952, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.699813961982727, + "num_tokens": 76071996.0, + "step": 2945 + }, + { + "epoch": 0.32352295189984626, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.2891557216644287, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7223646640777588, + "num_tokens": 76095474.0, + "step": 2946 + }, + { + "epoch": 0.3236327696024599, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 2.3729817867279053, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7098990678787231, + "num_tokens": 76119182.0, + "step": 2947 + }, + { + "epoch": 0.3237425873050736, + "ewc_loss": 1.1846423149108887e-06, + "grad_norm": 2.308788537979126, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7138780355453491, + "num_tokens": 76141518.0, + "step": 2948 + }, + { + "epoch": 0.32385240500768725, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.1348984241485596, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6987345218658447, + "num_tokens": 76166749.0, + "step": 2949 + }, + { + "epoch": 0.3239622227103009, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.1189823150634766, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7210004329681396, + "num_tokens": 76194031.0, + "step": 2950 + }, + { + "epoch": 0.32407204041291454, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.2775306701660156, + "learning_rate": 1e-06, + "loss": 1.1272, + "mean_token_accuracy": 0.6716739535331726, + "num_tokens": 76220946.0, + "step": 2951 + }, + { + "epoch": 0.32418185811552824, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.2061767578125, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.7002867460250854, + "num_tokens": 76247852.0, + "step": 2952 + }, + { + "epoch": 0.3242916758181419, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.681027412414551, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7461444139480591, + "num_tokens": 76265952.0, + "step": 2953 + }, + { + "epoch": 0.32440149352075554, + "ewc_loss": 1.1920928955078125e-06, + "grad_norm": 2.09277081489563, + "learning_rate": 1e-06, + "loss": 1.0652, + "mean_token_accuracy": 0.6883766055107117, + "num_tokens": 76295479.0, + "step": 2954 + }, + { + "epoch": 0.3245113112233692, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.41926646232605, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7125416398048401, + "num_tokens": 76319832.0, + "step": 2955 + }, + { + "epoch": 0.3246211289259829, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.2466983795166016, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6980642080307007, + "num_tokens": 76346144.0, + "step": 2956 + }, + { + "epoch": 0.32473094662859653, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.1838204860687256, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7144454717636108, + "num_tokens": 76373994.0, + "step": 2957 + }, + { + "epoch": 0.3248407643312102, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.131307363510132, + "learning_rate": 1e-06, + "loss": 1.1143, + "mean_token_accuracy": 0.6681758761405945, + "num_tokens": 76404158.0, + "step": 2958 + }, + { + "epoch": 0.3249505820338239, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 1.8860008716583252, + "learning_rate": 1e-06, + "loss": 1.0586, + "mean_token_accuracy": 0.6898481845855713, + "num_tokens": 76440125.0, + "step": 2959 + }, + { + "epoch": 0.3250603997364375, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.415109634399414, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.683756947517395, + "num_tokens": 76464590.0, + "step": 2960 + }, + { + "epoch": 0.32517021743905117, + "ewc_loss": 1.1995434761047363e-06, + "grad_norm": 2.3234355449676514, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6855936050415039, + "num_tokens": 76488566.0, + "step": 2961 + }, + { + "epoch": 0.3252800351416648, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.0985450744628906, + "learning_rate": 1e-06, + "loss": 1.1039, + "mean_token_accuracy": 0.6829738616943359, + "num_tokens": 76517572.0, + "step": 2962 + }, + { + "epoch": 0.3253898528442785, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.4291269779205322, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6909065842628479, + "num_tokens": 76539776.0, + "step": 2963 + }, + { + "epoch": 0.32549967054689216, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.1814422607421875, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7206212282180786, + "num_tokens": 76566605.0, + "step": 2964 + }, + { + "epoch": 0.3256094882495058, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.206486940383911, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7111744284629822, + "num_tokens": 76591486.0, + "step": 2965 + }, + { + "epoch": 0.3257193059521195, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.4929490089416504, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7058588862419128, + "num_tokens": 76611866.0, + "step": 2966 + }, + { + "epoch": 0.32582912365473315, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.253103733062744, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6984819173812866, + "num_tokens": 76637126.0, + "step": 2967 + }, + { + "epoch": 0.3259389413573468, + "ewc_loss": 1.2069940567016602e-06, + "grad_norm": 2.451963424682617, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7000436186790466, + "num_tokens": 76659032.0, + "step": 2968 + }, + { + "epoch": 0.32604875905996045, + "ewc_loss": 1.214444637298584e-06, + "grad_norm": 2.5365471839904785, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7045325040817261, + "num_tokens": 76681382.0, + "step": 2969 + }, + { + "epoch": 0.32615857676257415, + "ewc_loss": 1.214444637298584e-06, + "grad_norm": 2.4609811305999756, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7015557289123535, + "num_tokens": 76703872.0, + "step": 2970 + }, + { + "epoch": 0.3262683944651878, + "ewc_loss": 1.214444637298584e-06, + "grad_norm": 2.103649854660034, + "learning_rate": 1e-06, + "loss": 1.1137, + "mean_token_accuracy": 0.6685163974761963, + "num_tokens": 76732208.0, + "step": 2971 + }, + { + "epoch": 0.32637821216780144, + "ewc_loss": 1.214444637298584e-06, + "grad_norm": 2.073545455932617, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6884510517120361, + "num_tokens": 76761485.0, + "step": 2972 + }, + { + "epoch": 0.3264880298704151, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.0868210792541504, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6998238563537598, + "num_tokens": 76790799.0, + "step": 2973 + }, + { + "epoch": 0.3265978475730288, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.3611929416656494, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6929032802581787, + "num_tokens": 76814140.0, + "step": 2974 + }, + { + "epoch": 0.32670766527564243, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.411022663116455, + "learning_rate": 1e-06, + "loss": 1.049, + "mean_token_accuracy": 0.6923643350601196, + "num_tokens": 76836439.0, + "step": 2975 + }, + { + "epoch": 0.3268174829782561, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.433856248855591, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7066229581832886, + "num_tokens": 76857722.0, + "step": 2976 + }, + { + "epoch": 0.3269273006808698, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.3496224880218506, + "learning_rate": 1e-06, + "loss": 1.1361, + "mean_token_accuracy": 0.6668030023574829, + "num_tokens": 76883370.0, + "step": 2977 + }, + { + "epoch": 0.3270371183834834, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.3754498958587646, + "learning_rate": 1e-06, + "loss": 1.0537, + "mean_token_accuracy": 0.6779351830482483, + "num_tokens": 76907397.0, + "step": 2978 + }, + { + "epoch": 0.3271469360860971, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.298696756362915, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6975849866867065, + "num_tokens": 76931225.0, + "step": 2979 + }, + { + "epoch": 0.3272567537887107, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.285571813583374, + "learning_rate": 1e-06, + "loss": 1.0875, + "mean_token_accuracy": 0.6711885929107666, + "num_tokens": 76956296.0, + "step": 2980 + }, + { + "epoch": 0.3273665714913244, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.252274751663208, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7147351503372192, + "num_tokens": 76979668.0, + "step": 2981 + }, + { + "epoch": 0.32747638919393807, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.588460683822632, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7010582685470581, + "num_tokens": 76999757.0, + "step": 2982 + }, + { + "epoch": 0.3275862068965517, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 1.8846925497055054, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7112076282501221, + "num_tokens": 77031664.0, + "step": 2983 + }, + { + "epoch": 0.32769602459916536, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.633525848388672, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.6992509365081787, + "num_tokens": 77050012.0, + "step": 2984 + }, + { + "epoch": 0.32780584230177906, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.453883647918701, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.6962865591049194, + "num_tokens": 77071442.0, + "step": 2985 + }, + { + "epoch": 0.3279156600043927, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.2844574451446533, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6911824941635132, + "num_tokens": 77098835.0, + "step": 2986 + }, + { + "epoch": 0.32802547770700635, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.1639795303344727, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7121633291244507, + "num_tokens": 77125012.0, + "step": 2987 + }, + { + "epoch": 0.32813529540962005, + "ewc_loss": 1.2218952178955078e-06, + "grad_norm": 2.3028764724731445, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7250732183456421, + "num_tokens": 77148158.0, + "step": 2988 + }, + { + "epoch": 0.3282451131122337, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.1900486946105957, + "learning_rate": 1e-06, + "loss": 1.0634, + "mean_token_accuracy": 0.6825750470161438, + "num_tokens": 77175883.0, + "step": 2989 + }, + { + "epoch": 0.32835493081484735, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.4601640701293945, + "learning_rate": 1e-06, + "loss": 1.1079, + "mean_token_accuracy": 0.6721547245979309, + "num_tokens": 77200164.0, + "step": 2990 + }, + { + "epoch": 0.328464748517461, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.2416391372680664, + "learning_rate": 1e-06, + "loss": 1.0912, + "mean_token_accuracy": 0.6790377497673035, + "num_tokens": 77226914.0, + "step": 2991 + }, + { + "epoch": 0.3285745662200747, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.2116072177886963, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6956503391265869, + "num_tokens": 77253262.0, + "step": 2992 + }, + { + "epoch": 0.32868438392268834, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.3242647647857666, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7307562828063965, + "num_tokens": 77280602.0, + "step": 2993 + }, + { + "epoch": 0.328794201625302, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.135622262954712, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7016308307647705, + "num_tokens": 77308632.0, + "step": 2994 + }, + { + "epoch": 0.3289040193279157, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.5358572006225586, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6956003904342651, + "num_tokens": 77331211.0, + "step": 2995 + }, + { + "epoch": 0.32901383703052933, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.11777400970459, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7162628173828125, + "num_tokens": 77358266.0, + "step": 2996 + }, + { + "epoch": 0.329123654733143, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.315871000289917, + "learning_rate": 1e-06, + "loss": 1.0871, + "mean_token_accuracy": 0.6889214515686035, + "num_tokens": 77382026.0, + "step": 2997 + }, + { + "epoch": 0.3292334724357566, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 1.9583427906036377, + "learning_rate": 1e-06, + "loss": 1.1228, + "mean_token_accuracy": 0.6723325252532959, + "num_tokens": 77420022.0, + "step": 2998 + }, + { + "epoch": 0.3293432901383703, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.024346351623535, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.6918313503265381, + "num_tokens": 77451856.0, + "step": 2999 + }, + { + "epoch": 0.32945310784098397, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.087980031967163, + "learning_rate": 1e-06, + "loss": 1.0498, + "mean_token_accuracy": 0.6831867098808289, + "num_tokens": 77481883.0, + "step": 3000 + }, + { + "epoch": 0.3295629255435976, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.6328370571136475, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7172970771789551, + "num_tokens": 77501260.0, + "step": 3001 + }, + { + "epoch": 0.32967274324621126, + "ewc_loss": 1.2293457984924316e-06, + "grad_norm": 2.4608006477355957, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.7063170671463013, + "num_tokens": 77522736.0, + "step": 3002 + }, + { + "epoch": 0.32978256094882497, + "ewc_loss": 1.2442469596862793e-06, + "grad_norm": 2.1187357902526855, + "learning_rate": 1e-06, + "loss": 1.1264, + "mean_token_accuracy": 0.6595407724380493, + "num_tokens": 77552652.0, + "step": 3003 + }, + { + "epoch": 0.3298923786514386, + "ewc_loss": 1.2442469596862793e-06, + "grad_norm": 2.266160011291504, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7249143123626709, + "num_tokens": 77577360.0, + "step": 3004 + }, + { + "epoch": 0.33000219635405226, + "ewc_loss": 1.2442469596862793e-06, + "grad_norm": 2.0401155948638916, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7103146314620972, + "num_tokens": 77606752.0, + "step": 3005 + }, + { + "epoch": 0.33011201405666596, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.403186798095703, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7100546956062317, + "num_tokens": 77628709.0, + "step": 3006 + }, + { + "epoch": 0.3302218317592796, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.14166522026062, + "learning_rate": 1e-06, + "loss": 1.0555, + "mean_token_accuracy": 0.6809656620025635, + "num_tokens": 77656252.0, + "step": 3007 + }, + { + "epoch": 0.33033164946189325, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.2465929985046387, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.7010927200317383, + "num_tokens": 77681961.0, + "step": 3008 + }, + { + "epoch": 0.3304414671645069, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.150798797607422, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7012438774108887, + "num_tokens": 77709226.0, + "step": 3009 + }, + { + "epoch": 0.3305512848671206, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.6289124488830566, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7184075713157654, + "num_tokens": 77726517.0, + "step": 3010 + }, + { + "epoch": 0.33066110256973424, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.2319223880767822, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6999556422233582, + "num_tokens": 77752521.0, + "step": 3011 + }, + { + "epoch": 0.3307709202723479, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.1935126781463623, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6977256536483765, + "num_tokens": 77779017.0, + "step": 3012 + }, + { + "epoch": 0.33088073797496154, + "ewc_loss": 1.2516975402832031e-06, + "grad_norm": 2.3204052448272705, + "learning_rate": 1e-06, + "loss": 1.0624, + "mean_token_accuracy": 0.6784282326698303, + "num_tokens": 77803179.0, + "step": 3013 + }, + { + "epoch": 0.33099055567757524, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.0968143939971924, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7154587507247925, + "num_tokens": 77830529.0, + "step": 3014 + }, + { + "epoch": 0.3311003733801889, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.3217709064483643, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6979817152023315, + "num_tokens": 77857041.0, + "step": 3015 + }, + { + "epoch": 0.33121019108280253, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.4379689693450928, + "learning_rate": 1e-06, + "loss": 1.0795, + "mean_token_accuracy": 0.6817842721939087, + "num_tokens": 77881178.0, + "step": 3016 + }, + { + "epoch": 0.33132000878541623, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.179777145385742, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6824105381965637, + "num_tokens": 77910440.0, + "step": 3017 + }, + { + "epoch": 0.3314298264880299, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.086390495300293, + "learning_rate": 1e-06, + "loss": 1.0835, + "mean_token_accuracy": 0.6774177551269531, + "num_tokens": 77941588.0, + "step": 3018 + }, + { + "epoch": 0.3315396441906435, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.481194019317627, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7132570743560791, + "num_tokens": 77962746.0, + "step": 3019 + }, + { + "epoch": 0.33164946189325717, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.463840961456299, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.68885338306427, + "num_tokens": 77985228.0, + "step": 3020 + }, + { + "epoch": 0.33175927959587087, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.1275949478149414, + "learning_rate": 1e-06, + "loss": 1.0786, + "mean_token_accuracy": 0.6757804155349731, + "num_tokens": 78015834.0, + "step": 3021 + }, + { + "epoch": 0.3318690972984845, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.4943394660949707, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.7035162448883057, + "num_tokens": 78038463.0, + "step": 3022 + }, + { + "epoch": 0.33197891500109816, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.065582275390625, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.696020245552063, + "num_tokens": 78068067.0, + "step": 3023 + }, + { + "epoch": 0.33208873270371186, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.1290171146392822, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.698434591293335, + "num_tokens": 78099716.0, + "step": 3024 + }, + { + "epoch": 0.3321985504063255, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.1966772079467773, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7297303080558777, + "num_tokens": 78123803.0, + "step": 3025 + }, + { + "epoch": 0.33230836810893916, + "ewc_loss": 1.259148120880127e-06, + "grad_norm": 2.31977915763855, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7089205980300903, + "num_tokens": 78148900.0, + "step": 3026 + }, + { + "epoch": 0.3324181858115528, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.5633597373962402, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7038150429725647, + "num_tokens": 78172082.0, + "step": 3027 + }, + { + "epoch": 0.3325280035141665, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.2632522583007812, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7151848077774048, + "num_tokens": 78197140.0, + "step": 3028 + }, + { + "epoch": 0.33263782121678015, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.009934425354004, + "learning_rate": 1e-06, + "loss": 1.077, + "mean_token_accuracy": 0.6775375008583069, + "num_tokens": 78229344.0, + "step": 3029 + }, + { + "epoch": 0.3327476389193938, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.1101863384246826, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.6961032748222351, + "num_tokens": 78257350.0, + "step": 3030 + }, + { + "epoch": 0.33285745662200744, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.0491418838500977, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7105181813240051, + "num_tokens": 78286663.0, + "step": 3031 + }, + { + "epoch": 0.33296727432462114, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.031519651412964, + "learning_rate": 1e-06, + "loss": 1.0461, + "mean_token_accuracy": 0.6884928941726685, + "num_tokens": 78318532.0, + "step": 3032 + }, + { + "epoch": 0.3330770920272348, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.2772324085235596, + "learning_rate": 1e-06, + "loss": 1.0931, + "mean_token_accuracy": 0.6727393865585327, + "num_tokens": 78342676.0, + "step": 3033 + }, + { + "epoch": 0.33318690972984844, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.344144105911255, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7048696279525757, + "num_tokens": 78364760.0, + "step": 3034 + }, + { + "epoch": 0.33329672743246214, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.3859379291534424, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6970239877700806, + "num_tokens": 78388601.0, + "step": 3035 + }, + { + "epoch": 0.3334065451350758, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.2257978916168213, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6960432529449463, + "num_tokens": 78415549.0, + "step": 3036 + }, + { + "epoch": 0.33351636283768943, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.185245990753174, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7044009566307068, + "num_tokens": 78442685.0, + "step": 3037 + }, + { + "epoch": 0.3336261805403031, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.19358229637146, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6972259283065796, + "num_tokens": 78469472.0, + "step": 3038 + }, + { + "epoch": 0.3337359982429168, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.0038506984710693, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7122461795806885, + "num_tokens": 78501465.0, + "step": 3039 + }, + { + "epoch": 0.3338458159455304, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.213136911392212, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7180505990982056, + "num_tokens": 78525782.0, + "step": 3040 + }, + { + "epoch": 0.33395563364814407, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.249453067779541, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6996239423751831, + "num_tokens": 78550247.0, + "step": 3041 + }, + { + "epoch": 0.33406545135075777, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.1487975120544434, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7056513428688049, + "num_tokens": 78576229.0, + "step": 3042 + }, + { + "epoch": 0.3341752690533714, + "ewc_loss": 1.2665987014770508e-06, + "grad_norm": 2.0883543491363525, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6965763568878174, + "num_tokens": 78603794.0, + "step": 3043 + }, + { + "epoch": 0.33428508675598506, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.4692203998565674, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.703364908695221, + "num_tokens": 78624317.0, + "step": 3044 + }, + { + "epoch": 0.3343949044585987, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.239995002746582, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.6830651760101318, + "num_tokens": 78649693.0, + "step": 3045 + }, + { + "epoch": 0.3345047221612124, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.059312582015991, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7277606725692749, + "num_tokens": 78676792.0, + "step": 3046 + }, + { + "epoch": 0.33461453986382605, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.413625717163086, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7138841152191162, + "num_tokens": 78698812.0, + "step": 3047 + }, + { + "epoch": 0.3347243575664397, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.426328659057617, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6994535326957703, + "num_tokens": 78723823.0, + "step": 3048 + }, + { + "epoch": 0.33483417526905335, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.1434242725372314, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6894775629043579, + "num_tokens": 78751359.0, + "step": 3049 + }, + { + "epoch": 0.33494399297166705, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.1287639141082764, + "learning_rate": 1e-06, + "loss": 1.1004, + "mean_token_accuracy": 0.6701425313949585, + "num_tokens": 78778325.0, + "step": 3050 + }, + { + "epoch": 0.3350538106742807, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.4605581760406494, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7212746739387512, + "num_tokens": 78800096.0, + "step": 3051 + }, + { + "epoch": 0.33516362837689434, + "ewc_loss": 1.2740492820739746e-06, + "grad_norm": 2.682978868484497, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7039918899536133, + "num_tokens": 78820036.0, + "step": 3052 + }, + { + "epoch": 0.33527344607950804, + "ewc_loss": 1.2814998626708984e-06, + "grad_norm": 2.2383790016174316, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6971730589866638, + "num_tokens": 78845369.0, + "step": 3053 + }, + { + "epoch": 0.3353832637821217, + "ewc_loss": 1.2814998626708984e-06, + "grad_norm": 2.0150094032287598, + "learning_rate": 1e-06, + "loss": 1.0943, + "mean_token_accuracy": 0.6759865880012512, + "num_tokens": 78878373.0, + "step": 3054 + }, + { + "epoch": 0.33549308148473533, + "ewc_loss": 1.2814998626708984e-06, + "grad_norm": 2.3702890872955322, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.69572514295578, + "num_tokens": 78900627.0, + "step": 3055 + }, + { + "epoch": 0.335602899187349, + "ewc_loss": 1.2814998626708984e-06, + "grad_norm": 1.938874363899231, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6919551491737366, + "num_tokens": 78932358.0, + "step": 3056 + }, + { + "epoch": 0.3357127168899627, + "ewc_loss": 1.2814998626708984e-06, + "grad_norm": 2.347860813140869, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.7018834948539734, + "num_tokens": 78954929.0, + "step": 3057 + }, + { + "epoch": 0.3358225345925763, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.0578553676605225, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7128492593765259, + "num_tokens": 78984236.0, + "step": 3058 + }, + { + "epoch": 0.33593235229519, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.1353378295898438, + "learning_rate": 1e-06, + "loss": 1.0427, + "mean_token_accuracy": 0.6947165131568909, + "num_tokens": 79013737.0, + "step": 3059 + }, + { + "epoch": 0.3360421699978036, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 1.9419927597045898, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.6796471476554871, + "num_tokens": 79047783.0, + "step": 3060 + }, + { + "epoch": 0.3361519877004173, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.4784669876098633, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.700403094291687, + "num_tokens": 79069512.0, + "step": 3061 + }, + { + "epoch": 0.33626180540303097, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.059145212173462, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7233521938323975, + "num_tokens": 79098817.0, + "step": 3062 + }, + { + "epoch": 0.3363716231056446, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.5499889850616455, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6882297992706299, + "num_tokens": 79119523.0, + "step": 3063 + }, + { + "epoch": 0.3364814408082583, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.164764642715454, + "learning_rate": 1e-06, + "loss": 1.092, + "mean_token_accuracy": 0.6761398315429688, + "num_tokens": 79148617.0, + "step": 3064 + }, + { + "epoch": 0.33659125851087196, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.1954338550567627, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7032997012138367, + "num_tokens": 79174100.0, + "step": 3065 + }, + { + "epoch": 0.3367010762134856, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.1883544921875, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7102998495101929, + "num_tokens": 79200804.0, + "step": 3066 + }, + { + "epoch": 0.33681089391609925, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.347905397415161, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7043005228042603, + "num_tokens": 79223988.0, + "step": 3067 + }, + { + "epoch": 0.33692071161871295, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.339402675628662, + "learning_rate": 1e-06, + "loss": 1.0817, + "mean_token_accuracy": 0.6795555353164673, + "num_tokens": 79248404.0, + "step": 3068 + }, + { + "epoch": 0.3370305293213266, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.0606930255889893, + "learning_rate": 1e-06, + "loss": 1.0809, + "mean_token_accuracy": 0.6859098672866821, + "num_tokens": 79277072.0, + "step": 3069 + }, + { + "epoch": 0.33714034702394025, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.159597635269165, + "learning_rate": 1e-06, + "loss": 1.0979, + "mean_token_accuracy": 0.6727623343467712, + "num_tokens": 79304757.0, + "step": 3070 + }, + { + "epoch": 0.33725016472655395, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.7496063709259033, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6898351907730103, + "num_tokens": 79339350.0, + "step": 3071 + }, + { + "epoch": 0.3373599824291676, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.095283031463623, + "learning_rate": 1e-06, + "loss": 1.0499, + "mean_token_accuracy": 0.6824795603752136, + "num_tokens": 79369818.0, + "step": 3072 + }, + { + "epoch": 0.33746980013178124, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.107057809829712, + "learning_rate": 1e-06, + "loss": 1.0866, + "mean_token_accuracy": 0.6821597218513489, + "num_tokens": 79401881.0, + "step": 3073 + }, + { + "epoch": 0.3375796178343949, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.5057103633880615, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.7040177583694458, + "num_tokens": 79420457.0, + "step": 3074 + }, + { + "epoch": 0.3376894355370086, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.1217284202575684, + "learning_rate": 1e-06, + "loss": 1.0409, + "mean_token_accuracy": 0.6997275948524475, + "num_tokens": 79450330.0, + "step": 3075 + }, + { + "epoch": 0.33779925323962223, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.201944589614868, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6984704732894897, + "num_tokens": 79477626.0, + "step": 3076 + }, + { + "epoch": 0.3379090709422359, + "ewc_loss": 1.2889504432678223e-06, + "grad_norm": 2.2915091514587402, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6955614686012268, + "num_tokens": 79500703.0, + "step": 3077 + }, + { + "epoch": 0.3380188886448495, + "ewc_loss": 1.296401023864746e-06, + "grad_norm": 2.2458508014678955, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.685283899307251, + "num_tokens": 79526708.0, + "step": 3078 + }, + { + "epoch": 0.3381287063474632, + "ewc_loss": 1.296401023864746e-06, + "grad_norm": 1.8875653743743896, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7023739814758301, + "num_tokens": 79561806.0, + "step": 3079 + }, + { + "epoch": 0.33823852405007687, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2351388931274414, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7051247954368591, + "num_tokens": 79587764.0, + "step": 3080 + }, + { + "epoch": 0.3383483417526905, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.1870312690734863, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7120970487594604, + "num_tokens": 79616144.0, + "step": 3081 + }, + { + "epoch": 0.3384581594553042, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.3675036430358887, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7089518308639526, + "num_tokens": 79638796.0, + "step": 3082 + }, + { + "epoch": 0.33856797715791787, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.269803524017334, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7194214463233948, + "num_tokens": 79662566.0, + "step": 3083 + }, + { + "epoch": 0.3386777948605315, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2118916511535645, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7057726383209229, + "num_tokens": 79691734.0, + "step": 3084 + }, + { + "epoch": 0.33878761256314516, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.082671880722046, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.7005448341369629, + "num_tokens": 79720234.0, + "step": 3085 + }, + { + "epoch": 0.33889743026575886, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.267817974090576, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6914751529693604, + "num_tokens": 79744274.0, + "step": 3086 + }, + { + "epoch": 0.3390072479683725, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.1456825733184814, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6963870525360107, + "num_tokens": 79771274.0, + "step": 3087 + }, + { + "epoch": 0.33911706567098615, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2815725803375244, + "learning_rate": 1e-06, + "loss": 1.0511, + "mean_token_accuracy": 0.6842427253723145, + "num_tokens": 79797808.0, + "step": 3088 + }, + { + "epoch": 0.3392268833735998, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.36289119720459, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7044985890388489, + "num_tokens": 79818490.0, + "step": 3089 + }, + { + "epoch": 0.3393367010762135, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.254650831222534, + "learning_rate": 1e-06, + "loss": 1.065, + "mean_token_accuracy": 0.6796472072601318, + "num_tokens": 79846912.0, + "step": 3090 + }, + { + "epoch": 0.33944651877882714, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.9106621742248535, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7165869474411011, + "num_tokens": 79862416.0, + "step": 3091 + }, + { + "epoch": 0.3395563364814408, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2779440879821777, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6884474158287048, + "num_tokens": 79888186.0, + "step": 3092 + }, + { + "epoch": 0.3396661541840545, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.21323561668396, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6932369470596313, + "num_tokens": 79914312.0, + "step": 3093 + }, + { + "epoch": 0.33977597188666814, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.35041880607605, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7087253928184509, + "num_tokens": 79938231.0, + "step": 3094 + }, + { + "epoch": 0.3398857895892818, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.11435866355896, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.715503454208374, + "num_tokens": 79965742.0, + "step": 3095 + }, + { + "epoch": 0.33999560729189543, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 1.9606573581695557, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.7001575231552124, + "num_tokens": 80000024.0, + "step": 3096 + }, + { + "epoch": 0.34010542499450913, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 1.9649851322174072, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6992607116699219, + "num_tokens": 80032041.0, + "step": 3097 + }, + { + "epoch": 0.3402152426971228, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2092196941375732, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6863404512405396, + "num_tokens": 80057633.0, + "step": 3098 + }, + { + "epoch": 0.3403250603997364, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.2383742332458496, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6972552537918091, + "num_tokens": 80083175.0, + "step": 3099 + }, + { + "epoch": 0.3404348781023501, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.112029552459717, + "learning_rate": 1e-06, + "loss": 1.0853, + "mean_token_accuracy": 0.6817188262939453, + "num_tokens": 80112222.0, + "step": 3100 + }, + { + "epoch": 0.34054469580496377, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.745687246322632, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7059088945388794, + "num_tokens": 80129502.0, + "step": 3101 + }, + { + "epoch": 0.3406545135075774, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.4396982192993164, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7131881713867188, + "num_tokens": 80152719.0, + "step": 3102 + }, + { + "epoch": 0.34076433121019106, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.1705801486968994, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7133324146270752, + "num_tokens": 80177076.0, + "step": 3103 + }, + { + "epoch": 0.34087414891280476, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.1354901790618896, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6969226598739624, + "num_tokens": 80204022.0, + "step": 3104 + }, + { + "epoch": 0.3409839666154184, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.1949989795684814, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7117913961410522, + "num_tokens": 80229727.0, + "step": 3105 + }, + { + "epoch": 0.34109378431803206, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.3169543743133545, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7202575206756592, + "num_tokens": 80251886.0, + "step": 3106 + }, + { + "epoch": 0.3412036020206457, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 1.9961434602737427, + "learning_rate": 1e-06, + "loss": 1.0592, + "mean_token_accuracy": 0.6812787652015686, + "num_tokens": 80283712.0, + "step": 3107 + }, + { + "epoch": 0.3413134197232594, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.5415759086608887, + "learning_rate": 1e-06, + "loss": 1.0809, + "mean_token_accuracy": 0.6844984292984009, + "num_tokens": 80311758.0, + "step": 3108 + }, + { + "epoch": 0.34142323742587305, + "ewc_loss": 1.30385160446167e-06, + "grad_norm": 2.421067476272583, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6962339878082275, + "num_tokens": 80333423.0, + "step": 3109 + }, + { + "epoch": 0.3415330551284867, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.330672025680542, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6829836368560791, + "num_tokens": 80358027.0, + "step": 3110 + }, + { + "epoch": 0.3416428728311004, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.1436069011688232, + "learning_rate": 1e-06, + "loss": 1.0664, + "mean_token_accuracy": 0.6987559795379639, + "num_tokens": 80385582.0, + "step": 3111 + }, + { + "epoch": 0.34175269053371404, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.1516573429107666, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7151569128036499, + "num_tokens": 80415533.0, + "step": 3112 + }, + { + "epoch": 0.3418625082363277, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.293311595916748, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.716548502445221, + "num_tokens": 80439223.0, + "step": 3113 + }, + { + "epoch": 0.34197232593894134, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.4707095623016357, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7018455266952515, + "num_tokens": 80462023.0, + "step": 3114 + }, + { + "epoch": 0.34208214364155504, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.483032703399658, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6955883502960205, + "num_tokens": 80484919.0, + "step": 3115 + }, + { + "epoch": 0.3421919613441687, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.2179055213928223, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7234747409820557, + "num_tokens": 80508170.0, + "step": 3116 + }, + { + "epoch": 0.34230177904678233, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.1561279296875, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6871172189712524, + "num_tokens": 80539222.0, + "step": 3117 + }, + { + "epoch": 0.34241159674939603, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.3063042163848877, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7118328809738159, + "num_tokens": 80562836.0, + "step": 3118 + }, + { + "epoch": 0.3425214144520097, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.1930816173553467, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7010103464126587, + "num_tokens": 80588438.0, + "step": 3119 + }, + { + "epoch": 0.3426312321546233, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.2379472255706787, + "learning_rate": 1e-06, + "loss": 1.0806, + "mean_token_accuracy": 0.6808139681816101, + "num_tokens": 80615939.0, + "step": 3120 + }, + { + "epoch": 0.34274104985723697, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.1757760047912598, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6903363466262817, + "num_tokens": 80642335.0, + "step": 3121 + }, + { + "epoch": 0.34285086755985067, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.3414294719696045, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6939042806625366, + "num_tokens": 80665781.0, + "step": 3122 + }, + { + "epoch": 0.3429606852624643, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.0861942768096924, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.698544442653656, + "num_tokens": 80696047.0, + "step": 3123 + }, + { + "epoch": 0.34307050296507796, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.4998509883880615, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7059780359268188, + "num_tokens": 80719122.0, + "step": 3124 + }, + { + "epoch": 0.3431803206676916, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.1002814769744873, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6988383531570435, + "num_tokens": 80747066.0, + "step": 3125 + }, + { + "epoch": 0.3432901383703053, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.125322103500366, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.7028168439865112, + "num_tokens": 80775529.0, + "step": 3126 + }, + { + "epoch": 0.34339995607291895, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.0600852966308594, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.689092755317688, + "num_tokens": 80806422.0, + "step": 3127 + }, + { + "epoch": 0.3435097737755326, + "ewc_loss": 1.3187527656555176e-06, + "grad_norm": 2.2486565113067627, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7009953856468201, + "num_tokens": 80829206.0, + "step": 3128 + }, + { + "epoch": 0.3436195914781463, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.018601894378662, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6913676261901855, + "num_tokens": 80859251.0, + "step": 3129 + }, + { + "epoch": 0.34372940918075995, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.5151240825653076, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7072175741195679, + "num_tokens": 80879585.0, + "step": 3130 + }, + { + "epoch": 0.3438392268833736, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.74112606048584, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.717669665813446, + "num_tokens": 80896459.0, + "step": 3131 + }, + { + "epoch": 0.34394904458598724, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.522595167160034, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.709419846534729, + "num_tokens": 80916349.0, + "step": 3132 + }, + { + "epoch": 0.34405886228860094, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.419640064239502, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7136394381523132, + "num_tokens": 80940274.0, + "step": 3133 + }, + { + "epoch": 0.3441686799912146, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.1584463119506836, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6967598795890808, + "num_tokens": 80967458.0, + "step": 3134 + }, + { + "epoch": 0.34427849769382823, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.581230640411377, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7094894647598267, + "num_tokens": 80987311.0, + "step": 3135 + }, + { + "epoch": 0.3443883153964419, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.5134174823760986, + "learning_rate": 1e-06, + "loss": 1.1241, + "mean_token_accuracy": 0.6649080514907837, + "num_tokens": 81009634.0, + "step": 3136 + }, + { + "epoch": 0.3444981330990556, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.533264636993408, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7042033672332764, + "num_tokens": 81029324.0, + "step": 3137 + }, + { + "epoch": 0.3446079508016692, + "ewc_loss": 1.3262033462524414e-06, + "grad_norm": 2.070892095565796, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6884971261024475, + "num_tokens": 81060117.0, + "step": 3138 + }, + { + "epoch": 0.3447177685042829, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 2.395587921142578, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7048508524894714, + "num_tokens": 81084163.0, + "step": 3139 + }, + { + "epoch": 0.3448275862068966, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 2.1237945556640625, + "learning_rate": 1e-06, + "loss": 1.0655, + "mean_token_accuracy": 0.6837080717086792, + "num_tokens": 81115321.0, + "step": 3140 + }, + { + "epoch": 0.3449374039095102, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 2.610933303833008, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7063993811607361, + "num_tokens": 81136307.0, + "step": 3141 + }, + { + "epoch": 0.34504722161212387, + "ewc_loss": 1.341104507446289e-06, + "grad_norm": 2.169796943664551, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6887032985687256, + "num_tokens": 81163479.0, + "step": 3142 + }, + { + "epoch": 0.3451570393147375, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.3693056106567383, + "learning_rate": 1e-06, + "loss": 1.0382, + "mean_token_accuracy": 0.6870101690292358, + "num_tokens": 81188105.0, + "step": 3143 + }, + { + "epoch": 0.3452668570173512, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.0797293186187744, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6996031999588013, + "num_tokens": 81217529.0, + "step": 3144 + }, + { + "epoch": 0.34537667471996486, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.084030866622925, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.706915020942688, + "num_tokens": 81245442.0, + "step": 3145 + }, + { + "epoch": 0.3454864924225785, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.0970354080200195, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7123250961303711, + "num_tokens": 81273374.0, + "step": 3146 + }, + { + "epoch": 0.3455963101251922, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.3641748428344727, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7055005431175232, + "num_tokens": 81299150.0, + "step": 3147 + }, + { + "epoch": 0.34570612782780585, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.3223562240600586, + "learning_rate": 1e-06, + "loss": 1.0616, + "mean_token_accuracy": 0.6843905448913574, + "num_tokens": 81323114.0, + "step": 3148 + }, + { + "epoch": 0.3458159455304195, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.5126688480377197, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7220146059989929, + "num_tokens": 81343642.0, + "step": 3149 + }, + { + "epoch": 0.34592576323303315, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.1644301414489746, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7154560089111328, + "num_tokens": 81371186.0, + "step": 3150 + }, + { + "epoch": 0.34603558093564685, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.1970040798187256, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.700875461101532, + "num_tokens": 81398073.0, + "step": 3151 + }, + { + "epoch": 0.3461453986382605, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.072843313217163, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7050430774688721, + "num_tokens": 81426255.0, + "step": 3152 + }, + { + "epoch": 0.34625521634087414, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.202394723892212, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6931907534599304, + "num_tokens": 81452329.0, + "step": 3153 + }, + { + "epoch": 0.3463650340434878, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.233534336090088, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7270544767379761, + "num_tokens": 81475750.0, + "step": 3154 + }, + { + "epoch": 0.3464748517461015, + "ewc_loss": 1.3485550880432129e-06, + "grad_norm": 2.5279810428619385, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6989414095878601, + "num_tokens": 81499098.0, + "step": 3155 + }, + { + "epoch": 0.34658466944871513, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.1710455417633057, + "learning_rate": 1e-06, + "loss": 1.032, + "mean_token_accuracy": 0.6862539052963257, + "num_tokens": 81525567.0, + "step": 3156 + }, + { + "epoch": 0.3466944871513288, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.2709298133850098, + "learning_rate": 1e-06, + "loss": 1.0779, + "mean_token_accuracy": 0.673271656036377, + "num_tokens": 81552757.0, + "step": 3157 + }, + { + "epoch": 0.3468043048539425, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.2430624961853027, + "learning_rate": 1e-06, + "loss": 1.0229, + "mean_token_accuracy": 0.6903867721557617, + "num_tokens": 81579060.0, + "step": 3158 + }, + { + "epoch": 0.3469141225565561, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.1809747219085693, + "learning_rate": 1e-06, + "loss": 1.0556, + "mean_token_accuracy": 0.6847259998321533, + "num_tokens": 81608542.0, + "step": 3159 + }, + { + "epoch": 0.34702394025916977, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 1.9876277446746826, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7030019760131836, + "num_tokens": 81639879.0, + "step": 3160 + }, + { + "epoch": 0.3471337579617834, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.1964352130889893, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7144214510917664, + "num_tokens": 81664867.0, + "step": 3161 + }, + { + "epoch": 0.3472435756643971, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.3615119457244873, + "learning_rate": 1e-06, + "loss": 1.0777, + "mean_token_accuracy": 0.6815734505653381, + "num_tokens": 81691172.0, + "step": 3162 + }, + { + "epoch": 0.34735339336701077, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 1.9612364768981934, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.7062002420425415, + "num_tokens": 81722600.0, + "step": 3163 + }, + { + "epoch": 0.3474632110696244, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.9383885860443115, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7232034206390381, + "num_tokens": 81737797.0, + "step": 3164 + }, + { + "epoch": 0.34757302877223806, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.5344955921173096, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7130652666091919, + "num_tokens": 81757397.0, + "step": 3165 + }, + { + "epoch": 0.34768284647485176, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.3858978748321533, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7181355357170105, + "num_tokens": 81778433.0, + "step": 3166 + }, + { + "epoch": 0.3477926641774654, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.0633156299591064, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6928246021270752, + "num_tokens": 81808813.0, + "step": 3167 + }, + { + "epoch": 0.34790248188007905, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.2625081539154053, + "learning_rate": 1e-06, + "loss": 1.0371, + "mean_token_accuracy": 0.683307945728302, + "num_tokens": 81835433.0, + "step": 3168 + }, + { + "epoch": 0.34801229958269275, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.552293539047241, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7076320648193359, + "num_tokens": 81855762.0, + "step": 3169 + }, + { + "epoch": 0.3481221172853064, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.3448426723480225, + "learning_rate": 1e-06, + "loss": 1.0868, + "mean_token_accuracy": 0.6733572483062744, + "num_tokens": 81880430.0, + "step": 3170 + }, + { + "epoch": 0.34823193498792004, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.353031635284424, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7152181267738342, + "num_tokens": 81903500.0, + "step": 3171 + }, + { + "epoch": 0.3483417526905337, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.111912250518799, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7136684656143188, + "num_tokens": 81935057.0, + "step": 3172 + }, + { + "epoch": 0.3484515703931474, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.4147188663482666, + "learning_rate": 1e-06, + "loss": 1.0692, + "mean_token_accuracy": 0.6767401099205017, + "num_tokens": 81959076.0, + "step": 3173 + }, + { + "epoch": 0.34856138809576104, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.4704463481903076, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7089149951934814, + "num_tokens": 81981974.0, + "step": 3174 + }, + { + "epoch": 0.3486712057983747, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 1.9547133445739746, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6861469149589539, + "num_tokens": 82015685.0, + "step": 3175 + }, + { + "epoch": 0.3487810235009884, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.486119508743286, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7081038951873779, + "num_tokens": 82037034.0, + "step": 3176 + }, + { + "epoch": 0.34889084120360203, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.3460681438446045, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7086822390556335, + "num_tokens": 82060036.0, + "step": 3177 + }, + { + "epoch": 0.3490006589062157, + "ewc_loss": 1.3634562492370605e-06, + "grad_norm": 2.3033480644226074, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.6982280611991882, + "num_tokens": 82086924.0, + "step": 3178 + }, + { + "epoch": 0.3491104766088293, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.416285753250122, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7069460153579712, + "num_tokens": 82110692.0, + "step": 3179 + }, + { + "epoch": 0.349220294311443, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.515700340270996, + "learning_rate": 1e-06, + "loss": 1.0521, + "mean_token_accuracy": 0.6904104948043823, + "num_tokens": 82133124.0, + "step": 3180 + }, + { + "epoch": 0.34933011201405667, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.104926347732544, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7355238795280457, + "num_tokens": 82160041.0, + "step": 3181 + }, + { + "epoch": 0.3494399297166703, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.416818618774414, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7057806253433228, + "num_tokens": 82183128.0, + "step": 3182 + }, + { + "epoch": 0.34954974741928396, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.343846559524536, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.705155611038208, + "num_tokens": 82206128.0, + "step": 3183 + }, + { + "epoch": 0.34965956512189766, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.4616949558258057, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7042286396026611, + "num_tokens": 82227187.0, + "step": 3184 + }, + { + "epoch": 0.3497693828245113, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.3382108211517334, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7115679383277893, + "num_tokens": 82250941.0, + "step": 3185 + }, + { + "epoch": 0.34987920052712496, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.1286611557006836, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6939529180526733, + "num_tokens": 82279220.0, + "step": 3186 + }, + { + "epoch": 0.34998901822973866, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.1212024688720703, + "learning_rate": 1e-06, + "loss": 1.0466, + "mean_token_accuracy": 0.6885417699813843, + "num_tokens": 82308794.0, + "step": 3187 + }, + { + "epoch": 0.3500988359323523, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.159079074859619, + "learning_rate": 1e-06, + "loss": 1.0229, + "mean_token_accuracy": 0.7048336267471313, + "num_tokens": 82333167.0, + "step": 3188 + }, + { + "epoch": 0.35020865363496595, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.38022780418396, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7166213989257812, + "num_tokens": 82356761.0, + "step": 3189 + }, + { + "epoch": 0.3503184713375796, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.3991503715515137, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7059474587440491, + "num_tokens": 82377746.0, + "step": 3190 + }, + { + "epoch": 0.3504282890401933, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.210280656814575, + "learning_rate": 1e-06, + "loss": 1.1178, + "mean_token_accuracy": 0.6707476377487183, + "num_tokens": 82404722.0, + "step": 3191 + }, + { + "epoch": 0.35053810674280694, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.197577714920044, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7063324451446533, + "num_tokens": 82431008.0, + "step": 3192 + }, + { + "epoch": 0.3506479244454206, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.1954917907714844, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6834632158279419, + "num_tokens": 82459447.0, + "step": 3193 + }, + { + "epoch": 0.3507577421480343, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.1605756282806396, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6898602247238159, + "num_tokens": 82485581.0, + "step": 3194 + }, + { + "epoch": 0.35086755985064794, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.04007625579834, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7048620581626892, + "num_tokens": 82512694.0, + "step": 3195 + }, + { + "epoch": 0.3509773775532616, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.23232364654541, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.715741753578186, + "num_tokens": 82537247.0, + "step": 3196 + }, + { + "epoch": 0.35108719525587523, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.4265940189361572, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7175397276878357, + "num_tokens": 82558570.0, + "step": 3197 + }, + { + "epoch": 0.35119701295848893, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 2.0987548828125, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.6944825649261475, + "num_tokens": 82587299.0, + "step": 3198 + }, + { + "epoch": 0.3513068306611026, + "ewc_loss": 1.3709068298339844e-06, + "grad_norm": 1.9457799196243286, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6927889585494995, + "num_tokens": 82619644.0, + "step": 3199 + }, + { + "epoch": 0.3514166483637162, + "ewc_loss": 1.3783574104309082e-06, + "grad_norm": 2.391611099243164, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7077536582946777, + "num_tokens": 82644335.0, + "step": 3200 + }, + { + "epoch": 0.35152646606632987, + "ewc_loss": 1.3783574104309082e-06, + "grad_norm": 2.0180141925811768, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6938952803611755, + "num_tokens": 82674506.0, + "step": 3201 + }, + { + "epoch": 0.35163628376894357, + "ewc_loss": 1.3783574104309082e-06, + "grad_norm": 2.158358097076416, + "learning_rate": 1e-06, + "loss": 1.0314, + "mean_token_accuracy": 0.687659740447998, + "num_tokens": 82702472.0, + "step": 3202 + }, + { + "epoch": 0.3517461014715572, + "ewc_loss": 1.3783574104309082e-06, + "grad_norm": 2.1038076877593994, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.68085116147995, + "num_tokens": 82730656.0, + "step": 3203 + }, + { + "epoch": 0.35185591917417086, + "ewc_loss": 1.3783574104309082e-06, + "grad_norm": 2.0567104816436768, + "learning_rate": 1e-06, + "loss": 1.0898, + "mean_token_accuracy": 0.6745163202285767, + "num_tokens": 82761816.0, + "step": 3204 + }, + { + "epoch": 0.35196573687678456, + "ewc_loss": 1.385807991027832e-06, + "grad_norm": 2.24680495262146, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7127585411071777, + "num_tokens": 82785818.0, + "step": 3205 + }, + { + "epoch": 0.3520755545793982, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.1449053287506104, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7030578851699829, + "num_tokens": 82813666.0, + "step": 3206 + }, + { + "epoch": 0.35218537228201185, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.229593276977539, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.7062896490097046, + "num_tokens": 82840296.0, + "step": 3207 + }, + { + "epoch": 0.3522951899846255, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.838440418243408, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7037827968597412, + "num_tokens": 82857720.0, + "step": 3208 + }, + { + "epoch": 0.3524050076872392, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.363729238510132, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6844642162322998, + "num_tokens": 82882989.0, + "step": 3209 + }, + { + "epoch": 0.35251482538985285, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.1605100631713867, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7256479263305664, + "num_tokens": 82908373.0, + "step": 3210 + }, + { + "epoch": 0.3526246430924665, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.0025458335876465, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7045300006866455, + "num_tokens": 82939675.0, + "step": 3211 + }, + { + "epoch": 0.35273446079508014, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.4257071018218994, + "learning_rate": 1e-06, + "loss": 1.0199, + "mean_token_accuracy": 0.7074849009513855, + "num_tokens": 82964508.0, + "step": 3212 + }, + { + "epoch": 0.35284427849769384, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.3316762447357178, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7230826616287231, + "num_tokens": 82986542.0, + "step": 3213 + }, + { + "epoch": 0.3529540962003075, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.291280508041382, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.7035000324249268, + "num_tokens": 83010692.0, + "step": 3214 + }, + { + "epoch": 0.35306391390292113, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.179323434829712, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7187856435775757, + "num_tokens": 83037047.0, + "step": 3215 + }, + { + "epoch": 0.35317373160553484, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.1434550285339355, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7002359628677368, + "num_tokens": 83063122.0, + "step": 3216 + }, + { + "epoch": 0.3532835493081485, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.125793218612671, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6967066526412964, + "num_tokens": 83091684.0, + "step": 3217 + }, + { + "epoch": 0.3533933670107621, + "ewc_loss": 1.3932585716247559e-06, + "grad_norm": 2.1013479232788086, + "learning_rate": 1e-06, + "loss": 1.0205, + "mean_token_accuracy": 0.6966531276702881, + "num_tokens": 83119244.0, + "step": 3218 + }, + { + "epoch": 0.3535031847133758, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.443225860595703, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.7109863758087158, + "num_tokens": 83141800.0, + "step": 3219 + }, + { + "epoch": 0.3536130024159895, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.140349864959717, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7003780603408813, + "num_tokens": 83170081.0, + "step": 3220 + }, + { + "epoch": 0.3537228201186031, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.5234110355377197, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7014853954315186, + "num_tokens": 83190689.0, + "step": 3221 + }, + { + "epoch": 0.35383263782121677, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.080158233642578, + "learning_rate": 1e-06, + "loss": 1.0757, + "mean_token_accuracy": 0.6767265796661377, + "num_tokens": 83223907.0, + "step": 3222 + }, + { + "epoch": 0.35394245552383047, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.48712420463562, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6992419362068176, + "num_tokens": 83245462.0, + "step": 3223 + }, + { + "epoch": 0.3540522732264441, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.4755618572235107, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7060035467147827, + "num_tokens": 83265293.0, + "step": 3224 + }, + { + "epoch": 0.35416209092905776, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.260808229446411, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7128702402114868, + "num_tokens": 83291068.0, + "step": 3225 + }, + { + "epoch": 0.3542719086316714, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.315023899078369, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6919406652450562, + "num_tokens": 83318510.0, + "step": 3226 + }, + { + "epoch": 0.3543817263342851, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 1.8869576454162598, + "learning_rate": 1e-06, + "loss": 1.0527, + "mean_token_accuracy": 0.6799083352088928, + "num_tokens": 83353576.0, + "step": 3227 + }, + { + "epoch": 0.35449154403689875, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.3917884826660156, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6945890784263611, + "num_tokens": 83376074.0, + "step": 3228 + }, + { + "epoch": 0.3546013617395124, + "ewc_loss": 1.4007091522216797e-06, + "grad_norm": 2.106966972351074, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.7005270719528198, + "num_tokens": 83401544.0, + "step": 3229 + }, + { + "epoch": 0.35471117944212605, + "ewc_loss": 1.4081597328186035e-06, + "grad_norm": 1.9295039176940918, + "learning_rate": 1e-06, + "loss": 1.0003, + "mean_token_accuracy": 0.6971008777618408, + "num_tokens": 83433777.0, + "step": 3230 + }, + { + "epoch": 0.35482099714473975, + "ewc_loss": 1.4081597328186035e-06, + "grad_norm": 2.195798873901367, + "learning_rate": 1e-06, + "loss": 1.0455, + "mean_token_accuracy": 0.6917308568954468, + "num_tokens": 83459923.0, + "step": 3231 + }, + { + "epoch": 0.3549308148473534, + "ewc_loss": 1.4081597328186035e-06, + "grad_norm": 2.2085938453674316, + "learning_rate": 1e-06, + "loss": 1.0595, + "mean_token_accuracy": 0.6783028841018677, + "num_tokens": 83487301.0, + "step": 3232 + }, + { + "epoch": 0.35504063254996704, + "ewc_loss": 1.4081597328186035e-06, + "grad_norm": 2.5112292766571045, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7101637721061707, + "num_tokens": 83507683.0, + "step": 3233 + }, + { + "epoch": 0.35515045025258074, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.2848076820373535, + "learning_rate": 1e-06, + "loss": 1.0419, + "mean_token_accuracy": 0.6832873225212097, + "num_tokens": 83533092.0, + "step": 3234 + }, + { + "epoch": 0.3552602679551944, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.150575876235962, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6871753334999084, + "num_tokens": 83562221.0, + "step": 3235 + }, + { + "epoch": 0.35537008565780803, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.2054672241210938, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7051358819007874, + "num_tokens": 83587927.0, + "step": 3236 + }, + { + "epoch": 0.3554799033604217, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 1.9565138816833496, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6954506039619446, + "num_tokens": 83619553.0, + "step": 3237 + }, + { + "epoch": 0.3555897210630354, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 1.9370802640914917, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7240462303161621, + "num_tokens": 83652859.0, + "step": 3238 + }, + { + "epoch": 0.355699538765649, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.151906728744507, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6937466859817505, + "num_tokens": 83679314.0, + "step": 3239 + }, + { + "epoch": 0.35580935646826267, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.161965847015381, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.6974312663078308, + "num_tokens": 83704588.0, + "step": 3240 + }, + { + "epoch": 0.3559191741708763, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.027708053588867, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7024625539779663, + "num_tokens": 83732073.0, + "step": 3241 + }, + { + "epoch": 0.35602899187349, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.0699639320373535, + "learning_rate": 1e-06, + "loss": 1.0589, + "mean_token_accuracy": 0.6819888353347778, + "num_tokens": 83761233.0, + "step": 3242 + }, + { + "epoch": 0.35613880957610367, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.241053819656372, + "learning_rate": 1e-06, + "loss": 1.065, + "mean_token_accuracy": 0.6750611066818237, + "num_tokens": 83789309.0, + "step": 3243 + }, + { + "epoch": 0.3562486272787173, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.092334747314453, + "learning_rate": 1e-06, + "loss": 1.0773, + "mean_token_accuracy": 0.6842731237411499, + "num_tokens": 83822253.0, + "step": 3244 + }, + { + "epoch": 0.356358444981331, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.0586459636688232, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7059088945388794, + "num_tokens": 83850718.0, + "step": 3245 + }, + { + "epoch": 0.35646826268394466, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.257225751876831, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6975607872009277, + "num_tokens": 83874658.0, + "step": 3246 + }, + { + "epoch": 0.3565780803865583, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.2677125930786133, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7214982509613037, + "num_tokens": 83899291.0, + "step": 3247 + }, + { + "epoch": 0.35668789808917195, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.223809003829956, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7215734720230103, + "num_tokens": 83922590.0, + "step": 3248 + }, + { + "epoch": 0.35679771579178565, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.1017191410064697, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.7004694938659668, + "num_tokens": 83950687.0, + "step": 3249 + }, + { + "epoch": 0.3569075334943993, + "ewc_loss": 1.4156103134155273e-06, + "grad_norm": 2.176771402359009, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6959123611450195, + "num_tokens": 83977564.0, + "step": 3250 + }, + { + "epoch": 0.35701735119701294, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.3623929023742676, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6824737787246704, + "num_tokens": 84003184.0, + "step": 3251 + }, + { + "epoch": 0.35712716889962665, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.05708384513855, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.7009615898132324, + "num_tokens": 84031308.0, + "step": 3252 + }, + { + "epoch": 0.3572369866022403, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.231396198272705, + "learning_rate": 1e-06, + "loss": 1.0627, + "mean_token_accuracy": 0.6766314506530762, + "num_tokens": 84057005.0, + "step": 3253 + }, + { + "epoch": 0.35734680430485394, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.211987257003784, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6870189309120178, + "num_tokens": 84084931.0, + "step": 3254 + }, + { + "epoch": 0.3574566220074676, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.3699963092803955, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6833839416503906, + "num_tokens": 84108380.0, + "step": 3255 + }, + { + "epoch": 0.3575664397100813, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.4639174938201904, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7215898633003235, + "num_tokens": 84130087.0, + "step": 3256 + }, + { + "epoch": 0.35767625741269493, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.1944048404693604, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7056170105934143, + "num_tokens": 84154892.0, + "step": 3257 + }, + { + "epoch": 0.3577860751153086, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.4236531257629395, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.709262490272522, + "num_tokens": 84176496.0, + "step": 3258 + }, + { + "epoch": 0.3578958928179222, + "ewc_loss": 1.4230608940124512e-06, + "grad_norm": 2.412285089492798, + "learning_rate": 1e-06, + "loss": 1.0417, + "mean_token_accuracy": 0.6881706714630127, + "num_tokens": 84200797.0, + "step": 3259 + }, + { + "epoch": 0.3580057105205359, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2445242404937744, + "learning_rate": 1e-06, + "loss": 1.088, + "mean_token_accuracy": 0.6885926127433777, + "num_tokens": 84226155.0, + "step": 3260 + }, + { + "epoch": 0.35811552822314957, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.3691515922546387, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7181475162506104, + "num_tokens": 84249673.0, + "step": 3261 + }, + { + "epoch": 0.3582253459257632, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.344860792160034, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6997228860855103, + "num_tokens": 84273724.0, + "step": 3262 + }, + { + "epoch": 0.3583351636283769, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.383526086807251, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6925269961357117, + "num_tokens": 84299628.0, + "step": 3263 + }, + { + "epoch": 0.35844498133099056, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2505605220794678, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7123591899871826, + "num_tokens": 84327527.0, + "step": 3264 + }, + { + "epoch": 0.3585547990336042, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2703914642333984, + "learning_rate": 1e-06, + "loss": 1.0495, + "mean_token_accuracy": 0.6907671093940735, + "num_tokens": 84354889.0, + "step": 3265 + }, + { + "epoch": 0.35866461673621786, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2742111682891846, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6986510157585144, + "num_tokens": 84380522.0, + "step": 3266 + }, + { + "epoch": 0.35877443443883156, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2500319480895996, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7130048274993896, + "num_tokens": 84405266.0, + "step": 3267 + }, + { + "epoch": 0.3588842521414452, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2959346771240234, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.718845784664154, + "num_tokens": 84428101.0, + "step": 3268 + }, + { + "epoch": 0.35899406984405885, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.2397849559783936, + "learning_rate": 1e-06, + "loss": 1.0807, + "mean_token_accuracy": 0.6799536943435669, + "num_tokens": 84454829.0, + "step": 3269 + }, + { + "epoch": 0.35910388754667255, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.357168197631836, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7254152297973633, + "num_tokens": 84475796.0, + "step": 3270 + }, + { + "epoch": 0.3592137052492862, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 1.991790771484375, + "learning_rate": 1e-06, + "loss": 1.0742, + "mean_token_accuracy": 0.6749968528747559, + "num_tokens": 84506712.0, + "step": 3271 + }, + { + "epoch": 0.35932352295189984, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.5039234161376953, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7074485421180725, + "num_tokens": 84527613.0, + "step": 3272 + }, + { + "epoch": 0.3594333406545135, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 2.373274564743042, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6953868269920349, + "num_tokens": 84552379.0, + "step": 3273 + }, + { + "epoch": 0.3595431583571272, + "ewc_loss": 1.430511474609375e-06, + "grad_norm": 1.9710752964019775, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6813086867332458, + "num_tokens": 84583957.0, + "step": 3274 + }, + { + "epoch": 0.35965297605974084, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.135657787322998, + "learning_rate": 1e-06, + "loss": 1.0803, + "mean_token_accuracy": 0.673725962638855, + "num_tokens": 84616061.0, + "step": 3275 + }, + { + "epoch": 0.3597627937623545, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.3479344844818115, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6948882937431335, + "num_tokens": 84639966.0, + "step": 3276 + }, + { + "epoch": 0.35987261146496813, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.0822482109069824, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.7015623450279236, + "num_tokens": 84670805.0, + "step": 3277 + }, + { + "epoch": 0.35998242916758183, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.3256993293762207, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6967016458511353, + "num_tokens": 84695492.0, + "step": 3278 + }, + { + "epoch": 0.3600922468701955, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.218970775604248, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7238775491714478, + "num_tokens": 84721170.0, + "step": 3279 + }, + { + "epoch": 0.3602020645728091, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.2151520252227783, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6936278343200684, + "num_tokens": 84746589.0, + "step": 3280 + }, + { + "epoch": 0.3603118822754228, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.267141103744507, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.697078287601471, + "num_tokens": 84770710.0, + "step": 3281 + }, + { + "epoch": 0.36042169997803647, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 1.9019688367843628, + "learning_rate": 1e-06, + "loss": 1.1261, + "mean_token_accuracy": 0.6658481359481812, + "num_tokens": 84806101.0, + "step": 3282 + }, + { + "epoch": 0.3605315176806501, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.0550639629364014, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7101039290428162, + "num_tokens": 84835710.0, + "step": 3283 + }, + { + "epoch": 0.36064133538326376, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.265892267227173, + "learning_rate": 1e-06, + "loss": 1.0349, + "mean_token_accuracy": 0.6934653520584106, + "num_tokens": 84863843.0, + "step": 3284 + }, + { + "epoch": 0.36075115308587746, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.362687587738037, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.687382698059082, + "num_tokens": 84886682.0, + "step": 3285 + }, + { + "epoch": 0.3608609707884911, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.5642125606536865, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6946671009063721, + "num_tokens": 84907655.0, + "step": 3286 + }, + { + "epoch": 0.36097078849110475, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.0819804668426514, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6943289041519165, + "num_tokens": 84936194.0, + "step": 3287 + }, + { + "epoch": 0.3610806061937184, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.0450053215026855, + "learning_rate": 1e-06, + "loss": 1.0477, + "mean_token_accuracy": 0.6862346529960632, + "num_tokens": 84964512.0, + "step": 3288 + }, + { + "epoch": 0.3611904238963321, + "ewc_loss": 1.4379620552062988e-06, + "grad_norm": 2.219675302505493, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6984397172927856, + "num_tokens": 84991502.0, + "step": 3289 + }, + { + "epoch": 0.36130024159894575, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.4445440769195557, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6889675855636597, + "num_tokens": 85013129.0, + "step": 3290 + }, + { + "epoch": 0.3614100593015594, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.333665370941162, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7158804535865784, + "num_tokens": 85036620.0, + "step": 3291 + }, + { + "epoch": 0.3615198770041731, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.3757686614990234, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6941759586334229, + "num_tokens": 85060171.0, + "step": 3292 + }, + { + "epoch": 0.36162969470678674, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.170224189758301, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.686066210269928, + "num_tokens": 85086553.0, + "step": 3293 + }, + { + "epoch": 0.3617395124094004, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.0280752182006836, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6935164928436279, + "num_tokens": 85118680.0, + "step": 3294 + }, + { + "epoch": 0.36184933011201403, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.3083975315093994, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7130761742591858, + "num_tokens": 85142298.0, + "step": 3295 + }, + { + "epoch": 0.36195914781462774, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.1004929542541504, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7115733027458191, + "num_tokens": 85170484.0, + "step": 3296 + }, + { + "epoch": 0.3620689655172414, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.169184923171997, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7196310758590698, + "num_tokens": 85196017.0, + "step": 3297 + }, + { + "epoch": 0.362178783219855, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.403219223022461, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.726292610168457, + "num_tokens": 85216845.0, + "step": 3298 + }, + { + "epoch": 0.36228860092246873, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.1418514251708984, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7120457887649536, + "num_tokens": 85243824.0, + "step": 3299 + }, + { + "epoch": 0.3623984186250824, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.14597225189209, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.7063437104225159, + "num_tokens": 85274198.0, + "step": 3300 + }, + { + "epoch": 0.362508236327696, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.1507434844970703, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6989949941635132, + "num_tokens": 85303239.0, + "step": 3301 + }, + { + "epoch": 0.36261805403030967, + "ewc_loss": 1.4454126358032227e-06, + "grad_norm": 2.04390549659729, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7076776027679443, + "num_tokens": 85330757.0, + "step": 3302 + }, + { + "epoch": 0.36272787173292337, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 2.3154549598693848, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7329286336898804, + "num_tokens": 85354065.0, + "step": 3303 + }, + { + "epoch": 0.362837689435537, + "ewc_loss": 1.4528632164001465e-06, + "grad_norm": 2.2730658054351807, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7395457029342651, + "num_tokens": 85377501.0, + "step": 3304 + }, + { + "epoch": 0.36294750713815066, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 2.159496545791626, + "learning_rate": 1e-06, + "loss": 1.0222, + "mean_token_accuracy": 0.6938391923904419, + "num_tokens": 85405009.0, + "step": 3305 + }, + { + "epoch": 0.3630573248407643, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 2.473567485809326, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7236312031745911, + "num_tokens": 85425905.0, + "step": 3306 + }, + { + "epoch": 0.363167142543378, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 2.241488456726074, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7078052759170532, + "num_tokens": 85452029.0, + "step": 3307 + }, + { + "epoch": 0.36327696024599165, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.337055206298828, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.7038164734840393, + "num_tokens": 85478947.0, + "step": 3308 + }, + { + "epoch": 0.3633867779486053, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.1112723350524902, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6987768411636353, + "num_tokens": 85510887.0, + "step": 3309 + }, + { + "epoch": 0.363496595651219, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.2343358993530273, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6933065056800842, + "num_tokens": 85539105.0, + "step": 3310 + }, + { + "epoch": 0.36360641335383265, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.3926830291748047, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7171828150749207, + "num_tokens": 85561827.0, + "step": 3311 + }, + { + "epoch": 0.3637162310564463, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 1.9803824424743652, + "learning_rate": 1e-06, + "loss": 1.0912, + "mean_token_accuracy": 0.6726439595222473, + "num_tokens": 85594193.0, + "step": 3312 + }, + { + "epoch": 0.36382604875905994, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.2040822505950928, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6936364769935608, + "num_tokens": 85619805.0, + "step": 3313 + }, + { + "epoch": 0.36393586646167364, + "ewc_loss": 1.475214958190918e-06, + "grad_norm": 2.580360174179077, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7027627229690552, + "num_tokens": 85640275.0, + "step": 3314 + }, + { + "epoch": 0.3640456841642873, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 2.293313980102539, + "learning_rate": 1e-06, + "loss": 1.0497, + "mean_token_accuracy": 0.6897977590560913, + "num_tokens": 85665126.0, + "step": 3315 + }, + { + "epoch": 0.36415550186690093, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 2.4027180671691895, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.702477216720581, + "num_tokens": 85689650.0, + "step": 3316 + }, + { + "epoch": 0.3642653195695146, + "ewc_loss": 1.4603137969970703e-06, + "grad_norm": 1.9873603582382202, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6945915222167969, + "num_tokens": 85720567.0, + "step": 3317 + }, + { + "epoch": 0.3643751372721283, + "ewc_loss": 1.4677643775939941e-06, + "grad_norm": 2.360926628112793, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.6934217810630798, + "num_tokens": 85744555.0, + "step": 3318 + }, + { + "epoch": 0.3644849549747419, + "ewc_loss": 1.4677643775939941e-06, + "grad_norm": 2.1746654510498047, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7008110284805298, + "num_tokens": 85769197.0, + "step": 3319 + }, + { + "epoch": 0.36459477267735557, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.1705429553985596, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7071452140808105, + "num_tokens": 85794927.0, + "step": 3320 + }, + { + "epoch": 0.3647045903799693, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.203207492828369, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7050456404685974, + "num_tokens": 85820362.0, + "step": 3321 + }, + { + "epoch": 0.3648144080825829, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.3963277339935303, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7237083911895752, + "num_tokens": 85841839.0, + "step": 3322 + }, + { + "epoch": 0.36492422578519657, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.216465473175049, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6904586553573608, + "num_tokens": 85867377.0, + "step": 3323 + }, + { + "epoch": 0.3650340434878102, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.2200703620910645, + "learning_rate": 1e-06, + "loss": 1.051, + "mean_token_accuracy": 0.6816987991333008, + "num_tokens": 85894728.0, + "step": 3324 + }, + { + "epoch": 0.3651438611904239, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.213820695877075, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7150346636772156, + "num_tokens": 85920460.0, + "step": 3325 + }, + { + "epoch": 0.36525367889303756, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.214717388153076, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6901756525039673, + "num_tokens": 85947417.0, + "step": 3326 + }, + { + "epoch": 0.3653634965956512, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.060295820236206, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6875066757202148, + "num_tokens": 85978475.0, + "step": 3327 + }, + { + "epoch": 0.3654733142982649, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.1115448474884033, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.6960728168487549, + "num_tokens": 86005268.0, + "step": 3328 + }, + { + "epoch": 0.36558313200087855, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.1786270141601562, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7127970457077026, + "num_tokens": 86032447.0, + "step": 3329 + }, + { + "epoch": 0.3656929497034922, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.3780651092529297, + "learning_rate": 1e-06, + "loss": 1.0439, + "mean_token_accuracy": 0.6848125457763672, + "num_tokens": 86055965.0, + "step": 3330 + }, + { + "epoch": 0.36580276740610584, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.155226707458496, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6789063811302185, + "num_tokens": 86083407.0, + "step": 3331 + }, + { + "epoch": 0.36591258510871955, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.3414113521575928, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.706753134727478, + "num_tokens": 86107077.0, + "step": 3332 + }, + { + "epoch": 0.3660224028113332, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.3283638954162598, + "learning_rate": 1e-06, + "loss": 1.0545, + "mean_token_accuracy": 0.6920875310897827, + "num_tokens": 86131381.0, + "step": 3333 + }, + { + "epoch": 0.36613222051394684, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.4451379776000977, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6928246021270752, + "num_tokens": 86154122.0, + "step": 3334 + }, + { + "epoch": 0.3662420382165605, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.279900550842285, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.6936521530151367, + "num_tokens": 86180267.0, + "step": 3335 + }, + { + "epoch": 0.3663518559191742, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 1.9151664972305298, + "learning_rate": 1e-06, + "loss": 1.0754, + "mean_token_accuracy": 0.6790177226066589, + "num_tokens": 86215594.0, + "step": 3336 + }, + { + "epoch": 0.36646167362178783, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 1.996935248374939, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6917663812637329, + "num_tokens": 86246071.0, + "step": 3337 + }, + { + "epoch": 0.3665714913244015, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.0914931297302246, + "learning_rate": 1e-06, + "loss": 1.0993, + "mean_token_accuracy": 0.671781063079834, + "num_tokens": 86277975.0, + "step": 3338 + }, + { + "epoch": 0.3666813090270152, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.1429481506347656, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6960321664810181, + "num_tokens": 86304784.0, + "step": 3339 + }, + { + "epoch": 0.3667911267296288, + "ewc_loss": 1.4826655387878418e-06, + "grad_norm": 2.14609432220459, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7122557163238525, + "num_tokens": 86334727.0, + "step": 3340 + }, + { + "epoch": 0.36690094443224247, + "ewc_loss": 1.4901161193847656e-06, + "grad_norm": 2.275132417678833, + "learning_rate": 1e-06, + "loss": 1.0835, + "mean_token_accuracy": 0.6954752206802368, + "num_tokens": 86360301.0, + "step": 3341 + }, + { + "epoch": 0.3670107621348561, + "ewc_loss": 1.4901161193847656e-06, + "grad_norm": 2.175365686416626, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7118746042251587, + "num_tokens": 86384913.0, + "step": 3342 + }, + { + "epoch": 0.3671205798374698, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.1875898838043213, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7025787830352783, + "num_tokens": 86410825.0, + "step": 3343 + }, + { + "epoch": 0.36723039754008346, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.2098422050476074, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6946227550506592, + "num_tokens": 86436639.0, + "step": 3344 + }, + { + "epoch": 0.3673402152426971, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.096780300140381, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.7005317807197571, + "num_tokens": 86465164.0, + "step": 3345 + }, + { + "epoch": 0.3674500329453108, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.381916046142578, + "learning_rate": 1e-06, + "loss": 1.0255, + "mean_token_accuracy": 0.6943659782409668, + "num_tokens": 86487983.0, + "step": 3346 + }, + { + "epoch": 0.36755985064792446, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.0555949211120605, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7063257694244385, + "num_tokens": 86515243.0, + "step": 3347 + }, + { + "epoch": 0.3676696683505381, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.287844657897949, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7097314596176147, + "num_tokens": 86540172.0, + "step": 3348 + }, + { + "epoch": 0.36777948605315175, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.048884868621826, + "learning_rate": 1e-06, + "loss": 1.0537, + "mean_token_accuracy": 0.6819790601730347, + "num_tokens": 86571138.0, + "step": 3349 + }, + { + "epoch": 0.36788930375576545, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 1.9816930294036865, + "learning_rate": 1e-06, + "loss": 1.0525, + "mean_token_accuracy": 0.68671715259552, + "num_tokens": 86604797.0, + "step": 3350 + }, + { + "epoch": 0.3679991214583791, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.2431962490081787, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7138055562973022, + "num_tokens": 86628636.0, + "step": 3351 + }, + { + "epoch": 0.36810893916099274, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 1.8797693252563477, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.695656418800354, + "num_tokens": 86663120.0, + "step": 3352 + }, + { + "epoch": 0.3682187568636064, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.1440234184265137, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7188817262649536, + "num_tokens": 86689552.0, + "step": 3353 + }, + { + "epoch": 0.3683285745662201, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.413870334625244, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.699663519859314, + "num_tokens": 86712863.0, + "step": 3354 + }, + { + "epoch": 0.36843839226883374, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.322556972503662, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.693917989730835, + "num_tokens": 86736563.0, + "step": 3355 + }, + { + "epoch": 0.3685482099714474, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.087986946105957, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.704350471496582, + "num_tokens": 86764256.0, + "step": 3356 + }, + { + "epoch": 0.3686580276740611, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.2960431575775146, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7190803289413452, + "num_tokens": 86787797.0, + "step": 3357 + }, + { + "epoch": 0.36876784537667473, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.1034398078918457, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7146272659301758, + "num_tokens": 86815390.0, + "step": 3358 + }, + { + "epoch": 0.3688776630792884, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 1.9821324348449707, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7344385981559753, + "num_tokens": 86846028.0, + "step": 3359 + }, + { + "epoch": 0.368987480781902, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.4188430309295654, + "learning_rate": 1e-06, + "loss": 1.063, + "mean_token_accuracy": 0.6835401654243469, + "num_tokens": 86869114.0, + "step": 3360 + }, + { + "epoch": 0.3690972984845157, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.3546717166900635, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6954933404922485, + "num_tokens": 86893563.0, + "step": 3361 + }, + { + "epoch": 0.36920711618712937, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.3532276153564453, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.704281210899353, + "num_tokens": 86916057.0, + "step": 3362 + }, + { + "epoch": 0.369316933889743, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.163175582885742, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6880592107772827, + "num_tokens": 86944193.0, + "step": 3363 + }, + { + "epoch": 0.36942675159235666, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.131241798400879, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6961912512779236, + "num_tokens": 86973130.0, + "step": 3364 + }, + { + "epoch": 0.36953656929497036, + "ewc_loss": 1.5050172805786133e-06, + "grad_norm": 2.3371849060058594, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7178163528442383, + "num_tokens": 86997045.0, + "step": 3365 + }, + { + "epoch": 0.369646386997584, + "ewc_loss": 1.5050172805786133e-06, + "grad_norm": 2.2932851314544678, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.6900418400764465, + "num_tokens": 87021295.0, + "step": 3366 + }, + { + "epoch": 0.36975620470019765, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.14310622215271, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.7020881772041321, + "num_tokens": 87046413.0, + "step": 3367 + }, + { + "epoch": 0.36986602240281136, + "ewc_loss": 1.4975666999816895e-06, + "grad_norm": 2.3587210178375244, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6941185593605042, + "num_tokens": 87073882.0, + "step": 3368 + }, + { + "epoch": 0.369975840105425, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.276621103286743, + "learning_rate": 1e-06, + "loss": 1.0762, + "mean_token_accuracy": 0.6861287951469421, + "num_tokens": 87100385.0, + "step": 3369 + }, + { + "epoch": 0.37008565780803865, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.2396152019500732, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7147925496101379, + "num_tokens": 87126593.0, + "step": 3370 + }, + { + "epoch": 0.3701954755106523, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 1.8466333150863647, + "learning_rate": 1e-06, + "loss": 1.0385, + "mean_token_accuracy": 0.6893649697303772, + "num_tokens": 87161071.0, + "step": 3371 + }, + { + "epoch": 0.370305293213266, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.2018887996673584, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6974967122077942, + "num_tokens": 87186272.0, + "step": 3372 + }, + { + "epoch": 0.37041511091587964, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.1997835636138916, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.709932804107666, + "num_tokens": 87212120.0, + "step": 3373 + }, + { + "epoch": 0.3705249286184933, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.563098192214966, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7099292278289795, + "num_tokens": 87231883.0, + "step": 3374 + }, + { + "epoch": 0.370634746321107, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.0646212100982666, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6859868168830872, + "num_tokens": 87260370.0, + "step": 3375 + }, + { + "epoch": 0.37074456402372064, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.1306822299957275, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6968222856521606, + "num_tokens": 87286894.0, + "step": 3376 + }, + { + "epoch": 0.3708543817263343, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 1.9706135988235474, + "learning_rate": 1e-06, + "loss": 1.0604, + "mean_token_accuracy": 0.684005618095398, + "num_tokens": 87321398.0, + "step": 3377 + }, + { + "epoch": 0.3709641994289479, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.4813594818115234, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7162728905677795, + "num_tokens": 87345279.0, + "step": 3378 + }, + { + "epoch": 0.37107401713156163, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.5293972492218018, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7054375410079956, + "num_tokens": 87369573.0, + "step": 3379 + }, + { + "epoch": 0.3711838348341753, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.3173131942749023, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6841247081756592, + "num_tokens": 87396176.0, + "step": 3380 + }, + { + "epoch": 0.3712936525367889, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 1.9720308780670166, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.6969636678695679, + "num_tokens": 87426871.0, + "step": 3381 + }, + { + "epoch": 0.37140347023940257, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.195787191390991, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.7034674882888794, + "num_tokens": 87453349.0, + "step": 3382 + }, + { + "epoch": 0.37151328794201627, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 1.9631142616271973, + "learning_rate": 1e-06, + "loss": 1.1008, + "mean_token_accuracy": 0.6972465515136719, + "num_tokens": 87484742.0, + "step": 3383 + }, + { + "epoch": 0.3716231056446299, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.445021152496338, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7154374122619629, + "num_tokens": 87506704.0, + "step": 3384 + }, + { + "epoch": 0.37173292334724356, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.3960976600646973, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.6931456327438354, + "num_tokens": 87529839.0, + "step": 3385 + }, + { + "epoch": 0.37184274104985726, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.168074607849121, + "learning_rate": 1e-06, + "loss": 1.0457, + "mean_token_accuracy": 0.692207932472229, + "num_tokens": 87558485.0, + "step": 3386 + }, + { + "epoch": 0.3719525587524709, + "ewc_loss": 1.5124678611755371e-06, + "grad_norm": 2.2079384326934814, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7076468467712402, + "num_tokens": 87586352.0, + "step": 3387 + }, + { + "epoch": 0.37206237645508455, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.161755323410034, + "learning_rate": 1e-06, + "loss": 1.0861, + "mean_token_accuracy": 0.6890953183174133, + "num_tokens": 87614672.0, + "step": 3388 + }, + { + "epoch": 0.3721721941576982, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.0339035987854004, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7139113545417786, + "num_tokens": 87643313.0, + "step": 3389 + }, + { + "epoch": 0.3722820118603119, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.1629638671875, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6978778839111328, + "num_tokens": 87670451.0, + "step": 3390 + }, + { + "epoch": 0.37239182956292555, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.483248472213745, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6862696409225464, + "num_tokens": 87692044.0, + "step": 3391 + }, + { + "epoch": 0.3725016472655392, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 1.9302419424057007, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7029104232788086, + "num_tokens": 87721089.0, + "step": 3392 + }, + { + "epoch": 0.37261146496815284, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.3148419857025146, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.7027401328086853, + "num_tokens": 87743663.0, + "step": 3393 + }, + { + "epoch": 0.37272128267076654, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.4928016662597656, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7109178304672241, + "num_tokens": 87764986.0, + "step": 3394 + }, + { + "epoch": 0.3728311003733802, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.252992868423462, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7079851627349854, + "num_tokens": 87790380.0, + "step": 3395 + }, + { + "epoch": 0.37294091807599383, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.316175699234009, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6909728050231934, + "num_tokens": 87814012.0, + "step": 3396 + }, + { + "epoch": 0.37305073577860753, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.308821439743042, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6912773251533508, + "num_tokens": 87838307.0, + "step": 3397 + }, + { + "epoch": 0.3731605534812212, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.2154459953308105, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.702957034111023, + "num_tokens": 87864332.0, + "step": 3398 + }, + { + "epoch": 0.3732703711838348, + "ewc_loss": 1.519918441772461e-06, + "grad_norm": 2.660856008529663, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6938245892524719, + "num_tokens": 87884903.0, + "step": 3399 + }, + { + "epoch": 0.37338018888644847, + "ewc_loss": 1.5273690223693848e-06, + "grad_norm": 2.268151044845581, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6972695589065552, + "num_tokens": 87908543.0, + "step": 3400 + }, + { + "epoch": 0.3734900065890622, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.4076128005981445, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7084622383117676, + "num_tokens": 87930022.0, + "step": 3401 + }, + { + "epoch": 0.3735998242916758, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 1.955848217010498, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6927512884140015, + "num_tokens": 87961748.0, + "step": 3402 + }, + { + "epoch": 0.37370964199428947, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 1.8798149824142456, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7070179581642151, + "num_tokens": 87995355.0, + "step": 3403 + }, + { + "epoch": 0.37381945969690317, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 1.9577784538269043, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6814521551132202, + "num_tokens": 88026996.0, + "step": 3404 + }, + { + "epoch": 0.3739292773995168, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.1722326278686523, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7216209173202515, + "num_tokens": 88052140.0, + "step": 3405 + }, + { + "epoch": 0.37403909510213046, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.4191782474517822, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6892294883728027, + "num_tokens": 88073640.0, + "step": 3406 + }, + { + "epoch": 0.3741489128047441, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.291124105453491, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.7001964449882507, + "num_tokens": 88096965.0, + "step": 3407 + }, + { + "epoch": 0.3742587305073578, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 1.9167635440826416, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6925223469734192, + "num_tokens": 88128966.0, + "step": 3408 + }, + { + "epoch": 0.37436854820997145, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.21749210357666, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.712769627571106, + "num_tokens": 88152934.0, + "step": 3409 + }, + { + "epoch": 0.3744783659125851, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.53452730178833, + "learning_rate": 1e-06, + "loss": 1.056, + "mean_token_accuracy": 0.6852321028709412, + "num_tokens": 88175382.0, + "step": 3410 + }, + { + "epoch": 0.37458818361519874, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.1173062324523926, + "learning_rate": 1e-06, + "loss": 1.0822, + "mean_token_accuracy": 0.6862820386886597, + "num_tokens": 88204938.0, + "step": 3411 + }, + { + "epoch": 0.37469800131781245, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.447559118270874, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7007615566253662, + "num_tokens": 88226259.0, + "step": 3412 + }, + { + "epoch": 0.3748078190204261, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.229199171066284, + "learning_rate": 1e-06, + "loss": 1.0675, + "mean_token_accuracy": 0.6743203401565552, + "num_tokens": 88251893.0, + "step": 3413 + }, + { + "epoch": 0.37491763672303974, + "ewc_loss": 1.5422701835632324e-06, + "grad_norm": 2.221276044845581, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7067379355430603, + "num_tokens": 88277786.0, + "step": 3414 + }, + { + "epoch": 0.37502745442565344, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.44069766998291, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.7040338516235352, + "num_tokens": 88300143.0, + "step": 3415 + }, + { + "epoch": 0.3751372721282671, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.420459747314453, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7043777704238892, + "num_tokens": 88323457.0, + "step": 3416 + }, + { + "epoch": 0.37524708983088073, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.594371795654297, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7141398191452026, + "num_tokens": 88343643.0, + "step": 3417 + }, + { + "epoch": 0.3753569075334944, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.0349884033203125, + "learning_rate": 1e-06, + "loss": 1.0824, + "mean_token_accuracy": 0.6753599047660828, + "num_tokens": 88373513.0, + "step": 3418 + }, + { + "epoch": 0.3754667252361081, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.0868582725524902, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6924095153808594, + "num_tokens": 88402245.0, + "step": 3419 + }, + { + "epoch": 0.3755765429387217, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.125232696533203, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6929936408996582, + "num_tokens": 88430919.0, + "step": 3420 + }, + { + "epoch": 0.37568636064133537, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.066117286682129, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7121115922927856, + "num_tokens": 88460174.0, + "step": 3421 + }, + { + "epoch": 0.37579617834394907, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.127795696258545, + "learning_rate": 1e-06, + "loss": 1.0263, + "mean_token_accuracy": 0.6905959844589233, + "num_tokens": 88488063.0, + "step": 3422 + }, + { + "epoch": 0.3759059960465627, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.081191301345825, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6946485042572021, + "num_tokens": 88520294.0, + "step": 3423 + }, + { + "epoch": 0.37601581374917636, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 1.9535272121429443, + "learning_rate": 1e-06, + "loss": 1.0656, + "mean_token_accuracy": 0.677994430065155, + "num_tokens": 88554544.0, + "step": 3424 + }, + { + "epoch": 0.37612563145179, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.2492799758911133, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7091630697250366, + "num_tokens": 88578934.0, + "step": 3425 + }, + { + "epoch": 0.3762354491544037, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.369558572769165, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7072768807411194, + "num_tokens": 88600495.0, + "step": 3426 + }, + { + "epoch": 0.37634526685701736, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.121194839477539, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7156243324279785, + "num_tokens": 88627034.0, + "step": 3427 + }, + { + "epoch": 0.376455084559631, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.2683963775634766, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7024657130241394, + "num_tokens": 88650799.0, + "step": 3428 + }, + { + "epoch": 0.37656490226224465, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.4683103561401367, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7073262929916382, + "num_tokens": 88670895.0, + "step": 3429 + }, + { + "epoch": 0.37667471996485835, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.264871835708618, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6954764127731323, + "num_tokens": 88693919.0, + "step": 3430 + }, + { + "epoch": 0.376784537667472, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.1584506034851074, + "learning_rate": 1e-06, + "loss": 1.0714, + "mean_token_accuracy": 0.6749931573867798, + "num_tokens": 88721418.0, + "step": 3431 + }, + { + "epoch": 0.37689435537008564, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.2845871448516846, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7216402888298035, + "num_tokens": 88745403.0, + "step": 3432 + }, + { + "epoch": 0.37700417307269934, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 1.9056587219238281, + "learning_rate": 1e-06, + "loss": 1.0602, + "mean_token_accuracy": 0.6855001449584961, + "num_tokens": 88780492.0, + "step": 3433 + }, + { + "epoch": 0.377113990775313, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.081483840942383, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6894459128379822, + "num_tokens": 88811625.0, + "step": 3434 + }, + { + "epoch": 0.37722380847792664, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.0525460243225098, + "learning_rate": 1e-06, + "loss": 1.1217, + "mean_token_accuracy": 0.678327739238739, + "num_tokens": 88843377.0, + "step": 3435 + }, + { + "epoch": 0.3773336261805403, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.127009868621826, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7045953273773193, + "num_tokens": 88871222.0, + "step": 3436 + }, + { + "epoch": 0.377443443883154, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 1.863357663154602, + "learning_rate": 1e-06, + "loss": 1.0776, + "mean_token_accuracy": 0.6749628782272339, + "num_tokens": 88906003.0, + "step": 3437 + }, + { + "epoch": 0.37755326158576763, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 2.1724703311920166, + "learning_rate": 1e-06, + "loss": 1.0669, + "mean_token_accuracy": 0.6844490766525269, + "num_tokens": 88934175.0, + "step": 3438 + }, + { + "epoch": 0.3776630792883813, + "ewc_loss": 1.5497207641601562e-06, + "grad_norm": 1.9696629047393799, + "learning_rate": 1e-06, + "loss": 1.0853, + "mean_token_accuracy": 0.6776332855224609, + "num_tokens": 88967305.0, + "step": 3439 + }, + { + "epoch": 0.3777728969909949, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 2.203646659851074, + "learning_rate": 1e-06, + "loss": 1.0888, + "mean_token_accuracy": 0.6718055605888367, + "num_tokens": 88995058.0, + "step": 3440 + }, + { + "epoch": 0.3778827146936086, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 2.5114989280700684, + "learning_rate": 1e-06, + "loss": 1.0446, + "mean_token_accuracy": 0.6862961053848267, + "num_tokens": 89016825.0, + "step": 3441 + }, + { + "epoch": 0.37799253239622227, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.2317750453948975, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6994641423225403, + "num_tokens": 89044034.0, + "step": 3442 + }, + { + "epoch": 0.3781023500988359, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.3088462352752686, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.70184326171875, + "num_tokens": 89068052.0, + "step": 3443 + }, + { + "epoch": 0.3782121678014496, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.1253488063812256, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.713324248790741, + "num_tokens": 89094888.0, + "step": 3444 + }, + { + "epoch": 0.37832198550406326, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.242802143096924, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7022268772125244, + "num_tokens": 89120885.0, + "step": 3445 + }, + { + "epoch": 0.3784318032066769, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.0257716178894043, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7170122265815735, + "num_tokens": 89148606.0, + "step": 3446 + }, + { + "epoch": 0.37854162090929055, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.4769320487976074, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7245962619781494, + "num_tokens": 89170649.0, + "step": 3447 + }, + { + "epoch": 0.37865143861190426, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.2197952270507812, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6908776760101318, + "num_tokens": 89196599.0, + "step": 3448 + }, + { + "epoch": 0.3787612563145179, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.4017486572265625, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7116345167160034, + "num_tokens": 89218130.0, + "step": 3449 + }, + { + "epoch": 0.37887107401713155, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.263000011444092, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6965879797935486, + "num_tokens": 89242478.0, + "step": 3450 + }, + { + "epoch": 0.37898089171974525, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.258371353149414, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6956495642662048, + "num_tokens": 89267931.0, + "step": 3451 + }, + { + "epoch": 0.3790907094223589, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 1.9615334272384644, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7232315540313721, + "num_tokens": 89298706.0, + "step": 3452 + }, + { + "epoch": 0.37920052712497254, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.2255070209503174, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.7105770707130432, + "num_tokens": 89325375.0, + "step": 3453 + }, + { + "epoch": 0.3793103448275862, + "ewc_loss": 1.564621925354004e-06, + "grad_norm": 2.4794669151306152, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7109429240226746, + "num_tokens": 89346526.0, + "step": 3454 + }, + { + "epoch": 0.3794201625301999, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.2229859828948975, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6935721635818481, + "num_tokens": 89372830.0, + "step": 3455 + }, + { + "epoch": 0.37952998023281354, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.261356830596924, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6998339891433716, + "num_tokens": 89396497.0, + "step": 3456 + }, + { + "epoch": 0.3796397979354272, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 2.227524995803833, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.7079578042030334, + "num_tokens": 89422444.0, + "step": 3457 + }, + { + "epoch": 0.3797496156380408, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 2.228647470474243, + "learning_rate": 1e-06, + "loss": 1.0581, + "mean_token_accuracy": 0.6840619444847107, + "num_tokens": 89447722.0, + "step": 3458 + }, + { + "epoch": 0.37985943334065453, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 2.1889712810516357, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7234759330749512, + "num_tokens": 89472695.0, + "step": 3459 + }, + { + "epoch": 0.3799692510432682, + "ewc_loss": 1.55717134475708e-06, + "grad_norm": 1.9913643598556519, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.692523717880249, + "num_tokens": 89504156.0, + "step": 3460 + }, + { + "epoch": 0.3800790687458818, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.013033151626587, + "learning_rate": 1e-06, + "loss": 1.045, + "mean_token_accuracy": 0.6933038234710693, + "num_tokens": 89533887.0, + "step": 3461 + }, + { + "epoch": 0.3801888864484955, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.3698806762695312, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7044177651405334, + "num_tokens": 89558518.0, + "step": 3462 + }, + { + "epoch": 0.38029870415110917, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.5491766929626465, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.6924710273742676, + "num_tokens": 89579890.0, + "step": 3463 + }, + { + "epoch": 0.3804085218537228, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.2703299522399902, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7052628993988037, + "num_tokens": 89604732.0, + "step": 3464 + }, + { + "epoch": 0.38051833955633646, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.4801812171936035, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7124853134155273, + "num_tokens": 89625793.0, + "step": 3465 + }, + { + "epoch": 0.38062815725895016, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.308120012283325, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.714678168296814, + "num_tokens": 89648859.0, + "step": 3466 + }, + { + "epoch": 0.3807379749615638, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.012855291366577, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6989380121231079, + "num_tokens": 89679663.0, + "step": 3467 + }, + { + "epoch": 0.38084779266417745, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.1947829723358154, + "learning_rate": 1e-06, + "loss": 1.0677, + "mean_token_accuracy": 0.6823490858078003, + "num_tokens": 89705345.0, + "step": 3468 + }, + { + "epoch": 0.3809576103667911, + "ewc_loss": 1.5720725059509277e-06, + "grad_norm": 2.239015817642212, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6925860643386841, + "num_tokens": 89730700.0, + "step": 3469 + }, + { + "epoch": 0.3810674280694048, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 1.9520679712295532, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7116310596466064, + "num_tokens": 89763152.0, + "step": 3470 + }, + { + "epoch": 0.38117724577201845, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.34301495552063, + "learning_rate": 1e-06, + "loss": 1.0203, + "mean_token_accuracy": 0.6916213631629944, + "num_tokens": 89786840.0, + "step": 3471 + }, + { + "epoch": 0.3812870634746321, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.2241854667663574, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7144350409507751, + "num_tokens": 89811531.0, + "step": 3472 + }, + { + "epoch": 0.3813968811772458, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.162400245666504, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6937924027442932, + "num_tokens": 89837392.0, + "step": 3473 + }, + { + "epoch": 0.38150669887985944, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.3110857009887695, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7153666019439697, + "num_tokens": 89861657.0, + "step": 3474 + }, + { + "epoch": 0.3816165165824731, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.233682155609131, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7000176906585693, + "num_tokens": 89888089.0, + "step": 3475 + }, + { + "epoch": 0.38172633428508673, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 1.8923590183258057, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.701515257358551, + "num_tokens": 89922136.0, + "step": 3476 + }, + { + "epoch": 0.38183615198770043, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.2433717250823975, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7080755233764648, + "num_tokens": 89946651.0, + "step": 3477 + }, + { + "epoch": 0.3819459696903141, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.4078667163848877, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7047240734100342, + "num_tokens": 89970358.0, + "step": 3478 + }, + { + "epoch": 0.3820557873929277, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.4044344425201416, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6916764974594116, + "num_tokens": 89994077.0, + "step": 3479 + }, + { + "epoch": 0.3821656050955414, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.4342079162597656, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.6878947615623474, + "num_tokens": 90017250.0, + "step": 3480 + }, + { + "epoch": 0.3822754227981551, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.128777503967285, + "learning_rate": 1e-06, + "loss": 1.0638, + "mean_token_accuracy": 0.6808546185493469, + "num_tokens": 90046866.0, + "step": 3481 + }, + { + "epoch": 0.3823852405007687, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.410085678100586, + "learning_rate": 1e-06, + "loss": 1.0912, + "mean_token_accuracy": 0.6819034218788147, + "num_tokens": 90070820.0, + "step": 3482 + }, + { + "epoch": 0.38249505820338237, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 1.9746001958847046, + "learning_rate": 1e-06, + "loss": 1.0489, + "mean_token_accuracy": 0.6844721436500549, + "num_tokens": 90103266.0, + "step": 3483 + }, + { + "epoch": 0.38260487590599607, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.3110973834991455, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.7071565985679626, + "num_tokens": 90128277.0, + "step": 3484 + }, + { + "epoch": 0.3827146936086097, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.080601453781128, + "learning_rate": 1e-06, + "loss": 1.046, + "mean_token_accuracy": 0.6929067373275757, + "num_tokens": 90158627.0, + "step": 3485 + }, + { + "epoch": 0.38282451131122336, + "ewc_loss": 1.5795230865478516e-06, + "grad_norm": 2.112196445465088, + "learning_rate": 1e-06, + "loss": 1.0682, + "mean_token_accuracy": 0.679018497467041, + "num_tokens": 90188177.0, + "step": 3486 + }, + { + "epoch": 0.382934329013837, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 1.999729037284851, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6890684962272644, + "num_tokens": 90219370.0, + "step": 3487 + }, + { + "epoch": 0.3830441467164507, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.5108389854431152, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7274714708328247, + "num_tokens": 90239794.0, + "step": 3488 + }, + { + "epoch": 0.38315396441906435, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.05617094039917, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6788987517356873, + "num_tokens": 90270780.0, + "step": 3489 + }, + { + "epoch": 0.383263782121678, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.0577893257141113, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7175979614257812, + "num_tokens": 90299170.0, + "step": 3490 + }, + { + "epoch": 0.3833735998242917, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.03642201423645, + "learning_rate": 1e-06, + "loss": 1.0902, + "mean_token_accuracy": 0.6767641305923462, + "num_tokens": 90330360.0, + "step": 3491 + }, + { + "epoch": 0.38348341752690535, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.5553953647613525, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.707395076751709, + "num_tokens": 90350344.0, + "step": 3492 + }, + { + "epoch": 0.383593235229519, + "ewc_loss": 1.5869736671447754e-06, + "grad_norm": 2.387619733810425, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7182759046554565, + "num_tokens": 90374020.0, + "step": 3493 + }, + { + "epoch": 0.38370305293213264, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.2905142307281494, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7032285928726196, + "num_tokens": 90398968.0, + "step": 3494 + }, + { + "epoch": 0.38381287063474634, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.284376859664917, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7344361543655396, + "num_tokens": 90424099.0, + "step": 3495 + }, + { + "epoch": 0.38392268833736, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.2995333671569824, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6989268064498901, + "num_tokens": 90447219.0, + "step": 3496 + }, + { + "epoch": 0.38403250603997363, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.0811383724212646, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7088576555252075, + "num_tokens": 90475098.0, + "step": 3497 + }, + { + "epoch": 0.38414232374258733, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.5146517753601074, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7374153137207031, + "num_tokens": 90493385.0, + "step": 3498 + }, + { + "epoch": 0.384252141445201, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.1081840991973877, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7043524980545044, + "num_tokens": 90524253.0, + "step": 3499 + }, + { + "epoch": 0.3843619591478146, + "ewc_loss": 1.5944242477416992e-06, + "grad_norm": 2.0539252758026123, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6929851770401001, + "num_tokens": 90554973.0, + "step": 3500 + }, + { + "epoch": 0.38447177685042827, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.134671449661255, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.7018439769744873, + "num_tokens": 90584515.0, + "step": 3501 + }, + { + "epoch": 0.38458159455304197, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.3595335483551025, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7007448077201843, + "num_tokens": 90608481.0, + "step": 3502 + }, + { + "epoch": 0.3846914122556556, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.1725926399230957, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6950796842575073, + "num_tokens": 90633367.0, + "step": 3503 + }, + { + "epoch": 0.38480122995826926, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.3449697494506836, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7017219066619873, + "num_tokens": 90657309.0, + "step": 3504 + }, + { + "epoch": 0.3849110476608829, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.4318063259124756, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.7002919912338257, + "num_tokens": 90681447.0, + "step": 3505 + }, + { + "epoch": 0.3850208653634966, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.330296039581299, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6887145638465881, + "num_tokens": 90705941.0, + "step": 3506 + }, + { + "epoch": 0.38513068306611026, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.0100905895233154, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7141255736351013, + "num_tokens": 90734216.0, + "step": 3507 + }, + { + "epoch": 0.3852405007687239, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.134247303009033, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7061347365379333, + "num_tokens": 90761487.0, + "step": 3508 + }, + { + "epoch": 0.3853503184713376, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.193814992904663, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6906514167785645, + "num_tokens": 90789622.0, + "step": 3509 + }, + { + "epoch": 0.38546013617395125, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.0393295288085938, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6927162408828735, + "num_tokens": 90819427.0, + "step": 3510 + }, + { + "epoch": 0.3855699538765649, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.192995548248291, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6879082918167114, + "num_tokens": 90846329.0, + "step": 3511 + }, + { + "epoch": 0.38567977157917854, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.060080051422119, + "learning_rate": 1e-06, + "loss": 1.0951, + "mean_token_accuracy": 0.6711599826812744, + "num_tokens": 90875571.0, + "step": 3512 + }, + { + "epoch": 0.38578958928179224, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.5299978256225586, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7338589429855347, + "num_tokens": 90895193.0, + "step": 3513 + }, + { + "epoch": 0.3858994069844059, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 1.9740419387817383, + "learning_rate": 1e-06, + "loss": 1.0899, + "mean_token_accuracy": 0.6781718730926514, + "num_tokens": 90926516.0, + "step": 3514 + }, + { + "epoch": 0.38600922468701954, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.198925495147705, + "learning_rate": 1e-06, + "loss": 1.0597, + "mean_token_accuracy": 0.6780112981796265, + "num_tokens": 90955106.0, + "step": 3515 + }, + { + "epoch": 0.3861190423896332, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.2281270027160645, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7019989490509033, + "num_tokens": 90980991.0, + "step": 3516 + }, + { + "epoch": 0.3862288600922469, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.449787139892578, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6983412504196167, + "num_tokens": 91004316.0, + "step": 3517 + }, + { + "epoch": 0.38633867779486053, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.240941047668457, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.72621750831604, + "num_tokens": 91027761.0, + "step": 3518 + }, + { + "epoch": 0.3864484954974742, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 2.3823838233947754, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6866647005081177, + "num_tokens": 91050506.0, + "step": 3519 + }, + { + "epoch": 0.3865583132000879, + "ewc_loss": 1.601874828338623e-06, + "grad_norm": 1.9656829833984375, + "learning_rate": 1e-06, + "loss": 1.0566, + "mean_token_accuracy": 0.6784647703170776, + "num_tokens": 91082101.0, + "step": 3520 + }, + { + "epoch": 0.3866681309027015, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.448643445968628, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7195948362350464, + "num_tokens": 91104010.0, + "step": 3521 + }, + { + "epoch": 0.38677794860531517, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.1861562728881836, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.7050600051879883, + "num_tokens": 91129990.0, + "step": 3522 + }, + { + "epoch": 0.3868877663079288, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.1241815090179443, + "learning_rate": 1e-06, + "loss": 1.0923, + "mean_token_accuracy": 0.6758324503898621, + "num_tokens": 91158919.0, + "step": 3523 + }, + { + "epoch": 0.3869975840105425, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.351106882095337, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7063534259796143, + "num_tokens": 91180887.0, + "step": 3524 + }, + { + "epoch": 0.38710740171315616, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.0352509021759033, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.6915160417556763, + "num_tokens": 91211256.0, + "step": 3525 + }, + { + "epoch": 0.3872172194157698, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.2561397552490234, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.7005705833435059, + "num_tokens": 91235786.0, + "step": 3526 + }, + { + "epoch": 0.3873270371183835, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.398566484451294, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7003559470176697, + "num_tokens": 91259530.0, + "step": 3527 + }, + { + "epoch": 0.38743685482099716, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.34852933883667, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.6823435425758362, + "num_tokens": 91284855.0, + "step": 3528 + }, + { + "epoch": 0.3875466725236108, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.1078498363494873, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7237422466278076, + "num_tokens": 91311673.0, + "step": 3529 + }, + { + "epoch": 0.38765649022622445, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.234123468399048, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6955760717391968, + "num_tokens": 91337540.0, + "step": 3530 + }, + { + "epoch": 0.38776630792883815, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.1984946727752686, + "learning_rate": 1e-06, + "loss": 0.9975, + "mean_token_accuracy": 0.6965932846069336, + "num_tokens": 91363407.0, + "step": 3531 + }, + { + "epoch": 0.3878761256314518, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.6315255165100098, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.702134907245636, + "num_tokens": 91382935.0, + "step": 3532 + }, + { + "epoch": 0.38798594333406544, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.3261075019836426, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6992532014846802, + "num_tokens": 91407494.0, + "step": 3533 + }, + { + "epoch": 0.3880957610366791, + "ewc_loss": 1.6167759895324707e-06, + "grad_norm": 2.103987693786621, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7209217548370361, + "num_tokens": 91434532.0, + "step": 3534 + }, + { + "epoch": 0.3882055787392928, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 2.720780849456787, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7299680113792419, + "num_tokens": 91453366.0, + "step": 3535 + }, + { + "epoch": 0.38831539644190644, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 2.2103545665740967, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6971977353096008, + "num_tokens": 91479978.0, + "step": 3536 + }, + { + "epoch": 0.3884252141445201, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 1.9392571449279785, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.692132830619812, + "num_tokens": 91512790.0, + "step": 3537 + }, + { + "epoch": 0.3885350318471338, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 2.1281509399414062, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7010031342506409, + "num_tokens": 91539101.0, + "step": 3538 + }, + { + "epoch": 0.38864484954974743, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 2.5022943019866943, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6949681043624878, + "num_tokens": 91561374.0, + "step": 3539 + }, + { + "epoch": 0.3887546672523611, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 1.9707932472229004, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7070780992507935, + "num_tokens": 91589580.0, + "step": 3540 + }, + { + "epoch": 0.3888644849549747, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 2.375588893890381, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6915267705917358, + "num_tokens": 91612399.0, + "step": 3541 + }, + { + "epoch": 0.3889743026575884, + "ewc_loss": 1.6316771507263184e-06, + "grad_norm": 1.985150694847107, + "learning_rate": 1e-06, + "loss": 1.0749, + "mean_token_accuracy": 0.6810387372970581, + "num_tokens": 91644091.0, + "step": 3542 + }, + { + "epoch": 0.38908412036020207, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.140707015991211, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7008949518203735, + "num_tokens": 91670530.0, + "step": 3543 + }, + { + "epoch": 0.3891939380628157, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.3999199867248535, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.706545352935791, + "num_tokens": 91694669.0, + "step": 3544 + }, + { + "epoch": 0.38930375576542936, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 1.9568729400634766, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6861901879310608, + "num_tokens": 91728597.0, + "step": 3545 + }, + { + "epoch": 0.38941357346804306, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.051720380783081, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6990846395492554, + "num_tokens": 91756801.0, + "step": 3546 + }, + { + "epoch": 0.3895233911706567, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.51088809967041, + "learning_rate": 1e-06, + "loss": 1.0151, + "mean_token_accuracy": 0.6882478594779968, + "num_tokens": 91778500.0, + "step": 3547 + }, + { + "epoch": 0.38963320887327035, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.2000088691711426, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6973766088485718, + "num_tokens": 91804944.0, + "step": 3548 + }, + { + "epoch": 0.38974302657588406, + "ewc_loss": 1.6391277313232422e-06, + "grad_norm": 2.1134841442108154, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7285184860229492, + "num_tokens": 91833600.0, + "step": 3549 + }, + { + "epoch": 0.3898528442784977, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.193819999694824, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7296584248542786, + "num_tokens": 91856897.0, + "step": 3550 + }, + { + "epoch": 0.38996266198111135, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.3713488578796387, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.69685959815979, + "num_tokens": 91878769.0, + "step": 3551 + }, + { + "epoch": 0.390072479683725, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.340506076812744, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7115738391876221, + "num_tokens": 91901129.0, + "step": 3552 + }, + { + "epoch": 0.3901822973863387, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.453164577484131, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7264268398284912, + "num_tokens": 91921789.0, + "step": 3553 + }, + { + "epoch": 0.39029211508895234, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.178992986679077, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6928097009658813, + "num_tokens": 91949252.0, + "step": 3554 + }, + { + "epoch": 0.390401932791566, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.3678319454193115, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7157717347145081, + "num_tokens": 91970893.0, + "step": 3555 + }, + { + "epoch": 0.3905117504941797, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.031248092651367, + "learning_rate": 1e-06, + "loss": 1.0862, + "mean_token_accuracy": 0.6820722818374634, + "num_tokens": 92003459.0, + "step": 3556 + }, + { + "epoch": 0.39062156819679333, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.183009624481201, + "learning_rate": 1e-06, + "loss": 1.0716, + "mean_token_accuracy": 0.6841567754745483, + "num_tokens": 92032949.0, + "step": 3557 + }, + { + "epoch": 0.390731385899407, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.249208927154541, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.699599027633667, + "num_tokens": 92058753.0, + "step": 3558 + }, + { + "epoch": 0.3908412036020206, + "ewc_loss": 1.646578311920166e-06, + "grad_norm": 2.297074317932129, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6756004095077515, + "num_tokens": 92086817.0, + "step": 3559 + }, + { + "epoch": 0.3909510213046343, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.150552749633789, + "learning_rate": 1e-06, + "loss": 1.0726, + "mean_token_accuracy": 0.6912717223167419, + "num_tokens": 92114863.0, + "step": 3560 + }, + { + "epoch": 0.391060839007248, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.1601600646972656, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.6931848526000977, + "num_tokens": 92143174.0, + "step": 3561 + }, + { + "epoch": 0.3911706567098616, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.20827317237854, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7060878276824951, + "num_tokens": 92169271.0, + "step": 3562 + }, + { + "epoch": 0.39128047441247527, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.0732603073120117, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7006319761276245, + "num_tokens": 92197530.0, + "step": 3563 + }, + { + "epoch": 0.39139029211508897, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.0642762184143066, + "learning_rate": 1e-06, + "loss": 1.0868, + "mean_token_accuracy": 0.6667013168334961, + "num_tokens": 92226522.0, + "step": 3564 + }, + { + "epoch": 0.3915001098177026, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 2.05537748336792, + "learning_rate": 1e-06, + "loss": 1.0444, + "mean_token_accuracy": 0.6898044943809509, + "num_tokens": 92255728.0, + "step": 3565 + }, + { + "epoch": 0.39160992752031626, + "ewc_loss": 1.6540288925170898e-06, + "grad_norm": 1.9234905242919922, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7082333564758301, + "num_tokens": 92288415.0, + "step": 3566 + }, + { + "epoch": 0.39171974522292996, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 2.292853593826294, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7073402404785156, + "num_tokens": 92310582.0, + "step": 3567 + }, + { + "epoch": 0.3918295629255436, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 2.238253355026245, + "learning_rate": 1e-06, + "loss": 1.0315, + "mean_token_accuracy": 0.6907598972320557, + "num_tokens": 92337192.0, + "step": 3568 + }, + { + "epoch": 0.39193938062815725, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 2.172189950942993, + "learning_rate": 1e-06, + "loss": 1.077, + "mean_token_accuracy": 0.6763126850128174, + "num_tokens": 92366075.0, + "step": 3569 + }, + { + "epoch": 0.3920491983307709, + "ewc_loss": 1.6614794731140137e-06, + "grad_norm": 2.284666061401367, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6958800554275513, + "num_tokens": 92391433.0, + "step": 3570 + }, + { + "epoch": 0.3921590160333846, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.2918643951416016, + "learning_rate": 1e-06, + "loss": 1.0977, + "mean_token_accuracy": 0.6754839420318604, + "num_tokens": 92419463.0, + "step": 3571 + }, + { + "epoch": 0.39226883373599825, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.3781116008758545, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7030699849128723, + "num_tokens": 92441456.0, + "step": 3572 + }, + { + "epoch": 0.3923786514386119, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.099928855895996, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7209398746490479, + "num_tokens": 92466494.0, + "step": 3573 + }, + { + "epoch": 0.3924884691412256, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.1693694591522217, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6872020959854126, + "num_tokens": 92491099.0, + "step": 3574 + }, + { + "epoch": 0.39259828684383924, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.2462005615234375, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7234911918640137, + "num_tokens": 92518006.0, + "step": 3575 + }, + { + "epoch": 0.3927081045464529, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.2010433673858643, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7066729664802551, + "num_tokens": 92542700.0, + "step": 3576 + }, + { + "epoch": 0.39281792224906653, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.3171637058258057, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6944196224212646, + "num_tokens": 92566279.0, + "step": 3577 + }, + { + "epoch": 0.39292773995168023, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.001763343811035, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6880068182945251, + "num_tokens": 92599121.0, + "step": 3578 + }, + { + "epoch": 0.3930375576542939, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.442063570022583, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.7019948959350586, + "num_tokens": 92624001.0, + "step": 3579 + }, + { + "epoch": 0.3931473753569075, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.132270336151123, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7035430073738098, + "num_tokens": 92652425.0, + "step": 3580 + }, + { + "epoch": 0.39325719305952117, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.4005303382873535, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6973268985748291, + "num_tokens": 92674724.0, + "step": 3581 + }, + { + "epoch": 0.39336701076213487, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.2484209537506104, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6966232061386108, + "num_tokens": 92699667.0, + "step": 3582 + }, + { + "epoch": 0.3934768284647485, + "ewc_loss": 1.6689300537109375e-06, + "grad_norm": 2.1808154582977295, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6894283294677734, + "num_tokens": 92727517.0, + "step": 3583 + }, + { + "epoch": 0.39358664616736216, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.6256093978881836, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7115066051483154, + "num_tokens": 92747373.0, + "step": 3584 + }, + { + "epoch": 0.39369646386997587, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.1224193572998047, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.708319902420044, + "num_tokens": 92775832.0, + "step": 3585 + }, + { + "epoch": 0.3938062815725895, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 1.9172732830047607, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.6989325881004333, + "num_tokens": 92807378.0, + "step": 3586 + }, + { + "epoch": 0.39391609927520316, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.310530662536621, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7060179710388184, + "num_tokens": 92832155.0, + "step": 3587 + }, + { + "epoch": 0.3940259169778168, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.326951742172241, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7022420167922974, + "num_tokens": 92856290.0, + "step": 3588 + }, + { + "epoch": 0.3941357346804305, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.2898476123809814, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6968240141868591, + "num_tokens": 92881250.0, + "step": 3589 + }, + { + "epoch": 0.39424555238304415, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.3673181533813477, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6948680877685547, + "num_tokens": 92904024.0, + "step": 3590 + }, + { + "epoch": 0.3943553700856578, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.0259289741516113, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7022997140884399, + "num_tokens": 92932387.0, + "step": 3591 + }, + { + "epoch": 0.39446518778827144, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.1806564331054688, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.682669997215271, + "num_tokens": 92959209.0, + "step": 3592 + }, + { + "epoch": 0.39457500549088514, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.6965537071228027, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7175183892250061, + "num_tokens": 92977796.0, + "step": 3593 + }, + { + "epoch": 0.3946848231934988, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.3106887340545654, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7189285755157471, + "num_tokens": 93002954.0, + "step": 3594 + }, + { + "epoch": 0.39479464089611244, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 1.999638319015503, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6897329092025757, + "num_tokens": 93033062.0, + "step": 3595 + }, + { + "epoch": 0.39490445859872614, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.243633985519409, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6923518180847168, + "num_tokens": 93059527.0, + "step": 3596 + }, + { + "epoch": 0.3950142763013398, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.1444077491760254, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7158553600311279, + "num_tokens": 93085994.0, + "step": 3597 + }, + { + "epoch": 0.39512409400395343, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.1819727420806885, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6916619539260864, + "num_tokens": 93110793.0, + "step": 3598 + }, + { + "epoch": 0.3952339117065671, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.2183446884155273, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7127100229263306, + "num_tokens": 93137040.0, + "step": 3599 + }, + { + "epoch": 0.3953437294091808, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.346676826477051, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7033411264419556, + "num_tokens": 93160651.0, + "step": 3600 + }, + { + "epoch": 0.3954535471117944, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.084951400756836, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7074127197265625, + "num_tokens": 93189203.0, + "step": 3601 + }, + { + "epoch": 0.39556336481440807, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.2334492206573486, + "learning_rate": 1e-06, + "loss": 1.0666, + "mean_token_accuracy": 0.6837785243988037, + "num_tokens": 93215152.0, + "step": 3602 + }, + { + "epoch": 0.39567318251702177, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.053962469100952, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7098914980888367, + "num_tokens": 93242481.0, + "step": 3603 + }, + { + "epoch": 0.3957830002196354, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 1.925870418548584, + "learning_rate": 1e-06, + "loss": 1.0579, + "mean_token_accuracy": 0.6798704862594604, + "num_tokens": 93273852.0, + "step": 3604 + }, + { + "epoch": 0.39589281792224906, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.2866082191467285, + "learning_rate": 1e-06, + "loss": 1.0229, + "mean_token_accuracy": 0.7005646228790283, + "num_tokens": 93297600.0, + "step": 3605 + }, + { + "epoch": 0.3960026356248627, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.2125253677368164, + "learning_rate": 1e-06, + "loss": 1.1052, + "mean_token_accuracy": 0.6703116297721863, + "num_tokens": 93326205.0, + "step": 3606 + }, + { + "epoch": 0.3961124533274764, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.149264097213745, + "learning_rate": 1e-06, + "loss": 1.0892, + "mean_token_accuracy": 0.6752523183822632, + "num_tokens": 93356719.0, + "step": 3607 + }, + { + "epoch": 0.39622227103009006, + "ewc_loss": 1.6763806343078613e-06, + "grad_norm": 2.36958909034729, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.72105872631073, + "num_tokens": 93379681.0, + "step": 3608 + }, + { + "epoch": 0.3963320887327037, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.141786575317383, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.694476842880249, + "num_tokens": 93407884.0, + "step": 3609 + }, + { + "epoch": 0.39644190643531735, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.0997047424316406, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.7011209726333618, + "num_tokens": 93434797.0, + "step": 3610 + }, + { + "epoch": 0.39655172413793105, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.11942195892334, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7018905282020569, + "num_tokens": 93461110.0, + "step": 3611 + }, + { + "epoch": 0.3966615418405447, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.0424256324768066, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6934537887573242, + "num_tokens": 93493577.0, + "step": 3612 + }, + { + "epoch": 0.39677135954315834, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.3676059246063232, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7110358476638794, + "num_tokens": 93516458.0, + "step": 3613 + }, + { + "epoch": 0.39688117724577204, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.1169917583465576, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7133938074111938, + "num_tokens": 93543958.0, + "step": 3614 + }, + { + "epoch": 0.3969909949483857, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.9545860290527344, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7086809277534485, + "num_tokens": 93560899.0, + "step": 3615 + }, + { + "epoch": 0.39710081265099934, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.594921350479126, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7152413129806519, + "num_tokens": 93579658.0, + "step": 3616 + }, + { + "epoch": 0.397210630353613, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.0687549114227295, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7452131509780884, + "num_tokens": 93605222.0, + "step": 3617 + }, + { + "epoch": 0.3973204480562267, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.2466962337493896, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7061859965324402, + "num_tokens": 93631550.0, + "step": 3618 + }, + { + "epoch": 0.39743026575884033, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 1.9950075149536133, + "learning_rate": 1e-06, + "loss": 1.037, + "mean_token_accuracy": 0.6918704509735107, + "num_tokens": 93661110.0, + "step": 3619 + }, + { + "epoch": 0.397540083461454, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.610567569732666, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7036243081092834, + "num_tokens": 93680548.0, + "step": 3620 + }, + { + "epoch": 0.3976499011640676, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.105104923248291, + "learning_rate": 1e-06, + "loss": 1.1535, + "mean_token_accuracy": 0.6583261489868164, + "num_tokens": 93710057.0, + "step": 3621 + }, + { + "epoch": 0.3977597188666813, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.0014350414276123, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6936013698577881, + "num_tokens": 93742099.0, + "step": 3622 + }, + { + "epoch": 0.39786953656929497, + "ewc_loss": 1.6838312149047852e-06, + "grad_norm": 2.125049591064453, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7232033014297485, + "num_tokens": 93767255.0, + "step": 3623 + }, + { + "epoch": 0.3979793542719086, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.5568506717681885, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7073280811309814, + "num_tokens": 93787179.0, + "step": 3624 + }, + { + "epoch": 0.3980891719745223, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 1.9779365062713623, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.720390260219574, + "num_tokens": 93818656.0, + "step": 3625 + }, + { + "epoch": 0.39819898967713596, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.280154228210449, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6986591815948486, + "num_tokens": 93842689.0, + "step": 3626 + }, + { + "epoch": 0.3983088073797496, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.3635168075561523, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7025704383850098, + "num_tokens": 93866051.0, + "step": 3627 + }, + { + "epoch": 0.39841862508236325, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 1.987406611442566, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7121068239212036, + "num_tokens": 93894339.0, + "step": 3628 + }, + { + "epoch": 0.39852844278497696, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.461557149887085, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6870584487915039, + "num_tokens": 93917128.0, + "step": 3629 + }, + { + "epoch": 0.3986382604875906, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.437589168548584, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7199332118034363, + "num_tokens": 93937068.0, + "step": 3630 + }, + { + "epoch": 0.39874807819020425, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.3201558589935303, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.707750678062439, + "num_tokens": 93961825.0, + "step": 3631 + }, + { + "epoch": 0.39885789589281795, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.4312164783477783, + "learning_rate": 1e-06, + "loss": 1.0773, + "mean_token_accuracy": 0.6748820543289185, + "num_tokens": 93986820.0, + "step": 3632 + }, + { + "epoch": 0.3989677135954316, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.4797511100769043, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6852425932884216, + "num_tokens": 94009183.0, + "step": 3633 + }, + { + "epoch": 0.39907753129804524, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.255305290222168, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7075573205947876, + "num_tokens": 94035169.0, + "step": 3634 + }, + { + "epoch": 0.3991873490006589, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.165710926055908, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.701630711555481, + "num_tokens": 94058915.0, + "step": 3635 + }, + { + "epoch": 0.3992971667032726, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.131711006164551, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.705324649810791, + "num_tokens": 94087673.0, + "step": 3636 + }, + { + "epoch": 0.39940698440588623, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.023507833480835, + "learning_rate": 1e-06, + "loss": 1.0734, + "mean_token_accuracy": 0.6801592707633972, + "num_tokens": 94117111.0, + "step": 3637 + }, + { + "epoch": 0.3995168021084999, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.3506174087524414, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7124553322792053, + "num_tokens": 94142061.0, + "step": 3638 + }, + { + "epoch": 0.3996266198111135, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.273510456085205, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7241372466087341, + "num_tokens": 94166473.0, + "step": 3639 + }, + { + "epoch": 0.3997364375137272, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.0156710147857666, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7226579189300537, + "num_tokens": 94195339.0, + "step": 3640 + }, + { + "epoch": 0.3998462552163409, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.2485334873199463, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7054718732833862, + "num_tokens": 94220333.0, + "step": 3641 + }, + { + "epoch": 0.3999560729189545, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.04862642288208, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7023591995239258, + "num_tokens": 94247191.0, + "step": 3642 + }, + { + "epoch": 0.4000658906215682, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.2731621265411377, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6803524494171143, + "num_tokens": 94272074.0, + "step": 3643 + }, + { + "epoch": 0.40017570832418187, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.460226535797119, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7238776683807373, + "num_tokens": 94292812.0, + "step": 3644 + }, + { + "epoch": 0.4002855260267955, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.4120631217956543, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6936199069023132, + "num_tokens": 94317623.0, + "step": 3645 + }, + { + "epoch": 0.40039534372940916, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.0579686164855957, + "learning_rate": 1e-06, + "loss": 1.081, + "mean_token_accuracy": 0.6928059458732605, + "num_tokens": 94348024.0, + "step": 3646 + }, + { + "epoch": 0.40050516143202286, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.3961520195007324, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.696459174156189, + "num_tokens": 94370827.0, + "step": 3647 + }, + { + "epoch": 0.4006149791346365, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.486483573913574, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6862894296646118, + "num_tokens": 94395141.0, + "step": 3648 + }, + { + "epoch": 0.40072479683725015, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 1.910642147064209, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6879674196243286, + "num_tokens": 94427195.0, + "step": 3649 + }, + { + "epoch": 0.4008346145398638, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.0958895683288574, + "learning_rate": 1e-06, + "loss": 1.0505, + "mean_token_accuracy": 0.6855623722076416, + "num_tokens": 94456986.0, + "step": 3650 + }, + { + "epoch": 0.4009444322424775, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.177225351333618, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6940836906433105, + "num_tokens": 94486412.0, + "step": 3651 + }, + { + "epoch": 0.40105424994509115, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.1731789112091064, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7007895708084106, + "num_tokens": 94513341.0, + "step": 3652 + }, + { + "epoch": 0.4011640676477048, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.290625810623169, + "learning_rate": 1e-06, + "loss": 1.0707, + "mean_token_accuracy": 0.6910480856895447, + "num_tokens": 94539325.0, + "step": 3653 + }, + { + "epoch": 0.4012738853503185, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.102691173553467, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.6915024518966675, + "num_tokens": 94569341.0, + "step": 3654 + }, + { + "epoch": 0.40138370305293214, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.3907968997955322, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7040454149246216, + "num_tokens": 94590125.0, + "step": 3655 + }, + { + "epoch": 0.4014935207555458, + "ewc_loss": 1.691281795501709e-06, + "grad_norm": 2.207156181335449, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6863315105438232, + "num_tokens": 94618712.0, + "step": 3656 + }, + { + "epoch": 0.40160333845815943, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 2.3469254970550537, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7221898436546326, + "num_tokens": 94639634.0, + "step": 3657 + }, + { + "epoch": 0.40171315616077313, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 2.2899155616760254, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6821086406707764, + "num_tokens": 94663874.0, + "step": 3658 + }, + { + "epoch": 0.4018229738633868, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 2.141360282897949, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6864326000213623, + "num_tokens": 94692854.0, + "step": 3659 + }, + { + "epoch": 0.4019327915660004, + "ewc_loss": 1.6987323760986328e-06, + "grad_norm": 2.630523681640625, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7080457210540771, + "num_tokens": 94711951.0, + "step": 3660 + }, + { + "epoch": 0.4020426092686141, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.264249086380005, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6990109086036682, + "num_tokens": 94738139.0, + "step": 3661 + }, + { + "epoch": 0.40215242697122777, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.0898044109344482, + "learning_rate": 1e-06, + "loss": 1.1142, + "mean_token_accuracy": 0.6691396236419678, + "num_tokens": 94769276.0, + "step": 3662 + }, + { + "epoch": 0.4022622446738414, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.359118938446045, + "learning_rate": 1e-06, + "loss": 1.1057, + "mean_token_accuracy": 0.6727347373962402, + "num_tokens": 94793987.0, + "step": 3663 + }, + { + "epoch": 0.40237206237645506, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.4016401767730713, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7072569727897644, + "num_tokens": 94815453.0, + "step": 3664 + }, + { + "epoch": 0.40248188007906877, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.015265703201294, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7163106799125671, + "num_tokens": 94842720.0, + "step": 3665 + }, + { + "epoch": 0.4025916977816824, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.2695322036743164, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.689547061920166, + "num_tokens": 94869987.0, + "step": 3666 + }, + { + "epoch": 0.40270151548429606, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.549396514892578, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7009475231170654, + "num_tokens": 94891386.0, + "step": 3667 + }, + { + "epoch": 0.4028113331869097, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.2974660396575928, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6968545913696289, + "num_tokens": 94915405.0, + "step": 3668 + }, + { + "epoch": 0.4029211508895234, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.2386057376861572, + "learning_rate": 1e-06, + "loss": 1.042, + "mean_token_accuracy": 0.6999822854995728, + "num_tokens": 94940529.0, + "step": 3669 + }, + { + "epoch": 0.40303096859213705, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.3113033771514893, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7043731212615967, + "num_tokens": 94963891.0, + "step": 3670 + }, + { + "epoch": 0.4031407862947507, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.372720241546631, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.6952154040336609, + "num_tokens": 94986480.0, + "step": 3671 + }, + { + "epoch": 0.4032506039973644, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.2630984783172607, + "learning_rate": 1e-06, + "loss": 1.0439, + "mean_token_accuracy": 0.683601438999176, + "num_tokens": 95012838.0, + "step": 3672 + }, + { + "epoch": 0.40336042169997804, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.3934693336486816, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.696397066116333, + "num_tokens": 95034266.0, + "step": 3673 + }, + { + "epoch": 0.4034702394025917, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.4588510990142822, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6908513307571411, + "num_tokens": 95055943.0, + "step": 3674 + }, + { + "epoch": 0.40358005710520534, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.179947853088379, + "learning_rate": 1e-06, + "loss": 1.0922, + "mean_token_accuracy": 0.6821674108505249, + "num_tokens": 95083037.0, + "step": 3675 + }, + { + "epoch": 0.40368987480781904, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.2234113216400146, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6909908056259155, + "num_tokens": 95109674.0, + "step": 3676 + }, + { + "epoch": 0.4037996925104327, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.6025984287261963, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7016191482543945, + "num_tokens": 95128170.0, + "step": 3677 + }, + { + "epoch": 0.40390951021304633, + "ewc_loss": 1.7061829566955566e-06, + "grad_norm": 2.078500747680664, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.6931411623954773, + "num_tokens": 95158536.0, + "step": 3678 + }, + { + "epoch": 0.40401932791566003, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 1.9602524042129517, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7050761580467224, + "num_tokens": 95188713.0, + "step": 3679 + }, + { + "epoch": 0.4041291456182737, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.337387800216675, + "learning_rate": 1e-06, + "loss": 1.0809, + "mean_token_accuracy": 0.6741541624069214, + "num_tokens": 95214405.0, + "step": 3680 + }, + { + "epoch": 0.4042389633208873, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 1.985812783241272, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7202421426773071, + "num_tokens": 95244165.0, + "step": 3681 + }, + { + "epoch": 0.40434878102350097, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.241865396499634, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7162526845932007, + "num_tokens": 95267477.0, + "step": 3682 + }, + { + "epoch": 0.40445859872611467, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.135272741317749, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6900043487548828, + "num_tokens": 95295044.0, + "step": 3683 + }, + { + "epoch": 0.4045684164287283, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.4090793132781982, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.6985311508178711, + "num_tokens": 95319298.0, + "step": 3684 + }, + { + "epoch": 0.40467823413134196, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.2021689414978027, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7115366458892822, + "num_tokens": 95344447.0, + "step": 3685 + }, + { + "epoch": 0.4047880518339556, + "ewc_loss": 1.7136335372924805e-06, + "grad_norm": 2.5395898818969727, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7027316689491272, + "num_tokens": 95364424.0, + "step": 3686 + }, + { + "epoch": 0.4048978695365693, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.4483182430267334, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6944648623466492, + "num_tokens": 95387166.0, + "step": 3687 + }, + { + "epoch": 0.40500768723918296, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.1807596683502197, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6913082599639893, + "num_tokens": 95414783.0, + "step": 3688 + }, + { + "epoch": 0.4051175049417966, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.2365832328796387, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6975243091583252, + "num_tokens": 95440817.0, + "step": 3689 + }, + { + "epoch": 0.4052273226444103, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.4128799438476562, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7040767073631287, + "num_tokens": 95463800.0, + "step": 3690 + }, + { + "epoch": 0.40533714034702395, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.0027120113372803, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7096081376075745, + "num_tokens": 95493112.0, + "step": 3691 + }, + { + "epoch": 0.4054469580496376, + "ewc_loss": 1.7210841178894043e-06, + "grad_norm": 2.1315226554870605, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7077604532241821, + "num_tokens": 95519677.0, + "step": 3692 + }, + { + "epoch": 0.40555677575225124, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.007160186767578, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.6958593130111694, + "num_tokens": 95549291.0, + "step": 3693 + }, + { + "epoch": 0.40566659345486494, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.2189509868621826, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7169051170349121, + "num_tokens": 95575323.0, + "step": 3694 + }, + { + "epoch": 0.4057764111574786, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.1848089694976807, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7100353837013245, + "num_tokens": 95601929.0, + "step": 3695 + }, + { + "epoch": 0.40588622886009224, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.007382392883301, + "learning_rate": 1e-06, + "loss": 1.0702, + "mean_token_accuracy": 0.6844685077667236, + "num_tokens": 95632256.0, + "step": 3696 + }, + { + "epoch": 0.4059960465627059, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.215862512588501, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7122191190719604, + "num_tokens": 95657228.0, + "step": 3697 + }, + { + "epoch": 0.4061058642653196, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.2064833641052246, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6822596788406372, + "num_tokens": 95683329.0, + "step": 3698 + }, + { + "epoch": 0.40621568196793323, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.091649293899536, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6928614377975464, + "num_tokens": 95712366.0, + "step": 3699 + }, + { + "epoch": 0.4063254996705469, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.0241308212280273, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6923862099647522, + "num_tokens": 95741651.0, + "step": 3700 + }, + { + "epoch": 0.4064353173731606, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 1.998280644416809, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.7019270062446594, + "num_tokens": 95773832.0, + "step": 3701 + }, + { + "epoch": 0.4065451350757742, + "ewc_loss": 1.735985279083252e-06, + "grad_norm": 2.1556572914123535, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.7022088766098022, + "num_tokens": 95801416.0, + "step": 3702 + }, + { + "epoch": 0.40665495277838787, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.2748351097106934, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.7034543752670288, + "num_tokens": 95826058.0, + "step": 3703 + }, + { + "epoch": 0.4067647704810015, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 1.923485279083252, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.6906867623329163, + "num_tokens": 95859076.0, + "step": 3704 + }, + { + "epoch": 0.4068745881836152, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.2415308952331543, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7037684917449951, + "num_tokens": 95884058.0, + "step": 3705 + }, + { + "epoch": 0.40698440588622886, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.5591514110565186, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7147092223167419, + "num_tokens": 95906494.0, + "step": 3706 + }, + { + "epoch": 0.4070942235888425, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.0791497230529785, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6907305717468262, + "num_tokens": 95933524.0, + "step": 3707 + }, + { + "epoch": 0.4072040412914562, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.2375035285949707, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7090414762496948, + "num_tokens": 95957723.0, + "step": 3708 + }, + { + "epoch": 0.40731385899406986, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.34015154838562, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7006331086158752, + "num_tokens": 95980974.0, + "step": 3709 + }, + { + "epoch": 0.4074236766966835, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.069453477859497, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6992563605308533, + "num_tokens": 96010658.0, + "step": 3710 + }, + { + "epoch": 0.40753349439929715, + "ewc_loss": 1.7434358596801758e-06, + "grad_norm": 2.265010118484497, + "learning_rate": 1e-06, + "loss": 1.0426, + "mean_token_accuracy": 0.6817899942398071, + "num_tokens": 96036358.0, + "step": 3711 + }, + { + "epoch": 0.40764331210191085, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.143839120864868, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.6953221559524536, + "num_tokens": 96062798.0, + "step": 3712 + }, + { + "epoch": 0.4077531298045245, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.1552603244781494, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.691032886505127, + "num_tokens": 96089971.0, + "step": 3713 + }, + { + "epoch": 0.40786294750713814, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.175356388092041, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.68930983543396, + "num_tokens": 96116747.0, + "step": 3714 + }, + { + "epoch": 0.4079727652097518, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.230761766433716, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6878305077552795, + "num_tokens": 96142612.0, + "step": 3715 + }, + { + "epoch": 0.4080825829123655, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.389537811279297, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6938084363937378, + "num_tokens": 96167683.0, + "step": 3716 + }, + { + "epoch": 0.40819240061497913, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.089459180831909, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7131558656692505, + "num_tokens": 96195221.0, + "step": 3717 + }, + { + "epoch": 0.4083022183175928, + "ewc_loss": 1.7508864402770996e-06, + "grad_norm": 2.142305850982666, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.713839054107666, + "num_tokens": 96221480.0, + "step": 3718 + }, + { + "epoch": 0.4084120360202065, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.510345220565796, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6997692584991455, + "num_tokens": 96242839.0, + "step": 3719 + }, + { + "epoch": 0.4085218537228201, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.516282320022583, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.692241907119751, + "num_tokens": 96264354.0, + "step": 3720 + }, + { + "epoch": 0.4086316714254338, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.0612363815307617, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7167421579360962, + "num_tokens": 96292974.0, + "step": 3721 + }, + { + "epoch": 0.4087414891280474, + "ewc_loss": 1.7657876014709473e-06, + "grad_norm": 2.3423733711242676, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7067173719406128, + "num_tokens": 96317311.0, + "step": 3722 + }, + { + "epoch": 0.4088513068306611, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1823668479919434, + "learning_rate": 1e-06, + "loss": 1.051, + "mean_token_accuracy": 0.676445484161377, + "num_tokens": 96344098.0, + "step": 3723 + }, + { + "epoch": 0.40896112453327477, + "ewc_loss": 1.7657876014709473e-06, + "grad_norm": 2.2841086387634277, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.6997005343437195, + "num_tokens": 96368361.0, + "step": 3724 + }, + { + "epoch": 0.4090709422358884, + "ewc_loss": 1.7657876014709473e-06, + "grad_norm": 2.1463940143585205, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7114934921264648, + "num_tokens": 96394469.0, + "step": 3725 + }, + { + "epoch": 0.40918075993850206, + "ewc_loss": 1.7657876014709473e-06, + "grad_norm": 2.2417776584625244, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.6915110349655151, + "num_tokens": 96419691.0, + "step": 3726 + }, + { + "epoch": 0.40929057764111576, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 1.920983076095581, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6993393301963806, + "num_tokens": 96453773.0, + "step": 3727 + }, + { + "epoch": 0.4094003953437294, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1955888271331787, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7211417555809021, + "num_tokens": 96478387.0, + "step": 3728 + }, + { + "epoch": 0.40951021304634305, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.20449161529541, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.69632887840271, + "num_tokens": 96504213.0, + "step": 3729 + }, + { + "epoch": 0.40962003074895675, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 1.9766578674316406, + "learning_rate": 1e-06, + "loss": 1.0498, + "mean_token_accuracy": 0.6871893405914307, + "num_tokens": 96536107.0, + "step": 3730 + }, + { + "epoch": 0.4097298484515704, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.205200672149658, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.7094841003417969, + "num_tokens": 96561775.0, + "step": 3731 + }, + { + "epoch": 0.40983966615418405, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.323343276977539, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7187528610229492, + "num_tokens": 96584036.0, + "step": 3732 + }, + { + "epoch": 0.4099494838567977, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.489081621170044, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7173876762390137, + "num_tokens": 96605831.0, + "step": 3733 + }, + { + "epoch": 0.4100593015594114, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.4861202239990234, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7049350738525391, + "num_tokens": 96627335.0, + "step": 3734 + }, + { + "epoch": 0.41016911926202504, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.180067539215088, + "learning_rate": 1e-06, + "loss": 1.0688, + "mean_token_accuracy": 0.6906827092170715, + "num_tokens": 96655776.0, + "step": 3735 + }, + { + "epoch": 0.4102789369646387, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.0212082862854004, + "learning_rate": 1e-06, + "loss": 1.0884, + "mean_token_accuracy": 0.6851540803909302, + "num_tokens": 96686329.0, + "step": 3736 + }, + { + "epoch": 0.4103887546672524, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 1.9198315143585205, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7112302780151367, + "num_tokens": 96717402.0, + "step": 3737 + }, + { + "epoch": 0.41049857236986603, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1199092864990234, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7137429714202881, + "num_tokens": 96743834.0, + "step": 3738 + }, + { + "epoch": 0.4106083900724797, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.2803773880004883, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.6979387402534485, + "num_tokens": 96766567.0, + "step": 3739 + }, + { + "epoch": 0.4107182077750933, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1911826133728027, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6923220157623291, + "num_tokens": 96793264.0, + "step": 3740 + }, + { + "epoch": 0.410828025477707, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 1.9982966184616089, + "learning_rate": 1e-06, + "loss": 1.0423, + "mean_token_accuracy": 0.6924045085906982, + "num_tokens": 96825333.0, + "step": 3741 + }, + { + "epoch": 0.41093784318032067, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.032521963119507, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7206254005432129, + "num_tokens": 96853811.0, + "step": 3742 + }, + { + "epoch": 0.4110476608829343, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.260931968688965, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.702471137046814, + "num_tokens": 96878822.0, + "step": 3743 + }, + { + "epoch": 0.41115747858554796, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.4166295528411865, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7296098470687866, + "num_tokens": 96901200.0, + "step": 3744 + }, + { + "epoch": 0.41126729628816167, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.5367698669433594, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.7025200128555298, + "num_tokens": 96920484.0, + "step": 3745 + }, + { + "epoch": 0.4113771139907753, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.261927366256714, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7045124769210815, + "num_tokens": 96944095.0, + "step": 3746 + }, + { + "epoch": 0.41148693169338896, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.5162460803985596, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7026115655899048, + "num_tokens": 96964442.0, + "step": 3747 + }, + { + "epoch": 0.41159674939600266, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.0497562885284424, + "learning_rate": 1e-06, + "loss": 1.0484, + "mean_token_accuracy": 0.6887139081954956, + "num_tokens": 96994310.0, + "step": 3748 + }, + { + "epoch": 0.4117065670986163, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.503382444381714, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7152849435806274, + "num_tokens": 97015086.0, + "step": 3749 + }, + { + "epoch": 0.41181638480122995, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.4608147144317627, + "learning_rate": 1e-06, + "loss": 1.0668, + "mean_token_accuracy": 0.6854496002197266, + "num_tokens": 97038406.0, + "step": 3750 + }, + { + "epoch": 0.4119262025038436, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.2681050300598145, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.6918047666549683, + "num_tokens": 97063504.0, + "step": 3751 + }, + { + "epoch": 0.4120360202064573, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.4558191299438477, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6784472465515137, + "num_tokens": 97087001.0, + "step": 3752 + }, + { + "epoch": 0.41214583790907094, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.0567996501922607, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7115150690078735, + "num_tokens": 97114967.0, + "step": 3753 + }, + { + "epoch": 0.4122556556116846, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.368983745574951, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7208652496337891, + "num_tokens": 97137484.0, + "step": 3754 + }, + { + "epoch": 0.4123654733142983, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1167595386505127, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7034111618995667, + "num_tokens": 97164287.0, + "step": 3755 + }, + { + "epoch": 0.41247529101691194, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.504129409790039, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7233006358146667, + "num_tokens": 97185153.0, + "step": 3756 + }, + { + "epoch": 0.4125851087195256, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1602983474731445, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.6976665258407593, + "num_tokens": 97211905.0, + "step": 3757 + }, + { + "epoch": 0.41269492642213923, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1798062324523926, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.6994202136993408, + "num_tokens": 97238603.0, + "step": 3758 + }, + { + "epoch": 0.41280474412475293, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 1.8975661993026733, + "learning_rate": 1e-06, + "loss": 1.0644, + "mean_token_accuracy": 0.679477334022522, + "num_tokens": 97271073.0, + "step": 3759 + }, + { + "epoch": 0.4129145618273666, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.2174153327941895, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7010178565979004, + "num_tokens": 97297350.0, + "step": 3760 + }, + { + "epoch": 0.4130243795299802, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.2238218784332275, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7049612998962402, + "num_tokens": 97321488.0, + "step": 3761 + }, + { + "epoch": 0.41313419723259387, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.1324732303619385, + "learning_rate": 1e-06, + "loss": 1.1037, + "mean_token_accuracy": 0.6814367771148682, + "num_tokens": 97349133.0, + "step": 3762 + }, + { + "epoch": 0.41324401493520757, + "ewc_loss": 1.773238182067871e-06, + "grad_norm": 2.24412202835083, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6990039348602295, + "num_tokens": 97373337.0, + "step": 3763 + }, + { + "epoch": 0.4133538326378212, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.5137369632720947, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7332723140716553, + "num_tokens": 97393305.0, + "step": 3764 + }, + { + "epoch": 0.41346365034043486, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.1769626140594482, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7008620500564575, + "num_tokens": 97420834.0, + "step": 3765 + }, + { + "epoch": 0.41357346804304856, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.2146918773651123, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6953980922698975, + "num_tokens": 97447875.0, + "step": 3766 + }, + { + "epoch": 0.4136832857456622, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.433645725250244, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7180774807929993, + "num_tokens": 97468715.0, + "step": 3767 + }, + { + "epoch": 0.41379310344827586, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.399578094482422, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6981725096702576, + "num_tokens": 97491436.0, + "step": 3768 + }, + { + "epoch": 0.4139029211508895, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.0583364963531494, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.6916959881782532, + "num_tokens": 97521771.0, + "step": 3769 + }, + { + "epoch": 0.4140127388535032, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.21274471282959, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6975996494293213, + "num_tokens": 97548123.0, + "step": 3770 + }, + { + "epoch": 0.41412255655611685, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.3549540042877197, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7194509506225586, + "num_tokens": 97569764.0, + "step": 3771 + }, + { + "epoch": 0.4142323742587305, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.0873305797576904, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.6994587182998657, + "num_tokens": 97598522.0, + "step": 3772 + }, + { + "epoch": 0.41434219196134414, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.1859564781188965, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7348504662513733, + "num_tokens": 97623658.0, + "step": 3773 + }, + { + "epoch": 0.41445200966395784, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 1.9677848815917969, + "learning_rate": 1e-06, + "loss": 1.0571, + "mean_token_accuracy": 0.6878445148468018, + "num_tokens": 97654071.0, + "step": 3774 + }, + { + "epoch": 0.4145618273665715, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.3373570442199707, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6961785554885864, + "num_tokens": 97677831.0, + "step": 3775 + }, + { + "epoch": 0.41467164506918514, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.3617255687713623, + "learning_rate": 1e-06, + "loss": 1.0443, + "mean_token_accuracy": 0.6866630911827087, + "num_tokens": 97702503.0, + "step": 3776 + }, + { + "epoch": 0.41478146277179884, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.4450278282165527, + "learning_rate": 1e-06, + "loss": 1.0455, + "mean_token_accuracy": 0.6897099614143372, + "num_tokens": 97724802.0, + "step": 3777 + }, + { + "epoch": 0.4148912804744125, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.3445487022399902, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7131889462471008, + "num_tokens": 97749234.0, + "step": 3778 + }, + { + "epoch": 0.41500109817702613, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.345656156539917, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.712427020072937, + "num_tokens": 97770698.0, + "step": 3779 + }, + { + "epoch": 0.4151109158796398, + "ewc_loss": 1.780688762664795e-06, + "grad_norm": 2.2447357177734375, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7016178965568542, + "num_tokens": 97794172.0, + "step": 3780 + }, + { + "epoch": 0.4152207335822535, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.460380792617798, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.7016599774360657, + "num_tokens": 97815066.0, + "step": 3781 + }, + { + "epoch": 0.4153305512848671, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.0809168815612793, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7275717258453369, + "num_tokens": 97841199.0, + "step": 3782 + }, + { + "epoch": 0.41544036898748077, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.3648009300231934, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6935142874717712, + "num_tokens": 97865472.0, + "step": 3783 + }, + { + "epoch": 0.41555018669009447, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.4047882556915283, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7127694487571716, + "num_tokens": 97886053.0, + "step": 3784 + }, + { + "epoch": 0.4156600043927081, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.583388566970825, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7036926746368408, + "num_tokens": 97908195.0, + "step": 3785 + }, + { + "epoch": 0.41576982209532176, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.2435407638549805, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7078958749771118, + "num_tokens": 97935081.0, + "step": 3786 + }, + { + "epoch": 0.4158796397979354, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 1.9770790338516235, + "learning_rate": 1e-06, + "loss": 1.0469, + "mean_token_accuracy": 0.6806138753890991, + "num_tokens": 97966325.0, + "step": 3787 + }, + { + "epoch": 0.4159894575005491, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.393214702606201, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6994292736053467, + "num_tokens": 97991457.0, + "step": 3788 + }, + { + "epoch": 0.41609927520316276, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.191213846206665, + "learning_rate": 1e-06, + "loss": 1.0435, + "mean_token_accuracy": 0.6871705055236816, + "num_tokens": 98018341.0, + "step": 3789 + }, + { + "epoch": 0.4162090929057764, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.135021686553955, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7083479762077332, + "num_tokens": 98045098.0, + "step": 3790 + }, + { + "epoch": 0.41631891060839005, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.327816963195801, + "learning_rate": 1e-06, + "loss": 1.0406, + "mean_token_accuracy": 0.6850515604019165, + "num_tokens": 98069357.0, + "step": 3791 + }, + { + "epoch": 0.41642872831100375, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.290477752685547, + "learning_rate": 1e-06, + "loss": 1.0694, + "mean_token_accuracy": 0.6822490096092224, + "num_tokens": 98094848.0, + "step": 3792 + }, + { + "epoch": 0.4165385460136174, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.3762624263763428, + "learning_rate": 1e-06, + "loss": 1.0463, + "mean_token_accuracy": 0.6862781047821045, + "num_tokens": 98118179.0, + "step": 3793 + }, + { + "epoch": 0.41664836371623104, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.2833597660064697, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6956813335418701, + "num_tokens": 98142211.0, + "step": 3794 + }, + { + "epoch": 0.41675818141884474, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.256300210952759, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6954355239868164, + "num_tokens": 98168425.0, + "step": 3795 + }, + { + "epoch": 0.4168679991214584, + "ewc_loss": 1.7881393432617188e-06, + "grad_norm": 2.4287185668945312, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7130125761032104, + "num_tokens": 98191140.0, + "step": 3796 + }, + { + "epoch": 0.41697781682407203, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.039923906326294, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7108001708984375, + "num_tokens": 98221268.0, + "step": 3797 + }, + { + "epoch": 0.4170876345266857, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.4052319526672363, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7126902341842651, + "num_tokens": 98243642.0, + "step": 3798 + }, + { + "epoch": 0.4171974522292994, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.3095240592956543, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7317849397659302, + "num_tokens": 98265808.0, + "step": 3799 + }, + { + "epoch": 0.417307269931913, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.0386524200439453, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7113350629806519, + "num_tokens": 98293589.0, + "step": 3800 + }, + { + "epoch": 0.4174170876345267, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.1414594650268555, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6952575445175171, + "num_tokens": 98321281.0, + "step": 3801 + }, + { + "epoch": 0.4175269053371403, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.1895687580108643, + "learning_rate": 1e-06, + "loss": 1.0726, + "mean_token_accuracy": 0.6774591207504272, + "num_tokens": 98347302.0, + "step": 3802 + }, + { + "epoch": 0.417636723039754, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.1772079467773438, + "learning_rate": 1e-06, + "loss": 1.0565, + "mean_token_accuracy": 0.6773156523704529, + "num_tokens": 98375310.0, + "step": 3803 + }, + { + "epoch": 0.41774654074236767, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 1.9043147563934326, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7162231206893921, + "num_tokens": 98405962.0, + "step": 3804 + }, + { + "epoch": 0.4178563584449813, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.250577688217163, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7347484230995178, + "num_tokens": 98429344.0, + "step": 3805 + }, + { + "epoch": 0.417966176147595, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.217552661895752, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7028008103370667, + "num_tokens": 98455018.0, + "step": 3806 + }, + { + "epoch": 0.41807599385020866, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.0940752029418945, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.7089799642562866, + "num_tokens": 98484221.0, + "step": 3807 + }, + { + "epoch": 0.4181858115528223, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 1.9996496438980103, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6963396072387695, + "num_tokens": 98515513.0, + "step": 3808 + }, + { + "epoch": 0.41829562925543595, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.438546657562256, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7150854468345642, + "num_tokens": 98537078.0, + "step": 3809 + }, + { + "epoch": 0.41840544695804965, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.5077805519104004, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7063307166099548, + "num_tokens": 98556703.0, + "step": 3810 + }, + { + "epoch": 0.4185152646606633, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.0888559818267822, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6908919215202332, + "num_tokens": 98586006.0, + "step": 3811 + }, + { + "epoch": 0.41862508236327695, + "ewc_loss": 1.8030405044555664e-06, + "grad_norm": 2.3528594970703125, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.6841403245925903, + "num_tokens": 98609590.0, + "step": 3812 + }, + { + "epoch": 0.41873490006589065, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 2.3304364681243896, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7072886228561401, + "num_tokens": 98632955.0, + "step": 3813 + }, + { + "epoch": 0.4188447177685043, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 2.528536319732666, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.6991830468177795, + "num_tokens": 98654276.0, + "step": 3814 + }, + { + "epoch": 0.41895453547111794, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 2.4253876209259033, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.7011195421218872, + "num_tokens": 98676588.0, + "step": 3815 + }, + { + "epoch": 0.4190643531737316, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 2.1085569858551025, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6938839554786682, + "num_tokens": 98704836.0, + "step": 3816 + }, + { + "epoch": 0.4191741708763453, + "ewc_loss": 1.8104910850524902e-06, + "grad_norm": 2.107813596725464, + "learning_rate": 1e-06, + "loss": 1.1, + "mean_token_accuracy": 0.6711581945419312, + "num_tokens": 98736500.0, + "step": 3817 + }, + { + "epoch": 0.41928398857895893, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.5400121212005615, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6981972455978394, + "num_tokens": 98758217.0, + "step": 3818 + }, + { + "epoch": 0.4193938062815726, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.207395076751709, + "learning_rate": 1e-06, + "loss": 1.045, + "mean_token_accuracy": 0.6804752349853516, + "num_tokens": 98782642.0, + "step": 3819 + }, + { + "epoch": 0.4195036239841862, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.1163389682769775, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7002137303352356, + "num_tokens": 98809715.0, + "step": 3820 + }, + { + "epoch": 0.4196134416867999, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.3181240558624268, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7088508605957031, + "num_tokens": 98832396.0, + "step": 3821 + }, + { + "epoch": 0.41972325938941357, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.3758389949798584, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.7075028419494629, + "num_tokens": 98855032.0, + "step": 3822 + }, + { + "epoch": 0.4198330770920272, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.0132203102111816, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6835615634918213, + "num_tokens": 98885502.0, + "step": 3823 + }, + { + "epoch": 0.4199428947946409, + "ewc_loss": 1.817941665649414e-06, + "grad_norm": 2.2966580390930176, + "learning_rate": 1e-06, + "loss": 1.0743, + "mean_token_accuracy": 0.6830140948295593, + "num_tokens": 98910833.0, + "step": 3824 + }, + { + "epoch": 0.42005271249725457, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.155094623565674, + "learning_rate": 1e-06, + "loss": 1.0742, + "mean_token_accuracy": 0.6780565977096558, + "num_tokens": 98939922.0, + "step": 3825 + }, + { + "epoch": 0.4201625301998682, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.209134340286255, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6978155374526978, + "num_tokens": 98966447.0, + "step": 3826 + }, + { + "epoch": 0.42027234790248186, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.1994802951812744, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7011052370071411, + "num_tokens": 98990069.0, + "step": 3827 + }, + { + "epoch": 0.42038216560509556, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.1685585975646973, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7127256393432617, + "num_tokens": 99018779.0, + "step": 3828 + }, + { + "epoch": 0.4204919833077092, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.163414478302002, + "learning_rate": 1e-06, + "loss": 1.0476, + "mean_token_accuracy": 0.6835682392120361, + "num_tokens": 99046989.0, + "step": 3829 + }, + { + "epoch": 0.42060180101032285, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.200350046157837, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6906502842903137, + "num_tokens": 99075830.0, + "step": 3830 + }, + { + "epoch": 0.42071161871293655, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.2265100479125977, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6929093599319458, + "num_tokens": 99104043.0, + "step": 3831 + }, + { + "epoch": 0.4208214364155502, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.094336748123169, + "learning_rate": 1e-06, + "loss": 1.043, + "mean_token_accuracy": 0.6915290355682373, + "num_tokens": 99134589.0, + "step": 3832 + }, + { + "epoch": 0.42093125411816384, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.2932486534118652, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.688388466835022, + "num_tokens": 99159009.0, + "step": 3833 + }, + { + "epoch": 0.4210410718207775, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.1112165451049805, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6876363754272461, + "num_tokens": 99188690.0, + "step": 3834 + }, + { + "epoch": 0.4211508895233912, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.4289495944976807, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7202664613723755, + "num_tokens": 99209626.0, + "step": 3835 + }, + { + "epoch": 0.42126070722600484, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.142540693283081, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6925906538963318, + "num_tokens": 99237195.0, + "step": 3836 + }, + { + "epoch": 0.4213705249286185, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.098545789718628, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.716636061668396, + "num_tokens": 99265505.0, + "step": 3837 + }, + { + "epoch": 0.42148034263123213, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.1780507564544678, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7176466584205627, + "num_tokens": 99290547.0, + "step": 3838 + }, + { + "epoch": 0.42159016033384583, + "ewc_loss": 1.8328428268432617e-06, + "grad_norm": 2.357623338699341, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7033321261405945, + "num_tokens": 99313701.0, + "step": 3839 + }, + { + "epoch": 0.4216999780364595, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.1209912300109863, + "learning_rate": 1e-06, + "loss": 1.0763, + "mean_token_accuracy": 0.6779540777206421, + "num_tokens": 99344109.0, + "step": 3840 + }, + { + "epoch": 0.4218097957390731, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2009470462799072, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7084445953369141, + "num_tokens": 99370893.0, + "step": 3841 + }, + { + "epoch": 0.4219196134416868, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.293020486831665, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7030057311058044, + "num_tokens": 99393832.0, + "step": 3842 + }, + { + "epoch": 0.42202943114430047, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.3891971111297607, + "learning_rate": 1e-06, + "loss": 1.1181, + "mean_token_accuracy": 0.6663635969161987, + "num_tokens": 99418060.0, + "step": 3843 + }, + { + "epoch": 0.4221392488469141, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.3242058753967285, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.710336446762085, + "num_tokens": 99441684.0, + "step": 3844 + }, + { + "epoch": 0.42224906654952776, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.093332529067993, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7034749984741211, + "num_tokens": 99469380.0, + "step": 3845 + }, + { + "epoch": 0.42235888425214146, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.8095996379852295, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7283424139022827, + "num_tokens": 99487015.0, + "step": 3846 + }, + { + "epoch": 0.4224687019547551, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2359752655029297, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6897568702697754, + "num_tokens": 99513084.0, + "step": 3847 + }, + { + "epoch": 0.42257851965736876, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2930123805999756, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6911804676055908, + "num_tokens": 99538790.0, + "step": 3848 + }, + { + "epoch": 0.4226883373599824, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2034220695495605, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7104982137680054, + "num_tokens": 99563439.0, + "step": 3849 + }, + { + "epoch": 0.4227981550625961, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2884793281555176, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7257742881774902, + "num_tokens": 99586655.0, + "step": 3850 + }, + { + "epoch": 0.42290797276520975, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.2448086738586426, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.6859520077705383, + "num_tokens": 99611237.0, + "step": 3851 + }, + { + "epoch": 0.4230177904678234, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 1.9347870349884033, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.6950746178627014, + "num_tokens": 99645810.0, + "step": 3852 + }, + { + "epoch": 0.4231276081704371, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.6299922466278076, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.699683666229248, + "num_tokens": 99664956.0, + "step": 3853 + }, + { + "epoch": 0.42323742587305074, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.0413172245025635, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7176734805107117, + "num_tokens": 99693325.0, + "step": 3854 + }, + { + "epoch": 0.4233472435756644, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.428811550140381, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7175678014755249, + "num_tokens": 99715867.0, + "step": 3855 + }, + { + "epoch": 0.42345706127827804, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.217689037322998, + "learning_rate": 1e-06, + "loss": 1.101, + "mean_token_accuracy": 0.6736465692520142, + "num_tokens": 99745782.0, + "step": 3856 + }, + { + "epoch": 0.42356687898089174, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.3780882358551025, + "learning_rate": 1e-06, + "loss": 1.052, + "mean_token_accuracy": 0.6840888261795044, + "num_tokens": 99768440.0, + "step": 3857 + }, + { + "epoch": 0.4236766966835054, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.5136187076568604, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.6944645643234253, + "num_tokens": 99789971.0, + "step": 3858 + }, + { + "epoch": 0.42378651438611903, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.3823282718658447, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6992625594139099, + "num_tokens": 99813570.0, + "step": 3859 + }, + { + "epoch": 0.42389633208873273, + "ewc_loss": 1.8477439880371094e-06, + "grad_norm": 2.40513277053833, + "learning_rate": 1e-06, + "loss": 1.0512, + "mean_token_accuracy": 0.6872438192367554, + "num_tokens": 99837657.0, + "step": 3860 + }, + { + "epoch": 0.4240061497913464, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.1098461151123047, + "learning_rate": 1e-06, + "loss": 1.1139, + "mean_token_accuracy": 0.6662435531616211, + "num_tokens": 99868728.0, + "step": 3861 + }, + { + "epoch": 0.42411596749396, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.206266403198242, + "learning_rate": 1e-06, + "loss": 1.0557, + "mean_token_accuracy": 0.6828813552856445, + "num_tokens": 99896969.0, + "step": 3862 + }, + { + "epoch": 0.42422578519657367, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.3580265045166016, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7098933458328247, + "num_tokens": 99919296.0, + "step": 3863 + }, + { + "epoch": 0.42433560289918737, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.498711347579956, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.709287166595459, + "num_tokens": 99941534.0, + "step": 3864 + }, + { + "epoch": 0.424445420601801, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.1903092861175537, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7074096202850342, + "num_tokens": 99968505.0, + "step": 3865 + }, + { + "epoch": 0.42455523830441466, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.127445936203003, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7308685779571533, + "num_tokens": 99993455.0, + "step": 3866 + }, + { + "epoch": 0.4246650560070283, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.340658187866211, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6983098983764648, + "num_tokens": 100018287.0, + "step": 3867 + }, + { + "epoch": 0.424774873709642, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.422067642211914, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7101539373397827, + "num_tokens": 100040541.0, + "step": 3868 + }, + { + "epoch": 0.42488469141225566, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.129868984222412, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7180137634277344, + "num_tokens": 100066442.0, + "step": 3869 + }, + { + "epoch": 0.4249945091148693, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.345747232437134, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7506636381149292, + "num_tokens": 100086709.0, + "step": 3870 + }, + { + "epoch": 0.425104326817483, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.3225502967834473, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7135394215583801, + "num_tokens": 100110542.0, + "step": 3871 + }, + { + "epoch": 0.42521414452009665, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.0129811763763428, + "learning_rate": 1e-06, + "loss": 1.1053, + "mean_token_accuracy": 0.6741622686386108, + "num_tokens": 100141320.0, + "step": 3872 + }, + { + "epoch": 0.4253239622227103, + "ewc_loss": 1.8551945686340332e-06, + "grad_norm": 2.4075965881347656, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.6846781969070435, + "num_tokens": 100164223.0, + "step": 3873 + }, + { + "epoch": 0.42543377992532394, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.4060795307159424, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7135382890701294, + "num_tokens": 100185941.0, + "step": 3874 + }, + { + "epoch": 0.42554359762793764, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.1900570392608643, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7104424238204956, + "num_tokens": 100210732.0, + "step": 3875 + }, + { + "epoch": 0.4256534153305513, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.1263439655303955, + "learning_rate": 1e-06, + "loss": 1.0649, + "mean_token_accuracy": 0.6823253631591797, + "num_tokens": 100238137.0, + "step": 3876 + }, + { + "epoch": 0.42576323303316493, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.153581142425537, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6919782161712646, + "num_tokens": 100264186.0, + "step": 3877 + }, + { + "epoch": 0.4258730507357786, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 1.9546432495117188, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6990736722946167, + "num_tokens": 100294331.0, + "step": 3878 + }, + { + "epoch": 0.4259828684383923, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.3038854598999023, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7343801856040955, + "num_tokens": 100316130.0, + "step": 3879 + }, + { + "epoch": 0.4260926861410059, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.345285415649414, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7011383771896362, + "num_tokens": 100340052.0, + "step": 3880 + }, + { + "epoch": 0.4262025038436196, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.3307976722717285, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.699702262878418, + "num_tokens": 100365026.0, + "step": 3881 + }, + { + "epoch": 0.4263123215462333, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.3895161151885986, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6963238716125488, + "num_tokens": 100386885.0, + "step": 3882 + }, + { + "epoch": 0.4264221392488469, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.228144884109497, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6839416027069092, + "num_tokens": 100413857.0, + "step": 3883 + }, + { + "epoch": 0.42653195695146057, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 1.8167335987091064, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7012180089950562, + "num_tokens": 100448443.0, + "step": 3884 + }, + { + "epoch": 0.4266417746540742, + "ewc_loss": 1.862645149230957e-06, + "grad_norm": 2.1939709186553955, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7282641530036926, + "num_tokens": 100473036.0, + "step": 3885 + }, + { + "epoch": 0.4267515923566879, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.5627400875091553, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6992413401603699, + "num_tokens": 100492808.0, + "step": 3886 + }, + { + "epoch": 0.42686141005930156, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.31964111328125, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6871684789657593, + "num_tokens": 100516357.0, + "step": 3887 + }, + { + "epoch": 0.4269712277619152, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.234370708465576, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7090998888015747, + "num_tokens": 100539907.0, + "step": 3888 + }, + { + "epoch": 0.4270810454645289, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 1.9295754432678223, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7160714864730835, + "num_tokens": 100570127.0, + "step": 3889 + }, + { + "epoch": 0.42719086316714255, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.7870490550994873, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7221713066101074, + "num_tokens": 100587307.0, + "step": 3890 + }, + { + "epoch": 0.4273006808697562, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.394840955734253, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6958723664283752, + "num_tokens": 100610619.0, + "step": 3891 + }, + { + "epoch": 0.42741049857236985, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.1199398040771484, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.694831132888794, + "num_tokens": 100639567.0, + "step": 3892 + }, + { + "epoch": 0.42752031627498355, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.4780170917510986, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7061243653297424, + "num_tokens": 100659524.0, + "step": 3893 + }, + { + "epoch": 0.4276301339775972, + "ewc_loss": 1.8775463104248047e-06, + "grad_norm": 2.3656609058380127, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.709114670753479, + "num_tokens": 100681819.0, + "step": 3894 + }, + { + "epoch": 0.42773995168021084, + "ewc_loss": 1.8849968910217285e-06, + "grad_norm": 1.916115164756775, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.7015969157218933, + "num_tokens": 100715905.0, + "step": 3895 + }, + { + "epoch": 0.4278497693828245, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 1.983782410621643, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6950141191482544, + "num_tokens": 100751228.0, + "step": 3896 + }, + { + "epoch": 0.4279595870854382, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.4472594261169434, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7036852836608887, + "num_tokens": 100773363.0, + "step": 3897 + }, + { + "epoch": 0.42806940478805183, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.043837308883667, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6890702247619629, + "num_tokens": 100801727.0, + "step": 3898 + }, + { + "epoch": 0.4281792224906655, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 1.9229363203048706, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7043316960334778, + "num_tokens": 100834173.0, + "step": 3899 + }, + { + "epoch": 0.4282890401932792, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.5291879177093506, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.7012617588043213, + "num_tokens": 100855435.0, + "step": 3900 + }, + { + "epoch": 0.4283988578958928, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.0552515983581543, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6911097168922424, + "num_tokens": 100884327.0, + "step": 3901 + }, + { + "epoch": 0.42850867559850647, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.1739389896392822, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.6768093109130859, + "num_tokens": 100910211.0, + "step": 3902 + }, + { + "epoch": 0.4286184933011201, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.2691338062286377, + "learning_rate": 1e-06, + "loss": 1.0615, + "mean_token_accuracy": 0.6834639310836792, + "num_tokens": 100935401.0, + "step": 3903 + }, + { + "epoch": 0.4287283110037338, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.3224852085113525, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7134451270103455, + "num_tokens": 100958877.0, + "step": 3904 + }, + { + "epoch": 0.42883812870634747, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.304647445678711, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7131598591804504, + "num_tokens": 100980312.0, + "step": 3905 + }, + { + "epoch": 0.4289479464089611, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.0184409618377686, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7174521684646606, + "num_tokens": 101008596.0, + "step": 3906 + }, + { + "epoch": 0.4290577641115748, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.1893255710601807, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7181936502456665, + "num_tokens": 101033996.0, + "step": 3907 + }, + { + "epoch": 0.42916758181418846, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.278062343597412, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7065103054046631, + "num_tokens": 101057963.0, + "step": 3908 + }, + { + "epoch": 0.4292773995168021, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.6100103855133057, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6949441432952881, + "num_tokens": 101077804.0, + "step": 3909 + }, + { + "epoch": 0.42938721721941575, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.8029472827911377, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6871069669723511, + "num_tokens": 101100214.0, + "step": 3910 + }, + { + "epoch": 0.42949703492202945, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.329662322998047, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7274045944213867, + "num_tokens": 101124305.0, + "step": 3911 + }, + { + "epoch": 0.4296068526246431, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.137362480163574, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7212215662002563, + "num_tokens": 101150289.0, + "step": 3912 + }, + { + "epoch": 0.42971667032725674, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.195119857788086, + "learning_rate": 1e-06, + "loss": 1.0523, + "mean_token_accuracy": 0.691327691078186, + "num_tokens": 101178339.0, + "step": 3913 + }, + { + "epoch": 0.4298264880298704, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.2802605628967285, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7155622243881226, + "num_tokens": 101202614.0, + "step": 3914 + }, + { + "epoch": 0.4299363057324841, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.3218822479248047, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7200926542282104, + "num_tokens": 101224528.0, + "step": 3915 + }, + { + "epoch": 0.43004612343509774, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.1020023822784424, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.733514666557312, + "num_tokens": 101249037.0, + "step": 3916 + }, + { + "epoch": 0.4301559411377114, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 1.987237572669983, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6953654885292053, + "num_tokens": 101282444.0, + "step": 3917 + }, + { + "epoch": 0.4302657588403251, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.13034725189209, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7080758810043335, + "num_tokens": 101313358.0, + "step": 3918 + }, + { + "epoch": 0.43037557654293873, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.070845365524292, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7004070281982422, + "num_tokens": 101340581.0, + "step": 3919 + }, + { + "epoch": 0.4304853942455524, + "ewc_loss": 1.8924474716186523e-06, + "grad_norm": 2.489325523376465, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7186075448989868, + "num_tokens": 101361533.0, + "step": 3920 + }, + { + "epoch": 0.430595211948166, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.3906452655792236, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7098783254623413, + "num_tokens": 101383326.0, + "step": 3921 + }, + { + "epoch": 0.4307050296507797, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.283928632736206, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.7052364945411682, + "num_tokens": 101409081.0, + "step": 3922 + }, + { + "epoch": 0.43081484735339337, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 3.0497331619262695, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.6982119083404541, + "num_tokens": 101438699.0, + "step": 3923 + }, + { + "epoch": 0.430924665056007, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.0596015453338623, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7140288949012756, + "num_tokens": 101464864.0, + "step": 3924 + }, + { + "epoch": 0.43103448275862066, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.400970935821533, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6891536116600037, + "num_tokens": 101488078.0, + "step": 3925 + }, + { + "epoch": 0.43114430046123436, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.3311901092529297, + "learning_rate": 1e-06, + "loss": 1.0873, + "mean_token_accuracy": 0.6820476651191711, + "num_tokens": 101514050.0, + "step": 3926 + }, + { + "epoch": 0.431254118163848, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.392115354537964, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7227234840393066, + "num_tokens": 101536723.0, + "step": 3927 + }, + { + "epoch": 0.43136393586646166, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.278047561645508, + "learning_rate": 1e-06, + "loss": 1.0487, + "mean_token_accuracy": 0.6886993646621704, + "num_tokens": 101561970.0, + "step": 3928 + }, + { + "epoch": 0.43147375356907536, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.149524450302124, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7036476135253906, + "num_tokens": 101589583.0, + "step": 3929 + }, + { + "epoch": 0.431583571271689, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.021021604537964, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.6841274499893188, + "num_tokens": 101621673.0, + "step": 3930 + }, + { + "epoch": 0.43169338897430265, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.416806697845459, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6920243501663208, + "num_tokens": 101644065.0, + "step": 3931 + }, + { + "epoch": 0.4318032066769163, + "ewc_loss": 1.8998980522155762e-06, + "grad_norm": 2.909316301345825, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7296363115310669, + "num_tokens": 101662750.0, + "step": 3932 + }, + { + "epoch": 0.43191302437953, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.1120622158050537, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7098326683044434, + "num_tokens": 101688990.0, + "step": 3933 + }, + { + "epoch": 0.43202284208214364, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.42708420753479, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.697832465171814, + "num_tokens": 101710686.0, + "step": 3934 + }, + { + "epoch": 0.4321326597847573, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.2637953758239746, + "learning_rate": 1e-06, + "loss": 1.0542, + "mean_token_accuracy": 0.6897218227386475, + "num_tokens": 101737899.0, + "step": 3935 + }, + { + "epoch": 0.432242477487371, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.1342926025390625, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6843150854110718, + "num_tokens": 101766454.0, + "step": 3936 + }, + { + "epoch": 0.43235229518998464, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.2442755699157715, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7144585847854614, + "num_tokens": 101791654.0, + "step": 3937 + }, + { + "epoch": 0.4324621128925983, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.195406675338745, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6986774206161499, + "num_tokens": 101817589.0, + "step": 3938 + }, + { + "epoch": 0.43257193059521193, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.016064405441284, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6935421824455261, + "num_tokens": 101847859.0, + "step": 3939 + }, + { + "epoch": 0.43268174829782563, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.2719433307647705, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.7005283236503601, + "num_tokens": 101871930.0, + "step": 3940 + }, + { + "epoch": 0.4327915660004393, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.0901076793670654, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7040988206863403, + "num_tokens": 101898638.0, + "step": 3941 + }, + { + "epoch": 0.4329013837030529, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.206035852432251, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6920969486236572, + "num_tokens": 101926998.0, + "step": 3942 + }, + { + "epoch": 0.43301120140566657, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.4762892723083496, + "learning_rate": 1e-06, + "loss": 1.0256, + "mean_token_accuracy": 0.6915023922920227, + "num_tokens": 101949123.0, + "step": 3943 + }, + { + "epoch": 0.43312101910828027, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.3252170085906982, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.714887261390686, + "num_tokens": 101971407.0, + "step": 3944 + }, + { + "epoch": 0.4332308368108939, + "ewc_loss": 1.9073486328125e-06, + "grad_norm": 2.193556547164917, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7046383619308472, + "num_tokens": 101997943.0, + "step": 3945 + }, + { + "epoch": 0.43334065451350756, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 2.070204019546509, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7080585956573486, + "num_tokens": 102025822.0, + "step": 3946 + }, + { + "epoch": 0.43345047221612126, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 2.089493989944458, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6904056072235107, + "num_tokens": 102053305.0, + "step": 3947 + }, + { + "epoch": 0.4335602899187349, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 2.2298192977905273, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.7027019262313843, + "num_tokens": 102078836.0, + "step": 3948 + }, + { + "epoch": 0.43367010762134856, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 2.1881399154663086, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7238268852233887, + "num_tokens": 102103001.0, + "step": 3949 + }, + { + "epoch": 0.4337799253239622, + "ewc_loss": 1.9222497940063477e-06, + "grad_norm": 2.1418213844299316, + "learning_rate": 1e-06, + "loss": 1.0452, + "mean_token_accuracy": 0.6836147308349609, + "num_tokens": 102131895.0, + "step": 3950 + }, + { + "epoch": 0.4338897430265759, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.350266456604004, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7209444046020508, + "num_tokens": 102156023.0, + "step": 3951 + }, + { + "epoch": 0.43399956072918955, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.631589412689209, + "learning_rate": 1e-06, + "loss": 1.0634, + "mean_token_accuracy": 0.686750054359436, + "num_tokens": 102176947.0, + "step": 3952 + }, + { + "epoch": 0.4341093784318032, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.283827066421509, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.699081301689148, + "num_tokens": 102203157.0, + "step": 3953 + }, + { + "epoch": 0.43421919613441684, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.4031710624694824, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7019818425178528, + "num_tokens": 102226106.0, + "step": 3954 + }, + { + "epoch": 0.43432901383703054, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.352492094039917, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6955801844596863, + "num_tokens": 102250220.0, + "step": 3955 + }, + { + "epoch": 0.4344388315396442, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1933863162994385, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6789961457252502, + "num_tokens": 102276373.0, + "step": 3956 + }, + { + "epoch": 0.43454864924225783, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.3456714153289795, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7089958786964417, + "num_tokens": 102298062.0, + "step": 3957 + }, + { + "epoch": 0.43465846694487154, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.203040599822998, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.700922966003418, + "num_tokens": 102324012.0, + "step": 3958 + }, + { + "epoch": 0.4347682846474852, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1672375202178955, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.6878384947776794, + "num_tokens": 102352815.0, + "step": 3959 + }, + { + "epoch": 0.4348781023500988, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.2502970695495605, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6985957026481628, + "num_tokens": 102378014.0, + "step": 3960 + }, + { + "epoch": 0.4349879200527125, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.22459077835083, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7280148267745972, + "num_tokens": 102402223.0, + "step": 3961 + }, + { + "epoch": 0.4350977377553262, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.087620258331299, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7355141639709473, + "num_tokens": 102429827.0, + "step": 3962 + }, + { + "epoch": 0.4352075554579398, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.0601558685302734, + "learning_rate": 1e-06, + "loss": 1.0335, + "mean_token_accuracy": 0.695465087890625, + "num_tokens": 102458769.0, + "step": 3963 + }, + { + "epoch": 0.43531737316055347, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.072094440460205, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6970691680908203, + "num_tokens": 102487677.0, + "step": 3964 + }, + { + "epoch": 0.43542719086316717, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.111502170562744, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.7044958472251892, + "num_tokens": 102516194.0, + "step": 3965 + }, + { + "epoch": 0.4355370085657808, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.6000053882598877, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7204053997993469, + "num_tokens": 102534115.0, + "step": 3966 + }, + { + "epoch": 0.43564682626839446, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.3425443172454834, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7151232957839966, + "num_tokens": 102557009.0, + "step": 3967 + }, + { + "epoch": 0.4357566439710081, + "ewc_loss": 1.9371509552001953e-06, + "grad_norm": 2.190028667449951, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6963468790054321, + "num_tokens": 102581438.0, + "step": 3968 + }, + { + "epoch": 0.4358664616736218, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.3644211292266846, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7165103554725647, + "num_tokens": 102603822.0, + "step": 3969 + }, + { + "epoch": 0.43597627937623545, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.3710358142852783, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.707314133644104, + "num_tokens": 102625582.0, + "step": 3970 + }, + { + "epoch": 0.4360860970788491, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1788480281829834, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6885976195335388, + "num_tokens": 102655237.0, + "step": 3971 + }, + { + "epoch": 0.43619591478146275, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1925787925720215, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7072684168815613, + "num_tokens": 102681610.0, + "step": 3972 + }, + { + "epoch": 0.43630573248407645, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.0836679935455322, + "learning_rate": 1e-06, + "loss": 1.0364, + "mean_token_accuracy": 0.6820621490478516, + "num_tokens": 102711636.0, + "step": 3973 + }, + { + "epoch": 0.4364155501866901, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.2245733737945557, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6799190640449524, + "num_tokens": 102740269.0, + "step": 3974 + }, + { + "epoch": 0.43652536788930374, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.192748546600342, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7297207117080688, + "num_tokens": 102765140.0, + "step": 3975 + }, + { + "epoch": 0.43663518559191744, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.354923725128174, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6954250931739807, + "num_tokens": 102790550.0, + "step": 3976 + }, + { + "epoch": 0.4367450032945311, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1113083362579346, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6927505731582642, + "num_tokens": 102818283.0, + "step": 3977 + }, + { + "epoch": 0.43685482099714473, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.4126484394073486, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6971985101699829, + "num_tokens": 102839697.0, + "step": 3978 + }, + { + "epoch": 0.4369646386997584, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.0763189792633057, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6916998624801636, + "num_tokens": 102868038.0, + "step": 3979 + }, + { + "epoch": 0.4370744564023721, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.4234771728515625, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6963419318199158, + "num_tokens": 102890261.0, + "step": 3980 + }, + { + "epoch": 0.4371842741049857, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.5078959465026855, + "learning_rate": 1e-06, + "loss": 1.0466, + "mean_token_accuracy": 0.6844359636306763, + "num_tokens": 102912602.0, + "step": 3981 + }, + { + "epoch": 0.43729409180759937, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.296518087387085, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7199277877807617, + "num_tokens": 102936652.0, + "step": 3982 + }, + { + "epoch": 0.4374039095102131, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.067835807800293, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6793205142021179, + "num_tokens": 102964564.0, + "step": 3983 + }, + { + "epoch": 0.4375137272128267, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1211764812469482, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7240635752677917, + "num_tokens": 102991026.0, + "step": 3984 + }, + { + "epoch": 0.43762354491544037, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1915881633758545, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6951494812965393, + "num_tokens": 103017666.0, + "step": 3985 + }, + { + "epoch": 0.437733362618054, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.495215892791748, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7209575772285461, + "num_tokens": 103038445.0, + "step": 3986 + }, + { + "epoch": 0.4378431803206677, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1123456954956055, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6873983144760132, + "num_tokens": 103069019.0, + "step": 3987 + }, + { + "epoch": 0.43795299802328136, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.5038352012634277, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7073475122451782, + "num_tokens": 103090281.0, + "step": 3988 + }, + { + "epoch": 0.438062815725895, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.026362895965576, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7310411334037781, + "num_tokens": 103118962.0, + "step": 3989 + }, + { + "epoch": 0.43817263342850865, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.16215443611145, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7190139293670654, + "num_tokens": 103145401.0, + "step": 3990 + }, + { + "epoch": 0.43828245113112235, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.157663345336914, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.701036810874939, + "num_tokens": 103172448.0, + "step": 3991 + }, + { + "epoch": 0.438392268833736, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.2359158992767334, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7068850994110107, + "num_tokens": 103198006.0, + "step": 3992 + }, + { + "epoch": 0.43850208653634964, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.295712471008301, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7073416709899902, + "num_tokens": 103222454.0, + "step": 3993 + }, + { + "epoch": 0.43861190423896335, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1901168823242188, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7091175317764282, + "num_tokens": 103249089.0, + "step": 3994 + }, + { + "epoch": 0.438721721941577, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.3729188442230225, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7297834157943726, + "num_tokens": 103270197.0, + "step": 3995 + }, + { + "epoch": 0.43883153964419064, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1743414402008057, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6905217170715332, + "num_tokens": 103297499.0, + "step": 3996 + }, + { + "epoch": 0.4389413573468043, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1929121017456055, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.715186357498169, + "num_tokens": 103324855.0, + "step": 3997 + }, + { + "epoch": 0.439051175049418, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1660685539245605, + "learning_rate": 1e-06, + "loss": 1.0946, + "mean_token_accuracy": 0.6807642579078674, + "num_tokens": 103352855.0, + "step": 3998 + }, + { + "epoch": 0.43916099275203163, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.267493963241577, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.6982898712158203, + "num_tokens": 103375974.0, + "step": 3999 + }, + { + "epoch": 0.4392708104546453, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.0069127082824707, + "learning_rate": 1e-06, + "loss": 1.1033, + "mean_token_accuracy": 0.6668400168418884, + "num_tokens": 103407738.0, + "step": 4000 + }, + { + "epoch": 0.4393806281572589, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.0915160179138184, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6966953277587891, + "num_tokens": 103437027.0, + "step": 4001 + }, + { + "epoch": 0.4394904458598726, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.346108913421631, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7057309746742249, + "num_tokens": 103460046.0, + "step": 4002 + }, + { + "epoch": 0.43960026356248627, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.2396352291107178, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7041407227516174, + "num_tokens": 103485621.0, + "step": 4003 + }, + { + "epoch": 0.4397100812650999, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 1.8854033946990967, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6834671497344971, + "num_tokens": 103518984.0, + "step": 4004 + }, + { + "epoch": 0.4398198989677136, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.4438977241516113, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.6944530606269836, + "num_tokens": 103541595.0, + "step": 4005 + }, + { + "epoch": 0.43992971667032726, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.3504951000213623, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6889374852180481, + "num_tokens": 103565995.0, + "step": 4006 + }, + { + "epoch": 0.4400395343729409, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.447047472000122, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6902203559875488, + "num_tokens": 103589961.0, + "step": 4007 + }, + { + "epoch": 0.44014935207555456, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 1.93858802318573, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6867669820785522, + "num_tokens": 103624015.0, + "step": 4008 + }, + { + "epoch": 0.44025916977816826, + "ewc_loss": 1.952052116394043e-06, + "grad_norm": 2.1710968017578125, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7172660231590271, + "num_tokens": 103648981.0, + "step": 4009 + }, + { + "epoch": 0.4403689874807819, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.3118906021118164, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7149057388305664, + "num_tokens": 103672992.0, + "step": 4010 + }, + { + "epoch": 0.44047880518339555, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.4629247188568115, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7099261283874512, + "num_tokens": 103693983.0, + "step": 4011 + }, + { + "epoch": 0.44058862288600925, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.300525665283203, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6872632503509521, + "num_tokens": 103717413.0, + "step": 4012 + }, + { + "epoch": 0.4406984405886229, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.4345812797546387, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7066659331321716, + "num_tokens": 103739793.0, + "step": 4013 + }, + { + "epoch": 0.44080825829123654, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.3194713592529297, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7140097618103027, + "num_tokens": 103763425.0, + "step": 4014 + }, + { + "epoch": 0.4409180759938502, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.4693069458007812, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7280755043029785, + "num_tokens": 103784225.0, + "step": 4015 + }, + { + "epoch": 0.4410278936964639, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.0993099212646484, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7043653726577759, + "num_tokens": 103811751.0, + "step": 4016 + }, + { + "epoch": 0.44113771139907754, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 1.9405497312545776, + "learning_rate": 1e-06, + "loss": 1.03, + "mean_token_accuracy": 0.6925470232963562, + "num_tokens": 103843323.0, + "step": 4017 + }, + { + "epoch": 0.4412475291016912, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.259308099746704, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.7100578546524048, + "num_tokens": 103869122.0, + "step": 4018 + }, + { + "epoch": 0.44135734680430483, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.268505573272705, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6911972761154175, + "num_tokens": 103892813.0, + "step": 4019 + }, + { + "epoch": 0.44146716450691853, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 1.8992152214050293, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6911212205886841, + "num_tokens": 103925498.0, + "step": 4020 + }, + { + "epoch": 0.4415769822095322, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.337120771408081, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.715471625328064, + "num_tokens": 103948200.0, + "step": 4021 + }, + { + "epoch": 0.4416867999121458, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.304203748703003, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7318012714385986, + "num_tokens": 103969306.0, + "step": 4022 + }, + { + "epoch": 0.4417966176147595, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.368548631668091, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7119487524032593, + "num_tokens": 103992620.0, + "step": 4023 + }, + { + "epoch": 0.44190643531737317, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.159188747406006, + "learning_rate": 1e-06, + "loss": 1.047, + "mean_token_accuracy": 0.6977636814117432, + "num_tokens": 104022069.0, + "step": 4024 + }, + { + "epoch": 0.4420162530199868, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.100813150405884, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6979731917381287, + "num_tokens": 104052384.0, + "step": 4025 + }, + { + "epoch": 0.44212607072260046, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.152099609375, + "learning_rate": 1e-06, + "loss": 1.0554, + "mean_token_accuracy": 0.6861989498138428, + "num_tokens": 104080332.0, + "step": 4026 + }, + { + "epoch": 0.44223588842521416, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.1771748065948486, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6920666694641113, + "num_tokens": 104106260.0, + "step": 4027 + }, + { + "epoch": 0.4423457061278278, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 1.9799840450286865, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6964039206504822, + "num_tokens": 104135474.0, + "step": 4028 + }, + { + "epoch": 0.44245552383044146, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.3653275966644287, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7030609846115112, + "num_tokens": 104157966.0, + "step": 4029 + }, + { + "epoch": 0.4425653415330551, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.4287030696868896, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6940348148345947, + "num_tokens": 104180901.0, + "step": 4030 + }, + { + "epoch": 0.4426751592356688, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.099932909011841, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7030795216560364, + "num_tokens": 104210427.0, + "step": 4031 + }, + { + "epoch": 0.44278497693828245, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.0129382610321045, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7114391326904297, + "num_tokens": 104240781.0, + "step": 4032 + }, + { + "epoch": 0.4428947946408961, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.270991325378418, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6996410489082336, + "num_tokens": 104265444.0, + "step": 4033 + }, + { + "epoch": 0.4430046123435098, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.2866811752319336, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7104527950286865, + "num_tokens": 104289851.0, + "step": 4034 + }, + { + "epoch": 0.44311443004612344, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.0707337856292725, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7082287073135376, + "num_tokens": 104318477.0, + "step": 4035 + }, + { + "epoch": 0.4432242477487371, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.2251811027526855, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.689026951789856, + "num_tokens": 104344467.0, + "step": 4036 + }, + { + "epoch": 0.44333406545135073, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.271055221557617, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6887024641036987, + "num_tokens": 104370130.0, + "step": 4037 + }, + { + "epoch": 0.44344388315396444, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.2512025833129883, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7219313383102417, + "num_tokens": 104393844.0, + "step": 4038 + }, + { + "epoch": 0.4435537008565781, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.589993715286255, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6983813047409058, + "num_tokens": 104413570.0, + "step": 4039 + }, + { + "epoch": 0.4436635185591917, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.1245594024658203, + "learning_rate": 1e-06, + "loss": 1.1053, + "mean_token_accuracy": 0.6675496101379395, + "num_tokens": 104443227.0, + "step": 4040 + }, + { + "epoch": 0.44377333626180543, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.299755334854126, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.704745352268219, + "num_tokens": 104467348.0, + "step": 4041 + }, + { + "epoch": 0.4438831539644191, + "ewc_loss": 1.9669532775878906e-06, + "grad_norm": 2.468292474746704, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7169619798660278, + "num_tokens": 104487199.0, + "step": 4042 + }, + { + "epoch": 0.4439929716670327, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3954873085021973, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7046859264373779, + "num_tokens": 104508389.0, + "step": 4043 + }, + { + "epoch": 0.44410278936964637, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 1.9798396825790405, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7083384990692139, + "num_tokens": 104541411.0, + "step": 4044 + }, + { + "epoch": 0.44421260707226007, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.09281587600708, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.713546872138977, + "num_tokens": 104568153.0, + "step": 4045 + }, + { + "epoch": 0.4443224247748737, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.426740884780884, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6928204298019409, + "num_tokens": 104591936.0, + "step": 4046 + }, + { + "epoch": 0.44443224247748736, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.1407155990600586, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6856244802474976, + "num_tokens": 104621841.0, + "step": 4047 + }, + { + "epoch": 0.444542060180101, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.1547393798828125, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6988312602043152, + "num_tokens": 104651647.0, + "step": 4048 + }, + { + "epoch": 0.4446518778827147, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.386756658554077, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6891153454780579, + "num_tokens": 104677572.0, + "step": 4049 + }, + { + "epoch": 0.44476169558532835, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.0915701389312744, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7157031297683716, + "num_tokens": 104706092.0, + "step": 4050 + }, + { + "epoch": 0.444871513287942, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.235048532485962, + "learning_rate": 1e-06, + "loss": 1.0566, + "mean_token_accuracy": 0.6853386163711548, + "num_tokens": 104732079.0, + "step": 4051 + }, + { + "epoch": 0.4449813309905557, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.197887897491455, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6874465942382812, + "num_tokens": 104760767.0, + "step": 4052 + }, + { + "epoch": 0.44509114869316935, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.1868958473205566, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.71656334400177, + "num_tokens": 104788430.0, + "step": 4053 + }, + { + "epoch": 0.445200966395783, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.2273104190826416, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6902586221694946, + "num_tokens": 104817342.0, + "step": 4054 + }, + { + "epoch": 0.44531078409839664, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.161118984222412, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7004135847091675, + "num_tokens": 104843385.0, + "step": 4055 + }, + { + "epoch": 0.44542060180101034, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.2452123165130615, + "learning_rate": 1e-06, + "loss": 1.0718, + "mean_token_accuracy": 0.6768316626548767, + "num_tokens": 104868452.0, + "step": 4056 + }, + { + "epoch": 0.445530419503624, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3190340995788574, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.722678542137146, + "num_tokens": 104890186.0, + "step": 4057 + }, + { + "epoch": 0.44564023720623763, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.2631077766418457, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.6937388777732849, + "num_tokens": 104914247.0, + "step": 4058 + }, + { + "epoch": 0.44575005490885133, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.1689980030059814, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6941639184951782, + "num_tokens": 104939335.0, + "step": 4059 + }, + { + "epoch": 0.445859872611465, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3398139476776123, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7110954523086548, + "num_tokens": 104962635.0, + "step": 4060 + }, + { + "epoch": 0.4459696903140786, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3871653079986572, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7254610657691956, + "num_tokens": 104983584.0, + "step": 4061 + }, + { + "epoch": 0.44607950801669227, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.389082908630371, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.721161961555481, + "num_tokens": 105004819.0, + "step": 4062 + }, + { + "epoch": 0.446189325719306, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3673946857452393, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.733428955078125, + "num_tokens": 105026420.0, + "step": 4063 + }, + { + "epoch": 0.4462991434219196, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.37058424949646, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7053465843200684, + "num_tokens": 105048663.0, + "step": 4064 + }, + { + "epoch": 0.44640896112453327, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.472092866897583, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.704609751701355, + "num_tokens": 105072117.0, + "step": 4065 + }, + { + "epoch": 0.4465187788271469, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.169593095779419, + "learning_rate": 1e-06, + "loss": 1.1142, + "mean_token_accuracy": 0.6728291511535645, + "num_tokens": 105100997.0, + "step": 4066 + }, + { + "epoch": 0.4466285965297606, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.209474802017212, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6911340951919556, + "num_tokens": 105126987.0, + "step": 4067 + }, + { + "epoch": 0.44673841423237426, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.2698144912719727, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.70974200963974, + "num_tokens": 105151037.0, + "step": 4068 + }, + { + "epoch": 0.4468482319349879, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.150332450866699, + "learning_rate": 1e-06, + "loss": 1.0621, + "mean_token_accuracy": 0.6813713908195496, + "num_tokens": 105179052.0, + "step": 4069 + }, + { + "epoch": 0.4469580496376016, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.306029796600342, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.711841344833374, + "num_tokens": 105202762.0, + "step": 4070 + }, + { + "epoch": 0.44706786734021525, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.098801851272583, + "learning_rate": 1e-06, + "loss": 1.061, + "mean_token_accuracy": 0.6821913719177246, + "num_tokens": 105231743.0, + "step": 4071 + }, + { + "epoch": 0.4471776850428289, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.1439778804779053, + "learning_rate": 1e-06, + "loss": 1.0646, + "mean_token_accuracy": 0.677168607711792, + "num_tokens": 105258755.0, + "step": 4072 + }, + { + "epoch": 0.44728750274544254, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.314772129058838, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6931942105293274, + "num_tokens": 105283134.0, + "step": 4073 + }, + { + "epoch": 0.44739732044805625, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.547731637954712, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7344608902931213, + "num_tokens": 105302217.0, + "step": 4074 + }, + { + "epoch": 0.4475071381506699, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3660168647766113, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.700744092464447, + "num_tokens": 105324669.0, + "step": 4075 + }, + { + "epoch": 0.44761695585328354, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1710846424102783, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6887451410293579, + "num_tokens": 105351150.0, + "step": 4076 + }, + { + "epoch": 0.4477267735558972, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.8720295429229736, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7011326551437378, + "num_tokens": 105382344.0, + "step": 4077 + }, + { + "epoch": 0.4478365912585109, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9650565385818481, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.714975893497467, + "num_tokens": 105412565.0, + "step": 4078 + }, + { + "epoch": 0.44794640896112453, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.43632173538208, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6979557871818542, + "num_tokens": 105434373.0, + "step": 4079 + }, + { + "epoch": 0.4480562266637382, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.3721654415130615, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.6940789222717285, + "num_tokens": 105457771.0, + "step": 4080 + }, + { + "epoch": 0.4481660443663519, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0401179790496826, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7187862396240234, + "num_tokens": 105483652.0, + "step": 4081 + }, + { + "epoch": 0.4482758620689655, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.3009586334228516, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7221683263778687, + "num_tokens": 105506506.0, + "step": 4082 + }, + { + "epoch": 0.44838567977157917, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.039170742034912, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7148540616035461, + "num_tokens": 105535039.0, + "step": 4083 + }, + { + "epoch": 0.4484954974741928, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1128101348876953, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6981487274169922, + "num_tokens": 105561456.0, + "step": 4084 + }, + { + "epoch": 0.4486053151768065, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.221583843231201, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6851378679275513, + "num_tokens": 105587627.0, + "step": 4085 + }, + { + "epoch": 0.44871513287942016, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.3344945907592773, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7076494097709656, + "num_tokens": 105613040.0, + "step": 4086 + }, + { + "epoch": 0.4488249505820338, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.215900182723999, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.7014249563217163, + "num_tokens": 105638821.0, + "step": 4087 + }, + { + "epoch": 0.4489347682846475, + "ewc_loss": 1.996755599975586e-06, + "grad_norm": 2.3583781719207764, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7000176906585693, + "num_tokens": 105661894.0, + "step": 4088 + }, + { + "epoch": 0.44904458598726116, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.8696303367614746, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6911582946777344, + "num_tokens": 105697736.0, + "step": 4089 + }, + { + "epoch": 0.4491544036898748, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.409332275390625, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7250190377235413, + "num_tokens": 105717630.0, + "step": 4090 + }, + { + "epoch": 0.44926422139248845, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.3464572429656982, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7082722783088684, + "num_tokens": 105739780.0, + "step": 4091 + }, + { + "epoch": 0.44937403909510215, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0947763919830322, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7158207297325134, + "num_tokens": 105767513.0, + "step": 4092 + }, + { + "epoch": 0.4494838567977158, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1987876892089844, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7000826597213745, + "num_tokens": 105793045.0, + "step": 4093 + }, + { + "epoch": 0.44959367450032944, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.130007743835449, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6954473257064819, + "num_tokens": 105818510.0, + "step": 4094 + }, + { + "epoch": 0.4497034922029431, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.608140707015991, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7094703912734985, + "num_tokens": 105836643.0, + "step": 4095 + }, + { + "epoch": 0.4498133099055568, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9891499280929565, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6851284503936768, + "num_tokens": 105865818.0, + "step": 4096 + }, + { + "epoch": 0.44992312760817044, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.204240560531616, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7133576273918152, + "num_tokens": 105891557.0, + "step": 4097 + }, + { + "epoch": 0.4500329453107841, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.381119966506958, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.6916264295578003, + "num_tokens": 105914959.0, + "step": 4098 + }, + { + "epoch": 0.4501427630133978, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2515742778778076, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6971559524536133, + "num_tokens": 105940788.0, + "step": 4099 + }, + { + "epoch": 0.45025258071601143, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.675791025161743, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7283827066421509, + "num_tokens": 105960218.0, + "step": 4100 + }, + { + "epoch": 0.4503623984186251, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.248019218444824, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7091858983039856, + "num_tokens": 105985451.0, + "step": 4101 + }, + { + "epoch": 0.4504722161212387, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9732147455215454, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7102110385894775, + "num_tokens": 106015411.0, + "step": 4102 + }, + { + "epoch": 0.4505820338238524, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2621359825134277, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6953458786010742, + "num_tokens": 106040921.0, + "step": 4103 + }, + { + "epoch": 0.45069185152646607, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2594356536865234, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7068097591400146, + "num_tokens": 106065656.0, + "step": 4104 + }, + { + "epoch": 0.4508016692290797, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.993533730506897, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.707720160484314, + "num_tokens": 106097280.0, + "step": 4105 + }, + { + "epoch": 0.45091148693169336, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0918023586273193, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7301549911499023, + "num_tokens": 106123047.0, + "step": 4106 + }, + { + "epoch": 0.45102130463430706, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1122326850891113, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7143874168395996, + "num_tokens": 106151119.0, + "step": 4107 + }, + { + "epoch": 0.4511311223369207, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9477248191833496, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6835678815841675, + "num_tokens": 106183173.0, + "step": 4108 + }, + { + "epoch": 0.45124094003953436, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.379781484603882, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7094004154205322, + "num_tokens": 106206388.0, + "step": 4109 + }, + { + "epoch": 0.45135075774214806, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0538346767425537, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.7084516286849976, + "num_tokens": 106235476.0, + "step": 4110 + }, + { + "epoch": 0.4514605754447617, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.162876605987549, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7092087268829346, + "num_tokens": 106261962.0, + "step": 4111 + }, + { + "epoch": 0.45157039314737535, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.299989700317383, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6839995980262756, + "num_tokens": 106288609.0, + "step": 4112 + }, + { + "epoch": 0.451680210849989, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0980517864227295, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7132875323295593, + "num_tokens": 106316201.0, + "step": 4113 + }, + { + "epoch": 0.4517900285526027, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9582207202911377, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.6868814826011658, + "num_tokens": 106349835.0, + "step": 4114 + }, + { + "epoch": 0.45189984625521634, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.157214879989624, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.688819408416748, + "num_tokens": 106376203.0, + "step": 4115 + }, + { + "epoch": 0.45200966395783, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1158740520477295, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7015652060508728, + "num_tokens": 106404493.0, + "step": 4116 + }, + { + "epoch": 0.4521194816604437, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0378167629241943, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7017527222633362, + "num_tokens": 106432988.0, + "step": 4117 + }, + { + "epoch": 0.45222929936305734, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.5399463176727295, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6968362331390381, + "num_tokens": 106453120.0, + "step": 4118 + }, + { + "epoch": 0.452339117065671, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2889301776885986, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7153202295303345, + "num_tokens": 106475765.0, + "step": 4119 + }, + { + "epoch": 0.4524489347682846, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1952106952667236, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7115815877914429, + "num_tokens": 106501456.0, + "step": 4120 + }, + { + "epoch": 0.45255875247089833, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9769119024276733, + "learning_rate": 1e-06, + "loss": 1.0917, + "mean_token_accuracy": 0.6682361364364624, + "num_tokens": 106534135.0, + "step": 4121 + }, + { + "epoch": 0.452668570173512, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2965543270111084, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7158210277557373, + "num_tokens": 106558166.0, + "step": 4122 + }, + { + "epoch": 0.4527783878761256, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1451847553253174, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6888402700424194, + "num_tokens": 106585004.0, + "step": 4123 + }, + { + "epoch": 0.45288820557873927, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1578593254089355, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7280724048614502, + "num_tokens": 106609383.0, + "step": 4124 + }, + { + "epoch": 0.45299802328135297, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.038442850112915, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6973327398300171, + "num_tokens": 106641061.0, + "step": 4125 + }, + { + "epoch": 0.4531078409839666, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9073494672775269, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7105867862701416, + "num_tokens": 106672179.0, + "step": 4126 + }, + { + "epoch": 0.45321765868658026, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.4018442630767822, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6930476427078247, + "num_tokens": 106695512.0, + "step": 4127 + }, + { + "epoch": 0.45332747638919396, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9608235359191895, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7276478409767151, + "num_tokens": 106723126.0, + "step": 4128 + }, + { + "epoch": 0.4534372940918076, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2985637187957764, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7065355777740479, + "num_tokens": 106747737.0, + "step": 4129 + }, + { + "epoch": 0.45354711179442125, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.976905107498169, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6933647394180298, + "num_tokens": 106778830.0, + "step": 4130 + }, + { + "epoch": 0.4536569294970349, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.194568157196045, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6918727159500122, + "num_tokens": 106803825.0, + "step": 4131 + }, + { + "epoch": 0.4537667471996486, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 1.9063332080841064, + "learning_rate": 1e-06, + "loss": 1.0945, + "mean_token_accuracy": 0.6762892007827759, + "num_tokens": 106836896.0, + "step": 4132 + }, + { + "epoch": 0.45387656490226225, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.175347328186035, + "learning_rate": 1e-06, + "loss": 1.0754, + "mean_token_accuracy": 0.675921618938446, + "num_tokens": 106862131.0, + "step": 4133 + }, + { + "epoch": 0.4539863826048759, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.192610025405884, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.7026051878929138, + "num_tokens": 106888132.0, + "step": 4134 + }, + { + "epoch": 0.4540962003074896, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0895400047302246, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7121493816375732, + "num_tokens": 106916637.0, + "step": 4135 + }, + { + "epoch": 0.45420601801010324, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1670525074005127, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.700964093208313, + "num_tokens": 106943004.0, + "step": 4136 + }, + { + "epoch": 0.4543158357127169, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.0795371532440186, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.697289228439331, + "num_tokens": 106972634.0, + "step": 4137 + }, + { + "epoch": 0.45442565341533053, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 2.3806686401367188, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6884680986404419, + "num_tokens": 106995022.0, + "step": 4138 + }, + { + "epoch": 0.45453547111794423, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 2.528360366821289, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7205194234848022, + "num_tokens": 107014998.0, + "step": 4139 + }, + { + "epoch": 0.4546452888205579, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 1.9824992418289185, + "learning_rate": 1e-06, + "loss": 1.0648, + "mean_token_accuracy": 0.6782110929489136, + "num_tokens": 107047636.0, + "step": 4140 + }, + { + "epoch": 0.4547551065231715, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 2.3898608684539795, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7076101303100586, + "num_tokens": 107073075.0, + "step": 4141 + }, + { + "epoch": 0.45486492422578517, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.457442045211792, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.7103232145309448, + "num_tokens": 107095074.0, + "step": 4142 + }, + { + "epoch": 0.4549747419283989, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2129290103912354, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6982675790786743, + "num_tokens": 107120478.0, + "step": 4143 + }, + { + "epoch": 0.4550845596310125, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.1143500804901123, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.701574444770813, + "num_tokens": 107149935.0, + "step": 4144 + }, + { + "epoch": 0.45519437733362617, + "ewc_loss": 2.0116567611694336e-06, + "grad_norm": 2.2897439002990723, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7014289498329163, + "num_tokens": 107176438.0, + "step": 4145 + }, + { + "epoch": 0.45530419503623987, + "ewc_loss": 2.0265579223632812e-06, + "grad_norm": 2.064115285873413, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7055079936981201, + "num_tokens": 107202811.0, + "step": 4146 + }, + { + "epoch": 0.4554140127388535, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.075732946395874, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6929174661636353, + "num_tokens": 107233423.0, + "step": 4147 + }, + { + "epoch": 0.45552383044146716, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1871094703674316, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.7009186148643494, + "num_tokens": 107259979.0, + "step": 4148 + }, + { + "epoch": 0.4556336481440808, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.099548816680908, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.6861239075660706, + "num_tokens": 107289683.0, + "step": 4149 + }, + { + "epoch": 0.4557434658466945, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2758476734161377, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7067307233810425, + "num_tokens": 107315025.0, + "step": 4150 + }, + { + "epoch": 0.45585328354930815, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.453495979309082, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7017683386802673, + "num_tokens": 107337420.0, + "step": 4151 + }, + { + "epoch": 0.4559631012519218, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.450972318649292, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6971736550331116, + "num_tokens": 107360034.0, + "step": 4152 + }, + { + "epoch": 0.45607291895453544, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0563535690307617, + "learning_rate": 1e-06, + "loss": 1.0311, + "mean_token_accuracy": 0.6869211196899414, + "num_tokens": 107389164.0, + "step": 4153 + }, + { + "epoch": 0.45618273665714915, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.213695526123047, + "learning_rate": 1e-06, + "loss": 1.1085, + "mean_token_accuracy": 0.6738709807395935, + "num_tokens": 107417239.0, + "step": 4154 + }, + { + "epoch": 0.4562925543597628, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.301387310028076, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6938551664352417, + "num_tokens": 107442241.0, + "step": 4155 + }, + { + "epoch": 0.45640237206237644, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.297926902770996, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7027937769889832, + "num_tokens": 107464525.0, + "step": 4156 + }, + { + "epoch": 0.45651218976499014, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.408287525177002, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7182971239089966, + "num_tokens": 107487863.0, + "step": 4157 + }, + { + "epoch": 0.4566220074676038, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.061149835586548, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6963671445846558, + "num_tokens": 107516463.0, + "step": 4158 + }, + { + "epoch": 0.45673182517021743, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0550050735473633, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.682672381401062, + "num_tokens": 107547372.0, + "step": 4159 + }, + { + "epoch": 0.4568416428728311, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3602724075317383, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7077494859695435, + "num_tokens": 107570039.0, + "step": 4160 + }, + { + "epoch": 0.4569514605754448, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2394204139709473, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6922515034675598, + "num_tokens": 107595999.0, + "step": 4161 + }, + { + "epoch": 0.4570612782780584, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0671560764312744, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.6987107992172241, + "num_tokens": 107624303.0, + "step": 4162 + }, + { + "epoch": 0.45717109598067207, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1279711723327637, + "learning_rate": 1e-06, + "loss": 1.0645, + "mean_token_accuracy": 0.6864851713180542, + "num_tokens": 107654015.0, + "step": 4163 + }, + { + "epoch": 0.4572809136832858, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0377447605133057, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7327656745910645, + "num_tokens": 107681066.0, + "step": 4164 + }, + { + "epoch": 0.4573907313858994, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.349398374557495, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7174772024154663, + "num_tokens": 107704738.0, + "step": 4165 + }, + { + "epoch": 0.45750054908851306, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.010009527206421, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.6781570911407471, + "num_tokens": 107738308.0, + "step": 4166 + }, + { + "epoch": 0.4576103667911267, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0243287086486816, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6846228837966919, + "num_tokens": 107770284.0, + "step": 4167 + }, + { + "epoch": 0.4577201844937404, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0753116607666016, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.6887539625167847, + "num_tokens": 107800283.0, + "step": 4168 + }, + { + "epoch": 0.45783000219635406, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2678658962249756, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7206700444221497, + "num_tokens": 107824988.0, + "step": 4169 + }, + { + "epoch": 0.4579398198989677, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3294873237609863, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.722581148147583, + "num_tokens": 107847507.0, + "step": 4170 + }, + { + "epoch": 0.45804963760158135, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9951554536819458, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.695031464099884, + "num_tokens": 107878874.0, + "step": 4171 + }, + { + "epoch": 0.45815945530419505, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0328752994537354, + "learning_rate": 1e-06, + "loss": 1.0642, + "mean_token_accuracy": 0.6917144060134888, + "num_tokens": 107907987.0, + "step": 4172 + }, + { + "epoch": 0.4582692730068087, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.005159616470337, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.6933658123016357, + "num_tokens": 107935921.0, + "step": 4173 + }, + { + "epoch": 0.45837909070942234, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9720678329467773, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7008822560310364, + "num_tokens": 107969167.0, + "step": 4174 + }, + { + "epoch": 0.45848890841203604, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.231498956680298, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7135083675384521, + "num_tokens": 107993529.0, + "step": 4175 + }, + { + "epoch": 0.4585987261146497, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.294506549835205, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6851619482040405, + "num_tokens": 108019805.0, + "step": 4176 + }, + { + "epoch": 0.45870854381726334, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.013498544692993, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6920936107635498, + "num_tokens": 108048592.0, + "step": 4177 + }, + { + "epoch": 0.458818361519877, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.960981845855713, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7162473201751709, + "num_tokens": 108079410.0, + "step": 4178 + }, + { + "epoch": 0.4589281792224907, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.181356906890869, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7122278809547424, + "num_tokens": 108106842.0, + "step": 4179 + }, + { + "epoch": 0.45903799692510433, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1323039531707764, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7002779245376587, + "num_tokens": 108134819.0, + "step": 4180 + }, + { + "epoch": 0.459147814627718, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2198410034179688, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7125527858734131, + "num_tokens": 108160211.0, + "step": 4181 + }, + { + "epoch": 0.4592576323303316, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3001654148101807, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7040358781814575, + "num_tokens": 108184231.0, + "step": 4182 + }, + { + "epoch": 0.4593674500329453, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.330638885498047, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6997724771499634, + "num_tokens": 108206916.0, + "step": 4183 + }, + { + "epoch": 0.45947726773555897, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3477838039398193, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7149389982223511, + "num_tokens": 108227851.0, + "step": 4184 + }, + { + "epoch": 0.4595870854381726, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.182941436767578, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.69490647315979, + "num_tokens": 108254622.0, + "step": 4185 + }, + { + "epoch": 0.4596969031407863, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1053874492645264, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6993412971496582, + "num_tokens": 108282366.0, + "step": 4186 + }, + { + "epoch": 0.45980672084339996, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9744724035263062, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.69048011302948, + "num_tokens": 108314359.0, + "step": 4187 + }, + { + "epoch": 0.4599165385460136, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.994710922241211, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7006255388259888, + "num_tokens": 108342960.0, + "step": 4188 + }, + { + "epoch": 0.46002635624862726, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1068739891052246, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6897531151771545, + "num_tokens": 108369224.0, + "step": 4189 + }, + { + "epoch": 0.46013617395124096, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1751949787139893, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.692451000213623, + "num_tokens": 108397077.0, + "step": 4190 + }, + { + "epoch": 0.4602459916538546, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2165629863739014, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.6957063674926758, + "num_tokens": 108421690.0, + "step": 4191 + }, + { + "epoch": 0.46035580935646825, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1305365562438965, + "learning_rate": 1e-06, + "loss": 1.0612, + "mean_token_accuracy": 0.679075300693512, + "num_tokens": 108449768.0, + "step": 4192 + }, + { + "epoch": 0.46046562705908195, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.869856357574463, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.707831621170044, + "num_tokens": 108466772.0, + "step": 4193 + }, + { + "epoch": 0.4605754447616956, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.419215440750122, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7217099070549011, + "num_tokens": 108488184.0, + "step": 4194 + }, + { + "epoch": 0.46068526246430924, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1483774185180664, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6989585161209106, + "num_tokens": 108514984.0, + "step": 4195 + }, + { + "epoch": 0.4607950801669229, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1245229244232178, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7116433382034302, + "num_tokens": 108540998.0, + "step": 4196 + }, + { + "epoch": 0.4609048978695366, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0032694339752197, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7076853513717651, + "num_tokens": 108570211.0, + "step": 4197 + }, + { + "epoch": 0.46101471557215024, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0192620754241943, + "learning_rate": 1e-06, + "loss": 1.0633, + "mean_token_accuracy": 0.6786016225814819, + "num_tokens": 108599725.0, + "step": 4198 + }, + { + "epoch": 0.4611245332747639, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2715585231781006, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6932532787322998, + "num_tokens": 108624771.0, + "step": 4199 + }, + { + "epoch": 0.4612343509773775, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.111713171005249, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6910199522972107, + "num_tokens": 108655965.0, + "step": 4200 + }, + { + "epoch": 0.46134416867999123, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3636655807495117, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7025672197341919, + "num_tokens": 108678648.0, + "step": 4201 + }, + { + "epoch": 0.4614539863826049, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.107250690460205, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.692929208278656, + "num_tokens": 108708088.0, + "step": 4202 + }, + { + "epoch": 0.4615638040852185, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9403163194656372, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7126908302307129, + "num_tokens": 108737497.0, + "step": 4203 + }, + { + "epoch": 0.4616736217878322, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1720757484436035, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7091031670570374, + "num_tokens": 108764609.0, + "step": 4204 + }, + { + "epoch": 0.46178343949044587, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0512588024139404, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.702721357345581, + "num_tokens": 108790774.0, + "step": 4205 + }, + { + "epoch": 0.4618932571930595, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9690711498260498, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7175049781799316, + "num_tokens": 108820188.0, + "step": 4206 + }, + { + "epoch": 0.46200307489567316, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.4355289936065674, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.72463059425354, + "num_tokens": 108841518.0, + "step": 4207 + }, + { + "epoch": 0.46211289259828686, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3981330394744873, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6941425204277039, + "num_tokens": 108865351.0, + "step": 4208 + }, + { + "epoch": 0.4622227103009005, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.5267281532287598, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6894216537475586, + "num_tokens": 108887902.0, + "step": 4209 + }, + { + "epoch": 0.46233252800351415, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9801896810531616, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7057028412818909, + "num_tokens": 108917553.0, + "step": 4210 + }, + { + "epoch": 0.46244234570612786, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.965356707572937, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6993359923362732, + "num_tokens": 108949999.0, + "step": 4211 + }, + { + "epoch": 0.4625521634087415, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.235100746154785, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.7049874663352966, + "num_tokens": 108976054.0, + "step": 4212 + }, + { + "epoch": 0.46266198111135515, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.170489549636841, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7157300710678101, + "num_tokens": 109001853.0, + "step": 4213 + }, + { + "epoch": 0.4627717988139688, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.7581844329833984, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7327761650085449, + "num_tokens": 109018515.0, + "step": 4214 + }, + { + "epoch": 0.4628816165165825, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3268604278564453, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7080327272415161, + "num_tokens": 109043524.0, + "step": 4215 + }, + { + "epoch": 0.46299143421919614, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3533220291137695, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6945626735687256, + "num_tokens": 109065870.0, + "step": 4216 + }, + { + "epoch": 0.4631012519218098, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.107513189315796, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6901893019676208, + "num_tokens": 109092819.0, + "step": 4217 + }, + { + "epoch": 0.46321106962442343, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1445822715759277, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7167912721633911, + "num_tokens": 109119366.0, + "step": 4218 + }, + { + "epoch": 0.46332088732703713, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.5218915939331055, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.6972606778144836, + "num_tokens": 109142027.0, + "step": 4219 + }, + { + "epoch": 0.4634307050296508, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.406158447265625, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.7038558125495911, + "num_tokens": 109164566.0, + "step": 4220 + }, + { + "epoch": 0.4635405227322644, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.4920504093170166, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7250251770019531, + "num_tokens": 109185471.0, + "step": 4221 + }, + { + "epoch": 0.4636503404348781, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3217222690582275, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7011345624923706, + "num_tokens": 109209593.0, + "step": 4222 + }, + { + "epoch": 0.4637601581374918, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3215456008911133, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6866995096206665, + "num_tokens": 109233085.0, + "step": 4223 + }, + { + "epoch": 0.4638699758401054, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3538713455200195, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7110920548439026, + "num_tokens": 109255439.0, + "step": 4224 + }, + { + "epoch": 0.46397979354271907, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1278469562530518, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6937417387962341, + "num_tokens": 109285240.0, + "step": 4225 + }, + { + "epoch": 0.46408961124533277, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.516592502593994, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7194975018501282, + "num_tokens": 109306534.0, + "step": 4226 + }, + { + "epoch": 0.4641994289479464, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.156902551651001, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6929618120193481, + "num_tokens": 109335144.0, + "step": 4227 + }, + { + "epoch": 0.46430924665056006, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.066190719604492, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6946621537208557, + "num_tokens": 109363223.0, + "step": 4228 + }, + { + "epoch": 0.4644190643531737, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3911080360412598, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7108368277549744, + "num_tokens": 109385976.0, + "step": 4229 + }, + { + "epoch": 0.4645288820557874, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.178891658782959, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6934439539909363, + "num_tokens": 109413991.0, + "step": 4230 + }, + { + "epoch": 0.46463869975840105, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.453111171722412, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7219406366348267, + "num_tokens": 109434389.0, + "step": 4231 + }, + { + "epoch": 0.4647485174610147, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9423705339431763, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6906464099884033, + "num_tokens": 109466391.0, + "step": 4232 + }, + { + "epoch": 0.4648583351636284, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.4060428142547607, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7119585275650024, + "num_tokens": 109489623.0, + "step": 4233 + }, + { + "epoch": 0.46496815286624205, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1919920444488525, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6898186206817627, + "num_tokens": 109515780.0, + "step": 4234 + }, + { + "epoch": 0.4650779705688557, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2862322330474854, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7064651250839233, + "num_tokens": 109539212.0, + "step": 4235 + }, + { + "epoch": 0.46518778827146934, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.479980230331421, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6886194944381714, + "num_tokens": 109560287.0, + "step": 4236 + }, + { + "epoch": 0.46529760597408304, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.059309482574463, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7051640748977661, + "num_tokens": 109588456.0, + "step": 4237 + }, + { + "epoch": 0.4654074236766967, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.114398717880249, + "learning_rate": 1e-06, + "loss": 1.0679, + "mean_token_accuracy": 0.6826562881469727, + "num_tokens": 109616581.0, + "step": 4238 + }, + { + "epoch": 0.46551724137931033, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.288120746612549, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.719484269618988, + "num_tokens": 109641331.0, + "step": 4239 + }, + { + "epoch": 0.46562705908192403, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0444090366363525, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6906114816665649, + "num_tokens": 109670886.0, + "step": 4240 + }, + { + "epoch": 0.4657368767845377, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1430039405822754, + "learning_rate": 1e-06, + "loss": 1.0361, + "mean_token_accuracy": 0.6942920684814453, + "num_tokens": 109697801.0, + "step": 4241 + }, + { + "epoch": 0.4658466944871513, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1533730030059814, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7040987610816956, + "num_tokens": 109724159.0, + "step": 4242 + }, + { + "epoch": 0.46595651218976497, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.296246290206909, + "learning_rate": 1e-06, + "loss": 1.065, + "mean_token_accuracy": 0.671545147895813, + "num_tokens": 109749383.0, + "step": 4243 + }, + { + "epoch": 0.4660663298923787, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.210397720336914, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7094897031784058, + "num_tokens": 109776098.0, + "step": 4244 + }, + { + "epoch": 0.4661761475949923, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3412280082702637, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7124421000480652, + "num_tokens": 109797559.0, + "step": 4245 + }, + { + "epoch": 0.46628596529760596, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9414701461791992, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7300289869308472, + "num_tokens": 109828678.0, + "step": 4246 + }, + { + "epoch": 0.4663957830002196, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0813486576080322, + "learning_rate": 1e-06, + "loss": 1.0497, + "mean_token_accuracy": 0.683360755443573, + "num_tokens": 109857652.0, + "step": 4247 + }, + { + "epoch": 0.4665056007028333, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.137406587600708, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6997173428535461, + "num_tokens": 109884554.0, + "step": 4248 + }, + { + "epoch": 0.46661541840544696, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0923004150390625, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7170976996421814, + "num_tokens": 109911682.0, + "step": 4249 + }, + { + "epoch": 0.4667252361080606, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3111412525177, + "learning_rate": 1e-06, + "loss": 1.067, + "mean_token_accuracy": 0.6858224868774414, + "num_tokens": 109936157.0, + "step": 4250 + }, + { + "epoch": 0.4668350538106743, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.060800075531006, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.6967179179191589, + "num_tokens": 109963461.0, + "step": 4251 + }, + { + "epoch": 0.46694487151328795, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3597192764282227, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7272825241088867, + "num_tokens": 109985239.0, + "step": 4252 + }, + { + "epoch": 0.4670546892159016, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.213913917541504, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6892626285552979, + "num_tokens": 110009774.0, + "step": 4253 + }, + { + "epoch": 0.46716450691851524, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.4536542892456055, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.7023983001708984, + "num_tokens": 110030532.0, + "step": 4254 + }, + { + "epoch": 0.46727432462112894, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9135979413986206, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6880723237991333, + "num_tokens": 110064065.0, + "step": 4255 + }, + { + "epoch": 0.4673841423237426, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.177337169647217, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7036662101745605, + "num_tokens": 110089461.0, + "step": 4256 + }, + { + "epoch": 0.46749396002635624, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.8794865608215332, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6962613463401794, + "num_tokens": 110124072.0, + "step": 4257 + }, + { + "epoch": 0.4676037777289699, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.48372483253479, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.7055506706237793, + "num_tokens": 110145663.0, + "step": 4258 + }, + { + "epoch": 0.4677135954315836, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0854227542877197, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.6879034042358398, + "num_tokens": 110175157.0, + "step": 4259 + }, + { + "epoch": 0.46782341313419723, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3549742698669434, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.708017110824585, + "num_tokens": 110198393.0, + "step": 4260 + }, + { + "epoch": 0.4679332308368109, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9106930494308472, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.695647656917572, + "num_tokens": 110231770.0, + "step": 4261 + }, + { + "epoch": 0.4680430485394246, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1096136569976807, + "learning_rate": 1e-06, + "loss": 1.0652, + "mean_token_accuracy": 0.6838653087615967, + "num_tokens": 110261755.0, + "step": 4262 + }, + { + "epoch": 0.4681528662420382, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.5419905185699463, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6867189407348633, + "num_tokens": 110281687.0, + "step": 4263 + }, + { + "epoch": 0.46826268394465187, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3531265258789062, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7146483659744263, + "num_tokens": 110305069.0, + "step": 4264 + }, + { + "epoch": 0.4683725016472655, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.415316343307495, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7390942573547363, + "num_tokens": 110324935.0, + "step": 4265 + }, + { + "epoch": 0.4684823193498792, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.442354679107666, + "learning_rate": 1e-06, + "loss": 1.0252, + "mean_token_accuracy": 0.6961961388587952, + "num_tokens": 110346955.0, + "step": 4266 + }, + { + "epoch": 0.46859213705249286, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.10107421875, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.714652419090271, + "num_tokens": 110373544.0, + "step": 4267 + }, + { + "epoch": 0.4687019547551065, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.034388542175293, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.6844009757041931, + "num_tokens": 110405117.0, + "step": 4268 + }, + { + "epoch": 0.4688117724577202, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.409011125564575, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6966749429702759, + "num_tokens": 110426767.0, + "step": 4269 + }, + { + "epoch": 0.46892159016033386, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.322972297668457, + "learning_rate": 1e-06, + "loss": 1.0128, + "mean_token_accuracy": 0.6897056102752686, + "num_tokens": 110450394.0, + "step": 4270 + }, + { + "epoch": 0.4690314078629475, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.077573776245117, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.685369074344635, + "num_tokens": 110479240.0, + "step": 4271 + }, + { + "epoch": 0.46914122556556115, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2332746982574463, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.710039496421814, + "num_tokens": 110503201.0, + "step": 4272 + }, + { + "epoch": 0.46925104326817485, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1083149909973145, + "learning_rate": 1e-06, + "loss": 1.0443, + "mean_token_accuracy": 0.6869306564331055, + "num_tokens": 110533873.0, + "step": 4273 + }, + { + "epoch": 0.4693608609707885, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.5745129585266113, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7134925127029419, + "num_tokens": 110553795.0, + "step": 4274 + }, + { + "epoch": 0.46947067867340214, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1306352615356445, + "learning_rate": 1e-06, + "loss": 1.0518, + "mean_token_accuracy": 0.6844430565834045, + "num_tokens": 110582884.0, + "step": 4275 + }, + { + "epoch": 0.4695804963760158, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.107919454574585, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7066195011138916, + "num_tokens": 110608619.0, + "step": 4276 + }, + { + "epoch": 0.4696903140786295, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.251953125, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.708351731300354, + "num_tokens": 110633119.0, + "step": 4277 + }, + { + "epoch": 0.46980013178124314, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2062971591949463, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6887239813804626, + "num_tokens": 110658497.0, + "step": 4278 + }, + { + "epoch": 0.4699099494838568, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.254333019256592, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7118074297904968, + "num_tokens": 110682326.0, + "step": 4279 + }, + { + "epoch": 0.4700197671864705, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.133512258529663, + "learning_rate": 1e-06, + "loss": 1.0606, + "mean_token_accuracy": 0.6886101961135864, + "num_tokens": 110712977.0, + "step": 4280 + }, + { + "epoch": 0.47012958488908413, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3752872943878174, + "learning_rate": 1e-06, + "loss": 1.0295, + "mean_token_accuracy": 0.6930810213088989, + "num_tokens": 110735079.0, + "step": 4281 + }, + { + "epoch": 0.4702394025916978, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1143062114715576, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6899280548095703, + "num_tokens": 110763270.0, + "step": 4282 + }, + { + "epoch": 0.4703492202943114, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.263174295425415, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6942555904388428, + "num_tokens": 110789102.0, + "step": 4283 + }, + { + "epoch": 0.4704590379969251, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.980176568031311, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7189958691596985, + "num_tokens": 110820768.0, + "step": 4284 + }, + { + "epoch": 0.47056885569953877, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.02685284614563, + "learning_rate": 1e-06, + "loss": 1.0861, + "mean_token_accuracy": 0.6748994588851929, + "num_tokens": 110852042.0, + "step": 4285 + }, + { + "epoch": 0.4706786734021524, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2683634757995605, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7181118130683899, + "num_tokens": 110877114.0, + "step": 4286 + }, + { + "epoch": 0.4707884911047661, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9568248987197876, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.741997480392456, + "num_tokens": 110905762.0, + "step": 4287 + }, + { + "epoch": 0.47089830880737976, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.399354934692383, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.6983045935630798, + "num_tokens": 110926649.0, + "step": 4288 + }, + { + "epoch": 0.4710081265099934, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3085861206054688, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6937428116798401, + "num_tokens": 110952555.0, + "step": 4289 + }, + { + "epoch": 0.47111794421260705, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2448949813842773, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6907039880752563, + "num_tokens": 110978068.0, + "step": 4290 + }, + { + "epoch": 0.47122776191522076, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.495607376098633, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.7122257947921753, + "num_tokens": 111000492.0, + "step": 4291 + }, + { + "epoch": 0.4713375796178344, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0868897438049316, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6884756088256836, + "num_tokens": 111031966.0, + "step": 4292 + }, + { + "epoch": 0.47144739732044805, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1769964694976807, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6839104890823364, + "num_tokens": 111060055.0, + "step": 4293 + }, + { + "epoch": 0.4715572150230617, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1039767265319824, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6914583444595337, + "num_tokens": 111088646.0, + "step": 4294 + }, + { + "epoch": 0.4716670327256754, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.273085355758667, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7062633037567139, + "num_tokens": 111113189.0, + "step": 4295 + }, + { + "epoch": 0.47177685042828904, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3008861541748047, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7037668228149414, + "num_tokens": 111136436.0, + "step": 4296 + }, + { + "epoch": 0.4718866681309027, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.18066143989563, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7105427980422974, + "num_tokens": 111162183.0, + "step": 4297 + }, + { + "epoch": 0.4719964858335164, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1073806285858154, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6883521676063538, + "num_tokens": 111191417.0, + "step": 4298 + }, + { + "epoch": 0.47210630353613003, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0534732341766357, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7043471336364746, + "num_tokens": 111219309.0, + "step": 4299 + }, + { + "epoch": 0.4722161212387437, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0939767360687256, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7181423902511597, + "num_tokens": 111247213.0, + "step": 4300 + }, + { + "epoch": 0.4723259389413573, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.4095253944396973, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.689824104309082, + "num_tokens": 111270004.0, + "step": 4301 + }, + { + "epoch": 0.472435756643971, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2974820137023926, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7097571492195129, + "num_tokens": 111294602.0, + "step": 4302 + }, + { + "epoch": 0.4725455743465847, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.262746810913086, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6882534027099609, + "num_tokens": 111321839.0, + "step": 4303 + }, + { + "epoch": 0.4726553920491983, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.26092267036438, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6894575953483582, + "num_tokens": 111348923.0, + "step": 4304 + }, + { + "epoch": 0.47276520975181197, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.9521207809448242, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6916533708572388, + "num_tokens": 111378979.0, + "step": 4305 + }, + { + "epoch": 0.47287502745442567, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3862991333007812, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.6985478401184082, + "num_tokens": 111402072.0, + "step": 4306 + }, + { + "epoch": 0.4729848451570393, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.323018789291382, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7085185050964355, + "num_tokens": 111424366.0, + "step": 4307 + }, + { + "epoch": 0.47309466285965296, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 1.8888651132583618, + "learning_rate": 1e-06, + "loss": 1.0698, + "mean_token_accuracy": 0.6809847354888916, + "num_tokens": 111457727.0, + "step": 4308 + }, + { + "epoch": 0.47320448056226666, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.356295585632324, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7149657011032104, + "num_tokens": 111479473.0, + "step": 4309 + }, + { + "epoch": 0.4733142982648803, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2066071033477783, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6867825984954834, + "num_tokens": 111505459.0, + "step": 4310 + }, + { + "epoch": 0.47342411596749395, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.177887201309204, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6956899762153625, + "num_tokens": 111529820.0, + "step": 4311 + }, + { + "epoch": 0.4735339336701076, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2693121433258057, + "learning_rate": 1e-06, + "loss": 1.0415, + "mean_token_accuracy": 0.6996885538101196, + "num_tokens": 111556109.0, + "step": 4312 + }, + { + "epoch": 0.4736437513727213, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2173962593078613, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7091572284698486, + "num_tokens": 111580721.0, + "step": 4313 + }, + { + "epoch": 0.47375356907533495, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.281054973602295, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7143686413764954, + "num_tokens": 111604459.0, + "step": 4314 + }, + { + "epoch": 0.4738633867779486, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3442442417144775, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7029462456703186, + "num_tokens": 111629930.0, + "step": 4315 + }, + { + "epoch": 0.4739732044805623, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.089909791946411, + "learning_rate": 1e-06, + "loss": 1.0612, + "mean_token_accuracy": 0.6836788654327393, + "num_tokens": 111657613.0, + "step": 4316 + }, + { + "epoch": 0.47408302218317594, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.5110316276550293, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7271213531494141, + "num_tokens": 111677344.0, + "step": 4317 + }, + { + "epoch": 0.4741928398857896, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.3242454528808594, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7151708602905273, + "num_tokens": 111701017.0, + "step": 4318 + }, + { + "epoch": 0.47430265758840323, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.490596294403076, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7177777290344238, + "num_tokens": 111721446.0, + "step": 4319 + }, + { + "epoch": 0.47441247529101693, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.142808675765991, + "learning_rate": 1e-06, + "loss": 1.0538, + "mean_token_accuracy": 0.686996579170227, + "num_tokens": 111749768.0, + "step": 4320 + }, + { + "epoch": 0.4745222929936306, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.1259071826934814, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6990731954574585, + "num_tokens": 111780595.0, + "step": 4321 + }, + { + "epoch": 0.4746321106962442, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0390725135803223, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6963670253753662, + "num_tokens": 111808453.0, + "step": 4322 + }, + { + "epoch": 0.47474192839885787, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.069345474243164, + "learning_rate": 1e-06, + "loss": 1.0665, + "mean_token_accuracy": 0.6791835427284241, + "num_tokens": 111838333.0, + "step": 4323 + }, + { + "epoch": 0.4748517461014716, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2698261737823486, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7030201554298401, + "num_tokens": 111862836.0, + "step": 4324 + }, + { + "epoch": 0.4749615638040852, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.2717182636260986, + "learning_rate": 1e-06, + "loss": 1.096, + "mean_token_accuracy": 0.6749742031097412, + "num_tokens": 111889887.0, + "step": 4325 + }, + { + "epoch": 0.47507138150669886, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.2906651496887207, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.6897717118263245, + "num_tokens": 111916363.0, + "step": 4326 + }, + { + "epoch": 0.47518119920931257, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.0169413089752197, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7151774764060974, + "num_tokens": 111947248.0, + "step": 4327 + }, + { + "epoch": 0.4752910169119262, + "ewc_loss": 2.041459083557129e-06, + "grad_norm": 2.0012404918670654, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6876633167266846, + "num_tokens": 111979006.0, + "step": 4328 + }, + { + "epoch": 0.47540083461453986, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.254073143005371, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7198827266693115, + "num_tokens": 112001590.0, + "step": 4329 + }, + { + "epoch": 0.4755106523171535, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.5160932540893555, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.710780680179596, + "num_tokens": 112022137.0, + "step": 4330 + }, + { + "epoch": 0.4756204700197672, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.2090091705322266, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7294304370880127, + "num_tokens": 112047019.0, + "step": 4331 + }, + { + "epoch": 0.47573028772238085, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.1312804222106934, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6929950714111328, + "num_tokens": 112073720.0, + "step": 4332 + }, + { + "epoch": 0.4758401054249945, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.225158452987671, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7035323977470398, + "num_tokens": 112100905.0, + "step": 4333 + }, + { + "epoch": 0.47594992312760814, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.0911991596221924, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7050399780273438, + "num_tokens": 112127185.0, + "step": 4334 + }, + { + "epoch": 0.47605974083022184, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.3691306114196777, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7158531546592712, + "num_tokens": 112148344.0, + "step": 4335 + }, + { + "epoch": 0.4761695585328355, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.297062635421753, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6965289115905762, + "num_tokens": 112172753.0, + "step": 4336 + }, + { + "epoch": 0.47627937623544914, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.268441915512085, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7093958854675293, + "num_tokens": 112195835.0, + "step": 4337 + }, + { + "epoch": 0.47638919393806284, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.0309910774230957, + "learning_rate": 1e-06, + "loss": 1.0753, + "mean_token_accuracy": 0.6760334968566895, + "num_tokens": 112225281.0, + "step": 4338 + }, + { + "epoch": 0.4764990116406765, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.2580223083496094, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7029721140861511, + "num_tokens": 112249829.0, + "step": 4339 + }, + { + "epoch": 0.47660882934329013, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 1.9829163551330566, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6928160190582275, + "num_tokens": 112279445.0, + "step": 4340 + }, + { + "epoch": 0.4767186470459038, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.0554137229919434, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6914772987365723, + "num_tokens": 112309786.0, + "step": 4341 + }, + { + "epoch": 0.4768284647485175, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.1202332973480225, + "learning_rate": 1e-06, + "loss": 1.0556, + "mean_token_accuracy": 0.6847994327545166, + "num_tokens": 112338737.0, + "step": 4342 + }, + { + "epoch": 0.4769382824511311, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 1.8206385374069214, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6849853992462158, + "num_tokens": 112376012.0, + "step": 4343 + }, + { + "epoch": 0.47704810015374477, + "ewc_loss": 2.0563602447509766e-06, + "grad_norm": 2.4957497119903564, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6915109157562256, + "num_tokens": 112398477.0, + "step": 4344 + }, + { + "epoch": 0.47715791785635847, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.319089412689209, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7038264274597168, + "num_tokens": 112421259.0, + "step": 4345 + }, + { + "epoch": 0.4772677355589721, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.296344518661499, + "learning_rate": 1e-06, + "loss": 1.0364, + "mean_token_accuracy": 0.6864749193191528, + "num_tokens": 112445827.0, + "step": 4346 + }, + { + "epoch": 0.47737755326158576, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.502481460571289, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.6843689680099487, + "num_tokens": 112468017.0, + "step": 4347 + }, + { + "epoch": 0.4774873709641994, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.2352840900421143, + "learning_rate": 1e-06, + "loss": 1.0443, + "mean_token_accuracy": 0.6812166571617126, + "num_tokens": 112493136.0, + "step": 4348 + }, + { + "epoch": 0.4775971886668131, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.123499631881714, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6922088861465454, + "num_tokens": 112517619.0, + "step": 4349 + }, + { + "epoch": 0.47770700636942676, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 1.9994142055511475, + "learning_rate": 1e-06, + "loss": 1.1214, + "mean_token_accuracy": 0.6697608828544617, + "num_tokens": 112553523.0, + "step": 4350 + }, + { + "epoch": 0.4778168240720404, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.2114107608795166, + "learning_rate": 1e-06, + "loss": 1.1573, + "mean_token_accuracy": 0.6569663286209106, + "num_tokens": 112581580.0, + "step": 4351 + }, + { + "epoch": 0.47792664177465405, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.049487829208374, + "learning_rate": 1e-06, + "loss": 1.0663, + "mean_token_accuracy": 0.6845825910568237, + "num_tokens": 112612031.0, + "step": 4352 + }, + { + "epoch": 0.47803645947726775, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 1.9541406631469727, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.708313524723053, + "num_tokens": 112642814.0, + "step": 4353 + }, + { + "epoch": 0.4781462771798814, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.4642443656921387, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7044568657875061, + "num_tokens": 112665706.0, + "step": 4354 + }, + { + "epoch": 0.47825609488249504, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 2.2771494388580322, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6927268505096436, + "num_tokens": 112690358.0, + "step": 4355 + }, + { + "epoch": 0.47836591258510874, + "ewc_loss": 2.0712614059448242e-06, + "grad_norm": 1.974177360534668, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.6959809064865112, + "num_tokens": 112719946.0, + "step": 4356 + }, + { + "epoch": 0.4784757302877224, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.2730352878570557, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6936880946159363, + "num_tokens": 112744322.0, + "step": 4357 + }, + { + "epoch": 0.47858554799033604, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.0496158599853516, + "learning_rate": 1e-06, + "loss": 1.0993, + "mean_token_accuracy": 0.6723564863204956, + "num_tokens": 112775838.0, + "step": 4358 + }, + { + "epoch": 0.4786953656929497, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.055103302001953, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7121021747589111, + "num_tokens": 112803649.0, + "step": 4359 + }, + { + "epoch": 0.4788051833955634, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1538736820220947, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6848683953285217, + "num_tokens": 112831812.0, + "step": 4360 + }, + { + "epoch": 0.47891500109817703, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.105351686477661, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7254219055175781, + "num_tokens": 112858557.0, + "step": 4361 + }, + { + "epoch": 0.4790248188007907, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1028616428375244, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7076576948165894, + "num_tokens": 112884800.0, + "step": 4362 + }, + { + "epoch": 0.4791346365034044, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.112171173095703, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7019924521446228, + "num_tokens": 112913848.0, + "step": 4363 + }, + { + "epoch": 0.479244454206018, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.223902940750122, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7060354948043823, + "num_tokens": 112939866.0, + "step": 4364 + }, + { + "epoch": 0.47935427190863167, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1283416748046875, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7355154752731323, + "num_tokens": 112965056.0, + "step": 4365 + }, + { + "epoch": 0.4794640896112453, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.0842294692993164, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.6875528693199158, + "num_tokens": 112993346.0, + "step": 4366 + }, + { + "epoch": 0.479573907313859, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.2120494842529297, + "learning_rate": 1e-06, + "loss": 1.0766, + "mean_token_accuracy": 0.6832085251808167, + "num_tokens": 113020155.0, + "step": 4367 + }, + { + "epoch": 0.47968372501647266, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.2038493156433105, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6967136859893799, + "num_tokens": 113044250.0, + "step": 4368 + }, + { + "epoch": 0.4797935427190863, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.133112668991089, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6959057450294495, + "num_tokens": 113072237.0, + "step": 4369 + }, + { + "epoch": 0.47990336042169995, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.254645824432373, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7078493237495422, + "num_tokens": 113096600.0, + "step": 4370 + }, + { + "epoch": 0.48001317812431366, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 1.9128459692001343, + "learning_rate": 1e-06, + "loss": 1.0256, + "mean_token_accuracy": 0.6946136355400085, + "num_tokens": 113131476.0, + "step": 4371 + }, + { + "epoch": 0.4801229958269273, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.065394401550293, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6880547404289246, + "num_tokens": 113161614.0, + "step": 4372 + }, + { + "epoch": 0.48023281352954095, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.362696409225464, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7000540494918823, + "num_tokens": 113184723.0, + "step": 4373 + }, + { + "epoch": 0.48034263123215465, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.2227742671966553, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.6966469287872314, + "num_tokens": 113207737.0, + "step": 4374 + }, + { + "epoch": 0.4804524489347683, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1342248916625977, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6988447904586792, + "num_tokens": 113236517.0, + "step": 4375 + }, + { + "epoch": 0.48056226663738194, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.2731127738952637, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.714148998260498, + "num_tokens": 113258955.0, + "step": 4376 + }, + { + "epoch": 0.4806720843399956, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1319010257720947, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6995518803596497, + "num_tokens": 113287178.0, + "step": 4377 + }, + { + "epoch": 0.4807819020426093, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.164454936981201, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.706432044506073, + "num_tokens": 113314384.0, + "step": 4378 + }, + { + "epoch": 0.48089171974522293, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.4705810546875, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6949077248573303, + "num_tokens": 113335624.0, + "step": 4379 + }, + { + "epoch": 0.4810015374478366, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.153141736984253, + "learning_rate": 1e-06, + "loss": 1.0592, + "mean_token_accuracy": 0.682991087436676, + "num_tokens": 113364171.0, + "step": 4380 + }, + { + "epoch": 0.4811113551504502, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.497985601425171, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7150280475616455, + "num_tokens": 113385080.0, + "step": 4381 + }, + { + "epoch": 0.4812211728530639, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1236796379089355, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.6991875767707825, + "num_tokens": 113413851.0, + "step": 4382 + }, + { + "epoch": 0.4813309905556776, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.131476879119873, + "learning_rate": 1e-06, + "loss": 1.0503, + "mean_token_accuracy": 0.6821836233139038, + "num_tokens": 113444131.0, + "step": 4383 + }, + { + "epoch": 0.4814408082582912, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.0295050144195557, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6884362697601318, + "num_tokens": 113473839.0, + "step": 4384 + }, + { + "epoch": 0.4815506259609049, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.344360113143921, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6929543018341064, + "num_tokens": 113496652.0, + "step": 4385 + }, + { + "epoch": 0.48166044366351857, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.4031050205230713, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7009128332138062, + "num_tokens": 113517813.0, + "step": 4386 + }, + { + "epoch": 0.4817702613661322, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1710801124572754, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6910697221755981, + "num_tokens": 113544589.0, + "step": 4387 + }, + { + "epoch": 0.48188007906874586, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.0332345962524414, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.69631427526474, + "num_tokens": 113575425.0, + "step": 4388 + }, + { + "epoch": 0.48198989677135956, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1298978328704834, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7141595482826233, + "num_tokens": 113600542.0, + "step": 4389 + }, + { + "epoch": 0.4820997144739732, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.1602978706359863, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7086930274963379, + "num_tokens": 113626938.0, + "step": 4390 + }, + { + "epoch": 0.48220953217658685, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.218226909637451, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.6947273015975952, + "num_tokens": 113653288.0, + "step": 4391 + }, + { + "epoch": 0.48231934987920055, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.141723155975342, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7012810707092285, + "num_tokens": 113680097.0, + "step": 4392 + }, + { + "epoch": 0.4824291675818142, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4027981758117676, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6940802931785583, + "num_tokens": 113703073.0, + "step": 4393 + }, + { + "epoch": 0.48253898528442785, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.32131028175354, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7030128240585327, + "num_tokens": 113727831.0, + "step": 4394 + }, + { + "epoch": 0.4826488029870415, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.498551845550537, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.7050002217292786, + "num_tokens": 113750652.0, + "step": 4395 + }, + { + "epoch": 0.4827586206896552, + "ewc_loss": 2.086162567138672e-06, + "grad_norm": 2.17572283744812, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.6998486518859863, + "num_tokens": 113777924.0, + "step": 4396 + }, + { + "epoch": 0.48286843839226884, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.2902793884277344, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7153835296630859, + "num_tokens": 113801616.0, + "step": 4397 + }, + { + "epoch": 0.4829782560948825, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.2552669048309326, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6865094304084778, + "num_tokens": 113827240.0, + "step": 4398 + }, + { + "epoch": 0.48308807379749613, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4357428550720215, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7333768606185913, + "num_tokens": 113847276.0, + "step": 4399 + }, + { + "epoch": 0.48319789150010983, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0373027324676514, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.7034043669700623, + "num_tokens": 113876130.0, + "step": 4400 + }, + { + "epoch": 0.4833077092027235, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.316422939300537, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.7052472829818726, + "num_tokens": 113900524.0, + "step": 4401 + }, + { + "epoch": 0.4834175269053371, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4987072944641113, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7156586050987244, + "num_tokens": 113920621.0, + "step": 4402 + }, + { + "epoch": 0.4835273446079508, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.285705804824829, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7229018807411194, + "num_tokens": 113942610.0, + "step": 4403 + }, + { + "epoch": 0.4836371623105645, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.1190431118011475, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6888940930366516, + "num_tokens": 113968029.0, + "step": 4404 + }, + { + "epoch": 0.4837469800131781, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.243962287902832, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7126826047897339, + "num_tokens": 113992927.0, + "step": 4405 + }, + { + "epoch": 0.48385679771579176, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.2368805408477783, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.6974902153015137, + "num_tokens": 114018573.0, + "step": 4406 + }, + { + "epoch": 0.48396661541840547, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.1245386600494385, + "learning_rate": 1e-06, + "loss": 1.0445, + "mean_token_accuracy": 0.6849402785301208, + "num_tokens": 114048655.0, + "step": 4407 + }, + { + "epoch": 0.4840764331210191, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.2728371620178223, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.696804940700531, + "num_tokens": 114071771.0, + "step": 4408 + }, + { + "epoch": 0.48418625082363276, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4086692333221436, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.7034560441970825, + "num_tokens": 114094334.0, + "step": 4409 + }, + { + "epoch": 0.4842960685262464, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3014259338378906, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7110774517059326, + "num_tokens": 114118168.0, + "step": 4410 + }, + { + "epoch": 0.4844058862288601, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.547943353652954, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6951687335968018, + "num_tokens": 114138611.0, + "step": 4411 + }, + { + "epoch": 0.48451570393147375, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.2861785888671875, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7191890478134155, + "num_tokens": 114160678.0, + "step": 4412 + }, + { + "epoch": 0.4846255216340874, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4907233715057373, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7313816547393799, + "num_tokens": 114180098.0, + "step": 4413 + }, + { + "epoch": 0.4847353393367011, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3416755199432373, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7212489247322083, + "num_tokens": 114203210.0, + "step": 4414 + }, + { + "epoch": 0.48484515703931474, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 1.9903942346572876, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.7042528390884399, + "num_tokens": 114232636.0, + "step": 4415 + }, + { + "epoch": 0.4849549747419284, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.276743173599243, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6977485418319702, + "num_tokens": 114256563.0, + "step": 4416 + }, + { + "epoch": 0.48506479244454204, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3579206466674805, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6988016366958618, + "num_tokens": 114282338.0, + "step": 4417 + }, + { + "epoch": 0.48517461014715574, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.252735137939453, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.692115068435669, + "num_tokens": 114306573.0, + "step": 4418 + }, + { + "epoch": 0.4852844278497694, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.160726547241211, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.712714672088623, + "num_tokens": 114330804.0, + "step": 4419 + }, + { + "epoch": 0.48539424555238303, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 1.9821401834487915, + "learning_rate": 1e-06, + "loss": 1.035, + "mean_token_accuracy": 0.6820175647735596, + "num_tokens": 114362071.0, + "step": 4420 + }, + { + "epoch": 0.48550406325499673, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3183376789093018, + "learning_rate": 1e-06, + "loss": 1.0363, + "mean_token_accuracy": 0.6851886510848999, + "num_tokens": 114386985.0, + "step": 4421 + }, + { + "epoch": 0.4856138809576104, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.4233129024505615, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7134695053100586, + "num_tokens": 114408362.0, + "step": 4422 + }, + { + "epoch": 0.485723698660224, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0912256240844727, + "learning_rate": 1e-06, + "loss": 0.9941, + "mean_token_accuracy": 0.6959561109542847, + "num_tokens": 114437239.0, + "step": 4423 + }, + { + "epoch": 0.48583351636283767, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0543086528778076, + "learning_rate": 1e-06, + "loss": 1.0561, + "mean_token_accuracy": 0.6797307729721069, + "num_tokens": 114466495.0, + "step": 4424 + }, + { + "epoch": 0.48594333406545137, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.5436172485351562, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7150000333786011, + "num_tokens": 114486989.0, + "step": 4425 + }, + { + "epoch": 0.486053151768065, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3514695167541504, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6988464593887329, + "num_tokens": 114509544.0, + "step": 4426 + }, + { + "epoch": 0.48616296947067866, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.457369089126587, + "learning_rate": 1e-06, + "loss": 1.0241, + "mean_token_accuracy": 0.6898296475410461, + "num_tokens": 114533663.0, + "step": 4427 + }, + { + "epoch": 0.4862727871732923, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0508828163146973, + "learning_rate": 1e-06, + "loss": 1.048, + "mean_token_accuracy": 0.6815539598464966, + "num_tokens": 114562909.0, + "step": 4428 + }, + { + "epoch": 0.486382604875906, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.344000816345215, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6861156225204468, + "num_tokens": 114585352.0, + "step": 4429 + }, + { + "epoch": 0.48649242257851966, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3662242889404297, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7181087732315063, + "num_tokens": 114607829.0, + "step": 4430 + }, + { + "epoch": 0.4866022402811333, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0334436893463135, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7205085158348083, + "num_tokens": 114635523.0, + "step": 4431 + }, + { + "epoch": 0.486712057983747, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.115612745285034, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7186594009399414, + "num_tokens": 114660260.0, + "step": 4432 + }, + { + "epoch": 0.48682187568636065, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.1166722774505615, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.6838204860687256, + "num_tokens": 114688684.0, + "step": 4433 + }, + { + "epoch": 0.4869316933889743, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.372958183288574, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7176376581192017, + "num_tokens": 114711677.0, + "step": 4434 + }, + { + "epoch": 0.48704151109158794, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.28182053565979, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.7000620365142822, + "num_tokens": 114735640.0, + "step": 4435 + }, + { + "epoch": 0.48715132879420164, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.046588897705078, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6877340078353882, + "num_tokens": 114765799.0, + "step": 4436 + }, + { + "epoch": 0.4872611464968153, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 1.9696484804153442, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6876134276390076, + "num_tokens": 114795705.0, + "step": 4437 + }, + { + "epoch": 0.48737096419942894, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3464434146881104, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6932699680328369, + "num_tokens": 114820558.0, + "step": 4438 + }, + { + "epoch": 0.48748078190204264, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.100046157836914, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6998161673545837, + "num_tokens": 114847518.0, + "step": 4439 + }, + { + "epoch": 0.4875905996046563, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.0324950218200684, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6935428977012634, + "num_tokens": 114875666.0, + "step": 4440 + }, + { + "epoch": 0.48770041730726993, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.1165220737457275, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.6993368268013, + "num_tokens": 114906232.0, + "step": 4441 + }, + { + "epoch": 0.4878102350098836, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3877086639404297, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7062707543373108, + "num_tokens": 114926526.0, + "step": 4442 + }, + { + "epoch": 0.4879200527124973, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.211514949798584, + "learning_rate": 1e-06, + "loss": 1.0684, + "mean_token_accuracy": 0.6876230239868164, + "num_tokens": 114954475.0, + "step": 4443 + }, + { + "epoch": 0.4880298704151109, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.3370985984802246, + "learning_rate": 1e-06, + "loss": 1.0396, + "mean_token_accuracy": 0.6912282705307007, + "num_tokens": 114981638.0, + "step": 4444 + }, + { + "epoch": 0.48813968811772457, + "ewc_loss": 2.1010637283325195e-06, + "grad_norm": 2.223435163497925, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7195408940315247, + "num_tokens": 115006462.0, + "step": 4445 + }, + { + "epoch": 0.4882495058203382, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2333076000213623, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7177283763885498, + "num_tokens": 115031636.0, + "step": 4446 + }, + { + "epoch": 0.4883593235229519, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1951236724853516, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7101561427116394, + "num_tokens": 115056335.0, + "step": 4447 + }, + { + "epoch": 0.48846914122556556, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.349565267562866, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6951935291290283, + "num_tokens": 115078790.0, + "step": 4448 + }, + { + "epoch": 0.4885789589281792, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4081315994262695, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7170263528823853, + "num_tokens": 115101699.0, + "step": 4449 + }, + { + "epoch": 0.4886887766307929, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2534122467041016, + "learning_rate": 1e-06, + "loss": 1.0848, + "mean_token_accuracy": 0.6746455430984497, + "num_tokens": 115128503.0, + "step": 4450 + }, + { + "epoch": 0.48879859433340656, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 1.8698772192001343, + "learning_rate": 1e-06, + "loss": 1.0768, + "mean_token_accuracy": 0.6852219104766846, + "num_tokens": 115164634.0, + "step": 4451 + }, + { + "epoch": 0.4889084120360202, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.376720428466797, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6901946067810059, + "num_tokens": 115187246.0, + "step": 4452 + }, + { + "epoch": 0.48901822973863385, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.118231773376465, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.705670952796936, + "num_tokens": 115213586.0, + "step": 4453 + }, + { + "epoch": 0.48912804744124755, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4209096431732178, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.724044144153595, + "num_tokens": 115235227.0, + "step": 4454 + }, + { + "epoch": 0.4892378651438612, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.039527177810669, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7056815028190613, + "num_tokens": 115263710.0, + "step": 4455 + }, + { + "epoch": 0.48934768284647484, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2684385776519775, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7123305201530457, + "num_tokens": 115289195.0, + "step": 4456 + }, + { + "epoch": 0.4894575005490885, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2101752758026123, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7023540139198303, + "num_tokens": 115313974.0, + "step": 4457 + }, + { + "epoch": 0.4895673182517022, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.247511863708496, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6992138028144836, + "num_tokens": 115339114.0, + "step": 4458 + }, + { + "epoch": 0.48967713595431583, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2149930000305176, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.6953592896461487, + "num_tokens": 115365289.0, + "step": 4459 + }, + { + "epoch": 0.4897869536569295, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.336702346801758, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7069997787475586, + "num_tokens": 115387444.0, + "step": 4460 + }, + { + "epoch": 0.4898967713595432, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3376359939575195, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6975821852684021, + "num_tokens": 115409230.0, + "step": 4461 + }, + { + "epoch": 0.4900065890621568, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4314463138580322, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7180610299110413, + "num_tokens": 115430694.0, + "step": 4462 + }, + { + "epoch": 0.4901164067647705, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2295315265655518, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6862810254096985, + "num_tokens": 115456685.0, + "step": 4463 + }, + { + "epoch": 0.4902262244673841, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1302196979522705, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6912099123001099, + "num_tokens": 115483755.0, + "step": 4464 + }, + { + "epoch": 0.4903360421699978, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.5559873580932617, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.72315913438797, + "num_tokens": 115502535.0, + "step": 4465 + }, + { + "epoch": 0.49044585987261147, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.281339168548584, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6999427676200867, + "num_tokens": 115526350.0, + "step": 4466 + }, + { + "epoch": 0.4905556775752251, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.454017400741577, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7142226696014404, + "num_tokens": 115548180.0, + "step": 4467 + }, + { + "epoch": 0.4906654952778388, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.093735456466675, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6959946155548096, + "num_tokens": 115575251.0, + "step": 4468 + }, + { + "epoch": 0.49077531298045246, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3808982372283936, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7285979986190796, + "num_tokens": 115595870.0, + "step": 4469 + }, + { + "epoch": 0.4908851306830661, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1622085571289062, + "learning_rate": 1e-06, + "loss": 1.034, + "mean_token_accuracy": 0.6891661882400513, + "num_tokens": 115622500.0, + "step": 4470 + }, + { + "epoch": 0.49099494838567975, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.449683904647827, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6898543238639832, + "num_tokens": 115644812.0, + "step": 4471 + }, + { + "epoch": 0.49110476608829345, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3560187816619873, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7111965417861938, + "num_tokens": 115668225.0, + "step": 4472 + }, + { + "epoch": 0.4912145837909071, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4562830924987793, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7235848903656006, + "num_tokens": 115690591.0, + "step": 4473 + }, + { + "epoch": 0.49132440149352075, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3163466453552246, + "learning_rate": 1e-06, + "loss": 1.0496, + "mean_token_accuracy": 0.6859076619148254, + "num_tokens": 115715069.0, + "step": 4474 + }, + { + "epoch": 0.4914342191961344, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4205245971679688, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7031847834587097, + "num_tokens": 115737035.0, + "step": 4475 + }, + { + "epoch": 0.4915440368987481, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1433403491973877, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.7062676548957825, + "num_tokens": 115762455.0, + "step": 4476 + }, + { + "epoch": 0.49165385460136174, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.297537326812744, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7073081731796265, + "num_tokens": 115786117.0, + "step": 4477 + }, + { + "epoch": 0.4917636723039754, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3312766551971436, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7270551323890686, + "num_tokens": 115809341.0, + "step": 4478 + }, + { + "epoch": 0.4918734900065891, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.317397117614746, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7065653800964355, + "num_tokens": 115832269.0, + "step": 4479 + }, + { + "epoch": 0.49198330770920273, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.199760913848877, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7011256814002991, + "num_tokens": 115859033.0, + "step": 4480 + }, + { + "epoch": 0.4920931254118164, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3466248512268066, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.706763744354248, + "num_tokens": 115881543.0, + "step": 4481 + }, + { + "epoch": 0.49220294311443, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.172595500946045, + "learning_rate": 1e-06, + "loss": 1.0806, + "mean_token_accuracy": 0.6831592917442322, + "num_tokens": 115908358.0, + "step": 4482 + }, + { + "epoch": 0.4923127608170437, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1695668697357178, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7035201787948608, + "num_tokens": 115934190.0, + "step": 4483 + }, + { + "epoch": 0.4924225785196574, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.2416133880615234, + "learning_rate": 1e-06, + "loss": 1.0473, + "mean_token_accuracy": 0.6827619075775146, + "num_tokens": 115958843.0, + "step": 4484 + }, + { + "epoch": 0.492532396222271, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1235618591308594, + "learning_rate": 1e-06, + "loss": 1.0384, + "mean_token_accuracy": 0.6839851140975952, + "num_tokens": 115984943.0, + "step": 4485 + }, + { + "epoch": 0.49264221392488466, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.033961772918701, + "learning_rate": 1e-06, + "loss": 1.1079, + "mean_token_accuracy": 0.6795681715011597, + "num_tokens": 116015263.0, + "step": 4486 + }, + { + "epoch": 0.49275203162749837, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.144425630569458, + "learning_rate": 1e-06, + "loss": 1.0521, + "mean_token_accuracy": 0.6909223794937134, + "num_tokens": 116042471.0, + "step": 4487 + }, + { + "epoch": 0.492861849330112, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3118631839752197, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.7053909301757812, + "num_tokens": 116067499.0, + "step": 4488 + }, + { + "epoch": 0.49297166703272566, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3642396926879883, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7188791036605835, + "num_tokens": 116088440.0, + "step": 4489 + }, + { + "epoch": 0.49308148473533936, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.4158432483673096, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6870650053024292, + "num_tokens": 116111112.0, + "step": 4490 + }, + { + "epoch": 0.493191302437953, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.255983352661133, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7015583515167236, + "num_tokens": 116135455.0, + "step": 4491 + }, + { + "epoch": 0.49330112014056665, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.276909112930298, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6901644468307495, + "num_tokens": 116160890.0, + "step": 4492 + }, + { + "epoch": 0.4934109378431803, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1392135620117188, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7138283252716064, + "num_tokens": 116187808.0, + "step": 4493 + }, + { + "epoch": 0.493520755545794, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.0642900466918945, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.6875604391098022, + "num_tokens": 116217838.0, + "step": 4494 + }, + { + "epoch": 0.49363057324840764, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.058727502822876, + "learning_rate": 1e-06, + "loss": 1.0647, + "mean_token_accuracy": 0.6877840757369995, + "num_tokens": 116246981.0, + "step": 4495 + }, + { + "epoch": 0.4937403909510213, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.0893192291259766, + "learning_rate": 1e-06, + "loss": 1.0712, + "mean_token_accuracy": 0.6789162158966064, + "num_tokens": 116276532.0, + "step": 4496 + }, + { + "epoch": 0.493850208653635, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.1534085273742676, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7087391018867493, + "num_tokens": 116302544.0, + "step": 4497 + }, + { + "epoch": 0.49396002635624864, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.3977625370025635, + "learning_rate": 1e-06, + "loss": 1.1021, + "mean_token_accuracy": 0.6736769676208496, + "num_tokens": 116326255.0, + "step": 4498 + }, + { + "epoch": 0.4940698440588623, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.22129487991333, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.705496072769165, + "num_tokens": 116351717.0, + "step": 4499 + }, + { + "epoch": 0.49417966176147593, + "ewc_loss": 2.115964889526367e-06, + "grad_norm": 2.086592674255371, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6898432970046997, + "num_tokens": 116381049.0, + "step": 4500 + }, + { + "epoch": 0.49428947946408963, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.319387435913086, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.713422417640686, + "num_tokens": 116404558.0, + "step": 4501 + }, + { + "epoch": 0.4943992971667033, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.243417978286743, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6955450177192688, + "num_tokens": 116430821.0, + "step": 4502 + }, + { + "epoch": 0.4945091148693169, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.1923561096191406, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7163411974906921, + "num_tokens": 116456843.0, + "step": 4503 + }, + { + "epoch": 0.49461893257193057, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 1.9882533550262451, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7019832730293274, + "num_tokens": 116488522.0, + "step": 4504 + }, + { + "epoch": 0.49472875027454427, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 1.946623682975769, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6904493570327759, + "num_tokens": 116518860.0, + "step": 4505 + }, + { + "epoch": 0.4948385679771579, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.259997844696045, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7127658128738403, + "num_tokens": 116545374.0, + "step": 4506 + }, + { + "epoch": 0.49494838567977156, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.2191874980926514, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7048325538635254, + "num_tokens": 116572772.0, + "step": 4507 + }, + { + "epoch": 0.49505820338238526, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.0408196449279785, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6964191794395447, + "num_tokens": 116601968.0, + "step": 4508 + }, + { + "epoch": 0.4951680210849989, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.4252922534942627, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6996732950210571, + "num_tokens": 116624913.0, + "step": 4509 + }, + { + "epoch": 0.49527783878761256, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.4879603385925293, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6994675397872925, + "num_tokens": 116646611.0, + "step": 4510 + }, + { + "epoch": 0.4953876564902262, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.402491807937622, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.695855438709259, + "num_tokens": 116669267.0, + "step": 4511 + }, + { + "epoch": 0.4954974741928399, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.2712323665618896, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7091158032417297, + "num_tokens": 116694288.0, + "step": 4512 + }, + { + "epoch": 0.49560729189545355, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.232963800430298, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7112851142883301, + "num_tokens": 116719128.0, + "step": 4513 + }, + { + "epoch": 0.4957171095980672, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.313585042953491, + "learning_rate": 1e-06, + "loss": 1.0765, + "mean_token_accuracy": 0.6864705085754395, + "num_tokens": 116745096.0, + "step": 4514 + }, + { + "epoch": 0.4958269273006809, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.6033291816711426, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7143718004226685, + "num_tokens": 116763826.0, + "step": 4515 + }, + { + "epoch": 0.49593674500329454, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.131664752960205, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7026972770690918, + "num_tokens": 116795641.0, + "step": 4516 + }, + { + "epoch": 0.4960465627059082, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.1301794052124023, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6898493766784668, + "num_tokens": 116822911.0, + "step": 4517 + }, + { + "epoch": 0.49615638040852184, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.336024761199951, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7027038931846619, + "num_tokens": 116846246.0, + "step": 4518 + }, + { + "epoch": 0.49626619811113554, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.0232179164886475, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.7007244825363159, + "num_tokens": 116874624.0, + "step": 4519 + }, + { + "epoch": 0.4963760158137492, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.1082797050476074, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.71631920337677, + "num_tokens": 116899231.0, + "step": 4520 + }, + { + "epoch": 0.49648583351636283, + "ewc_loss": 2.130866050720215e-06, + "grad_norm": 2.352168083190918, + "learning_rate": 1e-06, + "loss": 1.0849, + "mean_token_accuracy": 0.6726815700531006, + "num_tokens": 116924433.0, + "step": 4521 + }, + { + "epoch": 0.4965956512189765, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.116542100906372, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6857713460922241, + "num_tokens": 116954796.0, + "step": 4522 + }, + { + "epoch": 0.4967054689215902, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 1.9510360956192017, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7039616107940674, + "num_tokens": 116984925.0, + "step": 4523 + }, + { + "epoch": 0.4968152866242038, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.090541124343872, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.6980825066566467, + "num_tokens": 117012418.0, + "step": 4524 + }, + { + "epoch": 0.49692510432681747, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.014543294906616, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7154786586761475, + "num_tokens": 117040062.0, + "step": 4525 + }, + { + "epoch": 0.49703492202943117, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 1.9758281707763672, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6875561475753784, + "num_tokens": 117069854.0, + "step": 4526 + }, + { + "epoch": 0.4971447397320448, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.0071089267730713, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7004120349884033, + "num_tokens": 117098015.0, + "step": 4527 + }, + { + "epoch": 0.49725455743465846, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2046029567718506, + "learning_rate": 1e-06, + "loss": 1.0674, + "mean_token_accuracy": 0.6818912029266357, + "num_tokens": 117127797.0, + "step": 4528 + }, + { + "epoch": 0.4973643751372721, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.4066383838653564, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.6959977149963379, + "num_tokens": 117147759.0, + "step": 4529 + }, + { + "epoch": 0.4974741928398858, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2572836875915527, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6940392255783081, + "num_tokens": 117171512.0, + "step": 4530 + }, + { + "epoch": 0.49758401054249946, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2794017791748047, + "learning_rate": 1e-06, + "loss": 1.0578, + "mean_token_accuracy": 0.6915896534919739, + "num_tokens": 117196904.0, + "step": 4531 + }, + { + "epoch": 0.4976938282451131, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.4964962005615234, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7076689600944519, + "num_tokens": 117218115.0, + "step": 4532 + }, + { + "epoch": 0.49780364594772675, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.0747740268707275, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7105908393859863, + "num_tokens": 117244201.0, + "step": 4533 + }, + { + "epoch": 0.49791346365034045, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2087743282318115, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.735336184501648, + "num_tokens": 117267186.0, + "step": 4534 + }, + { + "epoch": 0.4980232813529541, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.0940356254577637, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6951806545257568, + "num_tokens": 117297522.0, + "step": 4535 + }, + { + "epoch": 0.49813309905556774, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.3155579566955566, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6939046382904053, + "num_tokens": 117323164.0, + "step": 4536 + }, + { + "epoch": 0.49824291675818144, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.3023335933685303, + "learning_rate": 1e-06, + "loss": 1.0003, + "mean_token_accuracy": 0.7012563943862915, + "num_tokens": 117347079.0, + "step": 4537 + }, + { + "epoch": 0.4983527344607951, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.264496088027954, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.7054006457328796, + "num_tokens": 117373678.0, + "step": 4538 + }, + { + "epoch": 0.49846255216340873, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.16213321685791, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6925017833709717, + "num_tokens": 117400721.0, + "step": 4539 + }, + { + "epoch": 0.4985723698660224, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2401552200317383, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7120754718780518, + "num_tokens": 117425568.0, + "step": 4540 + }, + { + "epoch": 0.4986821875686361, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.3264319896698, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.6994752883911133, + "num_tokens": 117448642.0, + "step": 4541 + }, + { + "epoch": 0.4987920052712497, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.4252126216888428, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.7018161416053772, + "num_tokens": 117471867.0, + "step": 4542 + }, + { + "epoch": 0.4989018229738634, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 1.9045052528381348, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.7008200883865356, + "num_tokens": 117503893.0, + "step": 4543 + }, + { + "epoch": 0.4990116406764771, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.131072759628296, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6938494443893433, + "num_tokens": 117532300.0, + "step": 4544 + }, + { + "epoch": 0.4991214583790907, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2508881092071533, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7115639448165894, + "num_tokens": 117557057.0, + "step": 4545 + }, + { + "epoch": 0.49923127608170437, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.070476770401001, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6949544548988342, + "num_tokens": 117586763.0, + "step": 4546 + }, + { + "epoch": 0.499341093784318, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.187955379486084, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.7071399092674255, + "num_tokens": 117611783.0, + "step": 4547 + }, + { + "epoch": 0.4994509114869317, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.1162760257720947, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7276449203491211, + "num_tokens": 117638085.0, + "step": 4548 + }, + { + "epoch": 0.49956072918954536, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.002934217453003, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6967888474464417, + "num_tokens": 117669103.0, + "step": 4549 + }, + { + "epoch": 0.499670546892159, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.310321807861328, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6887718439102173, + "num_tokens": 117693830.0, + "step": 4550 + }, + { + "epoch": 0.49978036459477265, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.489267587661743, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.6919873952865601, + "num_tokens": 117714230.0, + "step": 4551 + }, + { + "epoch": 0.49989018229738635, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 1.911312460899353, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.6936344504356384, + "num_tokens": 117747833.0, + "step": 4552 + }, + { + "epoch": 0.5, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.2970023155212402, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6852063536643982, + "num_tokens": 117773052.0, + "step": 4553 + }, + { + "epoch": 0.5001098177026136, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.1404027938842773, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7065354585647583, + "num_tokens": 117798234.0, + "step": 4554 + }, + { + "epoch": 0.5002196354052273, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.502998113632202, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7097245454788208, + "num_tokens": 117820361.0, + "step": 4555 + }, + { + "epoch": 0.5003294531078409, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.243947744369507, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6882040500640869, + "num_tokens": 117846772.0, + "step": 4556 + }, + { + "epoch": 0.5004392708104547, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 1.9773550033569336, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6908529996871948, + "num_tokens": 117877864.0, + "step": 4557 + }, + { + "epoch": 0.5005490885130683, + "ewc_loss": 2.16066837310791e-06, + "grad_norm": 2.168344259262085, + "learning_rate": 1e-06, + "loss": 1.0515, + "mean_token_accuracy": 0.6815313100814819, + "num_tokens": 117905908.0, + "step": 4558 + }, + { + "epoch": 0.500658906215682, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.307469606399536, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6996493339538574, + "num_tokens": 117931453.0, + "step": 4559 + }, + { + "epoch": 0.5007687239182956, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.4210364818573, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7078418731689453, + "num_tokens": 117953672.0, + "step": 4560 + }, + { + "epoch": 0.5008785416209093, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.1938729286193848, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6875497102737427, + "num_tokens": 117979254.0, + "step": 4561 + }, + { + "epoch": 0.5009883593235229, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.0913867950439453, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7043165564537048, + "num_tokens": 118007249.0, + "step": 4562 + }, + { + "epoch": 0.5010981770261366, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.1231162548065186, + "learning_rate": 1e-06, + "loss": 1.0606, + "mean_token_accuracy": 0.6828622221946716, + "num_tokens": 118038228.0, + "step": 4563 + }, + { + "epoch": 0.5012079947287503, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.121286392211914, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.7036720514297485, + "num_tokens": 118064982.0, + "step": 4564 + }, + { + "epoch": 0.501317812431364, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 2.226398229598999, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7042789459228516, + "num_tokens": 118088105.0, + "step": 4565 + }, + { + "epoch": 0.5014276301339776, + "ewc_loss": 2.175569534301758e-06, + "grad_norm": 1.87200129032135, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6967158317565918, + "num_tokens": 118123176.0, + "step": 4566 + }, + { + "epoch": 0.5015374478365913, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1250064373016357, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7020537853240967, + "num_tokens": 118150578.0, + "step": 4567 + }, + { + "epoch": 0.5016472655392049, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.203885078430176, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7234059572219849, + "num_tokens": 118177358.0, + "step": 4568 + }, + { + "epoch": 0.5017570832418186, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1181888580322266, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7005480527877808, + "num_tokens": 118207315.0, + "step": 4569 + }, + { + "epoch": 0.5018669009444322, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.2297425270080566, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7000778317451477, + "num_tokens": 118231938.0, + "step": 4570 + }, + { + "epoch": 0.5019767186470458, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.016200304031372, + "learning_rate": 1e-06, + "loss": 1.0989, + "mean_token_accuracy": 0.6702859401702881, + "num_tokens": 118264767.0, + "step": 4571 + }, + { + "epoch": 0.5020865363496596, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3340301513671875, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6876800060272217, + "num_tokens": 118287546.0, + "step": 4572 + }, + { + "epoch": 0.5021963540522733, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3033838272094727, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7108113765716553, + "num_tokens": 118310519.0, + "step": 4573 + }, + { + "epoch": 0.5023061717548869, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1589701175689697, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7025796175003052, + "num_tokens": 118335584.0, + "step": 4574 + }, + { + "epoch": 0.5024159894575005, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.4263250827789307, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7171164155006409, + "num_tokens": 118355882.0, + "step": 4575 + }, + { + "epoch": 0.5025258071601142, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.274707555770874, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.698782205581665, + "num_tokens": 118379758.0, + "step": 4576 + }, + { + "epoch": 0.5026356248627278, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.0642664432525635, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6940321922302246, + "num_tokens": 118409998.0, + "step": 4577 + }, + { + "epoch": 0.5027454425653415, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.0529282093048096, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7027977108955383, + "num_tokens": 118437895.0, + "step": 4578 + }, + { + "epoch": 0.5028552602679552, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.9985944032669067, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6958112716674805, + "num_tokens": 118466254.0, + "step": 4579 + }, + { + "epoch": 0.5029650779705689, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.2894420623779297, + "learning_rate": 1e-06, + "loss": 1.0656, + "mean_token_accuracy": 0.6869750618934631, + "num_tokens": 118491266.0, + "step": 4580 + }, + { + "epoch": 0.5030748956731825, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.4319961071014404, + "learning_rate": 1e-06, + "loss": 1.0517, + "mean_token_accuracy": 0.6879163980484009, + "num_tokens": 118513453.0, + "step": 4581 + }, + { + "epoch": 0.5031847133757962, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3633785247802734, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6883610486984253, + "num_tokens": 118536240.0, + "step": 4582 + }, + { + "epoch": 0.5032945310784098, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.0503056049346924, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.6966861486434937, + "num_tokens": 118566445.0, + "step": 4583 + }, + { + "epoch": 0.5034043487810235, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.242630958557129, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6941407918930054, + "num_tokens": 118591213.0, + "step": 4584 + }, + { + "epoch": 0.5035141664836371, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1443066596984863, + "learning_rate": 1e-06, + "loss": 1.1232, + "mean_token_accuracy": 0.6709581017494202, + "num_tokens": 118621747.0, + "step": 4585 + }, + { + "epoch": 0.5036239841862509, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.7424964904785156, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6963832378387451, + "num_tokens": 118660496.0, + "step": 4586 + }, + { + "epoch": 0.5037338018888645, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.4067373275756836, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.713874340057373, + "num_tokens": 118682445.0, + "step": 4587 + }, + { + "epoch": 0.5038436195914782, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1086230278015137, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.7016296982765198, + "num_tokens": 118709958.0, + "step": 4588 + }, + { + "epoch": 0.5039534372940918, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.176267147064209, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.7066856622695923, + "num_tokens": 118736291.0, + "step": 4589 + }, + { + "epoch": 0.5040632549967055, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.086496353149414, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6886275410652161, + "num_tokens": 118763755.0, + "step": 4590 + }, + { + "epoch": 0.5041730726993191, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3929007053375244, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7110932469367981, + "num_tokens": 118785592.0, + "step": 4591 + }, + { + "epoch": 0.5042828904019327, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.913764476776123, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6988142728805542, + "num_tokens": 118815809.0, + "step": 4592 + }, + { + "epoch": 0.5043927081045465, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.9588178396224976, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7023988962173462, + "num_tokens": 118844118.0, + "step": 4593 + }, + { + "epoch": 0.5045025258071602, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.8603726625442505, + "learning_rate": 1e-06, + "loss": 1.1346, + "mean_token_accuracy": 0.6598721146583557, + "num_tokens": 118878278.0, + "step": 4594 + }, + { + "epoch": 0.5046123435097738, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1060097217559814, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6934123039245605, + "num_tokens": 118904888.0, + "step": 4595 + }, + { + "epoch": 0.5047221612123874, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.986104965209961, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6998805999755859, + "num_tokens": 118936516.0, + "step": 4596 + }, + { + "epoch": 0.5048319789150011, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.9734243154525757, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6879050731658936, + "num_tokens": 118965982.0, + "step": 4597 + }, + { + "epoch": 0.5049417966176147, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.2319650650024414, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7165104746818542, + "num_tokens": 118989660.0, + "step": 4598 + }, + { + "epoch": 0.5050516143202284, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.290318489074707, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7123880982398987, + "num_tokens": 119014512.0, + "step": 4599 + }, + { + "epoch": 0.505161432022842, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.9671225547790527, + "learning_rate": 1e-06, + "loss": 1.0589, + "mean_token_accuracy": 0.6804248690605164, + "num_tokens": 119046114.0, + "step": 4600 + }, + { + "epoch": 0.5052712497254558, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.501150369644165, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6997426748275757, + "num_tokens": 119069967.0, + "step": 4601 + }, + { + "epoch": 0.5053810674280694, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.052539348602295, + "learning_rate": 1e-06, + "loss": 1.0758, + "mean_token_accuracy": 0.6805386543273926, + "num_tokens": 119102712.0, + "step": 4602 + }, + { + "epoch": 0.5054908851306831, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.12970232963562, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6990092992782593, + "num_tokens": 119131270.0, + "step": 4603 + }, + { + "epoch": 0.5056007028332967, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.938159704208374, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7168784737586975, + "num_tokens": 119147580.0, + "step": 4604 + }, + { + "epoch": 0.5057105205359104, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.0937070846557617, + "learning_rate": 1e-06, + "loss": 1.062, + "mean_token_accuracy": 0.6810779571533203, + "num_tokens": 119176404.0, + "step": 4605 + }, + { + "epoch": 0.505820338238524, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1814076900482178, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6996230483055115, + "num_tokens": 119203393.0, + "step": 4606 + }, + { + "epoch": 0.5059301559411377, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.1353917121887207, + "learning_rate": 1e-06, + "loss": 1.091, + "mean_token_accuracy": 0.6755242347717285, + "num_tokens": 119231833.0, + "step": 4607 + }, + { + "epoch": 0.5060399736437514, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3495423793792725, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.701810896396637, + "num_tokens": 119256386.0, + "step": 4608 + }, + { + "epoch": 0.5061497913463651, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.0731489658355713, + "learning_rate": 1e-06, + "loss": 1.0335, + "mean_token_accuracy": 0.6830570101737976, + "num_tokens": 119285024.0, + "step": 4609 + }, + { + "epoch": 0.5062596090489787, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.271111488342285, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.721383273601532, + "num_tokens": 119308733.0, + "step": 4610 + }, + { + "epoch": 0.5063694267515924, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3147382736206055, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.6945642828941345, + "num_tokens": 119332389.0, + "step": 4611 + }, + { + "epoch": 0.506479244454206, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.141578435897827, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6935892105102539, + "num_tokens": 119360140.0, + "step": 4612 + }, + { + "epoch": 0.5065890621568196, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.4629409313201904, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.717972993850708, + "num_tokens": 119381098.0, + "step": 4613 + }, + { + "epoch": 0.5066988798594333, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.175558090209961, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7280569076538086, + "num_tokens": 119406309.0, + "step": 4614 + }, + { + "epoch": 0.506808697562047, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.3195643424987793, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.724819004535675, + "num_tokens": 119429642.0, + "step": 4615 + }, + { + "epoch": 0.5069185152646607, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 1.9363056421279907, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6919698715209961, + "num_tokens": 119461475.0, + "step": 4616 + }, + { + "epoch": 0.5070283329672743, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.296093463897705, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7121412754058838, + "num_tokens": 119485600.0, + "step": 4617 + }, + { + "epoch": 0.507138150669888, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.203518867492676, + "learning_rate": 1e-06, + "loss": 1.0813, + "mean_token_accuracy": 0.6840327382087708, + "num_tokens": 119511824.0, + "step": 4618 + }, + { + "epoch": 0.5072479683725016, + "ewc_loss": 2.1904706954956055e-06, + "grad_norm": 2.4579930305480957, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.709929883480072, + "num_tokens": 119533133.0, + "step": 4619 + }, + { + "epoch": 0.5073577860751153, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.286054849624634, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7021902799606323, + "num_tokens": 119557080.0, + "step": 4620 + }, + { + "epoch": 0.5074676037777289, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.478609800338745, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6912225484848022, + "num_tokens": 119577569.0, + "step": 4621 + }, + { + "epoch": 0.5075774214803427, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.204594135284424, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6935102343559265, + "num_tokens": 119605645.0, + "step": 4622 + }, + { + "epoch": 0.5076872391829563, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.442176580429077, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7077928781509399, + "num_tokens": 119626742.0, + "step": 4623 + }, + { + "epoch": 0.50779705688557, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.3938865661621094, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7009531259536743, + "num_tokens": 119648331.0, + "step": 4624 + }, + { + "epoch": 0.5079068745881836, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.3161706924438477, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7081756591796875, + "num_tokens": 119671073.0, + "step": 4625 + }, + { + "epoch": 0.5080166922907973, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.2194974422454834, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7062718272209167, + "num_tokens": 119697145.0, + "step": 4626 + }, + { + "epoch": 0.5081265099934109, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.4366087913513184, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6928225755691528, + "num_tokens": 119719279.0, + "step": 4627 + }, + { + "epoch": 0.5082363276960246, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.089569568634033, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7079597115516663, + "num_tokens": 119746924.0, + "step": 4628 + }, + { + "epoch": 0.5083461453986382, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.485142230987549, + "learning_rate": 1e-06, + "loss": 1.0317, + "mean_token_accuracy": 0.6823042035102844, + "num_tokens": 119769037.0, + "step": 4629 + }, + { + "epoch": 0.508455963101252, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.026857376098633, + "learning_rate": 1e-06, + "loss": 1.039, + "mean_token_accuracy": 0.6867164373397827, + "num_tokens": 119797640.0, + "step": 4630 + }, + { + "epoch": 0.5085657808038656, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.163515090942383, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7146826982498169, + "num_tokens": 119821460.0, + "step": 4631 + }, + { + "epoch": 0.5086755985064793, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.0921480655670166, + "learning_rate": 1e-06, + "loss": 1.0264, + "mean_token_accuracy": 0.6845179796218872, + "num_tokens": 119850542.0, + "step": 4632 + }, + { + "epoch": 0.5087854162090929, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.0511388778686523, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7078430652618408, + "num_tokens": 119877816.0, + "step": 4633 + }, + { + "epoch": 0.5088952339117065, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.0504486560821533, + "learning_rate": 1e-06, + "loss": 1.047, + "mean_token_accuracy": 0.6837260127067566, + "num_tokens": 119907607.0, + "step": 4634 + }, + { + "epoch": 0.5090050516143202, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.156022548675537, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6980019211769104, + "num_tokens": 119933810.0, + "step": 4635 + }, + { + "epoch": 0.5091148693169338, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.2423505783081055, + "learning_rate": 1e-06, + "loss": 1.0462, + "mean_token_accuracy": 0.6800662875175476, + "num_tokens": 119960256.0, + "step": 4636 + }, + { + "epoch": 0.5092246870195476, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.0323946475982666, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.7022062540054321, + "num_tokens": 119988437.0, + "step": 4637 + }, + { + "epoch": 0.5093345047221612, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.2473223209381104, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7029943466186523, + "num_tokens": 120012553.0, + "step": 4638 + }, + { + "epoch": 0.5094443224247749, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.166975975036621, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6989507675170898, + "num_tokens": 120037367.0, + "step": 4639 + }, + { + "epoch": 0.5095541401273885, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.5514957904815674, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7201123237609863, + "num_tokens": 120056816.0, + "step": 4640 + }, + { + "epoch": 0.5096639578300022, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 1.9231653213500977, + "learning_rate": 1e-06, + "loss": 1.0581, + "mean_token_accuracy": 0.6828649044036865, + "num_tokens": 120090690.0, + "step": 4641 + }, + { + "epoch": 0.5097737755326158, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.2658655643463135, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7061299085617065, + "num_tokens": 120114819.0, + "step": 4642 + }, + { + "epoch": 0.5098835932352295, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.0403454303741455, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7068281769752502, + "num_tokens": 120144814.0, + "step": 4643 + }, + { + "epoch": 0.5099934109378432, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.2443203926086426, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7214457392692566, + "num_tokens": 120168876.0, + "step": 4644 + }, + { + "epoch": 0.5101032286404569, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.1773622035980225, + "learning_rate": 1e-06, + "loss": 1.0585, + "mean_token_accuracy": 0.6818958520889282, + "num_tokens": 120196920.0, + "step": 4645 + }, + { + "epoch": 0.5102130463430705, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.3092739582061768, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6975786685943604, + "num_tokens": 120222102.0, + "step": 4646 + }, + { + "epoch": 0.5103228640456842, + "ewc_loss": 2.2202730178833008e-06, + "grad_norm": 2.1899571418762207, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7034540176391602, + "num_tokens": 120247712.0, + "step": 4647 + }, + { + "epoch": 0.5104326817482978, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 1.92250394821167, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6899300813674927, + "num_tokens": 120281315.0, + "step": 4648 + }, + { + "epoch": 0.5105424994509115, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.1219534873962402, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7036553621292114, + "num_tokens": 120309377.0, + "step": 4649 + }, + { + "epoch": 0.5106523171535251, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.087958335876465, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.688279390335083, + "num_tokens": 120337949.0, + "step": 4650 + }, + { + "epoch": 0.5107621348561389, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.170957088470459, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7077100872993469, + "num_tokens": 120362736.0, + "step": 4651 + }, + { + "epoch": 0.5108719525587525, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.1180031299591064, + "learning_rate": 1e-06, + "loss": 1.0355, + "mean_token_accuracy": 0.6917418241500854, + "num_tokens": 120394003.0, + "step": 4652 + }, + { + "epoch": 0.5109817702613662, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.3901383876800537, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.6934648752212524, + "num_tokens": 120417078.0, + "step": 4653 + }, + { + "epoch": 0.5110915879639798, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.2414934635162354, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7230915427207947, + "num_tokens": 120440605.0, + "step": 4654 + }, + { + "epoch": 0.5112014056665934, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.3808324337005615, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7055050134658813, + "num_tokens": 120463428.0, + "step": 4655 + }, + { + "epoch": 0.5113112233692071, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.1138925552368164, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.7028912305831909, + "num_tokens": 120493073.0, + "step": 4656 + }, + { + "epoch": 0.5114210410718207, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.208343267440796, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6997360587120056, + "num_tokens": 120518143.0, + "step": 4657 + }, + { + "epoch": 0.5115308587744345, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.298935890197754, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7052109241485596, + "num_tokens": 120542359.0, + "step": 4658 + }, + { + "epoch": 0.5116406764770481, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.153010606765747, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7028035521507263, + "num_tokens": 120571395.0, + "step": 4659 + }, + { + "epoch": 0.5117504941796618, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 1.856059193611145, + "learning_rate": 1e-06, + "loss": 1.0541, + "mean_token_accuracy": 0.6823788285255432, + "num_tokens": 120607912.0, + "step": 4660 + }, + { + "epoch": 0.5118603118822754, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 1.9929344654083252, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.700410783290863, + "num_tokens": 120639223.0, + "step": 4661 + }, + { + "epoch": 0.5119701295848891, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.5354506969451904, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7087554335594177, + "num_tokens": 120659595.0, + "step": 4662 + }, + { + "epoch": 0.5120799472875027, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 1.8823530673980713, + "learning_rate": 1e-06, + "loss": 1.0488, + "mean_token_accuracy": 0.6872054934501648, + "num_tokens": 120694465.0, + "step": 4663 + }, + { + "epoch": 0.5121897649901164, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.1514790058135986, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6983429193496704, + "num_tokens": 120722417.0, + "step": 4664 + }, + { + "epoch": 0.51229958269273, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.1175949573516846, + "learning_rate": 1e-06, + "loss": 1.0736, + "mean_token_accuracy": 0.6816563606262207, + "num_tokens": 120750601.0, + "step": 4665 + }, + { + "epoch": 0.5124094003953438, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.0113401412963867, + "learning_rate": 1e-06, + "loss": 1.0631, + "mean_token_accuracy": 0.6770697832107544, + "num_tokens": 120781378.0, + "step": 4666 + }, + { + "epoch": 0.5125192180979574, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.1957051753997803, + "learning_rate": 1e-06, + "loss": 1.025, + "mean_token_accuracy": 0.6936280727386475, + "num_tokens": 120807616.0, + "step": 4667 + }, + { + "epoch": 0.5126290358005711, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.2689242362976074, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6936872005462646, + "num_tokens": 120832422.0, + "step": 4668 + }, + { + "epoch": 0.5127388535031847, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.1443099975585938, + "learning_rate": 1e-06, + "loss": 1.0677, + "mean_token_accuracy": 0.676755428314209, + "num_tokens": 120859714.0, + "step": 4669 + }, + { + "epoch": 0.5128486712057984, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.322545289993286, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7137376070022583, + "num_tokens": 120883835.0, + "step": 4670 + }, + { + "epoch": 0.512958488908412, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.2285852432250977, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6965652704238892, + "num_tokens": 120907148.0, + "step": 4671 + }, + { + "epoch": 0.5130683066110256, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.0395076274871826, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7020800709724426, + "num_tokens": 120937197.0, + "step": 4672 + }, + { + "epoch": 0.5131781243136394, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.291990041732788, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7114439010620117, + "num_tokens": 120960227.0, + "step": 4673 + }, + { + "epoch": 0.513287942016253, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.0728759765625, + "learning_rate": 1e-06, + "loss": 1.0642, + "mean_token_accuracy": 0.6980551481246948, + "num_tokens": 120987671.0, + "step": 4674 + }, + { + "epoch": 0.5133977597188667, + "ewc_loss": 2.2351741790771484e-06, + "grad_norm": 2.2042462825775146, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7173883318901062, + "num_tokens": 121014518.0, + "step": 4675 + }, + { + "epoch": 0.5135075774214803, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.075498104095459, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6905303001403809, + "num_tokens": 121046518.0, + "step": 4676 + }, + { + "epoch": 0.513617395124094, + "ewc_loss": 2.250075340270996e-06, + "grad_norm": 2.3086087703704834, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7069624066352844, + "num_tokens": 121070381.0, + "step": 4677 + }, + { + "epoch": 0.5137272128267076, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.235476016998291, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.706378698348999, + "num_tokens": 121093702.0, + "step": 4678 + }, + { + "epoch": 0.5138370305293213, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.654686212539673, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.6961709856987, + "num_tokens": 121114919.0, + "step": 4679 + }, + { + "epoch": 0.513946848231935, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.2791757583618164, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6990779042243958, + "num_tokens": 121137886.0, + "step": 4680 + }, + { + "epoch": 0.5140566659345487, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.1445858478546143, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.7015913128852844, + "num_tokens": 121164767.0, + "step": 4681 + }, + { + "epoch": 0.5141664836371623, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.039278745651245, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.6908737421035767, + "num_tokens": 121194080.0, + "step": 4682 + }, + { + "epoch": 0.514276301339776, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 1.991102933883667, + "learning_rate": 1e-06, + "loss": 1.1088, + "mean_token_accuracy": 0.6682894229888916, + "num_tokens": 121226531.0, + "step": 4683 + }, + { + "epoch": 0.5143861190423896, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.2502098083496094, + "learning_rate": 1e-06, + "loss": 1.0955, + "mean_token_accuracy": 0.6757124662399292, + "num_tokens": 121254072.0, + "step": 4684 + }, + { + "epoch": 0.5144959367450033, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.277775764465332, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6996594667434692, + "num_tokens": 121279731.0, + "step": 4685 + }, + { + "epoch": 0.5146057544476169, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 1.8545308113098145, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7182536125183105, + "num_tokens": 121312266.0, + "step": 4686 + }, + { + "epoch": 0.5147155721502307, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.2163593769073486, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6980985403060913, + "num_tokens": 121337329.0, + "step": 4687 + }, + { + "epoch": 0.5148253898528443, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 1.9797556400299072, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6928640007972717, + "num_tokens": 121369123.0, + "step": 4688 + }, + { + "epoch": 0.514935207555458, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.3270506858825684, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7066236734390259, + "num_tokens": 121391602.0, + "step": 4689 + }, + { + "epoch": 0.5150450252580716, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.115513324737549, + "learning_rate": 1e-06, + "loss": 1.014, + "mean_token_accuracy": 0.687630295753479, + "num_tokens": 121418720.0, + "step": 4690 + }, + { + "epoch": 0.5151548429606853, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.1894443035125732, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7190207839012146, + "num_tokens": 121444116.0, + "step": 4691 + }, + { + "epoch": 0.5152646606632989, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 1.9788926839828491, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6998804807662964, + "num_tokens": 121474478.0, + "step": 4692 + }, + { + "epoch": 0.5153744783659125, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.0666801929473877, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6965633034706116, + "num_tokens": 121502744.0, + "step": 4693 + }, + { + "epoch": 0.5154842960685262, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.099062919616699, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7225658893585205, + "num_tokens": 121530615.0, + "step": 4694 + }, + { + "epoch": 0.51559411377114, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.0611138343811035, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7006913423538208, + "num_tokens": 121559108.0, + "step": 4695 + }, + { + "epoch": 0.5157039314737536, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.0631988048553467, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.7080358266830444, + "num_tokens": 121588392.0, + "step": 4696 + }, + { + "epoch": 0.5158137491763672, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.008931875228882, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.708940327167511, + "num_tokens": 121617083.0, + "step": 4697 + }, + { + "epoch": 0.5159235668789809, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.267049789428711, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7130211591720581, + "num_tokens": 121641301.0, + "step": 4698 + }, + { + "epoch": 0.5160333845815945, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.019944190979004, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7050487995147705, + "num_tokens": 121671011.0, + "step": 4699 + }, + { + "epoch": 0.5161432022842082, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.144268035888672, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7136732339859009, + "num_tokens": 121697278.0, + "step": 4700 + }, + { + "epoch": 0.5162530199868218, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.0969810485839844, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.7036843299865723, + "num_tokens": 121724693.0, + "step": 4701 + }, + { + "epoch": 0.5163628376894356, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.1120712757110596, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6980619430541992, + "num_tokens": 121753616.0, + "step": 4702 + }, + { + "epoch": 0.5164726553920492, + "ewc_loss": 2.2649765014648438e-06, + "grad_norm": 2.0647192001342773, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7057599425315857, + "num_tokens": 121781118.0, + "step": 4703 + }, + { + "epoch": 0.5165824730946629, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0918784141540527, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7213847637176514, + "num_tokens": 121807666.0, + "step": 4704 + }, + { + "epoch": 0.5166922907972765, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.502106189727783, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7316885590553284, + "num_tokens": 121826186.0, + "step": 4705 + }, + { + "epoch": 0.5168021084998902, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.186551094055176, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7089020013809204, + "num_tokens": 121853749.0, + "step": 4706 + }, + { + "epoch": 0.5169119262025038, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.247669219970703, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6984855532646179, + "num_tokens": 121879214.0, + "step": 4707 + }, + { + "epoch": 0.5170217439051175, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1940746307373047, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.7043284177780151, + "num_tokens": 121905782.0, + "step": 4708 + }, + { + "epoch": 0.5171315616077312, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0768232345581055, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6909365057945251, + "num_tokens": 121935314.0, + "step": 4709 + }, + { + "epoch": 0.5172413793103449, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.9129765033721924, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7150253653526306, + "num_tokens": 121965963.0, + "step": 4710 + }, + { + "epoch": 0.5173511970129585, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.9896774291992188, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7072878479957581, + "num_tokens": 121994548.0, + "step": 4711 + }, + { + "epoch": 0.5174610147155722, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.134080648422241, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7167038917541504, + "num_tokens": 122021964.0, + "step": 4712 + }, + { + "epoch": 0.5175708324181858, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.5411863327026367, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7158349752426147, + "num_tokens": 122043758.0, + "step": 4713 + }, + { + "epoch": 0.5176806501207994, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.255282163619995, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7129074335098267, + "num_tokens": 122068302.0, + "step": 4714 + }, + { + "epoch": 0.5177904678234131, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2154252529144287, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7051824331283569, + "num_tokens": 122095220.0, + "step": 4715 + }, + { + "epoch": 0.5179002855260268, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3706579208374023, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7014714479446411, + "num_tokens": 122117264.0, + "step": 4716 + }, + { + "epoch": 0.5180101032286405, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.4779915809631348, + "learning_rate": 1e-06, + "loss": 1.1104, + "mean_token_accuracy": 0.6613934636116028, + "num_tokens": 122140503.0, + "step": 4717 + }, + { + "epoch": 0.5181199209312541, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.294370174407959, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6985186338424683, + "num_tokens": 122162514.0, + "step": 4718 + }, + { + "epoch": 0.5182297386338678, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3560237884521484, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6954935193061829, + "num_tokens": 122184999.0, + "step": 4719 + }, + { + "epoch": 0.5183395563364814, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.239259719848633, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7121440768241882, + "num_tokens": 122213918.0, + "step": 4720 + }, + { + "epoch": 0.5184493740390951, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3429250717163086, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7057214975357056, + "num_tokens": 122237734.0, + "step": 4721 + }, + { + "epoch": 0.5185591917417087, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2986769676208496, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.700671911239624, + "num_tokens": 122260160.0, + "step": 4722 + }, + { + "epoch": 0.5186690094443224, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0884134769439697, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7203423976898193, + "num_tokens": 122286833.0, + "step": 4723 + }, + { + "epoch": 0.5187788271469361, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1702685356140137, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.735014796257019, + "num_tokens": 122310433.0, + "step": 4724 + }, + { + "epoch": 0.5188886448495498, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.247405529022217, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.6982287764549255, + "num_tokens": 122334006.0, + "step": 4725 + }, + { + "epoch": 0.5189984625521634, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1582016944885254, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7058655023574829, + "num_tokens": 122361228.0, + "step": 4726 + }, + { + "epoch": 0.5191082802547771, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.188408136367798, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7223899364471436, + "num_tokens": 122386834.0, + "step": 4727 + }, + { + "epoch": 0.5192180979573907, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2999346256256104, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6993480920791626, + "num_tokens": 122411805.0, + "step": 4728 + }, + { + "epoch": 0.5193279156600044, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.9464105367660522, + "learning_rate": 1e-06, + "loss": 1.1039, + "mean_token_accuracy": 0.6779435276985168, + "num_tokens": 122446482.0, + "step": 4729 + }, + { + "epoch": 0.519437733362618, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.18166184425354, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7071311473846436, + "num_tokens": 122473380.0, + "step": 4730 + }, + { + "epoch": 0.5195475510652318, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3544130325317383, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7215979099273682, + "num_tokens": 122497542.0, + "step": 4731 + }, + { + "epoch": 0.5196573687678454, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.039705753326416, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6926858425140381, + "num_tokens": 122526094.0, + "step": 4732 + }, + { + "epoch": 0.519767186470459, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.044828414916992, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6912151575088501, + "num_tokens": 122555978.0, + "step": 4733 + }, + { + "epoch": 0.5198770041730727, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.089475154876709, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7043150067329407, + "num_tokens": 122584360.0, + "step": 4734 + }, + { + "epoch": 0.5199868218756863, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3431689739227295, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6956390738487244, + "num_tokens": 122606846.0, + "step": 4735 + }, + { + "epoch": 0.5200966395783, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.5733656883239746, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7164170742034912, + "num_tokens": 122626112.0, + "step": 4736 + }, + { + "epoch": 0.5202064572809136, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.291771650314331, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.6930224895477295, + "num_tokens": 122650732.0, + "step": 4737 + }, + { + "epoch": 0.5203162749835274, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2772154808044434, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7112442255020142, + "num_tokens": 122674126.0, + "step": 4738 + }, + { + "epoch": 0.520426092686141, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.407130479812622, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7096216678619385, + "num_tokens": 122697488.0, + "step": 4739 + }, + { + "epoch": 0.5205359103887547, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3531107902526855, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7007540464401245, + "num_tokens": 122722311.0, + "step": 4740 + }, + { + "epoch": 0.5206457280913683, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.280749559402466, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7082546949386597, + "num_tokens": 122748210.0, + "step": 4741 + }, + { + "epoch": 0.520755545793982, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3431437015533447, + "learning_rate": 1e-06, + "loss": 1.0616, + "mean_token_accuracy": 0.6870827674865723, + "num_tokens": 122772245.0, + "step": 4742 + }, + { + "epoch": 0.5208653634965956, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.15398907661438, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6997997760772705, + "num_tokens": 122798903.0, + "step": 4743 + }, + { + "epoch": 0.5209751811992093, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.6894421577453613, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6994801759719849, + "num_tokens": 122818671.0, + "step": 4744 + }, + { + "epoch": 0.521084998901823, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.118825674057007, + "learning_rate": 1e-06, + "loss": 1.1346, + "mean_token_accuracy": 0.6671395897865295, + "num_tokens": 122849140.0, + "step": 4745 + }, + { + "epoch": 0.5211948166044367, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0627849102020264, + "learning_rate": 1e-06, + "loss": 1.0694, + "mean_token_accuracy": 0.6945382356643677, + "num_tokens": 122879171.0, + "step": 4746 + }, + { + "epoch": 0.5213046343070503, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1326210498809814, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6967481970787048, + "num_tokens": 122908443.0, + "step": 4747 + }, + { + "epoch": 0.521414452009664, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.39060640335083, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7175795435905457, + "num_tokens": 122928803.0, + "step": 4748 + }, + { + "epoch": 0.5215242697122776, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2074315547943115, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7020013332366943, + "num_tokens": 122954480.0, + "step": 4749 + }, + { + "epoch": 0.5216340874148913, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1122350692749023, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7114661335945129, + "num_tokens": 122982025.0, + "step": 4750 + }, + { + "epoch": 0.5217439051175049, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.279222249984741, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.707197904586792, + "num_tokens": 123009122.0, + "step": 4751 + }, + { + "epoch": 0.5218537228201185, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2443010807037354, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7249648571014404, + "num_tokens": 123034418.0, + "step": 4752 + }, + { + "epoch": 0.5219635405227323, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.415433883666992, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7044360041618347, + "num_tokens": 123055576.0, + "step": 4753 + }, + { + "epoch": 0.522073358225346, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.6550793647766113, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.704254686832428, + "num_tokens": 123074540.0, + "step": 4754 + }, + { + "epoch": 0.5221831759279596, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.4849801063537598, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7076291441917419, + "num_tokens": 123097449.0, + "step": 4755 + }, + { + "epoch": 0.5222929936305732, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1909961700439453, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6912519931793213, + "num_tokens": 123125290.0, + "step": 4756 + }, + { + "epoch": 0.5224028113331869, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.976567268371582, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.6989680528640747, + "num_tokens": 123155920.0, + "step": 4757 + }, + { + "epoch": 0.5225126290358005, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3442909717559814, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.718281090259552, + "num_tokens": 123176959.0, + "step": 4758 + }, + { + "epoch": 0.5226224467384142, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1972100734710693, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6910808086395264, + "num_tokens": 123202287.0, + "step": 4759 + }, + { + "epoch": 0.5227322644410279, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.360363245010376, + "learning_rate": 1e-06, + "loss": 1.0666, + "mean_token_accuracy": 0.6888542175292969, + "num_tokens": 123225282.0, + "step": 4760 + }, + { + "epoch": 0.5228420821436416, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2959821224212646, + "learning_rate": 1e-06, + "loss": 1.064, + "mean_token_accuracy": 0.6815970540046692, + "num_tokens": 123251470.0, + "step": 4761 + }, + { + "epoch": 0.5229518998462552, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.050955295562744, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.687843918800354, + "num_tokens": 123280914.0, + "step": 4762 + }, + { + "epoch": 0.5230617175488689, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.189068555831909, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7109825015068054, + "num_tokens": 123307373.0, + "step": 4763 + }, + { + "epoch": 0.5231715352514825, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.9426478147506714, + "learning_rate": 1e-06, + "loss": 1.0608, + "mean_token_accuracy": 0.6804769039154053, + "num_tokens": 123341593.0, + "step": 4764 + }, + { + "epoch": 0.5232813529540962, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.23276686668396, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7117254734039307, + "num_tokens": 123367583.0, + "step": 4765 + }, + { + "epoch": 0.5233911706567098, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 1.8437132835388184, + "learning_rate": 1e-06, + "loss": 1.0496, + "mean_token_accuracy": 0.6886054277420044, + "num_tokens": 123404988.0, + "step": 4766 + }, + { + "epoch": 0.5235009883593236, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1253511905670166, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7148120999336243, + "num_tokens": 123429995.0, + "step": 4767 + }, + { + "epoch": 0.5236108060619372, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.102519989013672, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.695290744304657, + "num_tokens": 123458296.0, + "step": 4768 + }, + { + "epoch": 0.5237206237645509, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.149662733078003, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.702027440071106, + "num_tokens": 123484696.0, + "step": 4769 + }, + { + "epoch": 0.5238304414671645, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.222968339920044, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6996716856956482, + "num_tokens": 123511954.0, + "step": 4770 + }, + { + "epoch": 0.5239402591697782, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.52227520942688, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7174900770187378, + "num_tokens": 123532302.0, + "step": 4771 + }, + { + "epoch": 0.5240500768723918, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.376624584197998, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7083344459533691, + "num_tokens": 123557394.0, + "step": 4772 + }, + { + "epoch": 0.5241598945750054, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.400540351867676, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6913337111473083, + "num_tokens": 123581035.0, + "step": 4773 + }, + { + "epoch": 0.5242697122776192, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.089460849761963, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7028485536575317, + "num_tokens": 123609927.0, + "step": 4774 + }, + { + "epoch": 0.5243795299802329, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2887630462646484, + "learning_rate": 1e-06, + "loss": 1.0516, + "mean_token_accuracy": 0.6855250000953674, + "num_tokens": 123636801.0, + "step": 4775 + }, + { + "epoch": 0.5244893476828465, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1286838054656982, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7241581678390503, + "num_tokens": 123662412.0, + "step": 4776 + }, + { + "epoch": 0.5245991653854601, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2055888175964355, + "learning_rate": 1e-06, + "loss": 1.0528, + "mean_token_accuracy": 0.6900372505187988, + "num_tokens": 123688200.0, + "step": 4777 + }, + { + "epoch": 0.5247089830880738, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.6386544704437256, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7024996280670166, + "num_tokens": 123707255.0, + "step": 4778 + }, + { + "epoch": 0.5248188007906874, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.369495391845703, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7108657956123352, + "num_tokens": 123730727.0, + "step": 4779 + }, + { + "epoch": 0.5249286184933011, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.409838914871216, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7438039779663086, + "num_tokens": 123750769.0, + "step": 4780 + }, + { + "epoch": 0.5250384361959147, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.191321849822998, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7182331681251526, + "num_tokens": 123774443.0, + "step": 4781 + }, + { + "epoch": 0.5251482538985285, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.475395441055298, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7029339075088501, + "num_tokens": 123797174.0, + "step": 4782 + }, + { + "epoch": 0.5252580716011421, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2392168045043945, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6961129903793335, + "num_tokens": 123823711.0, + "step": 4783 + }, + { + "epoch": 0.5253678893037558, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.399001359939575, + "learning_rate": 1e-06, + "loss": 1.0003, + "mean_token_accuracy": 0.6917856931686401, + "num_tokens": 123847343.0, + "step": 4784 + }, + { + "epoch": 0.5254777070063694, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.26170015335083, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.6858131289482117, + "num_tokens": 123872690.0, + "step": 4785 + }, + { + "epoch": 0.5255875247089831, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2855303287506104, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7033243179321289, + "num_tokens": 123897410.0, + "step": 4786 + }, + { + "epoch": 0.5256973424115967, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.3324711322784424, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6932516694068909, + "num_tokens": 123921194.0, + "step": 4787 + }, + { + "epoch": 0.5258071601142104, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1555917263031006, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7090427875518799, + "num_tokens": 123946528.0, + "step": 4788 + }, + { + "epoch": 0.5259169778168241, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0797266960144043, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7048261165618896, + "num_tokens": 123974449.0, + "step": 4789 + }, + { + "epoch": 0.5260267955194378, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.440413236618042, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7182167768478394, + "num_tokens": 123994838.0, + "step": 4790 + }, + { + "epoch": 0.5261366132220514, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1500439643859863, + "learning_rate": 1e-06, + "loss": 1.1191, + "mean_token_accuracy": 0.6646131277084351, + "num_tokens": 124025635.0, + "step": 4791 + }, + { + "epoch": 0.526246430924665, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.196342706680298, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.7014559507369995, + "num_tokens": 124052257.0, + "step": 4792 + }, + { + "epoch": 0.5263562486272787, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2972371578216553, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7047433853149414, + "num_tokens": 124076722.0, + "step": 4793 + }, + { + "epoch": 0.5264660663298923, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.686774730682373, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6918391585350037, + "num_tokens": 124095514.0, + "step": 4794 + }, + { + "epoch": 0.526575884032506, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1700549125671387, + "learning_rate": 1e-06, + "loss": 1.0574, + "mean_token_accuracy": 0.6759313344955444, + "num_tokens": 124123063.0, + "step": 4795 + }, + { + "epoch": 0.5266857017351197, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.24692964553833, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6892930269241333, + "num_tokens": 124146795.0, + "step": 4796 + }, + { + "epoch": 0.5267955194377334, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.2868118286132812, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.707748293876648, + "num_tokens": 124170669.0, + "step": 4797 + }, + { + "epoch": 0.526905337140347, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.1837270259857178, + "learning_rate": 1e-06, + "loss": 1.009, + "mean_token_accuracy": 0.6888207197189331, + "num_tokens": 124195844.0, + "step": 4798 + }, + { + "epoch": 0.5270151548429607, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.188431739807129, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6851651668548584, + "num_tokens": 124223053.0, + "step": 4799 + }, + { + "epoch": 0.5271249725455743, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.17319655418396, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6980533003807068, + "num_tokens": 124248218.0, + "step": 4800 + }, + { + "epoch": 0.527234790248188, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.174670934677124, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7178660035133362, + "num_tokens": 124273859.0, + "step": 4801 + }, + { + "epoch": 0.5273446079508016, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.5202419757843018, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7160391807556152, + "num_tokens": 124294213.0, + "step": 4802 + }, + { + "epoch": 0.5274544256534154, + "ewc_loss": 2.2798776626586914e-06, + "grad_norm": 2.0623586177825928, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.6994931101799011, + "num_tokens": 124319793.0, + "step": 4803 + }, + { + "epoch": 0.527564243356029, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.132946729660034, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7078176140785217, + "num_tokens": 124346866.0, + "step": 4804 + }, + { + "epoch": 0.5276740610586427, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.424550771713257, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7062312960624695, + "num_tokens": 124368693.0, + "step": 4805 + }, + { + "epoch": 0.5277838787612563, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.518228054046631, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6864544153213501, + "num_tokens": 124388582.0, + "step": 4806 + }, + { + "epoch": 0.52789369646387, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.1652798652648926, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7150964736938477, + "num_tokens": 124413481.0, + "step": 4807 + }, + { + "epoch": 0.5280035141664836, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.3261966705322266, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7103553414344788, + "num_tokens": 124436784.0, + "step": 4808 + }, + { + "epoch": 0.5281133318690973, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.267402172088623, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6884480118751526, + "num_tokens": 124460433.0, + "step": 4809 + }, + { + "epoch": 0.528223149571711, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.2971127033233643, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6906954050064087, + "num_tokens": 124484380.0, + "step": 4810 + }, + { + "epoch": 0.5283329672743247, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 2.225867986679077, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7138658761978149, + "num_tokens": 124509622.0, + "step": 4811 + }, + { + "epoch": 0.5284427849769383, + "ewc_loss": 2.294778823852539e-06, + "grad_norm": 1.9610843658447266, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6900968551635742, + "num_tokens": 124540590.0, + "step": 4812 + }, + { + "epoch": 0.528552602679552, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 2.5874602794647217, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7092272043228149, + "num_tokens": 124562283.0, + "step": 4813 + }, + { + "epoch": 0.5286624203821656, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 2.0702905654907227, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6836316585540771, + "num_tokens": 124594488.0, + "step": 4814 + }, + { + "epoch": 0.5287722380847792, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 2.1943600177764893, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7137566804885864, + "num_tokens": 124621489.0, + "step": 4815 + }, + { + "epoch": 0.5288820557873929, + "ewc_loss": 2.3245811462402344e-06, + "grad_norm": 1.9747668504714966, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7407520413398743, + "num_tokens": 124649436.0, + "step": 4816 + }, + { + "epoch": 0.5289918734900065, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.3982229232788086, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7132772207260132, + "num_tokens": 124671012.0, + "step": 4817 + }, + { + "epoch": 0.5291016911926203, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.25630521774292, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6956959962844849, + "num_tokens": 124694958.0, + "step": 4818 + }, + { + "epoch": 0.5292115088952339, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.5190656185150146, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.7065693140029907, + "num_tokens": 124715344.0, + "step": 4819 + }, + { + "epoch": 0.5293213265978476, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.1356306076049805, + "learning_rate": 1e-06, + "loss": 1.1324, + "mean_token_accuracy": 0.6657326817512512, + "num_tokens": 124743751.0, + "step": 4820 + }, + { + "epoch": 0.5294311443004612, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.244777202606201, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7250757217407227, + "num_tokens": 124769264.0, + "step": 4821 + }, + { + "epoch": 0.5295409620030749, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.238921642303467, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6949328184127808, + "num_tokens": 124796623.0, + "step": 4822 + }, + { + "epoch": 0.5296507797056885, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.1375081539154053, + "learning_rate": 1e-06, + "loss": 1.0907, + "mean_token_accuracy": 0.6724094748497009, + "num_tokens": 124825236.0, + "step": 4823 + }, + { + "epoch": 0.5297605974083022, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.2471742630004883, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6946870684623718, + "num_tokens": 124849570.0, + "step": 4824 + }, + { + "epoch": 0.5298704151109159, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.1543760299682617, + "learning_rate": 1e-06, + "loss": 1.0526, + "mean_token_accuracy": 0.6876327991485596, + "num_tokens": 124876140.0, + "step": 4825 + }, + { + "epoch": 0.5299802328135296, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.084459066390991, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6908570528030396, + "num_tokens": 124902546.0, + "step": 4826 + }, + { + "epoch": 0.5300900505161432, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.2808868885040283, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.7031830549240112, + "num_tokens": 124926641.0, + "step": 4827 + }, + { + "epoch": 0.5301998682187569, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.0541534423828125, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7005904912948608, + "num_tokens": 124955451.0, + "step": 4828 + }, + { + "epoch": 0.5303096859213705, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.0462846755981445, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7143490314483643, + "num_tokens": 124983402.0, + "step": 4829 + }, + { + "epoch": 0.5304195036239842, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.224182367324829, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.7016631364822388, + "num_tokens": 125008433.0, + "step": 4830 + }, + { + "epoch": 0.5305293213265978, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.058950662612915, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7004203796386719, + "num_tokens": 125036151.0, + "step": 4831 + }, + { + "epoch": 0.5306391390292116, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.3983476161956787, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7214139103889465, + "num_tokens": 125058766.0, + "step": 4832 + }, + { + "epoch": 0.5307489567318252, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.143688440322876, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.714272677898407, + "num_tokens": 125084381.0, + "step": 4833 + }, + { + "epoch": 0.5308587744344389, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.172224283218384, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.687647819519043, + "num_tokens": 125108921.0, + "step": 4834 + }, + { + "epoch": 0.5309685921370525, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.003676176071167, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.7023241519927979, + "num_tokens": 125140054.0, + "step": 4835 + }, + { + "epoch": 0.5310784098396661, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.1093432903289795, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6864293217658997, + "num_tokens": 125167042.0, + "step": 4836 + }, + { + "epoch": 0.5311882275422798, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.0014209747314453, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7161579132080078, + "num_tokens": 125195924.0, + "step": 4837 + }, + { + "epoch": 0.5312980452448934, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.222201347351074, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.6983256936073303, + "num_tokens": 125221340.0, + "step": 4838 + }, + { + "epoch": 0.5314078629475072, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.3089358806610107, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6990793943405151, + "num_tokens": 125243894.0, + "step": 4839 + }, + { + "epoch": 0.5315176806501208, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.2548019886016846, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.711824893951416, + "num_tokens": 125268549.0, + "step": 4840 + }, + { + "epoch": 0.5316274983527345, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.142869472503662, + "learning_rate": 1e-06, + "loss": 1.0301, + "mean_token_accuracy": 0.6993497610092163, + "num_tokens": 125295550.0, + "step": 4841 + }, + { + "epoch": 0.5317373160553481, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.2059690952301025, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7016520500183105, + "num_tokens": 125320744.0, + "step": 4842 + }, + { + "epoch": 0.5318471337579618, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.1579437255859375, + "learning_rate": 1e-06, + "loss": 1.051, + "mean_token_accuracy": 0.6875044107437134, + "num_tokens": 125348066.0, + "step": 4843 + }, + { + "epoch": 0.5319569514605754, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 1.9245686531066895, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6954364776611328, + "num_tokens": 125378429.0, + "step": 4844 + }, + { + "epoch": 0.5320667691631891, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.1923105716705322, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6960076093673706, + "num_tokens": 125403980.0, + "step": 4845 + }, + { + "epoch": 0.5321765868658027, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 1.969617247581482, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.6921371221542358, + "num_tokens": 125434953.0, + "step": 4846 + }, + { + "epoch": 0.5322864045684165, + "ewc_loss": 2.3543834686279297e-06, + "grad_norm": 2.228055000305176, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7026776075363159, + "num_tokens": 125461614.0, + "step": 4847 + }, + { + "epoch": 0.5323962222710301, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 1.9924728870391846, + "learning_rate": 1e-06, + "loss": 1.0822, + "mean_token_accuracy": 0.6787105798721313, + "num_tokens": 125494269.0, + "step": 4848 + }, + { + "epoch": 0.5325060399736438, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.146367073059082, + "learning_rate": 1e-06, + "loss": 1.0814, + "mean_token_accuracy": 0.671654999256134, + "num_tokens": 125523086.0, + "step": 4849 + }, + { + "epoch": 0.5326158576762574, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.3341329097747803, + "learning_rate": 1e-06, + "loss": 1.0422, + "mean_token_accuracy": 0.6912687420845032, + "num_tokens": 125548116.0, + "step": 4850 + }, + { + "epoch": 0.532725675378871, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.14538836479187, + "learning_rate": 1e-06, + "loss": 1.08, + "mean_token_accuracy": 0.674770712852478, + "num_tokens": 125576092.0, + "step": 4851 + }, + { + "epoch": 0.5328354930814847, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.100346803665161, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7036716938018799, + "num_tokens": 125606678.0, + "step": 4852 + }, + { + "epoch": 0.5329453107840983, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.3589396476745605, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.706417977809906, + "num_tokens": 125628470.0, + "step": 4853 + }, + { + "epoch": 0.5330551284867121, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.466930627822876, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6932938098907471, + "num_tokens": 125649194.0, + "step": 4854 + }, + { + "epoch": 0.5331649461893258, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.0590696334838867, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6885759234428406, + "num_tokens": 125681786.0, + "step": 4855 + }, + { + "epoch": 0.5332747638919394, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.126006841659546, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7066351175308228, + "num_tokens": 125708871.0, + "step": 4856 + }, + { + "epoch": 0.533384581594553, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.1008195877075195, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7076889276504517, + "num_tokens": 125736407.0, + "step": 4857 + }, + { + "epoch": 0.5334943992971667, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.0673725605010986, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6962866187095642, + "num_tokens": 125764625.0, + "step": 4858 + }, + { + "epoch": 0.5336042169997803, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.3090829849243164, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6995594501495361, + "num_tokens": 125788811.0, + "step": 4859 + }, + { + "epoch": 0.533714034702394, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.050917863845825, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.6966491937637329, + "num_tokens": 125816450.0, + "step": 4860 + }, + { + "epoch": 0.5338238524050077, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.2163422107696533, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.7028907537460327, + "num_tokens": 125843742.0, + "step": 4861 + }, + { + "epoch": 0.5339336701076214, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.4411208629608154, + "learning_rate": 1e-06, + "loss": 1.0295, + "mean_token_accuracy": 0.6871181726455688, + "num_tokens": 125867806.0, + "step": 4862 + }, + { + "epoch": 0.534043487810235, + "ewc_loss": 2.3692846298217773e-06, + "grad_norm": 2.043607234954834, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.6926532983779907, + "num_tokens": 125898897.0, + "step": 4863 + }, + { + "epoch": 0.5341533055128487, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1197149753570557, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7139273881912231, + "num_tokens": 125926333.0, + "step": 4864 + }, + { + "epoch": 0.5342631232154623, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1790425777435303, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6818797588348389, + "num_tokens": 125955239.0, + "step": 4865 + }, + { + "epoch": 0.534372940918076, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9954205751419067, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7109785079956055, + "num_tokens": 125984817.0, + "step": 4866 + }, + { + "epoch": 0.5344827586206896, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.4601216316223145, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.7062261700630188, + "num_tokens": 126007118.0, + "step": 4867 + }, + { + "epoch": 0.5345925763233034, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.140130043029785, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7041108012199402, + "num_tokens": 126033833.0, + "step": 4868 + }, + { + "epoch": 0.534702394025917, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1802496910095215, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.6970668435096741, + "num_tokens": 126058969.0, + "step": 4869 + }, + { + "epoch": 0.5348122117285307, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1087889671325684, + "learning_rate": 1e-06, + "loss": 1.0451, + "mean_token_accuracy": 0.6822258234024048, + "num_tokens": 126087601.0, + "step": 4870 + }, + { + "epoch": 0.5349220294311443, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.5015883445739746, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7000598907470703, + "num_tokens": 126108863.0, + "step": 4871 + }, + { + "epoch": 0.535031847133758, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.0997564792633057, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7151391506195068, + "num_tokens": 126135520.0, + "step": 4872 + }, + { + "epoch": 0.5351416648363716, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2451984882354736, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7013844847679138, + "num_tokens": 126159142.0, + "step": 4873 + }, + { + "epoch": 0.5352514825389852, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.055776834487915, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7124688029289246, + "num_tokens": 126185981.0, + "step": 4874 + }, + { + "epoch": 0.5353613002415989, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2418441772460938, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6962193846702576, + "num_tokens": 126211810.0, + "step": 4875 + }, + { + "epoch": 0.5354711179442126, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3105082511901855, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7053852677345276, + "num_tokens": 126236974.0, + "step": 4876 + }, + { + "epoch": 0.5355809356468263, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3046257495880127, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7023659944534302, + "num_tokens": 126261310.0, + "step": 4877 + }, + { + "epoch": 0.5356907533494399, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.6447064876556396, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6943168640136719, + "num_tokens": 126281643.0, + "step": 4878 + }, + { + "epoch": 0.5358005710520536, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.064098596572876, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6896553635597229, + "num_tokens": 126312906.0, + "step": 4879 + }, + { + "epoch": 0.5359103887546672, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9575001001358032, + "learning_rate": 1e-06, + "loss": 1.1091, + "mean_token_accuracy": 0.6682941913604736, + "num_tokens": 126345375.0, + "step": 4880 + }, + { + "epoch": 0.5360202064572809, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.0339977741241455, + "learning_rate": 1e-06, + "loss": 1.0578, + "mean_token_accuracy": 0.6816021203994751, + "num_tokens": 126376856.0, + "step": 4881 + }, + { + "epoch": 0.5361300241598945, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2578985691070557, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6986111402511597, + "num_tokens": 126401481.0, + "step": 4882 + }, + { + "epoch": 0.5362398418625083, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.626098155975342, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6991274952888489, + "num_tokens": 126421653.0, + "step": 4883 + }, + { + "epoch": 0.5363496595651219, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2826406955718994, + "learning_rate": 1e-06, + "loss": 1.0257, + "mean_token_accuracy": 0.6905009746551514, + "num_tokens": 126446537.0, + "step": 4884 + }, + { + "epoch": 0.5364594772677356, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2233452796936035, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7286804914474487, + "num_tokens": 126472324.0, + "step": 4885 + }, + { + "epoch": 0.5365692949703492, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1468565464019775, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7272365093231201, + "num_tokens": 126498138.0, + "step": 4886 + }, + { + "epoch": 0.5366791126729629, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3087594509124756, + "learning_rate": 1e-06, + "loss": 1.0608, + "mean_token_accuracy": 0.6760133504867554, + "num_tokens": 126524152.0, + "step": 4887 + }, + { + "epoch": 0.5367889303755765, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.144517183303833, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7022374868392944, + "num_tokens": 126551253.0, + "step": 4888 + }, + { + "epoch": 0.5368987480781902, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9170392751693726, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6908057332038879, + "num_tokens": 126583556.0, + "step": 4889 + }, + { + "epoch": 0.5370085657808039, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.5560920238494873, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7166935205459595, + "num_tokens": 126602628.0, + "step": 4890 + }, + { + "epoch": 0.5371183834834176, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.021470069885254, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7100626826286316, + "num_tokens": 126630945.0, + "step": 4891 + }, + { + "epoch": 0.5372282011860312, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.0142483711242676, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.7042950987815857, + "num_tokens": 126660562.0, + "step": 4892 + }, + { + "epoch": 0.5373380188886449, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9659370183944702, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7032915949821472, + "num_tokens": 126691431.0, + "step": 4893 + }, + { + "epoch": 0.5374478365912585, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.0650317668914795, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7043351531028748, + "num_tokens": 126718942.0, + "step": 4894 + }, + { + "epoch": 0.5375576542938721, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.7627203464508057, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7291570901870728, + "num_tokens": 126737774.0, + "step": 4895 + }, + { + "epoch": 0.5376674719964858, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.262186288833618, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.6963059902191162, + "num_tokens": 126764676.0, + "step": 4896 + }, + { + "epoch": 0.5377772896990995, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3079071044921875, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.698602557182312, + "num_tokens": 126790046.0, + "step": 4897 + }, + { + "epoch": 0.5378871074017132, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.678941011428833, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7363739013671875, + "num_tokens": 126808339.0, + "step": 4898 + }, + { + "epoch": 0.5379969251043268, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2155728340148926, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7266901731491089, + "num_tokens": 126832548.0, + "step": 4899 + }, + { + "epoch": 0.5381067428069405, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9576239585876465, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6949924230575562, + "num_tokens": 126864910.0, + "step": 4900 + }, + { + "epoch": 0.5382165605095541, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.130204200744629, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.693679928779602, + "num_tokens": 126893270.0, + "step": 4901 + }, + { + "epoch": 0.5383263782121678, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1341803073883057, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6979130506515503, + "num_tokens": 126921753.0, + "step": 4902 + }, + { + "epoch": 0.5384361959147814, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9456160068511963, + "learning_rate": 1e-06, + "loss": 1.05, + "mean_token_accuracy": 0.6863971948623657, + "num_tokens": 126955918.0, + "step": 4903 + }, + { + "epoch": 0.5385460136173951, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.347081184387207, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7105447053909302, + "num_tokens": 126978163.0, + "step": 4904 + }, + { + "epoch": 0.5386558313200088, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3864028453826904, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.7060377597808838, + "num_tokens": 127001935.0, + "step": 4905 + }, + { + "epoch": 0.5387656490226225, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1808764934539795, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7046578526496887, + "num_tokens": 127026364.0, + "step": 4906 + }, + { + "epoch": 0.5388754667252361, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.3155949115753174, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6876125335693359, + "num_tokens": 127051022.0, + "step": 4907 + }, + { + "epoch": 0.5389852844278498, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.2860312461853027, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6968433856964111, + "num_tokens": 127077302.0, + "step": 4908 + }, + { + "epoch": 0.5390951021304634, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9788658618927002, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.6967077255249023, + "num_tokens": 127107594.0, + "step": 4909 + }, + { + "epoch": 0.5392049198330771, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.426189661026001, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7058238983154297, + "num_tokens": 127129814.0, + "step": 4910 + }, + { + "epoch": 0.5393147375356907, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.4565420150756836, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7156482934951782, + "num_tokens": 127148691.0, + "step": 4911 + }, + { + "epoch": 0.5394245552383045, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.308455467224121, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7351557016372681, + "num_tokens": 127168865.0, + "step": 4912 + }, + { + "epoch": 0.5395343729409181, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.0063915252685547, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6963698267936707, + "num_tokens": 127199139.0, + "step": 4913 + }, + { + "epoch": 0.5396441906435318, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.5154120922088623, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7096922397613525, + "num_tokens": 127219551.0, + "step": 4914 + }, + { + "epoch": 0.5397540083461454, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1439411640167236, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7073113918304443, + "num_tokens": 127247441.0, + "step": 4915 + }, + { + "epoch": 0.539863826048759, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 1.9043139219284058, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7086609601974487, + "num_tokens": 127280922.0, + "step": 4916 + }, + { + "epoch": 0.5399736437513727, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1694509983062744, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7233520746231079, + "num_tokens": 127303976.0, + "step": 4917 + }, + { + "epoch": 0.5400834614539863, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.1607613563537598, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7242294549942017, + "num_tokens": 127329285.0, + "step": 4918 + }, + { + "epoch": 0.5401932791566001, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.572880983352661, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7194610834121704, + "num_tokens": 127352261.0, + "step": 4919 + }, + { + "epoch": 0.5403030968592137, + "ewc_loss": 2.384185791015625e-06, + "grad_norm": 2.133039951324463, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.7028590440750122, + "num_tokens": 127380086.0, + "step": 4920 + }, + { + "epoch": 0.5404129145618274, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.4326090812683105, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7165545225143433, + "num_tokens": 127402010.0, + "step": 4921 + }, + { + "epoch": 0.540522732264441, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.214280605316162, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.6918110847473145, + "num_tokens": 127429070.0, + "step": 4922 + }, + { + "epoch": 0.5406325499670547, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.9938839673995972, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7113542556762695, + "num_tokens": 127458094.0, + "step": 4923 + }, + { + "epoch": 0.5407423676696683, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.046438217163086, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7025845050811768, + "num_tokens": 127487961.0, + "step": 4924 + }, + { + "epoch": 0.540852185372282, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.229733467102051, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7024276852607727, + "num_tokens": 127514020.0, + "step": 4925 + }, + { + "epoch": 0.5409620030748957, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.045070171356201, + "learning_rate": 1e-06, + "loss": 1.0138, + "mean_token_accuracy": 0.6908973455429077, + "num_tokens": 127542668.0, + "step": 4926 + }, + { + "epoch": 0.5410718207775094, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.531426429748535, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7048988938331604, + "num_tokens": 127562630.0, + "step": 4927 + }, + { + "epoch": 0.541181638480123, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.1049916744232178, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7041758298873901, + "num_tokens": 127592931.0, + "step": 4928 + }, + { + "epoch": 0.5412914561827367, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.09014630317688, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6927547454833984, + "num_tokens": 127621487.0, + "step": 4929 + }, + { + "epoch": 0.5414012738853503, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.9986952543258667, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7293304204940796, + "num_tokens": 127650196.0, + "step": 4930 + }, + { + "epoch": 0.541511091587964, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.340695381164551, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.717700719833374, + "num_tokens": 127673970.0, + "step": 4931 + }, + { + "epoch": 0.5416209092905776, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.0750131607055664, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6863559484481812, + "num_tokens": 127703486.0, + "step": 4932 + }, + { + "epoch": 0.5417307269931912, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.0624046325683594, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7055716514587402, + "num_tokens": 127731693.0, + "step": 4933 + }, + { + "epoch": 0.541840544695805, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.282480001449585, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.726266086101532, + "num_tokens": 127754534.0, + "step": 4934 + }, + { + "epoch": 0.5419503623984187, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.9839987754821777, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.6974495649337769, + "num_tokens": 127785708.0, + "step": 4935 + }, + { + "epoch": 0.5420601801010323, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.3878262042999268, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7254278063774109, + "num_tokens": 127807870.0, + "step": 4936 + }, + { + "epoch": 0.5421699978036459, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.3529021739959717, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6953422427177429, + "num_tokens": 127830538.0, + "step": 4937 + }, + { + "epoch": 0.5422798155062596, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.0110955238342285, + "learning_rate": 1e-06, + "loss": 1.0695, + "mean_token_accuracy": 0.676996111869812, + "num_tokens": 127861733.0, + "step": 4938 + }, + { + "epoch": 0.5423896332088732, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.2565629482269287, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.6884242296218872, + "num_tokens": 127886190.0, + "step": 4939 + }, + { + "epoch": 0.5424994509114869, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.161497116088867, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6965838074684143, + "num_tokens": 127914203.0, + "step": 4940 + }, + { + "epoch": 0.5426092686141006, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.0971152782440186, + "learning_rate": 1e-06, + "loss": 1.0582, + "mean_token_accuracy": 0.6835839748382568, + "num_tokens": 127942783.0, + "step": 4941 + }, + { + "epoch": 0.5427190863167143, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.272692918777466, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6973809003829956, + "num_tokens": 127967032.0, + "step": 4942 + }, + { + "epoch": 0.5428289040193279, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.4501640796661377, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7156707644462585, + "num_tokens": 127988188.0, + "step": 4943 + }, + { + "epoch": 0.5429387217219416, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.3895230293273926, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6903256177902222, + "num_tokens": 128012088.0, + "step": 4944 + }, + { + "epoch": 0.5430485394245552, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.233017683029175, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6986255049705505, + "num_tokens": 128037208.0, + "step": 4945 + }, + { + "epoch": 0.5431583571271689, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.1987483501434326, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7322638034820557, + "num_tokens": 128062799.0, + "step": 4946 + }, + { + "epoch": 0.5432681748297825, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.9865740537643433, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7152585387229919, + "num_tokens": 128093529.0, + "step": 4947 + }, + { + "epoch": 0.5433779925323963, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.9905420541763306, + "learning_rate": 1e-06, + "loss": 1.0529, + "mean_token_accuracy": 0.6791554689407349, + "num_tokens": 128123690.0, + "step": 4948 + }, + { + "epoch": 0.5434878102350099, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.286736011505127, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6930072903633118, + "num_tokens": 128148433.0, + "step": 4949 + }, + { + "epoch": 0.5435976279376236, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.2391395568847656, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6870810985565186, + "num_tokens": 128173520.0, + "step": 4950 + }, + { + "epoch": 0.5437074456402372, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.970367670059204, + "learning_rate": 1e-06, + "loss": 1.0364, + "mean_token_accuracy": 0.7008570432662964, + "num_tokens": 128205469.0, + "step": 4951 + }, + { + "epoch": 0.5438172633428509, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.2660810947418213, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7284085154533386, + "num_tokens": 128228251.0, + "step": 4952 + }, + { + "epoch": 0.5439270810454645, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.0555260181427, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7115243673324585, + "num_tokens": 128255630.0, + "step": 4953 + }, + { + "epoch": 0.5440368987480781, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.360004186630249, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7077347040176392, + "num_tokens": 128278625.0, + "step": 4954 + }, + { + "epoch": 0.5441467164506919, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.3656585216522217, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7165762186050415, + "num_tokens": 128299970.0, + "step": 4955 + }, + { + "epoch": 0.5442565341533055, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.1814675331115723, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.711065948009491, + "num_tokens": 128325057.0, + "step": 4956 + }, + { + "epoch": 0.5443663518559192, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.384753942489624, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.700486421585083, + "num_tokens": 128348579.0, + "step": 4957 + }, + { + "epoch": 0.5444761695585328, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.152430534362793, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7134442329406738, + "num_tokens": 128374047.0, + "step": 4958 + }, + { + "epoch": 0.5445859872611465, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.5988330841064453, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7058238983154297, + "num_tokens": 128394609.0, + "step": 4959 + }, + { + "epoch": 0.5446958049637601, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 2.1539833545684814, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.7004389762878418, + "num_tokens": 128419846.0, + "step": 4960 + }, + { + "epoch": 0.5448056226663738, + "ewc_loss": 2.3990869522094727e-06, + "grad_norm": 1.8655028343200684, + "learning_rate": 1e-06, + "loss": 1.0768, + "mean_token_accuracy": 0.6779041290283203, + "num_tokens": 128453276.0, + "step": 4961 + }, + { + "epoch": 0.5449154403689875, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.243579149246216, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6976011991500854, + "num_tokens": 128476588.0, + "step": 4962 + }, + { + "epoch": 0.5450252580716012, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.140202045440674, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7014594078063965, + "num_tokens": 128503964.0, + "step": 4963 + }, + { + "epoch": 0.5451350757742148, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.0673630237579346, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7173931002616882, + "num_tokens": 128531796.0, + "step": 4964 + }, + { + "epoch": 0.5452448934768285, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.2688236236572266, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.709518551826477, + "num_tokens": 128554987.0, + "step": 4965 + }, + { + "epoch": 0.5453547111794421, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.3734121322631836, + "learning_rate": 1e-06, + "loss": 1.0415, + "mean_token_accuracy": 0.6892080307006836, + "num_tokens": 128580111.0, + "step": 4966 + }, + { + "epoch": 0.5454645288820558, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 1.977094054222107, + "learning_rate": 1e-06, + "loss": 1.0756, + "mean_token_accuracy": 0.676986813545227, + "num_tokens": 128612332.0, + "step": 4967 + }, + { + "epoch": 0.5455743465846694, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.3515803813934326, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6892123222351074, + "num_tokens": 128636853.0, + "step": 4968 + }, + { + "epoch": 0.5456841642872831, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.0752856731414795, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7240471839904785, + "num_tokens": 128664436.0, + "step": 4969 + }, + { + "epoch": 0.5457939819898968, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.09089732170105, + "learning_rate": 1e-06, + "loss": 1.0598, + "mean_token_accuracy": 0.6850048899650574, + "num_tokens": 128694240.0, + "step": 4970 + }, + { + "epoch": 0.5459037996925105, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.3697657585144043, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6963617205619812, + "num_tokens": 128719182.0, + "step": 4971 + }, + { + "epoch": 0.5460136173951241, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.0707292556762695, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7018570303916931, + "num_tokens": 128746729.0, + "step": 4972 + }, + { + "epoch": 0.5461234350977378, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.2676620483398438, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6935807466506958, + "num_tokens": 128772278.0, + "step": 4973 + }, + { + "epoch": 0.5462332528003514, + "ewc_loss": 2.4139881134033203e-06, + "grad_norm": 2.215080738067627, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6907958984375, + "num_tokens": 128799048.0, + "step": 4974 + }, + { + "epoch": 0.546343070502965, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.1083850860595703, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7033958435058594, + "num_tokens": 128825845.0, + "step": 4975 + }, + { + "epoch": 0.5464528882055787, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.1971631050109863, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.706516683101654, + "num_tokens": 128854929.0, + "step": 4976 + }, + { + "epoch": 0.5465627059081924, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.3396573066711426, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7061023712158203, + "num_tokens": 128878255.0, + "step": 4977 + }, + { + "epoch": 0.5466725236108061, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 1.9949806928634644, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6916208267211914, + "num_tokens": 128910671.0, + "step": 4978 + }, + { + "epoch": 0.5467823413134197, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.1065924167633057, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.6931766271591187, + "num_tokens": 128941001.0, + "step": 4979 + }, + { + "epoch": 0.5468921590160334, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.233534574508667, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7198520302772522, + "num_tokens": 128965677.0, + "step": 4980 + }, + { + "epoch": 0.547001976718647, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.3259520530700684, + "learning_rate": 1e-06, + "loss": 1.0705, + "mean_token_accuracy": 0.6790528297424316, + "num_tokens": 128992259.0, + "step": 4981 + }, + { + "epoch": 0.5471117944212607, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 1.864931344985962, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7144853472709656, + "num_tokens": 129026500.0, + "step": 4982 + }, + { + "epoch": 0.5472216121238743, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.1269593238830566, + "learning_rate": 1e-06, + "loss": 1.081, + "mean_token_accuracy": 0.6772828698158264, + "num_tokens": 129055142.0, + "step": 4983 + }, + { + "epoch": 0.5473314298264881, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 1.932929515838623, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7064257264137268, + "num_tokens": 129087650.0, + "step": 4984 + }, + { + "epoch": 0.5474412475291017, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 1.945151925086975, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7007161378860474, + "num_tokens": 129117894.0, + "step": 4985 + }, + { + "epoch": 0.5475510652317154, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.143385648727417, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6987630724906921, + "num_tokens": 129144524.0, + "step": 4986 + }, + { + "epoch": 0.547660882934329, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.4318363666534424, + "learning_rate": 1e-06, + "loss": 1.0425, + "mean_token_accuracy": 0.692476749420166, + "num_tokens": 129167161.0, + "step": 4987 + }, + { + "epoch": 0.5477707006369427, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.2942187786102295, + "learning_rate": 1e-06, + "loss": 1.0895, + "mean_token_accuracy": 0.6785368919372559, + "num_tokens": 129192187.0, + "step": 4988 + }, + { + "epoch": 0.5478805183395563, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.0933103561401367, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.7015998959541321, + "num_tokens": 129220666.0, + "step": 4989 + }, + { + "epoch": 0.54799033604217, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.3527169227600098, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7094709873199463, + "num_tokens": 129242823.0, + "step": 4990 + }, + { + "epoch": 0.5481001537447837, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.320490598678589, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.6894992589950562, + "num_tokens": 129267928.0, + "step": 4991 + }, + { + "epoch": 0.5482099714473974, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.3559043407440186, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6890830993652344, + "num_tokens": 129292838.0, + "step": 4992 + }, + { + "epoch": 0.548319789150011, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.3701515197753906, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7095392942428589, + "num_tokens": 129314953.0, + "step": 4993 + }, + { + "epoch": 0.5484296068526247, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.2006821632385254, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7197531461715698, + "num_tokens": 129339660.0, + "step": 4994 + }, + { + "epoch": 0.5485394245552383, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.4013125896453857, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7144551277160645, + "num_tokens": 129363111.0, + "step": 4995 + }, + { + "epoch": 0.5486492422578519, + "ewc_loss": 2.428889274597168e-06, + "grad_norm": 2.2293004989624023, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6865700483322144, + "num_tokens": 129390318.0, + "step": 4996 + }, + { + "epoch": 0.5487590599604656, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.238917589187622, + "learning_rate": 1e-06, + "loss": 1.0389, + "mean_token_accuracy": 0.6941885948181152, + "num_tokens": 129416078.0, + "step": 4997 + }, + { + "epoch": 0.5488688776630792, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.621983766555786, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6903896331787109, + "num_tokens": 129437106.0, + "step": 4998 + }, + { + "epoch": 0.548978695365693, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 1.9650040864944458, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7041498422622681, + "num_tokens": 129469652.0, + "step": 4999 + }, + { + "epoch": 0.5490885130683066, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.1382625102996826, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7118077874183655, + "num_tokens": 129496588.0, + "step": 5000 + }, + { + "epoch": 0.5491983307709203, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.227689027786255, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7062453031539917, + "num_tokens": 129521502.0, + "step": 5001 + }, + { + "epoch": 0.5493081484735339, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.3010950088500977, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7209023237228394, + "num_tokens": 129544748.0, + "step": 5002 + }, + { + "epoch": 0.5494179661761476, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.0324950218200684, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7019672393798828, + "num_tokens": 129574117.0, + "step": 5003 + }, + { + "epoch": 0.5495277838787612, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.780195474624634, + "learning_rate": 1e-06, + "loss": 1.1055, + "mean_token_accuracy": 0.6858114004135132, + "num_tokens": 129595211.0, + "step": 5004 + }, + { + "epoch": 0.5496376015813749, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.3192086219787598, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7000707387924194, + "num_tokens": 129619967.0, + "step": 5005 + }, + { + "epoch": 0.5497474192839886, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.162386894226074, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7121405601501465, + "num_tokens": 129645294.0, + "step": 5006 + }, + { + "epoch": 0.5498572369866023, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.5338473320007324, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7041534185409546, + "num_tokens": 129664580.0, + "step": 5007 + }, + { + "epoch": 0.5499670546892159, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.110360622406006, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.6948383450508118, + "num_tokens": 129691874.0, + "step": 5008 + }, + { + "epoch": 0.5500768723918296, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.179002285003662, + "learning_rate": 1e-06, + "loss": 1.0584, + "mean_token_accuracy": 0.6851989030838013, + "num_tokens": 129719829.0, + "step": 5009 + }, + { + "epoch": 0.5501866900944432, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.1963396072387695, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7293673157691956, + "num_tokens": 129743967.0, + "step": 5010 + }, + { + "epoch": 0.5502965077970569, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.325795888900757, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.716953456401825, + "num_tokens": 129767756.0, + "step": 5011 + }, + { + "epoch": 0.5504063254996705, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.3221540451049805, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7084894180297852, + "num_tokens": 129790423.0, + "step": 5012 + }, + { + "epoch": 0.5505161432022843, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.278921127319336, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6928368806838989, + "num_tokens": 129815734.0, + "step": 5013 + }, + { + "epoch": 0.5506259609048979, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.0816490650177, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.713604748249054, + "num_tokens": 129842341.0, + "step": 5014 + }, + { + "epoch": 0.5507357786075116, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 1.9013782739639282, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.7028273940086365, + "num_tokens": 129873370.0, + "step": 5015 + }, + { + "epoch": 0.5508455963101252, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.3331568241119385, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7029513716697693, + "num_tokens": 129897220.0, + "step": 5016 + }, + { + "epoch": 0.5509554140127388, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.182899236679077, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7003453969955444, + "num_tokens": 129922265.0, + "step": 5017 + }, + { + "epoch": 0.5510652317153525, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.343609571456909, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.7038474082946777, + "num_tokens": 129946103.0, + "step": 5018 + }, + { + "epoch": 0.5511750494179661, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.149630546569824, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.699589729309082, + "num_tokens": 129973655.0, + "step": 5019 + }, + { + "epoch": 0.5512848671205799, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.202901840209961, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.72672438621521, + "num_tokens": 129998339.0, + "step": 5020 + }, + { + "epoch": 0.5513946848231935, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.014028787612915, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6906449794769287, + "num_tokens": 130031600.0, + "step": 5021 + }, + { + "epoch": 0.5515045025258072, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.2524161338806152, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6967098712921143, + "num_tokens": 130055600.0, + "step": 5022 + }, + { + "epoch": 0.5516143202284208, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.363905429840088, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.719422459602356, + "num_tokens": 130077860.0, + "step": 5023 + }, + { + "epoch": 0.5517241379310345, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.2647318840026855, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.7131586670875549, + "num_tokens": 130102686.0, + "step": 5024 + }, + { + "epoch": 0.5518339556336481, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 2.417996644973755, + "learning_rate": 1e-06, + "loss": 1.0363, + "mean_token_accuracy": 0.6916632652282715, + "num_tokens": 130125007.0, + "step": 5025 + }, + { + "epoch": 0.5519437733362618, + "ewc_loss": 2.4586915969848633e-06, + "grad_norm": 1.9566407203674316, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.7019705772399902, + "num_tokens": 130154385.0, + "step": 5026 + }, + { + "epoch": 0.5520535910388754, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.2379369735717773, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7180273532867432, + "num_tokens": 130179220.0, + "step": 5027 + }, + { + "epoch": 0.5521634087414892, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.296708106994629, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7030309438705444, + "num_tokens": 130204245.0, + "step": 5028 + }, + { + "epoch": 0.5522732264441028, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 1.922279953956604, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7206991910934448, + "num_tokens": 130237468.0, + "step": 5029 + }, + { + "epoch": 0.5523830441467165, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.259538173675537, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.69648677110672, + "num_tokens": 130261337.0, + "step": 5030 + }, + { + "epoch": 0.5524928618493301, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 1.9830033779144287, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6896867156028748, + "num_tokens": 130292001.0, + "step": 5031 + }, + { + "epoch": 0.5526026795519438, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.398784875869751, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6994234323501587, + "num_tokens": 130315471.0, + "step": 5032 + }, + { + "epoch": 0.5527124972545574, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.157341241836548, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6967560648918152, + "num_tokens": 130343018.0, + "step": 5033 + }, + { + "epoch": 0.552822314957171, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.2194411754608154, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6929260492324829, + "num_tokens": 130368875.0, + "step": 5034 + }, + { + "epoch": 0.5529321326597848, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.0581839084625244, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7000003457069397, + "num_tokens": 130396092.0, + "step": 5035 + }, + { + "epoch": 0.5530419503623984, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.3142077922821045, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6933398246765137, + "num_tokens": 130420468.0, + "step": 5036 + }, + { + "epoch": 0.5531517680650121, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.517015218734741, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7052222490310669, + "num_tokens": 130440923.0, + "step": 5037 + }, + { + "epoch": 0.5532615857676257, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.152174711227417, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7298434972763062, + "num_tokens": 130465065.0, + "step": 5038 + }, + { + "epoch": 0.5533714034702394, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.484283685684204, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6873797178268433, + "num_tokens": 130487004.0, + "step": 5039 + }, + { + "epoch": 0.553481221172853, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.2935216426849365, + "learning_rate": 1e-06, + "loss": 1.0529, + "mean_token_accuracy": 0.6775685548782349, + "num_tokens": 130514627.0, + "step": 5040 + }, + { + "epoch": 0.5535910388754667, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.4377734661102295, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.6889194250106812, + "num_tokens": 130536856.0, + "step": 5041 + }, + { + "epoch": 0.5537008565780804, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.1450042724609375, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7067230939865112, + "num_tokens": 130562912.0, + "step": 5042 + }, + { + "epoch": 0.5538106742806941, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.113409996032715, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7063241004943848, + "num_tokens": 130589178.0, + "step": 5043 + }, + { + "epoch": 0.5539204919833077, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.1222846508026123, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7016834020614624, + "num_tokens": 130617557.0, + "step": 5044 + }, + { + "epoch": 0.5540303096859214, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.212820529937744, + "learning_rate": 1e-06, + "loss": 1.0503, + "mean_token_accuracy": 0.6826318502426147, + "num_tokens": 130645531.0, + "step": 5045 + }, + { + "epoch": 0.554140127388535, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.548609972000122, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7049046754837036, + "num_tokens": 130666322.0, + "step": 5046 + }, + { + "epoch": 0.5542499450911487, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.354344129562378, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7163673639297485, + "num_tokens": 130689313.0, + "step": 5047 + }, + { + "epoch": 0.5543597627937623, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.061246156692505, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7065917253494263, + "num_tokens": 130717803.0, + "step": 5048 + }, + { + "epoch": 0.5544695804963761, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 1.9725826978683472, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.6929495334625244, + "num_tokens": 130747161.0, + "step": 5049 + }, + { + "epoch": 0.5545793981989897, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.418684720993042, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6990262269973755, + "num_tokens": 130768925.0, + "step": 5050 + }, + { + "epoch": 0.5546892159016034, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.316075563430786, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7049146294593811, + "num_tokens": 130792964.0, + "step": 5051 + }, + { + "epoch": 0.554799033604217, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.525341033935547, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7183217406272888, + "num_tokens": 130813028.0, + "step": 5052 + }, + { + "epoch": 0.5549088513068307, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.1716320514678955, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7033053040504456, + "num_tokens": 130838883.0, + "step": 5053 + }, + { + "epoch": 0.5550186690094443, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.1470611095428467, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6953810453414917, + "num_tokens": 130866937.0, + "step": 5054 + }, + { + "epoch": 0.555128486712058, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.2418651580810547, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6895183324813843, + "num_tokens": 130891268.0, + "step": 5055 + }, + { + "epoch": 0.5552383044146716, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.158763885498047, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6974624991416931, + "num_tokens": 130920607.0, + "step": 5056 + }, + { + "epoch": 0.5553481221172853, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.184743881225586, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.7034624814987183, + "num_tokens": 130947053.0, + "step": 5057 + }, + { + "epoch": 0.555457939819899, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.575986623764038, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6935470104217529, + "num_tokens": 130968392.0, + "step": 5058 + }, + { + "epoch": 0.5555677575225126, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.3827970027923584, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7287542223930359, + "num_tokens": 130990351.0, + "step": 5059 + }, + { + "epoch": 0.5556775752251263, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.152634859085083, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7092543840408325, + "num_tokens": 131016161.0, + "step": 5060 + }, + { + "epoch": 0.5557873929277399, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.4415194988250732, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7098684310913086, + "num_tokens": 131037117.0, + "step": 5061 + }, + { + "epoch": 0.5558972106303536, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.363368511199951, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6988633275032043, + "num_tokens": 131060013.0, + "step": 5062 + }, + { + "epoch": 0.5560070283329672, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.368171453475952, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6957962512969971, + "num_tokens": 131084337.0, + "step": 5063 + }, + { + "epoch": 0.556116846035581, + "ewc_loss": 2.473592758178711e-06, + "grad_norm": 2.051285743713379, + "learning_rate": 1e-06, + "loss": 1.1057, + "mean_token_accuracy": 0.6710140705108643, + "num_tokens": 131116657.0, + "step": 5064 + }, + { + "epoch": 0.5562266637381946, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.2756848335266113, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7060832381248474, + "num_tokens": 131140951.0, + "step": 5065 + }, + { + "epoch": 0.5563364814408083, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.1890807151794434, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7016569972038269, + "num_tokens": 131166105.0, + "step": 5066 + }, + { + "epoch": 0.5564462991434219, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.078842878341675, + "learning_rate": 1e-06, + "loss": 1.0229, + "mean_token_accuracy": 0.693670392036438, + "num_tokens": 131197081.0, + "step": 5067 + }, + { + "epoch": 0.5565561168460356, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.455000400543213, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7274539470672607, + "num_tokens": 131218222.0, + "step": 5068 + }, + { + "epoch": 0.5566659345486492, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.2272825241088867, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6950826644897461, + "num_tokens": 131242354.0, + "step": 5069 + }, + { + "epoch": 0.5567757522512629, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.1920063495635986, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7070148587226868, + "num_tokens": 131267390.0, + "step": 5070 + }, + { + "epoch": 0.5568855699538766, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.10681414604187, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7032158374786377, + "num_tokens": 131295486.0, + "step": 5071 + }, + { + "epoch": 0.5569953876564903, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.359654664993286, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7060160636901855, + "num_tokens": 131317985.0, + "step": 5072 + }, + { + "epoch": 0.5571052053591039, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.6820714473724365, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7212262153625488, + "num_tokens": 131336640.0, + "step": 5073 + }, + { + "epoch": 0.5572150230617176, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.3213608264923096, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6904940009117126, + "num_tokens": 131361817.0, + "step": 5074 + }, + { + "epoch": 0.5573248407643312, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.101519823074341, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7299569845199585, + "num_tokens": 131388140.0, + "step": 5075 + }, + { + "epoch": 0.5574346584669448, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.1637122631073, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6843645572662354, + "num_tokens": 131414579.0, + "step": 5076 + }, + { + "epoch": 0.5575444761695585, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 1.9736106395721436, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7194415926933289, + "num_tokens": 131445760.0, + "step": 5077 + }, + { + "epoch": 0.5576542938721722, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.130614757537842, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6989598870277405, + "num_tokens": 131472746.0, + "step": 5078 + }, + { + "epoch": 0.5577641115747859, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.2871482372283936, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7150155305862427, + "num_tokens": 131494842.0, + "step": 5079 + }, + { + "epoch": 0.5578739292773995, + "ewc_loss": 2.5033950805664062e-06, + "grad_norm": 2.4035403728485107, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7012050151824951, + "num_tokens": 131517296.0, + "step": 5080 + }, + { + "epoch": 0.5579837469800132, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.0457091331481934, + "learning_rate": 1e-06, + "loss": 1.0541, + "mean_token_accuracy": 0.6813795566558838, + "num_tokens": 131546387.0, + "step": 5081 + }, + { + "epoch": 0.5580935646826268, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.655778169631958, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7119234800338745, + "num_tokens": 131564509.0, + "step": 5082 + }, + { + "epoch": 0.5582033823852405, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.3986804485321045, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.6996920108795166, + "num_tokens": 131588213.0, + "step": 5083 + }, + { + "epoch": 0.5583132000878541, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1384670734405518, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6939562559127808, + "num_tokens": 131616361.0, + "step": 5084 + }, + { + "epoch": 0.5584230177904678, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.050624132156372, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6907800436019897, + "num_tokens": 131646562.0, + "step": 5085 + }, + { + "epoch": 0.5585328354930815, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2359211444854736, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7105367183685303, + "num_tokens": 131670099.0, + "step": 5086 + }, + { + "epoch": 0.5586426531956952, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.093156576156616, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6915498375892639, + "num_tokens": 131697822.0, + "step": 5087 + }, + { + "epoch": 0.5587524708983088, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.246760368347168, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7124844789505005, + "num_tokens": 131721692.0, + "step": 5088 + }, + { + "epoch": 0.5588622886009225, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.463139772415161, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7065814137458801, + "num_tokens": 131743548.0, + "step": 5089 + }, + { + "epoch": 0.5589721063035361, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2431375980377197, + "learning_rate": 1e-06, + "loss": 1.135, + "mean_token_accuracy": 0.6674596667289734, + "num_tokens": 131769790.0, + "step": 5090 + }, + { + "epoch": 0.5590819240061498, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.0220398902893066, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.695425808429718, + "num_tokens": 131799241.0, + "step": 5091 + }, + { + "epoch": 0.5591917417087634, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 1.7521218061447144, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6880489587783813, + "num_tokens": 131839024.0, + "step": 5092 + }, + { + "epoch": 0.5593015594113772, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1077353954315186, + "learning_rate": 1e-06, + "loss": 1.0427, + "mean_token_accuracy": 0.6833795309066772, + "num_tokens": 131866780.0, + "step": 5093 + }, + { + "epoch": 0.5594113771139908, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.134875774383545, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.6981077790260315, + "num_tokens": 131894181.0, + "step": 5094 + }, + { + "epoch": 0.5595211948166045, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.6938111782073975, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6938880085945129, + "num_tokens": 131914061.0, + "step": 5095 + }, + { + "epoch": 0.5596310125192181, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.09336519241333, + "learning_rate": 1e-06, + "loss": 1.0926, + "mean_token_accuracy": 0.6755690574645996, + "num_tokens": 131944439.0, + "step": 5096 + }, + { + "epoch": 0.5597408302218317, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1487624645233154, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7247267961502075, + "num_tokens": 131969212.0, + "step": 5097 + }, + { + "epoch": 0.5598506479244454, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.4356470108032227, + "learning_rate": 1e-06, + "loss": 1.1421, + "mean_token_accuracy": 0.6582849025726318, + "num_tokens": 131998108.0, + "step": 5098 + }, + { + "epoch": 0.559960465627059, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 1.922438383102417, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.677790641784668, + "num_tokens": 132031655.0, + "step": 5099 + }, + { + "epoch": 0.5600702833296728, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.480109691619873, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7155280113220215, + "num_tokens": 132051282.0, + "step": 5100 + }, + { + "epoch": 0.5601801010322864, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1415786743164062, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6970845460891724, + "num_tokens": 132075903.0, + "step": 5101 + }, + { + "epoch": 0.5602899187349001, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1876866817474365, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6977026462554932, + "num_tokens": 132101575.0, + "step": 5102 + }, + { + "epoch": 0.5603997364375137, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1883275508880615, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7005568146705627, + "num_tokens": 132127711.0, + "step": 5103 + }, + { + "epoch": 0.5605095541401274, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2923202514648438, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7168837189674377, + "num_tokens": 132150607.0, + "step": 5104 + }, + { + "epoch": 0.560619371842741, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.228745222091675, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.7000918388366699, + "num_tokens": 132174942.0, + "step": 5105 + }, + { + "epoch": 0.5607291895453547, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.3022079467773438, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7194600105285645, + "num_tokens": 132197226.0, + "step": 5106 + }, + { + "epoch": 0.5608390072479684, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1237590312957764, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7184227705001831, + "num_tokens": 132224700.0, + "step": 5107 + }, + { + "epoch": 0.5609488249505821, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.26596736907959, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6904853582382202, + "num_tokens": 132249511.0, + "step": 5108 + }, + { + "epoch": 0.5610586426531957, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2124369144439697, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6870192289352417, + "num_tokens": 132276127.0, + "step": 5109 + }, + { + "epoch": 0.5611684603558094, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.0108742713928223, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6762310266494751, + "num_tokens": 132306672.0, + "step": 5110 + }, + { + "epoch": 0.561278278058423, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1685843467712402, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.7031622529029846, + "num_tokens": 132333910.0, + "step": 5111 + }, + { + "epoch": 0.5613880957610367, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.123722791671753, + "learning_rate": 1e-06, + "loss": 1.0683, + "mean_token_accuracy": 0.68686842918396, + "num_tokens": 132363479.0, + "step": 5112 + }, + { + "epoch": 0.5614979134636503, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.1520538330078125, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7052634358406067, + "num_tokens": 132390880.0, + "step": 5113 + }, + { + "epoch": 0.5616077311662641, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2469494342803955, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7041729092597961, + "num_tokens": 132416496.0, + "step": 5114 + }, + { + "epoch": 0.5617175488688777, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.220216751098633, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7104719877243042, + "num_tokens": 132441545.0, + "step": 5115 + }, + { + "epoch": 0.5618273665714913, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.118128538131714, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7196124792098999, + "num_tokens": 132466047.0, + "step": 5116 + }, + { + "epoch": 0.561937184274105, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.2831997871398926, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.722750186920166, + "num_tokens": 132487881.0, + "step": 5117 + }, + { + "epoch": 0.5620470019767186, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.178406238555908, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7186959385871887, + "num_tokens": 132515071.0, + "step": 5118 + }, + { + "epoch": 0.5621568196793323, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.3350305557250977, + "learning_rate": 1e-06, + "loss": 1.0501, + "mean_token_accuracy": 0.6823960542678833, + "num_tokens": 132539234.0, + "step": 5119 + }, + { + "epoch": 0.5622666373819459, + "ewc_loss": 2.518296241760254e-06, + "grad_norm": 2.133727788925171, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6887580752372742, + "num_tokens": 132569368.0, + "step": 5120 + }, + { + "epoch": 0.5623764550845596, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0818772315979004, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7045493721961975, + "num_tokens": 132595696.0, + "step": 5121 + }, + { + "epoch": 0.5624862727871733, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.9725719690322876, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.69315105676651, + "num_tokens": 132627634.0, + "step": 5122 + }, + { + "epoch": 0.562596090489787, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.222562551498413, + "learning_rate": 1e-06, + "loss": 0.9908, + "mean_token_accuracy": 0.6994161605834961, + "num_tokens": 132652802.0, + "step": 5123 + }, + { + "epoch": 0.5627059081924006, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1358344554901123, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7263868451118469, + "num_tokens": 132677330.0, + "step": 5124 + }, + { + "epoch": 0.5628157258950143, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.108222246170044, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.7040497064590454, + "num_tokens": 132703704.0, + "step": 5125 + }, + { + "epoch": 0.5629255435976279, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.5359702110290527, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7130734920501709, + "num_tokens": 132731686.0, + "step": 5126 + }, + { + "epoch": 0.5630353613002416, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0671005249023438, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7127070426940918, + "num_tokens": 132761771.0, + "step": 5127 + }, + { + "epoch": 0.5631451790028552, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.2222464084625244, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6929128766059875, + "num_tokens": 132786439.0, + "step": 5128 + }, + { + "epoch": 0.563254996705469, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1525216102600098, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.726039707660675, + "num_tokens": 132811005.0, + "step": 5129 + }, + { + "epoch": 0.5633648144080826, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1697750091552734, + "learning_rate": 1e-06, + "loss": 1.1087, + "mean_token_accuracy": 0.671410083770752, + "num_tokens": 132839399.0, + "step": 5130 + }, + { + "epoch": 0.5634746321106963, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.031672239303589, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.6893376111984253, + "num_tokens": 132867618.0, + "step": 5131 + }, + { + "epoch": 0.5635844498133099, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.9134091138839722, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.6895818114280701, + "num_tokens": 132901567.0, + "step": 5132 + }, + { + "epoch": 0.5636942675159236, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.2639694213867188, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7157709002494812, + "num_tokens": 132925044.0, + "step": 5133 + }, + { + "epoch": 0.5638040852185372, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.526275157928467, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7234009504318237, + "num_tokens": 132944209.0, + "step": 5134 + }, + { + "epoch": 0.5639139029211508, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0217506885528564, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.7017963528633118, + "num_tokens": 132973354.0, + "step": 5135 + }, + { + "epoch": 0.5640237206237646, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.515040159225464, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7325447797775269, + "num_tokens": 132992898.0, + "step": 5136 + }, + { + "epoch": 0.5641335383263782, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.501392126083374, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.7002291679382324, + "num_tokens": 133013935.0, + "step": 5137 + }, + { + "epoch": 0.5642433560289919, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.758798599243164, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7254591584205627, + "num_tokens": 133033719.0, + "step": 5138 + }, + { + "epoch": 0.5643531737316055, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0587944984436035, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.6977183818817139, + "num_tokens": 133062200.0, + "step": 5139 + }, + { + "epoch": 0.5644629914342192, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0947353839874268, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.697317361831665, + "num_tokens": 133090360.0, + "step": 5140 + }, + { + "epoch": 0.5645728091368328, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1896114349365234, + "learning_rate": 1e-06, + "loss": 1.0442, + "mean_token_accuracy": 0.6932919025421143, + "num_tokens": 133117850.0, + "step": 5141 + }, + { + "epoch": 0.5646826268394465, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.2373313903808594, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7149505615234375, + "num_tokens": 133142324.0, + "step": 5142 + }, + { + "epoch": 0.5647924445420602, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.4188342094421387, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7422621250152588, + "num_tokens": 133163820.0, + "step": 5143 + }, + { + "epoch": 0.5649022622446739, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.147076368331909, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7148170471191406, + "num_tokens": 133190171.0, + "step": 5144 + }, + { + "epoch": 0.5650120799472875, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.39363169670105, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.715833306312561, + "num_tokens": 133211488.0, + "step": 5145 + }, + { + "epoch": 0.5651218976499012, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.486511707305908, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7058674097061157, + "num_tokens": 133232643.0, + "step": 5146 + }, + { + "epoch": 0.5652317153525148, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.6662697792053223, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7115578651428223, + "num_tokens": 133250041.0, + "step": 5147 + }, + { + "epoch": 0.5653415330551285, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.2493479251861572, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6925886273384094, + "num_tokens": 133274638.0, + "step": 5148 + }, + { + "epoch": 0.5654513507577421, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1382601261138916, + "learning_rate": 1e-06, + "loss": 1.0519, + "mean_token_accuracy": 0.6914935111999512, + "num_tokens": 133307670.0, + "step": 5149 + }, + { + "epoch": 0.5655611684603558, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.15429425239563, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.70448237657547, + "num_tokens": 133332989.0, + "step": 5150 + }, + { + "epoch": 0.5656709861629695, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.076294183731079, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7080482840538025, + "num_tokens": 133360939.0, + "step": 5151 + }, + { + "epoch": 0.5657808038655832, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0914323329925537, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7118646502494812, + "num_tokens": 133388730.0, + "step": 5152 + }, + { + "epoch": 0.5658906215681968, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.061840772628784, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7008875608444214, + "num_tokens": 133416971.0, + "step": 5153 + }, + { + "epoch": 0.5660004392708105, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1164069175720215, + "learning_rate": 1e-06, + "loss": 1.0412, + "mean_token_accuracy": 0.6890527009963989, + "num_tokens": 133444888.0, + "step": 5154 + }, + { + "epoch": 0.5661102569734241, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.184385061264038, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6831108927726746, + "num_tokens": 133471556.0, + "step": 5155 + }, + { + "epoch": 0.5662200746760377, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.9671928882598877, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6953707337379456, + "num_tokens": 133504244.0, + "step": 5156 + }, + { + "epoch": 0.5663298923786514, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.001885175704956, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6856808662414551, + "num_tokens": 133533015.0, + "step": 5157 + }, + { + "epoch": 0.5664397100812651, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.5582165718078613, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6944651007652283, + "num_tokens": 133553724.0, + "step": 5158 + }, + { + "epoch": 0.5665495277838788, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.9264957904815674, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6869219541549683, + "num_tokens": 133586088.0, + "step": 5159 + }, + { + "epoch": 0.5666593454864924, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.2516367435455322, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6978106498718262, + "num_tokens": 133610842.0, + "step": 5160 + }, + { + "epoch": 0.5667691631891061, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.3822109699249268, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.7004562616348267, + "num_tokens": 133632718.0, + "step": 5161 + }, + { + "epoch": 0.5668789808917197, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.1378138065338135, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6954766511917114, + "num_tokens": 133657697.0, + "step": 5162 + }, + { + "epoch": 0.5669887985943334, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.220433473587036, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.7016494274139404, + "num_tokens": 133680641.0, + "step": 5163 + }, + { + "epoch": 0.567098616296947, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0379104614257812, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.722706139087677, + "num_tokens": 133707898.0, + "step": 5164 + }, + { + "epoch": 0.5672084339995608, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.3869011402130127, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.7040219902992249, + "num_tokens": 133729956.0, + "step": 5165 + }, + { + "epoch": 0.5673182517021744, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.0232486724853516, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7331119775772095, + "num_tokens": 133756836.0, + "step": 5166 + }, + { + "epoch": 0.5674280694047881, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.3527331352233887, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7195538282394409, + "num_tokens": 133778886.0, + "step": 5167 + }, + { + "epoch": 0.5675378871074017, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 2.104465961456299, + "learning_rate": 1e-06, + "loss": 1.0529, + "mean_token_accuracy": 0.6837100982666016, + "num_tokens": 133809059.0, + "step": 5168 + }, + { + "epoch": 0.5676477048100154, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.908521056175232, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7053576707839966, + "num_tokens": 133839123.0, + "step": 5169 + }, + { + "epoch": 0.567757522512629, + "ewc_loss": 2.5331974029541016e-06, + "grad_norm": 1.9943428039550781, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7353572845458984, + "num_tokens": 133866533.0, + "step": 5170 + }, + { + "epoch": 0.5678673402152427, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.0343563556671143, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.6887420415878296, + "num_tokens": 133897919.0, + "step": 5171 + }, + { + "epoch": 0.5679771579178564, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 1.9744679927825928, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6959866285324097, + "num_tokens": 133928244.0, + "step": 5172 + }, + { + "epoch": 0.5680869756204701, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.110347032546997, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.6918675899505615, + "num_tokens": 133956029.0, + "step": 5173 + }, + { + "epoch": 0.5681967933230837, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.0818450450897217, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.6833081841468811, + "num_tokens": 133986082.0, + "step": 5174 + }, + { + "epoch": 0.5683066110256974, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.237706184387207, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6975237131118774, + "num_tokens": 134011899.0, + "step": 5175 + }, + { + "epoch": 0.568416428728311, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.4380836486816406, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7067536115646362, + "num_tokens": 134032637.0, + "step": 5176 + }, + { + "epoch": 0.5685262464309246, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.2761144638061523, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7071604132652283, + "num_tokens": 134056987.0, + "step": 5177 + }, + { + "epoch": 0.5686360641335383, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.4170453548431396, + "learning_rate": 1e-06, + "loss": 1.0254, + "mean_token_accuracy": 0.6910302639007568, + "num_tokens": 134081097.0, + "step": 5178 + }, + { + "epoch": 0.5687458818361519, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.3599398136138916, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.7093605995178223, + "num_tokens": 134104706.0, + "step": 5179 + }, + { + "epoch": 0.5688556995387657, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.1748602390289307, + "learning_rate": 1e-06, + "loss": 1.0205, + "mean_token_accuracy": 0.6854572296142578, + "num_tokens": 134131050.0, + "step": 5180 + }, + { + "epoch": 0.5689655172413793, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.2292556762695312, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6934112310409546, + "num_tokens": 134157262.0, + "step": 5181 + }, + { + "epoch": 0.569075334943993, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.0194883346557617, + "learning_rate": 1e-06, + "loss": 1.1322, + "mean_token_accuracy": 0.66797935962677, + "num_tokens": 134188743.0, + "step": 5182 + }, + { + "epoch": 0.5691851526466066, + "ewc_loss": 2.5480985641479492e-06, + "grad_norm": 2.1319615840911865, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.704247772693634, + "num_tokens": 134215418.0, + "step": 5183 + }, + { + "epoch": 0.5692949703492203, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.8274586200714111, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7090063691139221, + "num_tokens": 134248000.0, + "step": 5184 + }, + { + "epoch": 0.5694047880518339, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3692545890808105, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7119089961051941, + "num_tokens": 134271086.0, + "step": 5185 + }, + { + "epoch": 0.5695146057544476, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.47338604927063, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7155898809432983, + "num_tokens": 134293514.0, + "step": 5186 + }, + { + "epoch": 0.5696244234570613, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1102468967437744, + "learning_rate": 1e-06, + "loss": 1.0486, + "mean_token_accuracy": 0.6862199902534485, + "num_tokens": 134324497.0, + "step": 5187 + }, + { + "epoch": 0.569734241159675, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.2791664600372314, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7082290649414062, + "num_tokens": 134348726.0, + "step": 5188 + }, + { + "epoch": 0.5698440588622886, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.133838653564453, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7165699005126953, + "num_tokens": 134375115.0, + "step": 5189 + }, + { + "epoch": 0.5699538765649023, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3881685733795166, + "learning_rate": 1e-06, + "loss": 1.0681, + "mean_token_accuracy": 0.6840380430221558, + "num_tokens": 134397922.0, + "step": 5190 + }, + { + "epoch": 0.5700636942675159, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1320459842681885, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7219514846801758, + "num_tokens": 134425011.0, + "step": 5191 + }, + { + "epoch": 0.5701735119701296, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.32204270362854, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.6947442889213562, + "num_tokens": 134452007.0, + "step": 5192 + }, + { + "epoch": 0.5702833296727432, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.094886302947998, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.7007637023925781, + "num_tokens": 134480460.0, + "step": 5193 + }, + { + "epoch": 0.570393147375357, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.227414608001709, + "learning_rate": 1e-06, + "loss": 1.0871, + "mean_token_accuracy": 0.6770938634872437, + "num_tokens": 134507478.0, + "step": 5194 + }, + { + "epoch": 0.5705029650779706, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.314908266067505, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6891523599624634, + "num_tokens": 134531741.0, + "step": 5195 + }, + { + "epoch": 0.5706127827805842, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1003479957580566, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7088798880577087, + "num_tokens": 134557636.0, + "step": 5196 + }, + { + "epoch": 0.5707226004831979, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3145642280578613, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7028837203979492, + "num_tokens": 134579402.0, + "step": 5197 + }, + { + "epoch": 0.5708324181858115, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.126112699508667, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7036468982696533, + "num_tokens": 134606547.0, + "step": 5198 + }, + { + "epoch": 0.5709422358884252, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.304372549057007, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7195314168930054, + "num_tokens": 134630304.0, + "step": 5199 + }, + { + "epoch": 0.5710520535910388, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.286670684814453, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7011212706565857, + "num_tokens": 134654483.0, + "step": 5200 + }, + { + "epoch": 0.5711618712936526, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.9666101932525635, + "learning_rate": 1e-06, + "loss": 1.0662, + "mean_token_accuracy": 0.6803947687149048, + "num_tokens": 134687064.0, + "step": 5201 + }, + { + "epoch": 0.5712716889962662, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.2278261184692383, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7081900835037231, + "num_tokens": 134713343.0, + "step": 5202 + }, + { + "epoch": 0.5713815066988799, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.4138143062591553, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6873661279678345, + "num_tokens": 134736151.0, + "step": 5203 + }, + { + "epoch": 0.5714913244014935, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.291297435760498, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7186115384101868, + "num_tokens": 134760650.0, + "step": 5204 + }, + { + "epoch": 0.5716011421041072, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.5763819217681885, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7099243402481079, + "num_tokens": 134780341.0, + "step": 5205 + }, + { + "epoch": 0.5717109598067208, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1949028968811035, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7012451887130737, + "num_tokens": 134804164.0, + "step": 5206 + }, + { + "epoch": 0.5718207775093345, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.155362367630005, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7022159099578857, + "num_tokens": 134830883.0, + "step": 5207 + }, + { + "epoch": 0.5719305952119481, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1946563720703125, + "learning_rate": 1e-06, + "loss": 1.0563, + "mean_token_accuracy": 0.684403657913208, + "num_tokens": 134857646.0, + "step": 5208 + }, + { + "epoch": 0.5720404129145619, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.050933361053467, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7063279151916504, + "num_tokens": 134886570.0, + "step": 5209 + }, + { + "epoch": 0.5721502306171755, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.336087226867676, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7186587452888489, + "num_tokens": 134909165.0, + "step": 5210 + }, + { + "epoch": 0.5722600483197892, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1643993854522705, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.7038484215736389, + "num_tokens": 134935723.0, + "step": 5211 + }, + { + "epoch": 0.5723698660224028, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.652257204055786, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.71120685338974, + "num_tokens": 134952760.0, + "step": 5212 + }, + { + "epoch": 0.5724796837250165, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.157696008682251, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7040925621986389, + "num_tokens": 134978712.0, + "step": 5213 + }, + { + "epoch": 0.5725895014276301, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1232752799987793, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6833348274230957, + "num_tokens": 135007948.0, + "step": 5214 + }, + { + "epoch": 0.5726993191302437, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.9426010847091675, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6883431673049927, + "num_tokens": 135039118.0, + "step": 5215 + }, + { + "epoch": 0.5728091368328575, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.8540595769882202, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7049249410629272, + "num_tokens": 135076922.0, + "step": 5216 + }, + { + "epoch": 0.5729189545354711, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.018744707107544, + "learning_rate": 1e-06, + "loss": 1.0623, + "mean_token_accuracy": 0.6858124732971191, + "num_tokens": 135106676.0, + "step": 5217 + }, + { + "epoch": 0.5730287722380848, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3665781021118164, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7173606157302856, + "num_tokens": 135127502.0, + "step": 5218 + }, + { + "epoch": 0.5731385899406984, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.196791648864746, + "learning_rate": 1e-06, + "loss": 1.0603, + "mean_token_accuracy": 0.6821378469467163, + "num_tokens": 135152895.0, + "step": 5219 + }, + { + "epoch": 0.5732484076433121, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.9009195566177368, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7068459987640381, + "num_tokens": 135185791.0, + "step": 5220 + }, + { + "epoch": 0.5733582253459257, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.957503318786621, + "learning_rate": 1e-06, + "loss": 1.0416, + "mean_token_accuracy": 0.6905120611190796, + "num_tokens": 135216237.0, + "step": 5221 + }, + { + "epoch": 0.5734680430485394, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.171283006668091, + "learning_rate": 1e-06, + "loss": 1.0693, + "mean_token_accuracy": 0.6819339990615845, + "num_tokens": 135241407.0, + "step": 5222 + }, + { + "epoch": 0.5735778607511531, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.4425649642944336, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7207201719284058, + "num_tokens": 135262741.0, + "step": 5223 + }, + { + "epoch": 0.5736876784537668, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.148852825164795, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7312201261520386, + "num_tokens": 135285964.0, + "step": 5224 + }, + { + "epoch": 0.5737974961563804, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1018807888031006, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6917358040809631, + "num_tokens": 135314741.0, + "step": 5225 + }, + { + "epoch": 0.5739073138589941, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3770530223846436, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.7013018131256104, + "num_tokens": 135336909.0, + "step": 5226 + }, + { + "epoch": 0.5740171315616077, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 1.9968068599700928, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7261956930160522, + "num_tokens": 135364698.0, + "step": 5227 + }, + { + "epoch": 0.5741269492642214, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.2534663677215576, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7033682465553284, + "num_tokens": 135389786.0, + "step": 5228 + }, + { + "epoch": 0.574236766966835, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3019840717315674, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.7173078060150146, + "num_tokens": 135413067.0, + "step": 5229 + }, + { + "epoch": 0.5743465846694488, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.4525301456451416, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.705608606338501, + "num_tokens": 135436186.0, + "step": 5230 + }, + { + "epoch": 0.5744564023720624, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.370852470397949, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7306833863258362, + "num_tokens": 135457201.0, + "step": 5231 + }, + { + "epoch": 0.5745662200746761, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.267603874206543, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7110855579376221, + "num_tokens": 135481982.0, + "step": 5232 + }, + { + "epoch": 0.5746760377772897, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.3903040885925293, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6936274170875549, + "num_tokens": 135505097.0, + "step": 5233 + }, + { + "epoch": 0.5747858554799034, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.1169936656951904, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6959856152534485, + "num_tokens": 135533256.0, + "step": 5234 + }, + { + "epoch": 0.574895673182517, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.0485587120056152, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.719896137714386, + "num_tokens": 135559313.0, + "step": 5235 + }, + { + "epoch": 0.5750054908851306, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.2478482723236084, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.710442841053009, + "num_tokens": 135583174.0, + "step": 5236 + }, + { + "epoch": 0.5751153085877443, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.082045793533325, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7034741640090942, + "num_tokens": 135610717.0, + "step": 5237 + }, + { + "epoch": 0.575225126290358, + "ewc_loss": 2.562999725341797e-06, + "grad_norm": 2.2601935863494873, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6974294185638428, + "num_tokens": 135635496.0, + "step": 5238 + }, + { + "epoch": 0.5753349439929717, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0326507091522217, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7202787399291992, + "num_tokens": 135662797.0, + "step": 5239 + }, + { + "epoch": 0.5754447616955853, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.6556222438812256, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7402095794677734, + "num_tokens": 135681463.0, + "step": 5240 + }, + { + "epoch": 0.575554579398199, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.000171422958374, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6961728930473328, + "num_tokens": 135712559.0, + "step": 5241 + }, + { + "epoch": 0.5756643971008126, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.311490058898926, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.714245617389679, + "num_tokens": 135735649.0, + "step": 5242 + }, + { + "epoch": 0.5757742148034263, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.289395809173584, + "learning_rate": 1e-06, + "loss": 1.0131, + "mean_token_accuracy": 0.6963012218475342, + "num_tokens": 135763058.0, + "step": 5243 + }, + { + "epoch": 0.5758840325060399, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2702131271362305, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7152984738349915, + "num_tokens": 135785873.0, + "step": 5244 + }, + { + "epoch": 0.5759938502086537, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2883360385894775, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7186933755874634, + "num_tokens": 135809402.0, + "step": 5245 + }, + { + "epoch": 0.5761036679112673, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.4403076171875, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6884010434150696, + "num_tokens": 135832842.0, + "step": 5246 + }, + { + "epoch": 0.576213485613881, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2270801067352295, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.6968584060668945, + "num_tokens": 135859345.0, + "step": 5247 + }, + { + "epoch": 0.5763233033164946, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.1918787956237793, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7044897079467773, + "num_tokens": 135885536.0, + "step": 5248 + }, + { + "epoch": 0.5764331210191083, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2099125385284424, + "learning_rate": 1e-06, + "loss": 1.0491, + "mean_token_accuracy": 0.6819283962249756, + "num_tokens": 135912140.0, + "step": 5249 + }, + { + "epoch": 0.5765429387217219, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.1111068725585938, + "learning_rate": 1e-06, + "loss": 1.1, + "mean_token_accuracy": 0.6805672645568848, + "num_tokens": 135942276.0, + "step": 5250 + }, + { + "epoch": 0.5766527564243356, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0541677474975586, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7080808877944946, + "num_tokens": 135969380.0, + "step": 5251 + }, + { + "epoch": 0.5767625741269493, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 1.9717170000076294, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7248764634132385, + "num_tokens": 135997935.0, + "step": 5252 + }, + { + "epoch": 0.576872391829563, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.1819262504577637, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6958544254302979, + "num_tokens": 136022158.0, + "step": 5253 + }, + { + "epoch": 0.5769822095321766, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0843346118927, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7083712816238403, + "num_tokens": 136048515.0, + "step": 5254 + }, + { + "epoch": 0.5770920272347903, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.3331243991851807, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7094647884368896, + "num_tokens": 136072347.0, + "step": 5255 + }, + { + "epoch": 0.5772018449374039, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.311988592147827, + "learning_rate": 1e-06, + "loss": 1.0558, + "mean_token_accuracy": 0.6819042563438416, + "num_tokens": 136098672.0, + "step": 5256 + }, + { + "epoch": 0.5773116626400175, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2363970279693604, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6989569664001465, + "num_tokens": 136123095.0, + "step": 5257 + }, + { + "epoch": 0.5774214803426312, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.1633927822113037, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6827870011329651, + "num_tokens": 136150673.0, + "step": 5258 + }, + { + "epoch": 0.577531298045245, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 1.892812728881836, + "learning_rate": 1e-06, + "loss": 1.0537, + "mean_token_accuracy": 0.6887927651405334, + "num_tokens": 136185183.0, + "step": 5259 + }, + { + "epoch": 0.5776411157478586, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 1.8653335571289062, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7100217342376709, + "num_tokens": 136217660.0, + "step": 5260 + }, + { + "epoch": 0.5777509334504722, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.31593656539917, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7140814661979675, + "num_tokens": 136240480.0, + "step": 5261 + }, + { + "epoch": 0.5778607511530859, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.3548381328582764, + "learning_rate": 1e-06, + "loss": 1.0438, + "mean_token_accuracy": 0.68743497133255, + "num_tokens": 136265258.0, + "step": 5262 + }, + { + "epoch": 0.5779705688556995, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2019708156585693, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7056190371513367, + "num_tokens": 136291840.0, + "step": 5263 + }, + { + "epoch": 0.5780803865583132, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.077357769012451, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7125474214553833, + "num_tokens": 136320057.0, + "step": 5264 + }, + { + "epoch": 0.5781902042609268, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0457987785339355, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6833089590072632, + "num_tokens": 136350046.0, + "step": 5265 + }, + { + "epoch": 0.5783000219635406, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.1381125450134277, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.7124859094619751, + "num_tokens": 136378683.0, + "step": 5266 + }, + { + "epoch": 0.5784098396661542, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0542616844177246, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.6814073324203491, + "num_tokens": 136408297.0, + "step": 5267 + }, + { + "epoch": 0.5785196573687679, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.0287904739379883, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.6784813404083252, + "num_tokens": 136438358.0, + "step": 5268 + }, + { + "epoch": 0.5786294750713815, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.143087863922119, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.7038222551345825, + "num_tokens": 136464614.0, + "step": 5269 + }, + { + "epoch": 0.5787392927739952, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.2008111476898193, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7037104368209839, + "num_tokens": 136487921.0, + "step": 5270 + }, + { + "epoch": 0.5788491104766088, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.278892993927002, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.700760006904602, + "num_tokens": 136512898.0, + "step": 5271 + }, + { + "epoch": 0.5789589281792225, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.162113904953003, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7258975505828857, + "num_tokens": 136539048.0, + "step": 5272 + }, + { + "epoch": 0.5790687458818361, + "ewc_loss": 2.5779008865356445e-06, + "grad_norm": 2.139747142791748, + "learning_rate": 1e-06, + "loss": 1.0678, + "mean_token_accuracy": 0.676283061504364, + "num_tokens": 136567431.0, + "step": 5273 + }, + { + "epoch": 0.5791785635844499, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.168431043624878, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7139106392860413, + "num_tokens": 136593366.0, + "step": 5274 + }, + { + "epoch": 0.5792883812870635, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.4952080249786377, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6921866536140442, + "num_tokens": 136614886.0, + "step": 5275 + }, + { + "epoch": 0.5793981989896771, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.0870394706726074, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6866164803504944, + "num_tokens": 136644948.0, + "step": 5276 + }, + { + "epoch": 0.5795080166922908, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.4611005783081055, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7057912349700928, + "num_tokens": 136667190.0, + "step": 5277 + }, + { + "epoch": 0.5796178343949044, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.17287540435791, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7199230194091797, + "num_tokens": 136691470.0, + "step": 5278 + }, + { + "epoch": 0.5797276520975181, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.014228343963623, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.7000962495803833, + "num_tokens": 136720589.0, + "step": 5279 + }, + { + "epoch": 0.5798374698001317, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 1.9315234422683716, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6869605779647827, + "num_tokens": 136751274.0, + "step": 5280 + }, + { + "epoch": 0.5799472875027455, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.6145758628845215, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6964697241783142, + "num_tokens": 136772880.0, + "step": 5281 + }, + { + "epoch": 0.5800571052053591, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.1017308235168457, + "learning_rate": 1e-06, + "loss": 1.0021, + "mean_token_accuracy": 0.6984636783599854, + "num_tokens": 136800349.0, + "step": 5282 + }, + { + "epoch": 0.5801669229079728, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.5794506072998047, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7221595048904419, + "num_tokens": 136819943.0, + "step": 5283 + }, + { + "epoch": 0.5802767406105864, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.016479015350342, + "learning_rate": 1e-06, + "loss": 1.0205, + "mean_token_accuracy": 0.7013413906097412, + "num_tokens": 136849882.0, + "step": 5284 + }, + { + "epoch": 0.5803865583132001, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.0719099044799805, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6754299998283386, + "num_tokens": 136878293.0, + "step": 5285 + }, + { + "epoch": 0.5804963760158137, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.145047903060913, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.6946823000907898, + "num_tokens": 136906377.0, + "step": 5286 + }, + { + "epoch": 0.5806061937184274, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.062793731689453, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6946092844009399, + "num_tokens": 136937276.0, + "step": 5287 + }, + { + "epoch": 0.5807160114210411, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.1849443912506104, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7015175819396973, + "num_tokens": 136962371.0, + "step": 5288 + }, + { + "epoch": 0.5808258291236548, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.0406343936920166, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7169253826141357, + "num_tokens": 136989852.0, + "step": 5289 + }, + { + "epoch": 0.5809356468262684, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.550219774246216, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7104456424713135, + "num_tokens": 137008127.0, + "step": 5290 + }, + { + "epoch": 0.5810454645288821, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.173902750015259, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7438258528709412, + "num_tokens": 137031663.0, + "step": 5291 + }, + { + "epoch": 0.5811552822314957, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.1883575916290283, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.6950039267539978, + "num_tokens": 137059073.0, + "step": 5292 + }, + { + "epoch": 0.5812650999341094, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.3946025371551514, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7286046147346497, + "num_tokens": 137079263.0, + "step": 5293 + }, + { + "epoch": 0.581374917636723, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 1.9993597269058228, + "learning_rate": 1e-06, + "loss": 1.0601, + "mean_token_accuracy": 0.6830638647079468, + "num_tokens": 137111579.0, + "step": 5294 + }, + { + "epoch": 0.5814847353393368, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.0990161895751953, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.721091628074646, + "num_tokens": 137136444.0, + "step": 5295 + }, + { + "epoch": 0.5815945530419504, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.1516213417053223, + "learning_rate": 1e-06, + "loss": 1.1288, + "mean_token_accuracy": 0.6594555377960205, + "num_tokens": 137165499.0, + "step": 5296 + }, + { + "epoch": 0.581704370744564, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 1.9971163272857666, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6960000991821289, + "num_tokens": 137194506.0, + "step": 5297 + }, + { + "epoch": 0.5818141884471777, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.1215553283691406, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7017039656639099, + "num_tokens": 137222222.0, + "step": 5298 + }, + { + "epoch": 0.5819240061497913, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.2211673259735107, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6899063587188721, + "num_tokens": 137248256.0, + "step": 5299 + }, + { + "epoch": 0.582033823852405, + "ewc_loss": 2.592802047729492e-06, + "grad_norm": 2.138627290725708, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7359839677810669, + "num_tokens": 137272022.0, + "step": 5300 + }, + { + "epoch": 0.5821436415550186, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.3390591144561768, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7026677131652832, + "num_tokens": 137294545.0, + "step": 5301 + }, + { + "epoch": 0.5822534592576323, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.5053038597106934, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7096264362335205, + "num_tokens": 137314838.0, + "step": 5302 + }, + { + "epoch": 0.582363276960246, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.134441375732422, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6999541521072388, + "num_tokens": 137340879.0, + "step": 5303 + }, + { + "epoch": 0.5824730946628597, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.198728084564209, + "learning_rate": 1e-06, + "loss": 1.0745, + "mean_token_accuracy": 0.693445086479187, + "num_tokens": 137368376.0, + "step": 5304 + }, + { + "epoch": 0.5825829123654733, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.085308313369751, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.6971579790115356, + "num_tokens": 137396813.0, + "step": 5305 + }, + { + "epoch": 0.582692730068087, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.2263104915618896, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7047311067581177, + "num_tokens": 137421814.0, + "step": 5306 + }, + { + "epoch": 0.5828025477707006, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.092712163925171, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7066643834114075, + "num_tokens": 137448987.0, + "step": 5307 + }, + { + "epoch": 0.5829123654733143, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 1.9227783679962158, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.70235276222229, + "num_tokens": 137478806.0, + "step": 5308 + }, + { + "epoch": 0.5830221831759279, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.218005895614624, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6844410300254822, + "num_tokens": 137504325.0, + "step": 5309 + }, + { + "epoch": 0.5831320008785417, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.5015079975128174, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7046209573745728, + "num_tokens": 137525204.0, + "step": 5310 + }, + { + "epoch": 0.5832418185811553, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.193256378173828, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.710196852684021, + "num_tokens": 137551618.0, + "step": 5311 + }, + { + "epoch": 0.583351636283769, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.1873135566711426, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7059003114700317, + "num_tokens": 137577095.0, + "step": 5312 + }, + { + "epoch": 0.5834614539863826, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 1.9831478595733643, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7160968780517578, + "num_tokens": 137606454.0, + "step": 5313 + }, + { + "epoch": 0.5835712716889963, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.0816543102264404, + "learning_rate": 1e-06, + "loss": 1.0772, + "mean_token_accuracy": 0.6725817918777466, + "num_tokens": 137634449.0, + "step": 5314 + }, + { + "epoch": 0.5836810893916099, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.1817705631256104, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.7020841240882874, + "num_tokens": 137658097.0, + "step": 5315 + }, + { + "epoch": 0.5837909070942235, + "ewc_loss": 2.60770320892334e-06, + "grad_norm": 2.3174612522125244, + "learning_rate": 1e-06, + "loss": 1.03, + "mean_token_accuracy": 0.686297595500946, + "num_tokens": 137683043.0, + "step": 5316 + }, + { + "epoch": 0.5839007247968373, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.4358644485473633, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7198065519332886, + "num_tokens": 137703335.0, + "step": 5317 + }, + { + "epoch": 0.584010542499451, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.5243186950683594, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7088758945465088, + "num_tokens": 137722898.0, + "step": 5318 + }, + { + "epoch": 0.5841203602020646, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 1.887980580329895, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7071521282196045, + "num_tokens": 137755357.0, + "step": 5319 + }, + { + "epoch": 0.5842301779046782, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.0254693031311035, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7023828029632568, + "num_tokens": 137785996.0, + "step": 5320 + }, + { + "epoch": 0.5843399956072919, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.1845040321350098, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7052983641624451, + "num_tokens": 137812226.0, + "step": 5321 + }, + { + "epoch": 0.5844498133099055, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 1.9420490264892578, + "learning_rate": 1e-06, + "loss": 1.1312, + "mean_token_accuracy": 0.6649749279022217, + "num_tokens": 137845770.0, + "step": 5322 + }, + { + "epoch": 0.5845596310125192, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.1011147499084473, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6962764263153076, + "num_tokens": 137874955.0, + "step": 5323 + }, + { + "epoch": 0.5846694487151329, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 1.8960635662078857, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7099751234054565, + "num_tokens": 137908104.0, + "step": 5324 + }, + { + "epoch": 0.5847792664177466, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.2458302974700928, + "learning_rate": 1e-06, + "loss": 1.0636, + "mean_token_accuracy": 0.6825292110443115, + "num_tokens": 137933143.0, + "step": 5325 + }, + { + "epoch": 0.5848890841203602, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.008342742919922, + "learning_rate": 1e-06, + "loss": 1.0541, + "mean_token_accuracy": 0.682195782661438, + "num_tokens": 137963416.0, + "step": 5326 + }, + { + "epoch": 0.5849989018229739, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.1803369522094727, + "learning_rate": 1e-06, + "loss": 1.0841, + "mean_token_accuracy": 0.6706541776657104, + "num_tokens": 137990308.0, + "step": 5327 + }, + { + "epoch": 0.5851087195255875, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.269653558731079, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7247793674468994, + "num_tokens": 138012509.0, + "step": 5328 + }, + { + "epoch": 0.5852185372282012, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.5144877433776855, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7200344800949097, + "num_tokens": 138032504.0, + "step": 5329 + }, + { + "epoch": 0.5853283549308148, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.111943006515503, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7270748615264893, + "num_tokens": 138057651.0, + "step": 5330 + }, + { + "epoch": 0.5854381726334285, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.0888993740081787, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7192912101745605, + "num_tokens": 138086199.0, + "step": 5331 + }, + { + "epoch": 0.5855479903360422, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.0967936515808105, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7158966064453125, + "num_tokens": 138113881.0, + "step": 5332 + }, + { + "epoch": 0.5856578080386559, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.431173086166382, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7197784781455994, + "num_tokens": 138135806.0, + "step": 5333 + }, + { + "epoch": 0.5857676257412695, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.310373067855835, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7112700343132019, + "num_tokens": 138159020.0, + "step": 5334 + }, + { + "epoch": 0.5858774434438832, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.108285427093506, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7034035325050354, + "num_tokens": 138184601.0, + "step": 5335 + }, + { + "epoch": 0.5859872611464968, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.1495604515075684, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6857306361198425, + "num_tokens": 138212102.0, + "step": 5336 + }, + { + "epoch": 0.5860970788491104, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.2605960369110107, + "learning_rate": 1e-06, + "loss": 1.0378, + "mean_token_accuracy": 0.6841859817504883, + "num_tokens": 138238813.0, + "step": 5337 + }, + { + "epoch": 0.5862068965517241, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.414576768875122, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6995853185653687, + "num_tokens": 138260141.0, + "step": 5338 + }, + { + "epoch": 0.5863167142543378, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.1813740730285645, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7265285849571228, + "num_tokens": 138283651.0, + "step": 5339 + }, + { + "epoch": 0.5864265319569515, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 1.8781706094741821, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.6980981826782227, + "num_tokens": 138315435.0, + "step": 5340 + }, + { + "epoch": 0.5865363496595651, + "ewc_loss": 2.6226043701171875e-06, + "grad_norm": 2.0644640922546387, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7100067138671875, + "num_tokens": 138342803.0, + "step": 5341 + }, + { + "epoch": 0.5866461673621788, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.3796119689941406, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6966369152069092, + "num_tokens": 138365491.0, + "step": 5342 + }, + { + "epoch": 0.5867559850647924, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 1.942961573600769, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.719763994216919, + "num_tokens": 138395065.0, + "step": 5343 + }, + { + "epoch": 0.5868658027674061, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.3327369689941406, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.682397186756134, + "num_tokens": 138420604.0, + "step": 5344 + }, + { + "epoch": 0.5869756204700197, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1456141471862793, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.7035030126571655, + "num_tokens": 138448634.0, + "step": 5345 + }, + { + "epoch": 0.5870854381726335, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.062394380569458, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6986452341079712, + "num_tokens": 138477350.0, + "step": 5346 + }, + { + "epoch": 0.5871952558752471, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.5175256729125977, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7071928977966309, + "num_tokens": 138499062.0, + "step": 5347 + }, + { + "epoch": 0.5873050735778608, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.380232572555542, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7035723924636841, + "num_tokens": 138520145.0, + "step": 5348 + }, + { + "epoch": 0.5874148912804744, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.4409799575805664, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7011510133743286, + "num_tokens": 138541908.0, + "step": 5349 + }, + { + "epoch": 0.5875247089830881, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0497748851776123, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7385616302490234, + "num_tokens": 138566648.0, + "step": 5350 + }, + { + "epoch": 0.5876345266857017, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0387790203094482, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.716328501701355, + "num_tokens": 138593706.0, + "step": 5351 + }, + { + "epoch": 0.5877443443883154, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.097067356109619, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7069149017333984, + "num_tokens": 138622965.0, + "step": 5352 + }, + { + "epoch": 0.5878541620909291, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.215512275695801, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6905433535575867, + "num_tokens": 138649693.0, + "step": 5353 + }, + { + "epoch": 0.5879639797935428, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 1.8182765245437622, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6961286067962646, + "num_tokens": 138681641.0, + "step": 5354 + }, + { + "epoch": 0.5880737974961564, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.2395622730255127, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6977776885032654, + "num_tokens": 138705096.0, + "step": 5355 + }, + { + "epoch": 0.58818361519877, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.4827451705932617, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7110389471054077, + "num_tokens": 138725318.0, + "step": 5356 + }, + { + "epoch": 0.5882934329013837, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1377787590026855, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7105551362037659, + "num_tokens": 138752583.0, + "step": 5357 + }, + { + "epoch": 0.5884032506039973, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.3046090602874756, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7124898433685303, + "num_tokens": 138775165.0, + "step": 5358 + }, + { + "epoch": 0.588513068306611, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0469765663146973, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7112820744514465, + "num_tokens": 138803464.0, + "step": 5359 + }, + { + "epoch": 0.5886228860092246, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.195023536682129, + "learning_rate": 1e-06, + "loss": 1.0173, + "mean_token_accuracy": 0.6942129135131836, + "num_tokens": 138830517.0, + "step": 5360 + }, + { + "epoch": 0.5887327037118384, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.4044713973999023, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.7026591897010803, + "num_tokens": 138852743.0, + "step": 5361 + }, + { + "epoch": 0.588842521414452, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.219226121902466, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7066792845726013, + "num_tokens": 138877770.0, + "step": 5362 + }, + { + "epoch": 0.5889523391170657, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0198800563812256, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6959472298622131, + "num_tokens": 138907114.0, + "step": 5363 + }, + { + "epoch": 0.5890621568196793, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0807619094848633, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7136578559875488, + "num_tokens": 138934234.0, + "step": 5364 + }, + { + "epoch": 0.589171974522293, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.561168909072876, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7087216377258301, + "num_tokens": 138953750.0, + "step": 5365 + }, + { + "epoch": 0.5892817922249066, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 1.922730803489685, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7033895254135132, + "num_tokens": 138985148.0, + "step": 5366 + }, + { + "epoch": 0.5893916099275203, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.206799268722534, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7212706804275513, + "num_tokens": 139009650.0, + "step": 5367 + }, + { + "epoch": 0.589501427630134, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.324068546295166, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.725924015045166, + "num_tokens": 139030423.0, + "step": 5368 + }, + { + "epoch": 0.5896112453327477, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.235291004180908, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7027666568756104, + "num_tokens": 139056646.0, + "step": 5369 + }, + { + "epoch": 0.5897210630353613, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.478747606277466, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7180376052856445, + "num_tokens": 139076987.0, + "step": 5370 + }, + { + "epoch": 0.589830880737975, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1213128566741943, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7249833345413208, + "num_tokens": 139101805.0, + "step": 5371 + }, + { + "epoch": 0.5899406984405886, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.076188325881958, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6987594366073608, + "num_tokens": 139129225.0, + "step": 5372 + }, + { + "epoch": 0.5900505161432023, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0997750759124756, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7124090194702148, + "num_tokens": 139155600.0, + "step": 5373 + }, + { + "epoch": 0.5901603338458159, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1968777179718018, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6894359588623047, + "num_tokens": 139183045.0, + "step": 5374 + }, + { + "epoch": 0.5902701515484297, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 1.8605080842971802, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7014840841293335, + "num_tokens": 139216042.0, + "step": 5375 + }, + { + "epoch": 0.5903799692510433, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.2630817890167236, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7072739601135254, + "num_tokens": 139240011.0, + "step": 5376 + }, + { + "epoch": 0.590489786953657, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1448585987091064, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7133577466011047, + "num_tokens": 139268069.0, + "step": 5377 + }, + { + "epoch": 0.5905996046562706, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.409661293029785, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.723294198513031, + "num_tokens": 139289186.0, + "step": 5378 + }, + { + "epoch": 0.5907094223588842, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.1186273097991943, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7067210674285889, + "num_tokens": 139318862.0, + "step": 5379 + }, + { + "epoch": 0.5908192400614979, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.360921859741211, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7044001221656799, + "num_tokens": 139342120.0, + "step": 5380 + }, + { + "epoch": 0.5909290577641115, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.0292763710021973, + "learning_rate": 1e-06, + "loss": 1.1163, + "mean_token_accuracy": 0.6655330061912537, + "num_tokens": 139374511.0, + "step": 5381 + }, + { + "epoch": 0.5910388754667253, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.439183473587036, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7257024645805359, + "num_tokens": 139395596.0, + "step": 5382 + }, + { + "epoch": 0.5911486931693389, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.2052268981933594, + "learning_rate": 1e-06, + "loss": 1.0213, + "mean_token_accuracy": 0.6945606470108032, + "num_tokens": 139420950.0, + "step": 5383 + }, + { + "epoch": 0.5912585108719526, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.134892463684082, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7137396335601807, + "num_tokens": 139449011.0, + "step": 5384 + }, + { + "epoch": 0.5913683285745662, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 1.9076834917068481, + "learning_rate": 1e-06, + "loss": 1.0531, + "mean_token_accuracy": 0.6902839541435242, + "num_tokens": 139480828.0, + "step": 5385 + }, + { + "epoch": 0.5914781462771799, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.2041101455688477, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.70262610912323, + "num_tokens": 139507020.0, + "step": 5386 + }, + { + "epoch": 0.5915879639797935, + "ewc_loss": 2.637505531311035e-06, + "grad_norm": 2.073793411254883, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7019324898719788, + "num_tokens": 139534335.0, + "step": 5387 + }, + { + "epoch": 0.5916977816824072, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.0728416442871094, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7085018157958984, + "num_tokens": 139561396.0, + "step": 5388 + }, + { + "epoch": 0.5918075993850208, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.0996410846710205, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6887158155441284, + "num_tokens": 139588667.0, + "step": 5389 + }, + { + "epoch": 0.5919174170876346, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.9717804193496704, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6870810985565186, + "num_tokens": 139621085.0, + "step": 5390 + }, + { + "epoch": 0.5920272347902482, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1961050033569336, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7249205112457275, + "num_tokens": 139644535.0, + "step": 5391 + }, + { + "epoch": 0.5921370524928619, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.9215725660324097, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.704583466053009, + "num_tokens": 139675145.0, + "step": 5392 + }, + { + "epoch": 0.5922468701954755, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2754087448120117, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7163316607475281, + "num_tokens": 139698271.0, + "step": 5393 + }, + { + "epoch": 0.5923566878980892, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.007117509841919, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7062826752662659, + "num_tokens": 139728259.0, + "step": 5394 + }, + { + "epoch": 0.5924665056007028, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1331231594085693, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7163227200508118, + "num_tokens": 139751081.0, + "step": 5395 + }, + { + "epoch": 0.5925763233033164, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.9321186542510986, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6972833871841431, + "num_tokens": 139783392.0, + "step": 5396 + }, + { + "epoch": 0.5926861410059302, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1468329429626465, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.707877516746521, + "num_tokens": 139809224.0, + "step": 5397 + }, + { + "epoch": 0.5927959587085438, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1586313247680664, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.732092559337616, + "num_tokens": 139835349.0, + "step": 5398 + }, + { + "epoch": 0.5929057764111575, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2481746673583984, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7234526872634888, + "num_tokens": 139858710.0, + "step": 5399 + }, + { + "epoch": 0.5930155941137711, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1849372386932373, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6936894655227661, + "num_tokens": 139883838.0, + "step": 5400 + }, + { + "epoch": 0.5931254118163848, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.3352174758911133, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6977005004882812, + "num_tokens": 139908596.0, + "step": 5401 + }, + { + "epoch": 0.5932352295189984, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.109034299850464, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.706143856048584, + "num_tokens": 139937133.0, + "step": 5402 + }, + { + "epoch": 0.5933450472216121, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2412631511688232, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.6899389624595642, + "num_tokens": 139962948.0, + "step": 5403 + }, + { + "epoch": 0.5934548649242258, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.0335452556610107, + "learning_rate": 1e-06, + "loss": 1.1104, + "mean_token_accuracy": 0.6635677218437195, + "num_tokens": 139993617.0, + "step": 5404 + }, + { + "epoch": 0.5935646826268395, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.0912108421325684, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6961585879325867, + "num_tokens": 140021652.0, + "step": 5405 + }, + { + "epoch": 0.5936745003294531, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.159627676010132, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6947991847991943, + "num_tokens": 140049633.0, + "step": 5406 + }, + { + "epoch": 0.5937843180320668, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.4699156284332275, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7384176254272461, + "num_tokens": 140070854.0, + "step": 5407 + }, + { + "epoch": 0.5938941357346804, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2609829902648926, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7093886137008667, + "num_tokens": 140094943.0, + "step": 5408 + }, + { + "epoch": 0.5940039534372941, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.4845876693725586, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7040600776672363, + "num_tokens": 140114169.0, + "step": 5409 + }, + { + "epoch": 0.5941137711399077, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.214041233062744, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.7011152505874634, + "num_tokens": 140139128.0, + "step": 5410 + }, + { + "epoch": 0.5942235888425215, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.15494966506958, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7181326150894165, + "num_tokens": 140165992.0, + "step": 5411 + }, + { + "epoch": 0.5943334065451351, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.3384151458740234, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6904077529907227, + "num_tokens": 140190420.0, + "step": 5412 + }, + { + "epoch": 0.5944432242477488, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.065260648727417, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.714881420135498, + "num_tokens": 140218466.0, + "step": 5413 + }, + { + "epoch": 0.5945530419503624, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2498433589935303, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7200939655303955, + "num_tokens": 140241243.0, + "step": 5414 + }, + { + "epoch": 0.594662859652976, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.11279559135437, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7105075120925903, + "num_tokens": 140269198.0, + "step": 5415 + }, + { + "epoch": 0.5947726773555897, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2619359493255615, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7267565727233887, + "num_tokens": 140291919.0, + "step": 5416 + }, + { + "epoch": 0.5948824950582033, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.3564467430114746, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7180060744285583, + "num_tokens": 140314477.0, + "step": 5417 + }, + { + "epoch": 0.5949923127608171, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.3208978176116943, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7066720724105835, + "num_tokens": 140337521.0, + "step": 5418 + }, + { + "epoch": 0.5951021304634307, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2812225818634033, + "learning_rate": 1e-06, + "loss": 1.0212, + "mean_token_accuracy": 0.7016282677650452, + "num_tokens": 140360715.0, + "step": 5419 + }, + { + "epoch": 0.5952119481660444, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2946617603302, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7060897946357727, + "num_tokens": 140384016.0, + "step": 5420 + }, + { + "epoch": 0.595321765868658, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2591893672943115, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7269680500030518, + "num_tokens": 140408011.0, + "step": 5421 + }, + { + "epoch": 0.5954315835712717, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.8898577690124512, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7036467790603638, + "num_tokens": 140439623.0, + "step": 5422 + }, + { + "epoch": 0.5955414012738853, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.220367193222046, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6878807544708252, + "num_tokens": 140463762.0, + "step": 5423 + }, + { + "epoch": 0.595651218976499, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1658289432525635, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6868634819984436, + "num_tokens": 140489649.0, + "step": 5424 + }, + { + "epoch": 0.5957610366791126, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.415393590927124, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7077195644378662, + "num_tokens": 140509941.0, + "step": 5425 + }, + { + "epoch": 0.5958708543817264, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1983566284179688, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6973932981491089, + "num_tokens": 140535871.0, + "step": 5426 + }, + { + "epoch": 0.59598067208434, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1378695964813232, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7085692882537842, + "num_tokens": 140562930.0, + "step": 5427 + }, + { + "epoch": 0.5960904897869537, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.0310702323913574, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7139009833335876, + "num_tokens": 140590643.0, + "step": 5428 + }, + { + "epoch": 0.5962003074895673, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2833244800567627, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6941210031509399, + "num_tokens": 140615200.0, + "step": 5429 + }, + { + "epoch": 0.596310125192181, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.9490450620651245, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7178221940994263, + "num_tokens": 140645253.0, + "step": 5430 + }, + { + "epoch": 0.5964199428947946, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2198004722595215, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7229436635971069, + "num_tokens": 140667899.0, + "step": 5431 + }, + { + "epoch": 0.5965297605974083, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.332656145095825, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.6910644769668579, + "num_tokens": 140691109.0, + "step": 5432 + }, + { + "epoch": 0.596639578300022, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1455307006835938, + "learning_rate": 1e-06, + "loss": 1.0232, + "mean_token_accuracy": 0.6954948902130127, + "num_tokens": 140717798.0, + "step": 5433 + }, + { + "epoch": 0.5967493960026357, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.948538064956665, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6914125680923462, + "num_tokens": 140749119.0, + "step": 5434 + }, + { + "epoch": 0.5968592137052493, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.157449960708618, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7144595384597778, + "num_tokens": 140774407.0, + "step": 5435 + }, + { + "epoch": 0.596969031407863, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1993350982666016, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.696713924407959, + "num_tokens": 140800384.0, + "step": 5436 + }, + { + "epoch": 0.5970788491104766, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.2874197959899902, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7192554473876953, + "num_tokens": 140823824.0, + "step": 5437 + }, + { + "epoch": 0.5971886668130902, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.3256595134735107, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7043633460998535, + "num_tokens": 140845377.0, + "step": 5438 + }, + { + "epoch": 0.5972984845157039, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.1149604320526123, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7180058360099792, + "num_tokens": 140869853.0, + "step": 5439 + }, + { + "epoch": 0.5974083022183176, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 2.227243661880493, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7220429182052612, + "num_tokens": 140893376.0, + "step": 5440 + }, + { + "epoch": 0.5975181199209313, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.8825442790985107, + "learning_rate": 1e-06, + "loss": 1.0707, + "mean_token_accuracy": 0.6727122068405151, + "num_tokens": 140930049.0, + "step": 5441 + }, + { + "epoch": 0.5976279376235449, + "ewc_loss": 2.652406692504883e-06, + "grad_norm": 1.9877684116363525, + "learning_rate": 1e-06, + "loss": 1.0724, + "mean_token_accuracy": 0.6784684062004089, + "num_tokens": 140960774.0, + "step": 5442 + }, + { + "epoch": 0.5977377553261586, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 1.8742196559906006, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6985687017440796, + "num_tokens": 140991688.0, + "step": 5443 + }, + { + "epoch": 0.5978475730287722, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.1417627334594727, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7119222283363342, + "num_tokens": 141018534.0, + "step": 5444 + }, + { + "epoch": 0.5979573907313859, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.471836805343628, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.6948065757751465, + "num_tokens": 141040847.0, + "step": 5445 + }, + { + "epoch": 0.5980672084339995, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.13338303565979, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6930814981460571, + "num_tokens": 141068941.0, + "step": 5446 + }, + { + "epoch": 0.5981770261366133, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.487187147140503, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7287949919700623, + "num_tokens": 141088818.0, + "step": 5447 + }, + { + "epoch": 0.5982868438392269, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.097364902496338, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7422444224357605, + "num_tokens": 141111050.0, + "step": 5448 + }, + { + "epoch": 0.5983966615418406, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.1749374866485596, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7243322730064392, + "num_tokens": 141134434.0, + "step": 5449 + }, + { + "epoch": 0.5985064792444542, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.3713200092315674, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7180878520011902, + "num_tokens": 141156074.0, + "step": 5450 + }, + { + "epoch": 0.5986162969470679, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.0961551666259766, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6984987258911133, + "num_tokens": 141184073.0, + "step": 5451 + }, + { + "epoch": 0.5987261146496815, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.2411317825317383, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7110787630081177, + "num_tokens": 141207261.0, + "step": 5452 + }, + { + "epoch": 0.5988359323522952, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.4740688800811768, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7191814184188843, + "num_tokens": 141227943.0, + "step": 5453 + }, + { + "epoch": 0.5989457500549088, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 1.9444245100021362, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6979498267173767, + "num_tokens": 141259938.0, + "step": 5454 + }, + { + "epoch": 0.5990555677575226, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.4006423950195312, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7047101259231567, + "num_tokens": 141281372.0, + "step": 5455 + }, + { + "epoch": 0.5991653854601362, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.4109556674957275, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7206701040267944, + "num_tokens": 141301792.0, + "step": 5456 + }, + { + "epoch": 0.5992752031627498, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.141249895095825, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7024163603782654, + "num_tokens": 141330241.0, + "step": 5457 + }, + { + "epoch": 0.5993850208653635, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.08860445022583, + "learning_rate": 1e-06, + "loss": 1.0536, + "mean_token_accuracy": 0.6930298805236816, + "num_tokens": 141360496.0, + "step": 5458 + }, + { + "epoch": 0.5994948385679771, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.1604208946228027, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.6906310319900513, + "num_tokens": 141386431.0, + "step": 5459 + }, + { + "epoch": 0.5996046562705908, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.091796875, + "learning_rate": 1e-06, + "loss": 1.1015, + "mean_token_accuracy": 0.6677894592285156, + "num_tokens": 141415944.0, + "step": 5460 + }, + { + "epoch": 0.5997144739732044, + "ewc_loss": 2.6673078536987305e-06, + "grad_norm": 2.060847520828247, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6994816064834595, + "num_tokens": 141444005.0, + "step": 5461 + }, + { + "epoch": 0.5998242916758182, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.146048069000244, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6899084448814392, + "num_tokens": 141470558.0, + "step": 5462 + }, + { + "epoch": 0.5999341093784318, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.056124687194824, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6890507936477661, + "num_tokens": 141499619.0, + "step": 5463 + }, + { + "epoch": 0.6000439270810455, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.4363958835601807, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7221379280090332, + "num_tokens": 141519971.0, + "step": 5464 + }, + { + "epoch": 0.6001537447836591, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.0929055213928223, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7230894565582275, + "num_tokens": 141547413.0, + "step": 5465 + }, + { + "epoch": 0.6002635624862728, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 1.8903239965438843, + "learning_rate": 1e-06, + "loss": 1.0671, + "mean_token_accuracy": 0.6795159578323364, + "num_tokens": 141582883.0, + "step": 5466 + }, + { + "epoch": 0.6003733801888864, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.2058589458465576, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7046264410018921, + "num_tokens": 141609111.0, + "step": 5467 + }, + { + "epoch": 0.6004831978915001, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.1417672634124756, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.723373293876648, + "num_tokens": 141632586.0, + "step": 5468 + }, + { + "epoch": 0.6005930155941138, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.4338700771331787, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7138749361038208, + "num_tokens": 141653030.0, + "step": 5469 + }, + { + "epoch": 0.6007028332967275, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.0617613792419434, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.6926478147506714, + "num_tokens": 141680762.0, + "step": 5470 + }, + { + "epoch": 0.6008126509993411, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.512293815612793, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6917830109596252, + "num_tokens": 141705139.0, + "step": 5471 + }, + { + "epoch": 0.6009224687019548, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.1855363845825195, + "learning_rate": 1e-06, + "loss": 1.0636, + "mean_token_accuracy": 0.6810510754585266, + "num_tokens": 141732248.0, + "step": 5472 + }, + { + "epoch": 0.6010322864045684, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.3704755306243896, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6878622770309448, + "num_tokens": 141756347.0, + "step": 5473 + }, + { + "epoch": 0.601142104107182, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 1.8510429859161377, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6892041563987732, + "num_tokens": 141790860.0, + "step": 5474 + }, + { + "epoch": 0.6012519218097957, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.100338935852051, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7034599781036377, + "num_tokens": 141818768.0, + "step": 5475 + }, + { + "epoch": 0.6013617395124095, + "ewc_loss": 2.682209014892578e-06, + "grad_norm": 2.000269889831543, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7170965075492859, + "num_tokens": 141846167.0, + "step": 5476 + }, + { + "epoch": 0.6014715572150231, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.4274094104766846, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7074980735778809, + "num_tokens": 141867109.0, + "step": 5477 + }, + { + "epoch": 0.6015813749176367, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.1095190048217773, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6889975070953369, + "num_tokens": 141894244.0, + "step": 5478 + }, + { + "epoch": 0.6016911926202504, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2793636322021484, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7138813734054565, + "num_tokens": 141918529.0, + "step": 5479 + }, + { + "epoch": 0.601801010322864, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2341983318328857, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7110501527786255, + "num_tokens": 141941287.0, + "step": 5480 + }, + { + "epoch": 0.6019108280254777, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2784817218780518, + "learning_rate": 1e-06, + "loss": 1.0237, + "mean_token_accuracy": 0.6855827569961548, + "num_tokens": 141965353.0, + "step": 5481 + }, + { + "epoch": 0.6020206457280913, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2282378673553467, + "learning_rate": 1e-06, + "loss": 1.1014, + "mean_token_accuracy": 0.6749944686889648, + "num_tokens": 141992824.0, + "step": 5482 + }, + { + "epoch": 0.602130463430705, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.3757147789001465, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.6907229423522949, + "num_tokens": 142016379.0, + "step": 5483 + }, + { + "epoch": 0.6022402811333187, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 1.9019882678985596, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6829097270965576, + "num_tokens": 142051236.0, + "step": 5484 + }, + { + "epoch": 0.6023500988359324, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2461559772491455, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6931910514831543, + "num_tokens": 142076418.0, + "step": 5485 + }, + { + "epoch": 0.602459916538546, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.5733962059020996, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7201439142227173, + "num_tokens": 142096091.0, + "step": 5486 + }, + { + "epoch": 0.6025697342411597, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.076335906982422, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7074790000915527, + "num_tokens": 142124345.0, + "step": 5487 + }, + { + "epoch": 0.6026795519437733, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.18210506439209, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6890705227851868, + "num_tokens": 142154203.0, + "step": 5488 + }, + { + "epoch": 0.602789369646387, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2442712783813477, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.6995386481285095, + "num_tokens": 142178440.0, + "step": 5489 + }, + { + "epoch": 0.6028991873490006, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.405327081680298, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7082086801528931, + "num_tokens": 142203260.0, + "step": 5490 + }, + { + "epoch": 0.6030090050516144, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.5908384323120117, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7131646275520325, + "num_tokens": 142222642.0, + "step": 5491 + }, + { + "epoch": 0.603118822754228, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.1385910511016846, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.712763786315918, + "num_tokens": 142248504.0, + "step": 5492 + }, + { + "epoch": 0.6032286404568417, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2081167697906494, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7073569297790527, + "num_tokens": 142275888.0, + "step": 5493 + }, + { + "epoch": 0.6033384581594553, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2001736164093018, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.696321964263916, + "num_tokens": 142302175.0, + "step": 5494 + }, + { + "epoch": 0.603448275862069, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.349367141723633, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7094395160675049, + "num_tokens": 142326863.0, + "step": 5495 + }, + { + "epoch": 0.6035580935646826, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.04807448387146, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7050997018814087, + "num_tokens": 142355356.0, + "step": 5496 + }, + { + "epoch": 0.6036679112672962, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.2617928981781006, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7068570852279663, + "num_tokens": 142379442.0, + "step": 5497 + }, + { + "epoch": 0.60377772896991, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.328677177429199, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7037407755851746, + "num_tokens": 142402486.0, + "step": 5498 + }, + { + "epoch": 0.6038875466725236, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.232530117034912, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6941899657249451, + "num_tokens": 142427756.0, + "step": 5499 + }, + { + "epoch": 0.6039973643751373, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.0961380004882812, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7052178382873535, + "num_tokens": 142456504.0, + "step": 5500 + }, + { + "epoch": 0.6041071820777509, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.059690237045288, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7192861437797546, + "num_tokens": 142481999.0, + "step": 5501 + }, + { + "epoch": 0.6042169997803646, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.1311588287353516, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7150107622146606, + "num_tokens": 142508085.0, + "step": 5502 + }, + { + "epoch": 0.6043268174829782, + "ewc_loss": 2.6971101760864258e-06, + "grad_norm": 2.331597089767456, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6820559501647949, + "num_tokens": 142533502.0, + "step": 5503 + }, + { + "epoch": 0.6044366351855919, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.116804838180542, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7076762318611145, + "num_tokens": 142562232.0, + "step": 5504 + }, + { + "epoch": 0.6045464528882056, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.275822877883911, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7195134162902832, + "num_tokens": 142584535.0, + "step": 5505 + }, + { + "epoch": 0.6046562705908193, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.059454917907715, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6980927586555481, + "num_tokens": 142613548.0, + "step": 5506 + }, + { + "epoch": 0.6047660882934329, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.411842107772827, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7036248445510864, + "num_tokens": 142634501.0, + "step": 5507 + }, + { + "epoch": 0.6048759059960466, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.1267647743225098, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7104179859161377, + "num_tokens": 142660053.0, + "step": 5508 + }, + { + "epoch": 0.6049857236986602, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.094428539276123, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.7019646167755127, + "num_tokens": 142686587.0, + "step": 5509 + }, + { + "epoch": 0.6050955414012739, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.2274253368377686, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.7003911733627319, + "num_tokens": 142710842.0, + "step": 5510 + }, + { + "epoch": 0.6052053591038875, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 1.8995945453643799, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.69837486743927, + "num_tokens": 142744140.0, + "step": 5511 + }, + { + "epoch": 0.6053151768065012, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.172478675842285, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7069358825683594, + "num_tokens": 142769131.0, + "step": 5512 + }, + { + "epoch": 0.6054249945091149, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.140874147415161, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.709666907787323, + "num_tokens": 142798422.0, + "step": 5513 + }, + { + "epoch": 0.6055348122117286, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.3787972927093506, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7110914587974548, + "num_tokens": 142820626.0, + "step": 5514 + }, + { + "epoch": 0.6056446299143422, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.205132007598877, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.714040219783783, + "num_tokens": 142842981.0, + "step": 5515 + }, + { + "epoch": 0.6057544476169558, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.494755506515503, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.709571123123169, + "num_tokens": 142862928.0, + "step": 5516 + }, + { + "epoch": 0.6058642653195695, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.1638011932373047, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.6936595439910889, + "num_tokens": 142891872.0, + "step": 5517 + }, + { + "epoch": 0.6059740830221831, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.0683000087738037, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7013592720031738, + "num_tokens": 142919028.0, + "step": 5518 + }, + { + "epoch": 0.6060839007247968, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 1.9639356136322021, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7179408669471741, + "num_tokens": 142947374.0, + "step": 5519 + }, + { + "epoch": 0.6061937184274105, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.191757917404175, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6871607899665833, + "num_tokens": 142974095.0, + "step": 5520 + }, + { + "epoch": 0.6063035361300242, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.4651339054107666, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6882659196853638, + "num_tokens": 142995967.0, + "step": 5521 + }, + { + "epoch": 0.6064133538326378, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.0760653018951416, + "learning_rate": 1e-06, + "loss": 1.0579, + "mean_token_accuracy": 0.6792210340499878, + "num_tokens": 143025221.0, + "step": 5522 + }, + { + "epoch": 0.6065231715352515, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.091670513153076, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7040873765945435, + "num_tokens": 143051527.0, + "step": 5523 + }, + { + "epoch": 0.6066329892378651, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.341254711151123, + "learning_rate": 1e-06, + "loss": 1.009, + "mean_token_accuracy": 0.6964006423950195, + "num_tokens": 143075544.0, + "step": 5524 + }, + { + "epoch": 0.6067428069404788, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.1571648120880127, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7207424640655518, + "num_tokens": 143099855.0, + "step": 5525 + }, + { + "epoch": 0.6068526246430924, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 1.9963756799697876, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7139668464660645, + "num_tokens": 143129817.0, + "step": 5526 + }, + { + "epoch": 0.6069624423457062, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.4635040760040283, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6973174810409546, + "num_tokens": 143150698.0, + "step": 5527 + }, + { + "epoch": 0.6070722600483198, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.4721717834472656, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.70567387342453, + "num_tokens": 143171043.0, + "step": 5528 + }, + { + "epoch": 0.6071820777509335, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.295860528945923, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7157956957817078, + "num_tokens": 143193800.0, + "step": 5529 + }, + { + "epoch": 0.6072918954535471, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.0629310607910156, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6968278884887695, + "num_tokens": 143222251.0, + "step": 5530 + }, + { + "epoch": 0.6074017131561608, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.4344193935394287, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7105831503868103, + "num_tokens": 143245623.0, + "step": 5531 + }, + { + "epoch": 0.6075115308587744, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.185520887374878, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.720474898815155, + "num_tokens": 143271328.0, + "step": 5532 + }, + { + "epoch": 0.607621348561388, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.544557809829712, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7175886034965515, + "num_tokens": 143290653.0, + "step": 5533 + }, + { + "epoch": 0.6077311662640018, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.13716197013855, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7233706712722778, + "num_tokens": 143316964.0, + "step": 5534 + }, + { + "epoch": 0.6078409839666155, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.4060187339782715, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7229755520820618, + "num_tokens": 143339087.0, + "step": 5535 + }, + { + "epoch": 0.6079508016692291, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 1.8804563283920288, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.7007861137390137, + "num_tokens": 143372999.0, + "step": 5536 + }, + { + "epoch": 0.6080606193718427, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.2307355403900146, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.7046693563461304, + "num_tokens": 143398034.0, + "step": 5537 + }, + { + "epoch": 0.6081704370744564, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.353362560272217, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7189971208572388, + "num_tokens": 143419511.0, + "step": 5538 + }, + { + "epoch": 0.60828025477707, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.0542991161346436, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6898633241653442, + "num_tokens": 143447620.0, + "step": 5539 + }, + { + "epoch": 0.6083900724796837, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.362240791320801, + "learning_rate": 1e-06, + "loss": 1.0418, + "mean_token_accuracy": 0.6957299709320068, + "num_tokens": 143470927.0, + "step": 5540 + }, + { + "epoch": 0.6084998901822973, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.1907193660736084, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7052643895149231, + "num_tokens": 143495362.0, + "step": 5541 + }, + { + "epoch": 0.6086097078849111, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.097642660140991, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7145213484764099, + "num_tokens": 143522298.0, + "step": 5542 + }, + { + "epoch": 0.6087195255875247, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.5934927463531494, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7194591164588928, + "num_tokens": 143540762.0, + "step": 5543 + }, + { + "epoch": 0.6088293432901384, + "ewc_loss": 2.7120113372802734e-06, + "grad_norm": 2.3862898349761963, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7303122282028198, + "num_tokens": 143562935.0, + "step": 5544 + }, + { + "epoch": 0.608939160992752, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.3700897693634033, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6958852410316467, + "num_tokens": 143586717.0, + "step": 5545 + }, + { + "epoch": 0.6090489786953657, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1444833278656006, + "learning_rate": 1e-06, + "loss": 1.0151, + "mean_token_accuracy": 0.6907926201820374, + "num_tokens": 143613936.0, + "step": 5546 + }, + { + "epoch": 0.6091587963979793, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1169042587280273, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7102082967758179, + "num_tokens": 143641181.0, + "step": 5547 + }, + { + "epoch": 0.609268614100593, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1533806324005127, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7038769125938416, + "num_tokens": 143666141.0, + "step": 5548 + }, + { + "epoch": 0.6093784318032067, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.0999457836151123, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7000300884246826, + "num_tokens": 143694096.0, + "step": 5549 + }, + { + "epoch": 0.6094882495058204, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.0170199871063232, + "learning_rate": 1e-06, + "loss": 1.0322, + "mean_token_accuracy": 0.6973809003829956, + "num_tokens": 143724483.0, + "step": 5550 + }, + { + "epoch": 0.609598067208434, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1853222846984863, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7101633548736572, + "num_tokens": 143748850.0, + "step": 5551 + }, + { + "epoch": 0.6097078849110477, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1137473583221436, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6987844705581665, + "num_tokens": 143776175.0, + "step": 5552 + }, + { + "epoch": 0.6098177026136613, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.207610845565796, + "learning_rate": 1e-06, + "loss": 1.0638, + "mean_token_accuracy": 0.6898045539855957, + "num_tokens": 143804042.0, + "step": 5553 + }, + { + "epoch": 0.609927520316275, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1642675399780273, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.707392692565918, + "num_tokens": 143831258.0, + "step": 5554 + }, + { + "epoch": 0.6100373380188886, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 1.9888676404953003, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6944440603256226, + "num_tokens": 143861686.0, + "step": 5555 + }, + { + "epoch": 0.6101471557215024, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.4137415885925293, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7007759809494019, + "num_tokens": 143883454.0, + "step": 5556 + }, + { + "epoch": 0.610256973424116, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.2436487674713135, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.707135021686554, + "num_tokens": 143907384.0, + "step": 5557 + }, + { + "epoch": 0.6103667911267296, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1875665187835693, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6941196918487549, + "num_tokens": 143934663.0, + "step": 5558 + }, + { + "epoch": 0.6104766088293433, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.0521247386932373, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7116796970367432, + "num_tokens": 143962719.0, + "step": 5559 + }, + { + "epoch": 0.6105864265319569, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.5577826499938965, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7120018005371094, + "num_tokens": 143983202.0, + "step": 5560 + }, + { + "epoch": 0.6106962442345706, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.42779541015625, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7142988443374634, + "num_tokens": 144004714.0, + "step": 5561 + }, + { + "epoch": 0.6108060619371842, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.5239906311035156, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6992219090461731, + "num_tokens": 144025861.0, + "step": 5562 + }, + { + "epoch": 0.610915879639798, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.318387508392334, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7146005034446716, + "num_tokens": 144048018.0, + "step": 5563 + }, + { + "epoch": 0.6110256973424116, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.10315203666687, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7261809706687927, + "num_tokens": 144073098.0, + "step": 5564 + }, + { + "epoch": 0.6111355150450253, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 1.9883579015731812, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6915035247802734, + "num_tokens": 144103692.0, + "step": 5565 + }, + { + "epoch": 0.6112453327476389, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.138198137283325, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6995301842689514, + "num_tokens": 144130071.0, + "step": 5566 + }, + { + "epoch": 0.6113551504502526, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 1.8823877573013306, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7124736905097961, + "num_tokens": 144160736.0, + "step": 5567 + }, + { + "epoch": 0.6114649681528662, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1549859046936035, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7165917754173279, + "num_tokens": 144187160.0, + "step": 5568 + }, + { + "epoch": 0.6115747858554799, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.1631298065185547, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6922538876533508, + "num_tokens": 144215373.0, + "step": 5569 + }, + { + "epoch": 0.6116846035580935, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 1.914121150970459, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7229788303375244, + "num_tokens": 144246010.0, + "step": 5570 + }, + { + "epoch": 0.6117944212607073, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.2621703147888184, + "learning_rate": 1e-06, + "loss": 1.0292, + "mean_token_accuracy": 0.6894223690032959, + "num_tokens": 144271534.0, + "step": 5571 + }, + { + "epoch": 0.6119042389633209, + "ewc_loss": 2.726912498474121e-06, + "grad_norm": 2.101273775100708, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.6989247798919678, + "num_tokens": 144296962.0, + "step": 5572 + }, + { + "epoch": 0.6120140566659346, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.3846187591552734, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7181207537651062, + "num_tokens": 144319660.0, + "step": 5573 + }, + { + "epoch": 0.6121238743685482, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.0504136085510254, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.706311047077179, + "num_tokens": 144345996.0, + "step": 5574 + }, + { + "epoch": 0.6122336920711619, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.2231605052948, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7190723419189453, + "num_tokens": 144370199.0, + "step": 5575 + }, + { + "epoch": 0.6123435097737755, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.1568222045898438, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6825998425483704, + "num_tokens": 144398358.0, + "step": 5576 + }, + { + "epoch": 0.6124533274763891, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.292280435562134, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7113641500473022, + "num_tokens": 144422666.0, + "step": 5577 + }, + { + "epoch": 0.6125631451790029, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 1.973989486694336, + "learning_rate": 1e-06, + "loss": 1.0232, + "mean_token_accuracy": 0.6911230087280273, + "num_tokens": 144453770.0, + "step": 5578 + }, + { + "epoch": 0.6126729628816165, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 1.9303126335144043, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6982068419456482, + "num_tokens": 144487826.0, + "step": 5579 + }, + { + "epoch": 0.6127827805842302, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.2550594806671143, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7207456231117249, + "num_tokens": 144510583.0, + "step": 5580 + }, + { + "epoch": 0.6128925982868438, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.1568591594696045, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7065473198890686, + "num_tokens": 144537090.0, + "step": 5581 + }, + { + "epoch": 0.6130024159894575, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 1.7803068161010742, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7072929739952087, + "num_tokens": 144573721.0, + "step": 5582 + }, + { + "epoch": 0.6131122336920711, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.0811140537261963, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.700888991355896, + "num_tokens": 144600228.0, + "step": 5583 + }, + { + "epoch": 0.6132220513946848, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.1988487243652344, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6982195377349854, + "num_tokens": 144624825.0, + "step": 5584 + }, + { + "epoch": 0.6133318690972985, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.0269694328308105, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7069356441497803, + "num_tokens": 144653870.0, + "step": 5585 + }, + { + "epoch": 0.6134416867999122, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.0265421867370605, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7000967860221863, + "num_tokens": 144682126.0, + "step": 5586 + }, + { + "epoch": 0.6135515045025258, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.079514265060425, + "learning_rate": 1e-06, + "loss": 1.0618, + "mean_token_accuracy": 0.6780481338500977, + "num_tokens": 144711108.0, + "step": 5587 + }, + { + "epoch": 0.6136613222051395, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.3157167434692383, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.737373948097229, + "num_tokens": 144732059.0, + "step": 5588 + }, + { + "epoch": 0.6137711399077531, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.2503509521484375, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7191581726074219, + "num_tokens": 144756402.0, + "step": 5589 + }, + { + "epoch": 0.6138809576103668, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.443399667739868, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.6996428966522217, + "num_tokens": 144778478.0, + "step": 5590 + }, + { + "epoch": 0.6139907753129804, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.1659462451934814, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6968897581100464, + "num_tokens": 144803171.0, + "step": 5591 + }, + { + "epoch": 0.6141005930155942, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.2122790813446045, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.696867823600769, + "num_tokens": 144829122.0, + "step": 5592 + }, + { + "epoch": 0.6142104107182078, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.30873441696167, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7041870355606079, + "num_tokens": 144853840.0, + "step": 5593 + }, + { + "epoch": 0.6143202284208215, + "ewc_loss": 2.7567148208618164e-06, + "grad_norm": 2.155874729156494, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7289143204689026, + "num_tokens": 144879799.0, + "step": 5594 + }, + { + "epoch": 0.6144300461234351, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.129624366760254, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7312175631523132, + "num_tokens": 144903876.0, + "step": 5595 + }, + { + "epoch": 0.6145398638260487, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.122615098953247, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.6888605356216431, + "num_tokens": 144932095.0, + "step": 5596 + }, + { + "epoch": 0.6146496815286624, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.315425157546997, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7162289023399353, + "num_tokens": 144955621.0, + "step": 5597 + }, + { + "epoch": 0.614759499231276, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.088273048400879, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7028439044952393, + "num_tokens": 144985577.0, + "step": 5598 + }, + { + "epoch": 0.6148693169338898, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.4618608951568604, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7305529117584229, + "num_tokens": 145005958.0, + "step": 5599 + }, + { + "epoch": 0.6149791346365034, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.4683825969696045, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7176417708396912, + "num_tokens": 145025793.0, + "step": 5600 + }, + { + "epoch": 0.6150889523391171, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.023716449737549, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7001165747642517, + "num_tokens": 145056561.0, + "step": 5601 + }, + { + "epoch": 0.6151987700417307, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.0305087566375732, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7202500700950623, + "num_tokens": 145084877.0, + "step": 5602 + }, + { + "epoch": 0.6153085877443444, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.2677953243255615, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7285678386688232, + "num_tokens": 145108660.0, + "step": 5603 + }, + { + "epoch": 0.615418405446958, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.261359214782715, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7083245515823364, + "num_tokens": 145132828.0, + "step": 5604 + }, + { + "epoch": 0.6155282231495717, + "ewc_loss": 2.7865171432495117e-06, + "grad_norm": 2.2804975509643555, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7017646431922913, + "num_tokens": 145154962.0, + "step": 5605 + }, + { + "epoch": 0.6156380408521853, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.3802490234375, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.683295726776123, + "num_tokens": 145182406.0, + "step": 5606 + }, + { + "epoch": 0.6157478585547991, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.1511855125427246, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6879479289054871, + "num_tokens": 145208430.0, + "step": 5607 + }, + { + "epoch": 0.6158576762574127, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.2327558994293213, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7096042037010193, + "num_tokens": 145236623.0, + "step": 5608 + }, + { + "epoch": 0.6159674939600264, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.19535493850708, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7083516120910645, + "num_tokens": 145261618.0, + "step": 5609 + }, + { + "epoch": 0.61607731166264, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.1204960346221924, + "learning_rate": 1e-06, + "loss": 1.0474, + "mean_token_accuracy": 0.6969650387763977, + "num_tokens": 145291071.0, + "step": 5610 + }, + { + "epoch": 0.6161871293652537, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 1.8525352478027344, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7352532744407654, + "num_tokens": 145323857.0, + "step": 5611 + }, + { + "epoch": 0.6162969470678673, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.139641046524048, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.6908001899719238, + "num_tokens": 145353905.0, + "step": 5612 + }, + { + "epoch": 0.616406764770481, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 1.9523965120315552, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7214221954345703, + "num_tokens": 145384056.0, + "step": 5613 + }, + { + "epoch": 0.6165165824730947, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.1637415885925293, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.726893424987793, + "num_tokens": 145410449.0, + "step": 5614 + }, + { + "epoch": 0.6166264001757084, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.2625741958618164, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7265257239341736, + "num_tokens": 145433886.0, + "step": 5615 + }, + { + "epoch": 0.616736217878322, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.434537410736084, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6885420083999634, + "num_tokens": 145455812.0, + "step": 5616 + }, + { + "epoch": 0.6168460355809356, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.1079132556915283, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6882779002189636, + "num_tokens": 145483198.0, + "step": 5617 + }, + { + "epoch": 0.6169558532835493, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.2120490074157715, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7080810070037842, + "num_tokens": 145508062.0, + "step": 5618 + }, + { + "epoch": 0.6170656709861629, + "ewc_loss": 2.8014183044433594e-06, + "grad_norm": 2.1911115646362305, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6956738829612732, + "num_tokens": 145533803.0, + "step": 5619 + }, + { + "epoch": 0.6171754886887766, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.0928783416748047, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6916359663009644, + "num_tokens": 145561163.0, + "step": 5620 + }, + { + "epoch": 0.6172853063913903, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.0753469467163086, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7058147192001343, + "num_tokens": 145587924.0, + "step": 5621 + }, + { + "epoch": 0.617395124094004, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.0649101734161377, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.71339350938797, + "num_tokens": 145614186.0, + "step": 5622 + }, + { + "epoch": 0.6175049417966176, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.219947099685669, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7158334255218506, + "num_tokens": 145636444.0, + "step": 5623 + }, + { + "epoch": 0.6176147594992313, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.0240254402160645, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.7032163739204407, + "num_tokens": 145665371.0, + "step": 5624 + }, + { + "epoch": 0.6177245772018449, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.180007219314575, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7033442258834839, + "num_tokens": 145689331.0, + "step": 5625 + }, + { + "epoch": 0.6178343949044586, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.444188117980957, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7016685605049133, + "num_tokens": 145709752.0, + "step": 5626 + }, + { + "epoch": 0.6179442126070722, + "ewc_loss": 2.816319465637207e-06, + "grad_norm": 2.3101632595062256, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7152919173240662, + "num_tokens": 145732554.0, + "step": 5627 + }, + { + "epoch": 0.618054030309686, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.8172941207885742, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6837859153747559, + "num_tokens": 145768213.0, + "step": 5628 + }, + { + "epoch": 0.6181638480122996, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.423703908920288, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7155199646949768, + "num_tokens": 145788714.0, + "step": 5629 + }, + { + "epoch": 0.6182736657149133, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.3844635486602783, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7198684811592102, + "num_tokens": 145811217.0, + "step": 5630 + }, + { + "epoch": 0.6183834834175269, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2900595664978027, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6973550915718079, + "num_tokens": 145834633.0, + "step": 5631 + }, + { + "epoch": 0.6184933011201406, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.4455924034118652, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7057422399520874, + "num_tokens": 145855250.0, + "step": 5632 + }, + { + "epoch": 0.6186031188227542, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9163910150527954, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6951999664306641, + "num_tokens": 145888420.0, + "step": 5633 + }, + { + "epoch": 0.6187129365253679, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.268068313598633, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7105032205581665, + "num_tokens": 145911814.0, + "step": 5634 + }, + { + "epoch": 0.6188227542279815, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1629862785339355, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7439512610435486, + "num_tokens": 145937059.0, + "step": 5635 + }, + { + "epoch": 0.6189325719305953, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2060117721557617, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.703045129776001, + "num_tokens": 145964732.0, + "step": 5636 + }, + { + "epoch": 0.6190423896332089, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2034308910369873, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7006258368492126, + "num_tokens": 145990592.0, + "step": 5637 + }, + { + "epoch": 0.6191522073358225, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9777514934539795, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6921755075454712, + "num_tokens": 146020581.0, + "step": 5638 + }, + { + "epoch": 0.6192620250384362, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2323429584503174, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6937121152877808, + "num_tokens": 146045852.0, + "step": 5639 + }, + { + "epoch": 0.6193718427410498, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2179856300354004, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7433717846870422, + "num_tokens": 146069988.0, + "step": 5640 + }, + { + "epoch": 0.6194816604436635, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.489114999771118, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7076494693756104, + "num_tokens": 146090334.0, + "step": 5641 + }, + { + "epoch": 0.6195914781462771, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.218824863433838, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.711787223815918, + "num_tokens": 146113724.0, + "step": 5642 + }, + { + "epoch": 0.6197012958488909, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2143139839172363, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7130531668663025, + "num_tokens": 146138973.0, + "step": 5643 + }, + { + "epoch": 0.6198111135515045, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9257962703704834, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7310120463371277, + "num_tokens": 146168630.0, + "step": 5644 + }, + { + "epoch": 0.6199209312541182, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.995010256767273, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6899683475494385, + "num_tokens": 146198815.0, + "step": 5645 + }, + { + "epoch": 0.6200307489567318, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.469183921813965, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7261874079704285, + "num_tokens": 146219566.0, + "step": 5646 + }, + { + "epoch": 0.6201405666593455, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2727153301239014, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6830730438232422, + "num_tokens": 146243718.0, + "step": 5647 + }, + { + "epoch": 0.6202503843619591, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.21640944480896, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.69724440574646, + "num_tokens": 146268636.0, + "step": 5648 + }, + { + "epoch": 0.6203602020645728, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.323164463043213, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7151364684104919, + "num_tokens": 146291335.0, + "step": 5649 + }, + { + "epoch": 0.6204700197671865, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.418846607208252, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7166118025779724, + "num_tokens": 146311255.0, + "step": 5650 + }, + { + "epoch": 0.6205798374698002, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2492878437042236, + "learning_rate": 1e-06, + "loss": 1.0089, + "mean_token_accuracy": 0.6986711025238037, + "num_tokens": 146335164.0, + "step": 5651 + }, + { + "epoch": 0.6206896551724138, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.5597760677337646, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7280091643333435, + "num_tokens": 146354287.0, + "step": 5652 + }, + { + "epoch": 0.6207994728750275, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.185906171798706, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7200765609741211, + "num_tokens": 146380335.0, + "step": 5653 + }, + { + "epoch": 0.6209092905776411, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.3685920238494873, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6993165016174316, + "num_tokens": 146401453.0, + "step": 5654 + }, + { + "epoch": 0.6210191082802548, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9891639947891235, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.6993204355239868, + "num_tokens": 146431847.0, + "step": 5655 + }, + { + "epoch": 0.6211289259828684, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.213226079940796, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.7050065994262695, + "num_tokens": 146456537.0, + "step": 5656 + }, + { + "epoch": 0.6212387436854822, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2506260871887207, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.7074207067489624, + "num_tokens": 146479809.0, + "step": 5657 + }, + { + "epoch": 0.6213485613880958, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1061673164367676, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.6962934136390686, + "num_tokens": 146505965.0, + "step": 5658 + }, + { + "epoch": 0.6214583790907094, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.240039348602295, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7090487480163574, + "num_tokens": 146531551.0, + "step": 5659 + }, + { + "epoch": 0.6215681967933231, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9308093786239624, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6807713508605957, + "num_tokens": 146563551.0, + "step": 5660 + }, + { + "epoch": 0.6216780144959367, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1427228450775146, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6927133202552795, + "num_tokens": 146591120.0, + "step": 5661 + }, + { + "epoch": 0.6217878321985504, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.4110612869262695, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.72051602602005, + "num_tokens": 146611965.0, + "step": 5662 + }, + { + "epoch": 0.621897649901164, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.8069205284118652, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7052322030067444, + "num_tokens": 146648461.0, + "step": 5663 + }, + { + "epoch": 0.6220074676037777, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1333494186401367, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7008816003799438, + "num_tokens": 146676427.0, + "step": 5664 + }, + { + "epoch": 0.6221172853063914, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1591897010803223, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7194761633872986, + "num_tokens": 146700165.0, + "step": 5665 + }, + { + "epoch": 0.6222271030090051, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.194305658340454, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7022610902786255, + "num_tokens": 146724996.0, + "step": 5666 + }, + { + "epoch": 0.6223369207116187, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2717623710632324, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7120288610458374, + "num_tokens": 146750905.0, + "step": 5667 + }, + { + "epoch": 0.6224467384142324, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.142713785171509, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7055583000183105, + "num_tokens": 146776383.0, + "step": 5668 + }, + { + "epoch": 0.622556556116846, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1389198303222656, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7335473895072937, + "num_tokens": 146800241.0, + "step": 5669 + }, + { + "epoch": 0.6226663738194597, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9670534133911133, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7085058689117432, + "num_tokens": 146831318.0, + "step": 5670 + }, + { + "epoch": 0.6227761915220733, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.3623409271240234, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7008650898933411, + "num_tokens": 146853834.0, + "step": 5671 + }, + { + "epoch": 0.6228860092246871, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.542121410369873, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.73506760597229, + "num_tokens": 146872499.0, + "step": 5672 + }, + { + "epoch": 0.6229958269273007, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.249399185180664, + "learning_rate": 1e-06, + "loss": 1.0645, + "mean_token_accuracy": 0.6827283501625061, + "num_tokens": 146897374.0, + "step": 5673 + }, + { + "epoch": 0.6231056446299144, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 1.9588268995285034, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6940239667892456, + "num_tokens": 146927243.0, + "step": 5674 + }, + { + "epoch": 0.623215462332528, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.2855799198150635, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7154733538627625, + "num_tokens": 146949952.0, + "step": 5675 + }, + { + "epoch": 0.6233252800351416, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.1506989002227783, + "learning_rate": 1e-06, + "loss": 1.0175, + "mean_token_accuracy": 0.6887905597686768, + "num_tokens": 146979182.0, + "step": 5676 + }, + { + "epoch": 0.6234350977377553, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.0136783123016357, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6976115107536316, + "num_tokens": 147009131.0, + "step": 5677 + }, + { + "epoch": 0.6235449154403689, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.0946569442749023, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.6978698372840881, + "num_tokens": 147038561.0, + "step": 5678 + }, + { + "epoch": 0.6236547331429827, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.659644365310669, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7316926717758179, + "num_tokens": 147055735.0, + "step": 5679 + }, + { + "epoch": 0.6237645508455963, + "ewc_loss": 2.8312206268310547e-06, + "grad_norm": 2.3508007526397705, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7168273329734802, + "num_tokens": 147076421.0, + "step": 5680 + }, + { + "epoch": 0.62387436854821, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.2544586658477783, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7098044157028198, + "num_tokens": 147099457.0, + "step": 5681 + }, + { + "epoch": 0.6239841862508236, + "ewc_loss": 2.8461217880249023e-06, + "grad_norm": 2.236114263534546, + "learning_rate": 1e-06, + "loss": 1.0199, + "mean_token_accuracy": 0.6954374313354492, + "num_tokens": 147125823.0, + "step": 5682 + }, + { + "epoch": 0.6240940039534373, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.0398571491241455, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7073872685432434, + "num_tokens": 147154551.0, + "step": 5683 + }, + { + "epoch": 0.6242038216560509, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.350677251815796, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7031604051589966, + "num_tokens": 147177596.0, + "step": 5684 + }, + { + "epoch": 0.6243136393586646, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.4211342334747314, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6966496706008911, + "num_tokens": 147199678.0, + "step": 5685 + }, + { + "epoch": 0.6244234570612783, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1781725883483887, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7293077111244202, + "num_tokens": 147223851.0, + "step": 5686 + }, + { + "epoch": 0.624533274763892, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.0794942378997803, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7195652723312378, + "num_tokens": 147251227.0, + "step": 5687 + }, + { + "epoch": 0.6246430924665056, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1732378005981445, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6883708238601685, + "num_tokens": 147277361.0, + "step": 5688 + }, + { + "epoch": 0.6247529101691193, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.347566604614258, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7046810388565063, + "num_tokens": 147297440.0, + "step": 5689 + }, + { + "epoch": 0.6248627278717329, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.124238967895508, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7081655263900757, + "num_tokens": 147324500.0, + "step": 5690 + }, + { + "epoch": 0.6249725455743466, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.129807949066162, + "learning_rate": 1e-06, + "loss": 1.0225, + "mean_token_accuracy": 0.6945933699607849, + "num_tokens": 147352353.0, + "step": 5691 + }, + { + "epoch": 0.6250823632769602, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.5410680770874023, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7141592502593994, + "num_tokens": 147371328.0, + "step": 5692 + }, + { + "epoch": 0.6251921809795739, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.4126312732696533, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7185367345809937, + "num_tokens": 147390884.0, + "step": 5693 + }, + { + "epoch": 0.6253019986821876, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.5669069290161133, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7203943729400635, + "num_tokens": 147410305.0, + "step": 5694 + }, + { + "epoch": 0.6254118163848013, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.2274885177612305, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7060520052909851, + "num_tokens": 147434813.0, + "step": 5695 + }, + { + "epoch": 0.6255216340874149, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 1.9591346979141235, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.7005304098129272, + "num_tokens": 147466456.0, + "step": 5696 + }, + { + "epoch": 0.6256314517900285, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.629263162612915, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7144538760185242, + "num_tokens": 147485784.0, + "step": 5697 + }, + { + "epoch": 0.6257412694926422, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1666553020477295, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7255440950393677, + "num_tokens": 147509484.0, + "step": 5698 + }, + { + "epoch": 0.6258510871952558, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.4157822132110596, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7060549855232239, + "num_tokens": 147530235.0, + "step": 5699 + }, + { + "epoch": 0.6259609048978695, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.2357981204986572, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6859551668167114, + "num_tokens": 147556317.0, + "step": 5700 + }, + { + "epoch": 0.6260707226004832, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.24731707572937, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.7090519666671753, + "num_tokens": 147580805.0, + "step": 5701 + }, + { + "epoch": 0.6261805403030969, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.245414972305298, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7145273685455322, + "num_tokens": 147605713.0, + "step": 5702 + }, + { + "epoch": 0.6262903580057105, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1426424980163574, + "learning_rate": 1e-06, + "loss": 0.994, + "mean_token_accuracy": 0.6955063343048096, + "num_tokens": 147632335.0, + "step": 5703 + }, + { + "epoch": 0.6264001757083242, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.065295934677124, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.7003837823867798, + "num_tokens": 147658970.0, + "step": 5704 + }, + { + "epoch": 0.6265099934109378, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1365835666656494, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.7005512714385986, + "num_tokens": 147686015.0, + "step": 5705 + }, + { + "epoch": 0.6266198111135515, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.437152624130249, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7073894739151001, + "num_tokens": 147708507.0, + "step": 5706 + }, + { + "epoch": 0.6267296288161651, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1574065685272217, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.6824831962585449, + "num_tokens": 147736217.0, + "step": 5707 + }, + { + "epoch": 0.6268394465187789, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.3191311359405518, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7051929235458374, + "num_tokens": 147760731.0, + "step": 5708 + }, + { + "epoch": 0.6269492642213925, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1714673042297363, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.689360499382019, + "num_tokens": 147786980.0, + "step": 5709 + }, + { + "epoch": 0.6270590819240062, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.2935190200805664, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.721753716468811, + "num_tokens": 147807755.0, + "step": 5710 + }, + { + "epoch": 0.6271688996266198, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.139240026473999, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6977018117904663, + "num_tokens": 147834274.0, + "step": 5711 + }, + { + "epoch": 0.6272787173292335, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.1070103645324707, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7200892567634583, + "num_tokens": 147860854.0, + "step": 5712 + }, + { + "epoch": 0.6273885350318471, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.226567268371582, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.6824202537536621, + "num_tokens": 147887369.0, + "step": 5713 + }, + { + "epoch": 0.6274983527344608, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.0844969749450684, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7239372134208679, + "num_tokens": 147915582.0, + "step": 5714 + }, + { + "epoch": 0.6276081704370745, + "ewc_loss": 2.86102294921875e-06, + "grad_norm": 2.0643651485443115, + "learning_rate": 1e-06, + "loss": 1.0372, + "mean_token_accuracy": 0.6932579278945923, + "num_tokens": 147943778.0, + "step": 5715 + }, + { + "epoch": 0.6277179881396882, + "ewc_loss": 2.8908252716064453e-06, + "grad_norm": 2.229163408279419, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.694458544254303, + "num_tokens": 147967687.0, + "step": 5716 + }, + { + "epoch": 0.6278278058423018, + "ewc_loss": 2.8908252716064453e-06, + "grad_norm": 2.122464418411255, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6983749270439148, + "num_tokens": 147996893.0, + "step": 5717 + }, + { + "epoch": 0.6279376235449154, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 1.9490206241607666, + "learning_rate": 1e-06, + "loss": 1.0613, + "mean_token_accuracy": 0.6805247068405151, + "num_tokens": 148030636.0, + "step": 5718 + }, + { + "epoch": 0.6280474412475291, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.2306954860687256, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7123973965644836, + "num_tokens": 148054012.0, + "step": 5719 + }, + { + "epoch": 0.6281572589501427, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 1.9773043394088745, + "learning_rate": 1e-06, + "loss": 1.0723, + "mean_token_accuracy": 0.6784288883209229, + "num_tokens": 148085309.0, + "step": 5720 + }, + { + "epoch": 0.6282670766527564, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.299790620803833, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.696053683757782, + "num_tokens": 148109119.0, + "step": 5721 + }, + { + "epoch": 0.62837689435537, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.385671615600586, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6973800659179688, + "num_tokens": 148132385.0, + "step": 5722 + }, + { + "epoch": 0.6284867120579838, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 1.9104492664337158, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7024767398834229, + "num_tokens": 148165336.0, + "step": 5723 + }, + { + "epoch": 0.6285965297605974, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.0095527172088623, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.699837327003479, + "num_tokens": 148195345.0, + "step": 5724 + }, + { + "epoch": 0.6287063474632111, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.301907777786255, + "learning_rate": 1e-06, + "loss": 1.0629, + "mean_token_accuracy": 0.6858166456222534, + "num_tokens": 148218406.0, + "step": 5725 + }, + { + "epoch": 0.6288161651658247, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.0162339210510254, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7063950300216675, + "num_tokens": 148249170.0, + "step": 5726 + }, + { + "epoch": 0.6289259828684384, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.4312362670898438, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7212299108505249, + "num_tokens": 148269680.0, + "step": 5727 + }, + { + "epoch": 0.629035800571052, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.3652102947235107, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7248883247375488, + "num_tokens": 148291251.0, + "step": 5728 + }, + { + "epoch": 0.6291456182736657, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.1920554637908936, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.71448814868927, + "num_tokens": 148318381.0, + "step": 5729 + }, + { + "epoch": 0.6292554359762794, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.0766921043395996, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.713415265083313, + "num_tokens": 148345808.0, + "step": 5730 + }, + { + "epoch": 0.6293652536788931, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.3571767807006836, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7226930260658264, + "num_tokens": 148367241.0, + "step": 5731 + }, + { + "epoch": 0.6294750713815067, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.226750373840332, + "learning_rate": 1e-06, + "loss": 1.028, + "mean_token_accuracy": 0.6867561340332031, + "num_tokens": 148394134.0, + "step": 5732 + }, + { + "epoch": 0.6295848890841204, + "ewc_loss": 2.9206275939941406e-06, + "grad_norm": 2.21925687789917, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7137547135353088, + "num_tokens": 148418898.0, + "step": 5733 + }, + { + "epoch": 0.629694706786734, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.1255898475646973, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7018631100654602, + "num_tokens": 148445447.0, + "step": 5734 + }, + { + "epoch": 0.6298045244893477, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.2995760440826416, + "learning_rate": 1e-06, + "loss": 0.9961, + "mean_token_accuracy": 0.698575496673584, + "num_tokens": 148468281.0, + "step": 5735 + }, + { + "epoch": 0.6299143421919613, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.1694602966308594, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7089487910270691, + "num_tokens": 148494389.0, + "step": 5736 + }, + { + "epoch": 0.630024159894575, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.230881452560425, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7058719992637634, + "num_tokens": 148521662.0, + "step": 5737 + }, + { + "epoch": 0.6301339775971887, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.5413432121276855, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7170058488845825, + "num_tokens": 148541457.0, + "step": 5738 + }, + { + "epoch": 0.6302437952998023, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 2.1365108489990234, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6880576610565186, + "num_tokens": 148569634.0, + "step": 5739 + }, + { + "epoch": 0.630353613002416, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 1.9763177633285522, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6998103857040405, + "num_tokens": 148600795.0, + "step": 5740 + }, + { + "epoch": 0.6304634307050296, + "ewc_loss": 2.950429916381836e-06, + "grad_norm": 1.9999278783798218, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7143158912658691, + "num_tokens": 148629434.0, + "step": 5741 + }, + { + "epoch": 0.6305732484076433, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0436227321624756, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7057106494903564, + "num_tokens": 148658548.0, + "step": 5742 + }, + { + "epoch": 0.6306830661102569, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.120098352432251, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.714668333530426, + "num_tokens": 148685449.0, + "step": 5743 + }, + { + "epoch": 0.6307928838128707, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.2082340717315674, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6986187696456909, + "num_tokens": 148709970.0, + "step": 5744 + }, + { + "epoch": 0.6309027015154843, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.165440082550049, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7044656276702881, + "num_tokens": 148737062.0, + "step": 5745 + }, + { + "epoch": 0.631012519218098, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.000042676925659, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7055647969245911, + "num_tokens": 148767671.0, + "step": 5746 + }, + { + "epoch": 0.6311223369207116, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0507500171661377, + "learning_rate": 1e-06, + "loss": 1.0697, + "mean_token_accuracy": 0.6754317283630371, + "num_tokens": 148798121.0, + "step": 5747 + }, + { + "epoch": 0.6312321546233253, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.187950372695923, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7434709072113037, + "num_tokens": 148822303.0, + "step": 5748 + }, + { + "epoch": 0.6313419723259389, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.2426247596740723, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.707493245601654, + "num_tokens": 148846091.0, + "step": 5749 + }, + { + "epoch": 0.6314517900285526, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0187742710113525, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7043827772140503, + "num_tokens": 148875284.0, + "step": 5750 + }, + { + "epoch": 0.6315616077311663, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.2381978034973145, + "learning_rate": 1e-06, + "loss": 1.0695, + "mean_token_accuracy": 0.692396342754364, + "num_tokens": 148903453.0, + "step": 5751 + }, + { + "epoch": 0.63167142543378, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.2021117210388184, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7032287120819092, + "num_tokens": 148930804.0, + "step": 5752 + }, + { + "epoch": 0.6317812431363936, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.1096034049987793, + "learning_rate": 1e-06, + "loss": 1.0604, + "mean_token_accuracy": 0.6824556589126587, + "num_tokens": 148961324.0, + "step": 5753 + }, + { + "epoch": 0.6318910608390073, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0616254806518555, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7078117728233337, + "num_tokens": 148989833.0, + "step": 5754 + }, + { + "epoch": 0.6320008785416209, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 1.9457507133483887, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6998468637466431, + "num_tokens": 149021071.0, + "step": 5755 + }, + { + "epoch": 0.6321106962442345, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.151103973388672, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7170658111572266, + "num_tokens": 149044826.0, + "step": 5756 + }, + { + "epoch": 0.6322205139468482, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0333385467529297, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6991640329360962, + "num_tokens": 149071338.0, + "step": 5757 + }, + { + "epoch": 0.6323303316494618, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.0562775135040283, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.7079877257347107, + "num_tokens": 149100176.0, + "step": 5758 + }, + { + "epoch": 0.6324401493520756, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.2240195274353027, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.6916109919548035, + "num_tokens": 149126871.0, + "step": 5759 + }, + { + "epoch": 0.6325499670546892, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.44454026222229, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7038946151733398, + "num_tokens": 149149003.0, + "step": 5760 + }, + { + "epoch": 0.6326597847573029, + "ewc_loss": 2.9653310775756836e-06, + "grad_norm": 2.500887632369995, + "learning_rate": 1e-06, + "loss": 1.0438, + "mean_token_accuracy": 0.6815483570098877, + "num_tokens": 149172255.0, + "step": 5761 + }, + { + "epoch": 0.6327696024599165, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3386082649230957, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.693764328956604, + "num_tokens": 149194487.0, + "step": 5762 + }, + { + "epoch": 0.6328794201625302, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1802427768707275, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7025402784347534, + "num_tokens": 149219638.0, + "step": 5763 + }, + { + "epoch": 0.6329892378651438, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.685234546661377, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.6963731050491333, + "num_tokens": 149238091.0, + "step": 5764 + }, + { + "epoch": 0.6330990555677575, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2610585689544678, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7056695818901062, + "num_tokens": 149261916.0, + "step": 5765 + }, + { + "epoch": 0.6332088732703712, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2958364486694336, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7257925271987915, + "num_tokens": 149283449.0, + "step": 5766 + }, + { + "epoch": 0.6333186909729849, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1189358234405518, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7101174592971802, + "num_tokens": 149309221.0, + "step": 5767 + }, + { + "epoch": 0.6334285086755985, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0171265602111816, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6918484568595886, + "num_tokens": 149338800.0, + "step": 5768 + }, + { + "epoch": 0.6335383263782122, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1046836376190186, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7047110795974731, + "num_tokens": 149367427.0, + "step": 5769 + }, + { + "epoch": 0.6336481440808258, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.347599983215332, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7061690092086792, + "num_tokens": 149391851.0, + "step": 5770 + }, + { + "epoch": 0.6337579617834395, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3195645809173584, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7137064933776855, + "num_tokens": 149415251.0, + "step": 5771 + }, + { + "epoch": 0.6338677794860531, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.061068296432495, + "learning_rate": 1e-06, + "loss": 1.0434, + "mean_token_accuracy": 0.6844855546951294, + "num_tokens": 149444650.0, + "step": 5772 + }, + { + "epoch": 0.6339775971886669, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 1.9414254426956177, + "learning_rate": 1e-06, + "loss": 1.0769, + "mean_token_accuracy": 0.673724353313446, + "num_tokens": 149475849.0, + "step": 5773 + }, + { + "epoch": 0.6340874148912805, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0602715015411377, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6929004192352295, + "num_tokens": 149503813.0, + "step": 5774 + }, + { + "epoch": 0.6341972325938942, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2909281253814697, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6957062482833862, + "num_tokens": 149527726.0, + "step": 5775 + }, + { + "epoch": 0.6343070502965078, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 1.945300817489624, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6939826607704163, + "num_tokens": 149560705.0, + "step": 5776 + }, + { + "epoch": 0.6344168679991214, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0159804821014404, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.703934907913208, + "num_tokens": 149587557.0, + "step": 5777 + }, + { + "epoch": 0.6345266857017351, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1469929218292236, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.6970480680465698, + "num_tokens": 149613308.0, + "step": 5778 + }, + { + "epoch": 0.6346365034043487, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2518694400787354, + "learning_rate": 1e-06, + "loss": 1.0605, + "mean_token_accuracy": 0.6761002540588379, + "num_tokens": 149639459.0, + "step": 5779 + }, + { + "epoch": 0.6347463211069625, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.033846855163574, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.702189564704895, + "num_tokens": 149669146.0, + "step": 5780 + }, + { + "epoch": 0.6348561388095761, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.100416660308838, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7082854509353638, + "num_tokens": 149695518.0, + "step": 5781 + }, + { + "epoch": 0.6349659565121898, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.4228515625, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7162079811096191, + "num_tokens": 149718383.0, + "step": 5782 + }, + { + "epoch": 0.6350757742148034, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3511674404144287, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7048355340957642, + "num_tokens": 149741179.0, + "step": 5783 + }, + { + "epoch": 0.6351855919174171, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3909411430358887, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6982514262199402, + "num_tokens": 149763981.0, + "step": 5784 + }, + { + "epoch": 0.6352954096200307, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 1.9826011657714844, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6977741718292236, + "num_tokens": 149794146.0, + "step": 5785 + }, + { + "epoch": 0.6354052273226444, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.250784158706665, + "learning_rate": 1e-06, + "loss": 1.0532, + "mean_token_accuracy": 0.6921422481536865, + "num_tokens": 149819811.0, + "step": 5786 + }, + { + "epoch": 0.635515045025258, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.055994987487793, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.6958104968070984, + "num_tokens": 149848238.0, + "step": 5787 + }, + { + "epoch": 0.6356248627278718, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.195543050765991, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7077181339263916, + "num_tokens": 149874801.0, + "step": 5788 + }, + { + "epoch": 0.6357346804304854, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.4168951511383057, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7102051973342896, + "num_tokens": 149897048.0, + "step": 5789 + }, + { + "epoch": 0.6358444981330991, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3895955085754395, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6907999515533447, + "num_tokens": 149919727.0, + "step": 5790 + }, + { + "epoch": 0.6359543158357127, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.143801689147949, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7048022747039795, + "num_tokens": 149945061.0, + "step": 5791 + }, + { + "epoch": 0.6360641335383264, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.442323684692383, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7189392447471619, + "num_tokens": 149966169.0, + "step": 5792 + }, + { + "epoch": 0.63617395124094, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0980684757232666, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7226101756095886, + "num_tokens": 149992607.0, + "step": 5793 + }, + { + "epoch": 0.6362837689435537, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.209214925765991, + "learning_rate": 1e-06, + "loss": 1.0265, + "mean_token_accuracy": 0.6854948997497559, + "num_tokens": 150019636.0, + "step": 5794 + }, + { + "epoch": 0.6363935866461674, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.106320381164551, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7000477313995361, + "num_tokens": 150046268.0, + "step": 5795 + }, + { + "epoch": 0.636503404348781, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.148252487182617, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6908130645751953, + "num_tokens": 150072283.0, + "step": 5796 + }, + { + "epoch": 0.6366132220513947, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.110466241836548, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7025349140167236, + "num_tokens": 150099010.0, + "step": 5797 + }, + { + "epoch": 0.6367230397540083, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1155683994293213, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7103251814842224, + "num_tokens": 150124807.0, + "step": 5798 + }, + { + "epoch": 0.636832857456622, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0989155769348145, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7102900743484497, + "num_tokens": 150153282.0, + "step": 5799 + }, + { + "epoch": 0.6369426751592356, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.394637107849121, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6953113675117493, + "num_tokens": 150175704.0, + "step": 5800 + }, + { + "epoch": 0.6370524928618493, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.173374891281128, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7154582738876343, + "num_tokens": 150201266.0, + "step": 5801 + }, + { + "epoch": 0.637162310564463, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2423901557922363, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7146472930908203, + "num_tokens": 150223486.0, + "step": 5802 + }, + { + "epoch": 0.6372721282670767, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.191802978515625, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7038089036941528, + "num_tokens": 150249127.0, + "step": 5803 + }, + { + "epoch": 0.6373819459696903, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.272531747817993, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7216548323631287, + "num_tokens": 150272652.0, + "step": 5804 + }, + { + "epoch": 0.637491763672304, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.3003334999084473, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7174080014228821, + "num_tokens": 150296163.0, + "step": 5805 + }, + { + "epoch": 0.6376015813749176, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.083078622817993, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6924836039543152, + "num_tokens": 150326248.0, + "step": 5806 + }, + { + "epoch": 0.6377113990775313, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.153599500656128, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7143781781196594, + "num_tokens": 150350743.0, + "step": 5807 + }, + { + "epoch": 0.6378212167801449, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2484281063079834, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7253702282905579, + "num_tokens": 150374351.0, + "step": 5808 + }, + { + "epoch": 0.6379310344827587, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0164077281951904, + "learning_rate": 1e-06, + "loss": 1.0426, + "mean_token_accuracy": 0.6847600936889648, + "num_tokens": 150403427.0, + "step": 5809 + }, + { + "epoch": 0.6380408521853723, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.323725700378418, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7196677923202515, + "num_tokens": 150425915.0, + "step": 5810 + }, + { + "epoch": 0.638150669887986, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.301673412322998, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7130618095397949, + "num_tokens": 150448372.0, + "step": 5811 + }, + { + "epoch": 0.6382604875905996, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.3030405044555664, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7219617962837219, + "num_tokens": 150470554.0, + "step": 5812 + }, + { + "epoch": 0.6383703052932133, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.370253801345825, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.684685468673706, + "num_tokens": 150493751.0, + "step": 5813 + }, + { + "epoch": 0.6384801229958269, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2653067111968994, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6931560039520264, + "num_tokens": 150518023.0, + "step": 5814 + }, + { + "epoch": 0.6385899406984406, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 1.9009829759597778, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7222577333450317, + "num_tokens": 150549241.0, + "step": 5815 + }, + { + "epoch": 0.6386997584010542, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0287678241729736, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7053131461143494, + "num_tokens": 150580556.0, + "step": 5816 + }, + { + "epoch": 0.638809576103668, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.050575017929077, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.6910172700881958, + "num_tokens": 150611486.0, + "step": 5817 + }, + { + "epoch": 0.6389193938062816, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2640581130981445, + "learning_rate": 1e-06, + "loss": 1.0598, + "mean_token_accuracy": 0.676575779914856, + "num_tokens": 150636022.0, + "step": 5818 + }, + { + "epoch": 0.6390292115088952, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.316206455230713, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6981789469718933, + "num_tokens": 150659070.0, + "step": 5819 + }, + { + "epoch": 0.6391390292115089, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1335411071777344, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7309708595275879, + "num_tokens": 150685946.0, + "step": 5820 + }, + { + "epoch": 0.6392488469141225, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2512331008911133, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.714786171913147, + "num_tokens": 150710552.0, + "step": 5821 + }, + { + "epoch": 0.6393586646167362, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.137890100479126, + "learning_rate": 1e-06, + "loss": 1.0729, + "mean_token_accuracy": 0.6871954202651978, + "num_tokens": 150739809.0, + "step": 5822 + }, + { + "epoch": 0.6394684823193498, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.472179412841797, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7093899846076965, + "num_tokens": 150761127.0, + "step": 5823 + }, + { + "epoch": 0.6395783000219636, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.0953447818756104, + "learning_rate": 1e-06, + "loss": 0.7721, + "mean_token_accuracy": 0.7595230937004089, + "num_tokens": 150786097.0, + "step": 5824 + }, + { + "epoch": 0.6396881177245772, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.429478645324707, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.6945346593856812, + "num_tokens": 150807894.0, + "step": 5825 + }, + { + "epoch": 0.6397979354271909, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.2895519733428955, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7077680826187134, + "num_tokens": 150831323.0, + "step": 5826 + }, + { + "epoch": 0.6399077531298045, + "ewc_loss": 2.995133399963379e-06, + "grad_norm": 2.1222031116485596, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.698129415512085, + "num_tokens": 150858194.0, + "step": 5827 + }, + { + "epoch": 0.6400175708324182, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.1899752616882324, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.709121823310852, + "num_tokens": 150882564.0, + "step": 5828 + }, + { + "epoch": 0.6401273885350318, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.1844396591186523, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7244974374771118, + "num_tokens": 150907187.0, + "step": 5829 + }, + { + "epoch": 0.6402372062376455, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.155639171600342, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6876230239868164, + "num_tokens": 150934895.0, + "step": 5830 + }, + { + "epoch": 0.6403470239402592, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.2800424098968506, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6958909034729004, + "num_tokens": 150958210.0, + "step": 5831 + }, + { + "epoch": 0.6404568416428729, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.2626101970672607, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7360080480575562, + "num_tokens": 150980567.0, + "step": 5832 + }, + { + "epoch": 0.6405666593454865, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.043887138366699, + "learning_rate": 1e-06, + "loss": 1.0158, + "mean_token_accuracy": 0.6942890882492065, + "num_tokens": 151009372.0, + "step": 5833 + }, + { + "epoch": 0.6406764770481002, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.208818197250366, + "learning_rate": 1e-06, + "loss": 1.0231, + "mean_token_accuracy": 0.6972866058349609, + "num_tokens": 151035712.0, + "step": 5834 + }, + { + "epoch": 0.6407862947507138, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.1122868061065674, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7248947620391846, + "num_tokens": 151061318.0, + "step": 5835 + }, + { + "epoch": 0.6408961124533274, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.11356520652771, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7032490372657776, + "num_tokens": 151087793.0, + "step": 5836 + }, + { + "epoch": 0.6410059301559411, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.2989182472229004, + "learning_rate": 1e-06, + "loss": 1.0427, + "mean_token_accuracy": 0.6946921348571777, + "num_tokens": 151111689.0, + "step": 5837 + }, + { + "epoch": 0.6411157478585549, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 1.9759011268615723, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6881628632545471, + "num_tokens": 151144752.0, + "step": 5838 + }, + { + "epoch": 0.6412255655611685, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 2.0013270378112793, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7108453512191772, + "num_tokens": 151175374.0, + "step": 5839 + }, + { + "epoch": 0.6413353832637821, + "ewc_loss": 3.0100345611572266e-06, + "grad_norm": 1.876529574394226, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6972319483757019, + "num_tokens": 151207581.0, + "step": 5840 + }, + { + "epoch": 0.6414452009663958, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.144447088241577, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7239658236503601, + "num_tokens": 151231272.0, + "step": 5841 + }, + { + "epoch": 0.6415550186690094, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.0193278789520264, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6970164775848389, + "num_tokens": 151259848.0, + "step": 5842 + }, + { + "epoch": 0.6416648363716231, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.2804641723632812, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7168384790420532, + "num_tokens": 151283535.0, + "step": 5843 + }, + { + "epoch": 0.6417746540742367, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.3458003997802734, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7172697186470032, + "num_tokens": 151307785.0, + "step": 5844 + }, + { + "epoch": 0.6418844717768504, + "ewc_loss": 3.0249357223510742e-06, + "grad_norm": 2.1979281902313232, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.6900578737258911, + "num_tokens": 151334314.0, + "step": 5845 + }, + { + "epoch": 0.6419942894794641, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.495767831802368, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7189339995384216, + "num_tokens": 151355616.0, + "step": 5846 + }, + { + "epoch": 0.6421041071820778, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.2737889289855957, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.727904736995697, + "num_tokens": 151377054.0, + "step": 5847 + }, + { + "epoch": 0.6422139248846914, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.17763352394104, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7113150358200073, + "num_tokens": 151402658.0, + "step": 5848 + }, + { + "epoch": 0.6423237425873051, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.1666650772094727, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.705144464969635, + "num_tokens": 151429171.0, + "step": 5849 + }, + { + "epoch": 0.6424335602899187, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.0584609508514404, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7034072875976562, + "num_tokens": 151458553.0, + "step": 5850 + }, + { + "epoch": 0.6425433779925324, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.272040367126465, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7174879312515259, + "num_tokens": 151481152.0, + "step": 5851 + }, + { + "epoch": 0.642653195695146, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 1.9357357025146484, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7076166868209839, + "num_tokens": 151511318.0, + "step": 5852 + }, + { + "epoch": 0.6427630133977598, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.2361671924591064, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7128229141235352, + "num_tokens": 151534627.0, + "step": 5853 + }, + { + "epoch": 0.6428728311003734, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.216508388519287, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6985419392585754, + "num_tokens": 151560458.0, + "step": 5854 + }, + { + "epoch": 0.6429826488029871, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.341357469558716, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7035053968429565, + "num_tokens": 151582610.0, + "step": 5855 + }, + { + "epoch": 0.6430924665056007, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.2432308197021484, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7207654714584351, + "num_tokens": 151608983.0, + "step": 5856 + }, + { + "epoch": 0.6432022842082143, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.5735862255096436, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7242558002471924, + "num_tokens": 151627493.0, + "step": 5857 + }, + { + "epoch": 0.643312101910828, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 1.7728729248046875, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7420530319213867, + "num_tokens": 151661357.0, + "step": 5858 + }, + { + "epoch": 0.6434219196134416, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 1.8926855325698853, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6967588663101196, + "num_tokens": 151696267.0, + "step": 5859 + }, + { + "epoch": 0.6435317373160554, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.033565044403076, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7036070227622986, + "num_tokens": 151724986.0, + "step": 5860 + }, + { + "epoch": 0.643641555018669, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.034390687942505, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7219075560569763, + "num_tokens": 151752918.0, + "step": 5861 + }, + { + "epoch": 0.6437513727212827, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.142646312713623, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7238088846206665, + "num_tokens": 151777243.0, + "step": 5862 + }, + { + "epoch": 0.6438611904238963, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.1754071712493896, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6993962526321411, + "num_tokens": 151805063.0, + "step": 5863 + }, + { + "epoch": 0.64397100812651, + "ewc_loss": 3.039836883544922e-06, + "grad_norm": 2.2245428562164307, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7108293175697327, + "num_tokens": 151831224.0, + "step": 5864 + }, + { + "epoch": 0.6440808258291236, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.0086517333984375, + "learning_rate": 1e-06, + "loss": 1.0422, + "mean_token_accuracy": 0.6799706220626831, + "num_tokens": 151862421.0, + "step": 5865 + }, + { + "epoch": 0.6441906435317373, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.296802282333374, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6973850727081299, + "num_tokens": 151888055.0, + "step": 5866 + }, + { + "epoch": 0.644300461234351, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.4624252319335938, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7280465364456177, + "num_tokens": 151907627.0, + "step": 5867 + }, + { + "epoch": 0.6444102789369647, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.257153034210205, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7067161798477173, + "num_tokens": 151931472.0, + "step": 5868 + }, + { + "epoch": 0.6445200966395783, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.2082061767578125, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7164679765701294, + "num_tokens": 151958960.0, + "step": 5869 + }, + { + "epoch": 0.644629914342192, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.0322446823120117, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7108319997787476, + "num_tokens": 151986428.0, + "step": 5870 + }, + { + "epoch": 0.6447397320448056, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.3374276161193848, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.706672191619873, + "num_tokens": 152009759.0, + "step": 5871 + }, + { + "epoch": 0.6448495497474193, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.0449953079223633, + "learning_rate": 1e-06, + "loss": 1.0342, + "mean_token_accuracy": 0.6870796084403992, + "num_tokens": 152039419.0, + "step": 5872 + }, + { + "epoch": 0.6449593674500329, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.294016122817993, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7113810777664185, + "num_tokens": 152061413.0, + "step": 5873 + }, + { + "epoch": 0.6450691851526466, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.2431862354278564, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6897138357162476, + "num_tokens": 152086909.0, + "step": 5874 + }, + { + "epoch": 0.6451790028552603, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.3066420555114746, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7119230628013611, + "num_tokens": 152110852.0, + "step": 5875 + }, + { + "epoch": 0.645288820557874, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.163306713104248, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6908893585205078, + "num_tokens": 152139170.0, + "step": 5876 + }, + { + "epoch": 0.6453986382604876, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.3427059650421143, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6968411207199097, + "num_tokens": 152162709.0, + "step": 5877 + }, + { + "epoch": 0.6455084559631012, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.1957194805145264, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6948572397232056, + "num_tokens": 152190478.0, + "step": 5878 + }, + { + "epoch": 0.6456182736657149, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.1108641624450684, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7021690607070923, + "num_tokens": 152218556.0, + "step": 5879 + }, + { + "epoch": 0.6457280913683285, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.268368721008301, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7103328704833984, + "num_tokens": 152242928.0, + "step": 5880 + }, + { + "epoch": 0.6458379090709422, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.326603651046753, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7112448811531067, + "num_tokens": 152266912.0, + "step": 5881 + }, + { + "epoch": 0.6459477267735559, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 1.980648398399353, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.6975828409194946, + "num_tokens": 152297859.0, + "step": 5882 + }, + { + "epoch": 0.6460575444761696, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.3755791187286377, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7071675062179565, + "num_tokens": 152320490.0, + "step": 5883 + }, + { + "epoch": 0.6461673621787832, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.145986795425415, + "learning_rate": 1e-06, + "loss": 1.0338, + "mean_token_accuracy": 0.6895521879196167, + "num_tokens": 152346999.0, + "step": 5884 + }, + { + "epoch": 0.6462771798813969, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.0736031532287598, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.705414891242981, + "num_tokens": 152377770.0, + "step": 5885 + }, + { + "epoch": 0.6463869975840105, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.0612242221832275, + "learning_rate": 1e-06, + "loss": 1.0449, + "mean_token_accuracy": 0.6843644380569458, + "num_tokens": 152408079.0, + "step": 5886 + }, + { + "epoch": 0.6464968152866242, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.327692747116089, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6976508498191833, + "num_tokens": 152430651.0, + "step": 5887 + }, + { + "epoch": 0.6466066329892378, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.1143388748168945, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6939873695373535, + "num_tokens": 152459557.0, + "step": 5888 + }, + { + "epoch": 0.6467164506918516, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.604595422744751, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7326698303222656, + "num_tokens": 152478085.0, + "step": 5889 + }, + { + "epoch": 0.6468262683944652, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.1250264644622803, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7250075340270996, + "num_tokens": 152506522.0, + "step": 5890 + }, + { + "epoch": 0.6469360860970789, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.075861930847168, + "learning_rate": 1e-06, + "loss": 1.0638, + "mean_token_accuracy": 0.6809128522872925, + "num_tokens": 152535998.0, + "step": 5891 + }, + { + "epoch": 0.6470459037996925, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.1553752422332764, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6923580169677734, + "num_tokens": 152562365.0, + "step": 5892 + }, + { + "epoch": 0.6471557215023062, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 1.991661548614502, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6861754655838013, + "num_tokens": 152591858.0, + "step": 5893 + }, + { + "epoch": 0.6472655392049198, + "ewc_loss": 3.069639205932617e-06, + "grad_norm": 2.245210886001587, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7285027503967285, + "num_tokens": 152616172.0, + "step": 5894 + }, + { + "epoch": 0.6473753569075335, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.2873151302337646, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.70029616355896, + "num_tokens": 152639601.0, + "step": 5895 + }, + { + "epoch": 0.6474851746101472, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1659793853759766, + "learning_rate": 1e-06, + "loss": 1.0597, + "mean_token_accuracy": 0.6759502291679382, + "num_tokens": 152667239.0, + "step": 5896 + }, + { + "epoch": 0.6475949923127609, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.0152573585510254, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6870934963226318, + "num_tokens": 152697231.0, + "step": 5897 + }, + { + "epoch": 0.6477048100153745, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.0402235984802246, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.695185661315918, + "num_tokens": 152728080.0, + "step": 5898 + }, + { + "epoch": 0.6478146277179881, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 1.843941569328308, + "learning_rate": 1e-06, + "loss": 1.0326, + "mean_token_accuracy": 0.6863954067230225, + "num_tokens": 152762345.0, + "step": 5899 + }, + { + "epoch": 0.6479244454206018, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.086820363998413, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6956764459609985, + "num_tokens": 152789650.0, + "step": 5900 + }, + { + "epoch": 0.6480342631232154, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 1.98414146900177, + "learning_rate": 1e-06, + "loss": 1.0846, + "mean_token_accuracy": 0.679284393787384, + "num_tokens": 152821126.0, + "step": 5901 + }, + { + "epoch": 0.6481440808258291, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.268160104751587, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.698800802230835, + "num_tokens": 152846609.0, + "step": 5902 + }, + { + "epoch": 0.6482538985284428, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1072967052459717, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7006538510322571, + "num_tokens": 152872815.0, + "step": 5903 + }, + { + "epoch": 0.6483637162310565, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.092806816101074, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.6907982230186462, + "num_tokens": 152900870.0, + "step": 5904 + }, + { + "epoch": 0.6484735339336701, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1587696075439453, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7210088968276978, + "num_tokens": 152926475.0, + "step": 5905 + }, + { + "epoch": 0.6485833516362838, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 1.9289361238479614, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.6927299499511719, + "num_tokens": 152957867.0, + "step": 5906 + }, + { + "epoch": 0.6486931693388974, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.179617404937744, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7146903276443481, + "num_tokens": 152982578.0, + "step": 5907 + }, + { + "epoch": 0.6488029870415111, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.2774224281311035, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7176839709281921, + "num_tokens": 153006120.0, + "step": 5908 + }, + { + "epoch": 0.6489128047441247, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1130168437957764, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7291940450668335, + "num_tokens": 153032777.0, + "step": 5909 + }, + { + "epoch": 0.6490226224467384, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1870357990264893, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6883940696716309, + "num_tokens": 153061135.0, + "step": 5910 + }, + { + "epoch": 0.6491324401493521, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 1.9802695512771606, + "learning_rate": 1e-06, + "loss": 1.0972, + "mean_token_accuracy": 0.6624830961227417, + "num_tokens": 153094626.0, + "step": 5911 + }, + { + "epoch": 0.6492422578519658, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.3250701427459717, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7018444538116455, + "num_tokens": 153120192.0, + "step": 5912 + }, + { + "epoch": 0.6493520755545794, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1881673336029053, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7078416347503662, + "num_tokens": 153146784.0, + "step": 5913 + }, + { + "epoch": 0.6494618932571931, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.218366861343384, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7224083542823792, + "num_tokens": 153170056.0, + "step": 5914 + }, + { + "epoch": 0.6495717109598067, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.154289484024048, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.721790075302124, + "num_tokens": 153195030.0, + "step": 5915 + }, + { + "epoch": 0.6496815286624203, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1216771602630615, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7148915529251099, + "num_tokens": 153221898.0, + "step": 5916 + }, + { + "epoch": 0.649791346365034, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 1.9170377254486084, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7002696394920349, + "num_tokens": 153255269.0, + "step": 5917 + }, + { + "epoch": 0.6499011640676478, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.1138880252838135, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7223198413848877, + "num_tokens": 153281596.0, + "step": 5918 + }, + { + "epoch": 0.6500109817702614, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.5979959964752197, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7196131944656372, + "num_tokens": 153299996.0, + "step": 5919 + }, + { + "epoch": 0.650120799472875, + "ewc_loss": 3.084540367126465e-06, + "grad_norm": 2.2336738109588623, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7066614031791687, + "num_tokens": 153324129.0, + "step": 5920 + }, + { + "epoch": 0.6502306171754887, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.248265027999878, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.7002855539321899, + "num_tokens": 153348145.0, + "step": 5921 + }, + { + "epoch": 0.6503404348781023, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.263498306274414, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.712932825088501, + "num_tokens": 153371737.0, + "step": 5922 + }, + { + "epoch": 0.650450252580716, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0209121704101562, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.702462375164032, + "num_tokens": 153402899.0, + "step": 5923 + }, + { + "epoch": 0.6505600702833296, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0658156871795654, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7315273880958557, + "num_tokens": 153428385.0, + "step": 5924 + }, + { + "epoch": 0.6506698879859434, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 1.903316617012024, + "learning_rate": 1e-06, + "loss": 1.0464, + "mean_token_accuracy": 0.6857249736785889, + "num_tokens": 153461679.0, + "step": 5925 + }, + { + "epoch": 0.650779705688557, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.1242165565490723, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.6994101405143738, + "num_tokens": 153491875.0, + "step": 5926 + }, + { + "epoch": 0.6508895233911707, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.468972682952881, + "learning_rate": 1e-06, + "loss": 0.9974, + "mean_token_accuracy": 0.698976993560791, + "num_tokens": 153512525.0, + "step": 5927 + }, + { + "epoch": 0.6509993410937843, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.237367868423462, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7066667079925537, + "num_tokens": 153537574.0, + "step": 5928 + }, + { + "epoch": 0.651109158796398, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.2979300022125244, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7137614488601685, + "num_tokens": 153561400.0, + "step": 5929 + }, + { + "epoch": 0.6512189764990116, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.216236114501953, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6935132741928101, + "num_tokens": 153588151.0, + "step": 5930 + }, + { + "epoch": 0.6513287942016253, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0983946323394775, + "learning_rate": 1e-06, + "loss": 1.1088, + "mean_token_accuracy": 0.6685836315155029, + "num_tokens": 153619776.0, + "step": 5931 + }, + { + "epoch": 0.651438611904239, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.5890986919403076, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7052115201950073, + "num_tokens": 153639527.0, + "step": 5932 + }, + { + "epoch": 0.6515484296068527, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 1.975361943244934, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.725429892539978, + "num_tokens": 153667505.0, + "step": 5933 + }, + { + "epoch": 0.6516582473094663, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.1036553382873535, + "learning_rate": 1e-06, + "loss": 0.9936, + "mean_token_accuracy": 0.6981024742126465, + "num_tokens": 153695390.0, + "step": 5934 + }, + { + "epoch": 0.65176806501208, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.4296066761016846, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7150592803955078, + "num_tokens": 153717638.0, + "step": 5935 + }, + { + "epoch": 0.6518778827146936, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.3669676780700684, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7153424620628357, + "num_tokens": 153738948.0, + "step": 5936 + }, + { + "epoch": 0.6519877004173072, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.6571109294891357, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7293939590454102, + "num_tokens": 153755757.0, + "step": 5937 + }, + { + "epoch": 0.6520975181199209, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.1227617263793945, + "learning_rate": 1e-06, + "loss": 1.045, + "mean_token_accuracy": 0.6855015754699707, + "num_tokens": 153782883.0, + "step": 5938 + }, + { + "epoch": 0.6522073358225345, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.1212329864501953, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7180496454238892, + "num_tokens": 153809530.0, + "step": 5939 + }, + { + "epoch": 0.6523171535251483, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.264331102371216, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7353062033653259, + "num_tokens": 153833445.0, + "step": 5940 + }, + { + "epoch": 0.6524269712277619, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0406715869903564, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.695845365524292, + "num_tokens": 153863494.0, + "step": 5941 + }, + { + "epoch": 0.6525367889303756, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.1828458309173584, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.7033067941665649, + "num_tokens": 153889125.0, + "step": 5942 + }, + { + "epoch": 0.6526466066329892, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.4040629863739014, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7023913264274597, + "num_tokens": 153911542.0, + "step": 5943 + }, + { + "epoch": 0.6527564243356029, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 1.9534882307052612, + "learning_rate": 1e-06, + "loss": 1.0628, + "mean_token_accuracy": 0.6870937943458557, + "num_tokens": 153942533.0, + "step": 5944 + }, + { + "epoch": 0.6528662420382165, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.2076778411865234, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.69516521692276, + "num_tokens": 153967534.0, + "step": 5945 + }, + { + "epoch": 0.6529760597408302, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0136983394622803, + "learning_rate": 1e-06, + "loss": 1.0835, + "mean_token_accuracy": 0.6760998964309692, + "num_tokens": 153998781.0, + "step": 5946 + }, + { + "epoch": 0.6530858774434439, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.555748701095581, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.699739933013916, + "num_tokens": 154019262.0, + "step": 5947 + }, + { + "epoch": 0.6531956951460576, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.0910863876342773, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6996897459030151, + "num_tokens": 154047477.0, + "step": 5948 + }, + { + "epoch": 0.6533055128486712, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.268850088119507, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7219061851501465, + "num_tokens": 154070596.0, + "step": 5949 + }, + { + "epoch": 0.6534153305512849, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.89315128326416, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.715910792350769, + "num_tokens": 154088080.0, + "step": 5950 + }, + { + "epoch": 0.6535251482538985, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.2990665435791016, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7186746597290039, + "num_tokens": 154111183.0, + "step": 5951 + }, + { + "epoch": 0.6536349659565122, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.2910008430480957, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7060357928276062, + "num_tokens": 154133988.0, + "step": 5952 + }, + { + "epoch": 0.6537447836591258, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.00758695602417, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7074872255325317, + "num_tokens": 154162949.0, + "step": 5953 + }, + { + "epoch": 0.6538546013617396, + "ewc_loss": 3.0994415283203125e-06, + "grad_norm": 2.057065010070801, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7035537958145142, + "num_tokens": 154191471.0, + "step": 5954 + }, + { + "epoch": 0.6539644190643532, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.2291877269744873, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7094489336013794, + "num_tokens": 154214934.0, + "step": 5955 + }, + { + "epoch": 0.6540742367669669, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0598514080047607, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7011868953704834, + "num_tokens": 154242775.0, + "step": 5956 + }, + { + "epoch": 0.6541840544695805, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.097454309463501, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6851046681404114, + "num_tokens": 154271659.0, + "step": 5957 + }, + { + "epoch": 0.6542938721721941, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.5042598247528076, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7356019616127014, + "num_tokens": 154291248.0, + "step": 5958 + }, + { + "epoch": 0.6544036898748078, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0660645961761475, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.7008719444274902, + "num_tokens": 154319496.0, + "step": 5959 + }, + { + "epoch": 0.6545135075774214, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0075652599334717, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7088453769683838, + "num_tokens": 154348439.0, + "step": 5960 + }, + { + "epoch": 0.6546233252800352, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.2176597118377686, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7278017997741699, + "num_tokens": 154372375.0, + "step": 5961 + }, + { + "epoch": 0.6547331429826488, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.1142139434814453, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7175784707069397, + "num_tokens": 154397081.0, + "step": 5962 + }, + { + "epoch": 0.6548429606852625, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.1974666118621826, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.6995417475700378, + "num_tokens": 154425629.0, + "step": 5963 + }, + { + "epoch": 0.6549527783878761, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.4367189407348633, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7279274463653564, + "num_tokens": 154447369.0, + "step": 5964 + }, + { + "epoch": 0.6550625960904898, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.471435785293579, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.7000610828399658, + "num_tokens": 154469352.0, + "step": 5965 + }, + { + "epoch": 0.6551724137931034, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.148050308227539, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6916482448577881, + "num_tokens": 154498264.0, + "step": 5966 + }, + { + "epoch": 0.6552822314957171, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.406886100769043, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7315965890884399, + "num_tokens": 154518816.0, + "step": 5967 + }, + { + "epoch": 0.6553920491983307, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 1.9213721752166748, + "learning_rate": 1e-06, + "loss": 1.0444, + "mean_token_accuracy": 0.6834051609039307, + "num_tokens": 154553216.0, + "step": 5968 + }, + { + "epoch": 0.6555018669009445, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.3550474643707275, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.738166868686676, + "num_tokens": 154573981.0, + "step": 5969 + }, + { + "epoch": 0.6556116846035581, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0508925914764404, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7018629312515259, + "num_tokens": 154602589.0, + "step": 5970 + }, + { + "epoch": 0.6557215023061718, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.599439859390259, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7050631642341614, + "num_tokens": 154621052.0, + "step": 5971 + }, + { + "epoch": 0.6558313200087854, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.2374427318573, + "learning_rate": 1e-06, + "loss": 1.0411, + "mean_token_accuracy": 0.6869435906410217, + "num_tokens": 154647699.0, + "step": 5972 + }, + { + "epoch": 0.6559411377113991, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.306968927383423, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7200117111206055, + "num_tokens": 154669577.0, + "step": 5973 + }, + { + "epoch": 0.6560509554140127, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.2460598945617676, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7135328650474548, + "num_tokens": 154695662.0, + "step": 5974 + }, + { + "epoch": 0.6561607731166264, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.1241674423217773, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6936978697776794, + "num_tokens": 154723545.0, + "step": 5975 + }, + { + "epoch": 0.6562705908192401, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0040652751922607, + "learning_rate": 1e-06, + "loss": 1.0224, + "mean_token_accuracy": 0.6870093941688538, + "num_tokens": 154755534.0, + "step": 5976 + }, + { + "epoch": 0.6563804085218538, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.17956805229187, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.6991140246391296, + "num_tokens": 154781348.0, + "step": 5977 + }, + { + "epoch": 0.6564902262244674, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.0847461223602295, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.696643590927124, + "num_tokens": 154810297.0, + "step": 5978 + }, + { + "epoch": 0.656600043927081, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.4055495262145996, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7129323482513428, + "num_tokens": 154831098.0, + "step": 5979 + }, + { + "epoch": 0.6567098616296947, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.1801254749298096, + "learning_rate": 1e-06, + "loss": 1.0345, + "mean_token_accuracy": 0.6820517182350159, + "num_tokens": 154857877.0, + "step": 5980 + }, + { + "epoch": 0.6568196793323083, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.2166600227355957, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7129040956497192, + "num_tokens": 154883350.0, + "step": 5981 + }, + { + "epoch": 0.656929497034922, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.1232709884643555, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.6936540603637695, + "num_tokens": 154909297.0, + "step": 5982 + }, + { + "epoch": 0.6570393147375357, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 2.014240264892578, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7105779051780701, + "num_tokens": 154936292.0, + "step": 5983 + }, + { + "epoch": 0.6571491324401494, + "ewc_loss": 3.11434268951416e-06, + "grad_norm": 1.9813646078109741, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6841756701469421, + "num_tokens": 154970211.0, + "step": 5984 + }, + { + "epoch": 0.657258950142763, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.520143985748291, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7190871834754944, + "num_tokens": 154989482.0, + "step": 5985 + }, + { + "epoch": 0.6573687678453767, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.104663133621216, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7072184085845947, + "num_tokens": 155017921.0, + "step": 5986 + }, + { + "epoch": 0.6574785855479903, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.2221145629882812, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6910864114761353, + "num_tokens": 155042883.0, + "step": 5987 + }, + { + "epoch": 0.657588403250604, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.087357521057129, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.6930201649665833, + "num_tokens": 155071816.0, + "step": 5988 + }, + { + "epoch": 0.6576982209532176, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.2699265480041504, + "learning_rate": 1e-06, + "loss": 1.046, + "mean_token_accuracy": 0.696247935295105, + "num_tokens": 155098747.0, + "step": 5989 + }, + { + "epoch": 0.6578080386558314, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.4770655632019043, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7204070091247559, + "num_tokens": 155119942.0, + "step": 5990 + }, + { + "epoch": 0.657917856358445, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.109351634979248, + "learning_rate": 1e-06, + "loss": 1.0503, + "mean_token_accuracy": 0.6844611763954163, + "num_tokens": 155146995.0, + "step": 5991 + }, + { + "epoch": 0.6580276740610587, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 1.9773484468460083, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6918841004371643, + "num_tokens": 155179232.0, + "step": 5992 + }, + { + "epoch": 0.6581374917636723, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.3467414379119873, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7020323276519775, + "num_tokens": 155202130.0, + "step": 5993 + }, + { + "epoch": 0.658247309466286, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.270899534225464, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6883591413497925, + "num_tokens": 155227196.0, + "step": 5994 + }, + { + "epoch": 0.6583571271688996, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.255847454071045, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6888974905014038, + "num_tokens": 155254350.0, + "step": 5995 + }, + { + "epoch": 0.6584669448715132, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.450204372406006, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.707961916923523, + "num_tokens": 155274055.0, + "step": 5996 + }, + { + "epoch": 0.6585767625741269, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.1241250038146973, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7051212787628174, + "num_tokens": 155299575.0, + "step": 5997 + }, + { + "epoch": 0.6586865802767407, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.432189702987671, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7055867910385132, + "num_tokens": 155322020.0, + "step": 5998 + }, + { + "epoch": 0.6587963979793543, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.025083065032959, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6932252645492554, + "num_tokens": 155351256.0, + "step": 5999 + }, + { + "epoch": 0.6589062156819679, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.0941050052642822, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6890835762023926, + "num_tokens": 155380192.0, + "step": 6000 + }, + { + "epoch": 0.6590160333845816, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.2084639072418213, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7013397812843323, + "num_tokens": 155404549.0, + "step": 6001 + }, + { + "epoch": 0.6591258510871952, + "ewc_loss": 3.129243850708008e-06, + "grad_norm": 2.080737590789795, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7023361921310425, + "num_tokens": 155432136.0, + "step": 6002 + }, + { + "epoch": 0.6592356687898089, + "ewc_loss": 3.1441450119018555e-06, + "grad_norm": 2.0328454971313477, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.7008086442947388, + "num_tokens": 155462102.0, + "step": 6003 + }, + { + "epoch": 0.6593454864924225, + "ewc_loss": 3.1441450119018555e-06, + "grad_norm": 2.1779351234436035, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7052116990089417, + "num_tokens": 155486045.0, + "step": 6004 + }, + { + "epoch": 0.6594553041950363, + "ewc_loss": 3.1441450119018555e-06, + "grad_norm": 2.297646999359131, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6989924907684326, + "num_tokens": 155510272.0, + "step": 6005 + }, + { + "epoch": 0.6595651218976499, + "ewc_loss": 3.1441450119018555e-06, + "grad_norm": 1.9089404344558716, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7240834832191467, + "num_tokens": 155541765.0, + "step": 6006 + }, + { + "epoch": 0.6596749396002636, + "ewc_loss": 3.1441450119018555e-06, + "grad_norm": 1.9315897226333618, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.702752947807312, + "num_tokens": 155574579.0, + "step": 6007 + }, + { + "epoch": 0.6597847573028772, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.017139196395874, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6928298473358154, + "num_tokens": 155603237.0, + "step": 6008 + }, + { + "epoch": 0.6598945750054909, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.4888010025024414, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7084188461303711, + "num_tokens": 155623349.0, + "step": 6009 + }, + { + "epoch": 0.6600043927081045, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.1587541103363037, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.684895396232605, + "num_tokens": 155651660.0, + "step": 6010 + }, + { + "epoch": 0.6601142104107182, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.219512939453125, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7045797109603882, + "num_tokens": 155676112.0, + "step": 6011 + }, + { + "epoch": 0.6602240281133319, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 1.8845064640045166, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7212080359458923, + "num_tokens": 155706650.0, + "step": 6012 + }, + { + "epoch": 0.6603338458159456, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 1.813033103942871, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7068621516227722, + "num_tokens": 155742884.0, + "step": 6013 + }, + { + "epoch": 0.6604436635185592, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.0301947593688965, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7272298336029053, + "num_tokens": 155769598.0, + "step": 6014 + }, + { + "epoch": 0.6605534812211729, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.2701942920684814, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6982440948486328, + "num_tokens": 155792342.0, + "step": 6015 + }, + { + "epoch": 0.6606632989237865, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.066011667251587, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7044745087623596, + "num_tokens": 155820917.0, + "step": 6016 + }, + { + "epoch": 0.6607731166264001, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.133246898651123, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.716564416885376, + "num_tokens": 155846192.0, + "step": 6017 + }, + { + "epoch": 0.6608829343290138, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.3057827949523926, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.693834125995636, + "num_tokens": 155869072.0, + "step": 6018 + }, + { + "epoch": 0.6609927520316276, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.146649122238159, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6938129663467407, + "num_tokens": 155898142.0, + "step": 6019 + }, + { + "epoch": 0.6611025697342412, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.097402811050415, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7202472686767578, + "num_tokens": 155925699.0, + "step": 6020 + }, + { + "epoch": 0.6612123874368548, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.273862600326538, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6939583420753479, + "num_tokens": 155950313.0, + "step": 6021 + }, + { + "epoch": 0.6613222051394685, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.04353928565979, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7183047533035278, + "num_tokens": 155977689.0, + "step": 6022 + }, + { + "epoch": 0.6614320228420821, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.1912264823913574, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7064178586006165, + "num_tokens": 156003254.0, + "step": 6023 + }, + { + "epoch": 0.6615418405446958, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.2199718952178955, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.72062087059021, + "num_tokens": 156027004.0, + "step": 6024 + }, + { + "epoch": 0.6616516582473094, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.086679697036743, + "learning_rate": 1e-06, + "loss": 1.0386, + "mean_token_accuracy": 0.6826652884483337, + "num_tokens": 156057555.0, + "step": 6025 + }, + { + "epoch": 0.6617614759499231, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.17157244682312, + "learning_rate": 1e-06, + "loss": 0.9956, + "mean_token_accuracy": 0.6940469741821289, + "num_tokens": 156083938.0, + "step": 6026 + }, + { + "epoch": 0.6618712936525368, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 1.9710556268692017, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7066234350204468, + "num_tokens": 156115291.0, + "step": 6027 + }, + { + "epoch": 0.6619811113551505, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.0378174781799316, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7179744243621826, + "num_tokens": 156141710.0, + "step": 6028 + }, + { + "epoch": 0.6620909290577641, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.1300902366638184, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6987959146499634, + "num_tokens": 156167447.0, + "step": 6029 + }, + { + "epoch": 0.6622007467603778, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.023829936981201, + "learning_rate": 1e-06, + "loss": 1.065, + "mean_token_accuracy": 0.6824718713760376, + "num_tokens": 156197415.0, + "step": 6030 + }, + { + "epoch": 0.6623105644629914, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 1.731727957725525, + "learning_rate": 1e-06, + "loss": 1.076, + "mean_token_accuracy": 0.6792298555374146, + "num_tokens": 156235692.0, + "step": 6031 + }, + { + "epoch": 0.6624203821656051, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.0578742027282715, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7159833908081055, + "num_tokens": 156263420.0, + "step": 6032 + }, + { + "epoch": 0.6625301998682187, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.3245301246643066, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7230023741722107, + "num_tokens": 156285855.0, + "step": 6033 + }, + { + "epoch": 0.6626400175708325, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.072373151779175, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.702775239944458, + "num_tokens": 156315182.0, + "step": 6034 + }, + { + "epoch": 0.6627498352734461, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.162768602371216, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7158746719360352, + "num_tokens": 156340739.0, + "step": 6035 + }, + { + "epoch": 0.6628596529760598, + "ewc_loss": 3.159046173095703e-06, + "grad_norm": 2.1410582065582275, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.698868453502655, + "num_tokens": 156367369.0, + "step": 6036 + }, + { + "epoch": 0.6629694706786734, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2139976024627686, + "learning_rate": 1e-06, + "loss": 1.0079, + "mean_token_accuracy": 0.694378137588501, + "num_tokens": 156393412.0, + "step": 6037 + }, + { + "epoch": 0.663079288381287, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.21098256111145, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7118067741394043, + "num_tokens": 156419102.0, + "step": 6038 + }, + { + "epoch": 0.6631891060839007, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.1033928394317627, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6937246322631836, + "num_tokens": 156446862.0, + "step": 6039 + }, + { + "epoch": 0.6632989237865143, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2464489936828613, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7070997953414917, + "num_tokens": 156471500.0, + "step": 6040 + }, + { + "epoch": 0.6634087414891281, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.150019884109497, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7083190679550171, + "num_tokens": 156497443.0, + "step": 6041 + }, + { + "epoch": 0.6635185591917417, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.0033981800079346, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.690544843673706, + "num_tokens": 156527999.0, + "step": 6042 + }, + { + "epoch": 0.6636283768943554, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.283940076828003, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7003103494644165, + "num_tokens": 156551836.0, + "step": 6043 + }, + { + "epoch": 0.663738194596969, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.110513925552368, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.6872810125350952, + "num_tokens": 156580317.0, + "step": 6044 + }, + { + "epoch": 0.6638480122995827, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.3258073329925537, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.717166006565094, + "num_tokens": 156603323.0, + "step": 6045 + }, + { + "epoch": 0.6639578300021963, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.0872554779052734, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7057455778121948, + "num_tokens": 156630293.0, + "step": 6046 + }, + { + "epoch": 0.66406764770481, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.0802431106567383, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6914640665054321, + "num_tokens": 156658155.0, + "step": 6047 + }, + { + "epoch": 0.6641774654074237, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.1109390258789062, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6965864896774292, + "num_tokens": 156685519.0, + "step": 6048 + }, + { + "epoch": 0.6642872831100374, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.0787696838378906, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.71571284532547, + "num_tokens": 156712257.0, + "step": 6049 + }, + { + "epoch": 0.664397100812651, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.258929967880249, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7029722332954407, + "num_tokens": 156735594.0, + "step": 6050 + }, + { + "epoch": 0.6645069185152647, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.15795636177063, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.698414146900177, + "num_tokens": 156760714.0, + "step": 6051 + }, + { + "epoch": 0.6646167362178783, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 1.9984780550003052, + "learning_rate": 1e-06, + "loss": 1.0951, + "mean_token_accuracy": 0.6812616586685181, + "num_tokens": 156791882.0, + "step": 6052 + }, + { + "epoch": 0.664726553920492, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2589352130889893, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6981067061424255, + "num_tokens": 156816170.0, + "step": 6053 + }, + { + "epoch": 0.6648363716231056, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2284252643585205, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7345259189605713, + "num_tokens": 156838617.0, + "step": 6054 + }, + { + "epoch": 0.6649461893257194, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.0318241119384766, + "learning_rate": 1e-06, + "loss": 1.0354, + "mean_token_accuracy": 0.6881067156791687, + "num_tokens": 156870103.0, + "step": 6055 + }, + { + "epoch": 0.665056007028333, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.237518787384033, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7084084749221802, + "num_tokens": 156895180.0, + "step": 6056 + }, + { + "epoch": 0.6651658247309467, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2955734729766846, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7061793208122253, + "num_tokens": 156919471.0, + "step": 6057 + }, + { + "epoch": 0.6652756424335603, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.140141010284424, + "learning_rate": 1e-06, + "loss": 1.0295, + "mean_token_accuracy": 0.7068925499916077, + "num_tokens": 156947861.0, + "step": 6058 + }, + { + "epoch": 0.665385460136174, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.5275156497955322, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7226204872131348, + "num_tokens": 156966379.0, + "step": 6059 + }, + { + "epoch": 0.6654952778387876, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.2763235569000244, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7394753694534302, + "num_tokens": 156988824.0, + "step": 6060 + }, + { + "epoch": 0.6656050955414012, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.3246750831604004, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6962131857872009, + "num_tokens": 157012835.0, + "step": 6061 + }, + { + "epoch": 0.6657149132440149, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.082576036453247, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7138444185256958, + "num_tokens": 157041090.0, + "step": 6062 + }, + { + "epoch": 0.6658247309466286, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.1618599891662598, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6910003423690796, + "num_tokens": 157069696.0, + "step": 6063 + }, + { + "epoch": 0.6659345486492423, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.4974100589752197, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7391660213470459, + "num_tokens": 157088566.0, + "step": 6064 + }, + { + "epoch": 0.6660443663518559, + "ewc_loss": 3.1739473342895508e-06, + "grad_norm": 2.1317052841186523, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7104974985122681, + "num_tokens": 157114916.0, + "step": 6065 + }, + { + "epoch": 0.6661541840544696, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.0922467708587646, + "learning_rate": 1e-06, + "loss": 1.0549, + "mean_token_accuracy": 0.6796132326126099, + "num_tokens": 157146212.0, + "step": 6066 + }, + { + "epoch": 0.6662640017570832, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4254610538482666, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7020256519317627, + "num_tokens": 157168951.0, + "step": 6067 + }, + { + "epoch": 0.6663738194596969, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.489898204803467, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7122334241867065, + "num_tokens": 157189670.0, + "step": 6068 + }, + { + "epoch": 0.6664836371623105, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4731900691986084, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7097488045692444, + "num_tokens": 157214174.0, + "step": 6069 + }, + { + "epoch": 0.6665934548649243, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1021275520324707, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6868475675582886, + "num_tokens": 157242797.0, + "step": 6070 + }, + { + "epoch": 0.6667032725675379, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.5699596405029297, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7056764960289001, + "num_tokens": 157262190.0, + "step": 6071 + }, + { + "epoch": 0.6668130902701516, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3352279663085938, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7165822982788086, + "num_tokens": 157283822.0, + "step": 6072 + }, + { + "epoch": 0.6669229079727652, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1929845809936523, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6961019039154053, + "num_tokens": 157308337.0, + "step": 6073 + }, + { + "epoch": 0.6670327256753789, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1002774238586426, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7112119197845459, + "num_tokens": 157332784.0, + "step": 6074 + }, + { + "epoch": 0.6671425433779925, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.149801731109619, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.688331663608551, + "num_tokens": 157360868.0, + "step": 6075 + }, + { + "epoch": 0.6672523610806061, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4279112815856934, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.698886513710022, + "num_tokens": 157382569.0, + "step": 6076 + }, + { + "epoch": 0.6673621787832199, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1050755977630615, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6946614384651184, + "num_tokens": 157409460.0, + "step": 6077 + }, + { + "epoch": 0.6674719964858336, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.0245919227600098, + "learning_rate": 1e-06, + "loss": 1.0397, + "mean_token_accuracy": 0.6882232427597046, + "num_tokens": 157440052.0, + "step": 6078 + }, + { + "epoch": 0.6675818141884472, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2961995601654053, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6957060098648071, + "num_tokens": 157464038.0, + "step": 6079 + }, + { + "epoch": 0.6676916318910608, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.987215280532837, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6979708075523376, + "num_tokens": 157492927.0, + "step": 6080 + }, + { + "epoch": 0.6678014495936745, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.9404295682907104, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7161927223205566, + "num_tokens": 157522518.0, + "step": 6081 + }, + { + "epoch": 0.6679112672962881, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2344369888305664, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7083353996276855, + "num_tokens": 157547170.0, + "step": 6082 + }, + { + "epoch": 0.6680210849989018, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.141233205795288, + "learning_rate": 1e-06, + "loss": 1.133, + "mean_token_accuracy": 0.6705778241157532, + "num_tokens": 157575044.0, + "step": 6083 + }, + { + "epoch": 0.6681309027015155, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.750751256942749, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6872001886367798, + "num_tokens": 157614962.0, + "step": 6084 + }, + { + "epoch": 0.6682407204041292, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.281761884689331, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7053548693656921, + "num_tokens": 157638797.0, + "step": 6085 + }, + { + "epoch": 0.6683505381067428, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.172745704650879, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7044950723648071, + "num_tokens": 157663629.0, + "step": 6086 + }, + { + "epoch": 0.6684603558093565, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3314168453216553, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7229228019714355, + "num_tokens": 157686851.0, + "step": 6087 + }, + { + "epoch": 0.6685701735119701, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.454047918319702, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7243509292602539, + "num_tokens": 157706324.0, + "step": 6088 + }, + { + "epoch": 0.6686799912145838, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3692145347595215, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.695267379283905, + "num_tokens": 157730442.0, + "step": 6089 + }, + { + "epoch": 0.6687898089171974, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.299224376678467, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7219929099082947, + "num_tokens": 157753769.0, + "step": 6090 + }, + { + "epoch": 0.6688996266198111, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.055264711380005, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6967576742172241, + "num_tokens": 157783437.0, + "step": 6091 + }, + { + "epoch": 0.6690094443224248, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.052284002304077, + "learning_rate": 1e-06, + "loss": 1.0715, + "mean_token_accuracy": 0.6730887293815613, + "num_tokens": 157813602.0, + "step": 6092 + }, + { + "epoch": 0.6691192620250385, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.0272347927093506, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7129061222076416, + "num_tokens": 157841665.0, + "step": 6093 + }, + { + "epoch": 0.6692290797276521, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3624565601348877, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7057985067367554, + "num_tokens": 157864011.0, + "step": 6094 + }, + { + "epoch": 0.6693388974302658, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.0173850059509277, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7096787691116333, + "num_tokens": 157893302.0, + "step": 6095 + }, + { + "epoch": 0.6694487151328794, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2535791397094727, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7214053869247437, + "num_tokens": 157917252.0, + "step": 6096 + }, + { + "epoch": 0.669558532835493, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.9847582578659058, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7067708969116211, + "num_tokens": 157946229.0, + "step": 6097 + }, + { + "epoch": 0.6696683505381067, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1992485523223877, + "learning_rate": 1e-06, + "loss": 1.0018, + "mean_token_accuracy": 0.6901103258132935, + "num_tokens": 157971749.0, + "step": 6098 + }, + { + "epoch": 0.6697781682407205, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.952790379524231, + "learning_rate": 1e-06, + "loss": 1.0515, + "mean_token_accuracy": 0.6820530891418457, + "num_tokens": 158004797.0, + "step": 6099 + }, + { + "epoch": 0.6698879859433341, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.126459836959839, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7031030654907227, + "num_tokens": 158032561.0, + "step": 6100 + }, + { + "epoch": 0.6699978036459477, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3220791816711426, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7296247482299805, + "num_tokens": 158054892.0, + "step": 6101 + }, + { + "epoch": 0.6701076213485614, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1407294273376465, + "learning_rate": 1e-06, + "loss": 1.0419, + "mean_token_accuracy": 0.6882915496826172, + "num_tokens": 158083531.0, + "step": 6102 + }, + { + "epoch": 0.670217439051175, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.434079885482788, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6946633458137512, + "num_tokens": 158105142.0, + "step": 6103 + }, + { + "epoch": 0.6703272567537887, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.596419334411621, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7074978351593018, + "num_tokens": 158124184.0, + "step": 6104 + }, + { + "epoch": 0.6704370744564023, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4536170959472656, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7187432646751404, + "num_tokens": 158144656.0, + "step": 6105 + }, + { + "epoch": 0.6705468921590161, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3561155796051025, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.706752598285675, + "num_tokens": 158168309.0, + "step": 6106 + }, + { + "epoch": 0.6706567098616297, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.389051914215088, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.722210168838501, + "num_tokens": 158189567.0, + "step": 6107 + }, + { + "epoch": 0.6707665275642434, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3976924419403076, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7090007066726685, + "num_tokens": 158210476.0, + "step": 6108 + }, + { + "epoch": 0.670876345266857, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.103358745574951, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7208524346351624, + "num_tokens": 158237155.0, + "step": 6109 + }, + { + "epoch": 0.6709861629694707, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.408884048461914, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6889129877090454, + "num_tokens": 158260792.0, + "step": 6110 + }, + { + "epoch": 0.6710959806720843, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4729185104370117, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7219332456588745, + "num_tokens": 158280874.0, + "step": 6111 + }, + { + "epoch": 0.671205798374698, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.276233196258545, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7200391888618469, + "num_tokens": 158304659.0, + "step": 6112 + }, + { + "epoch": 0.6713156160773117, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.0717926025390625, + "learning_rate": 1e-06, + "loss": 1.0564, + "mean_token_accuracy": 0.685141921043396, + "num_tokens": 158334472.0, + "step": 6113 + }, + { + "epoch": 0.6714254337799254, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.9910961389541626, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6872053146362305, + "num_tokens": 158367217.0, + "step": 6114 + }, + { + "epoch": 0.671535251482539, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.9930436611175537, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.6964578628540039, + "num_tokens": 158396304.0, + "step": 6115 + }, + { + "epoch": 0.6716450691851527, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1389715671539307, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6944389343261719, + "num_tokens": 158422496.0, + "step": 6116 + }, + { + "epoch": 0.6717548868877663, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3853087425231934, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7046787738800049, + "num_tokens": 158445506.0, + "step": 6117 + }, + { + "epoch": 0.67186470459038, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.027100086212158, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7143299579620361, + "num_tokens": 158473721.0, + "step": 6118 + }, + { + "epoch": 0.6719745222929936, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3089168071746826, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7159217596054077, + "num_tokens": 158496941.0, + "step": 6119 + }, + { + "epoch": 0.6720843399956072, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.042301654815674, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7211840152740479, + "num_tokens": 158525505.0, + "step": 6120 + }, + { + "epoch": 0.672194157698221, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2503888607025146, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7048750519752502, + "num_tokens": 158548342.0, + "step": 6121 + }, + { + "epoch": 0.6723039754008346, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.032628059387207, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6934928297996521, + "num_tokens": 158577693.0, + "step": 6122 + }, + { + "epoch": 0.6724137931034483, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2062113285064697, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7175784707069397, + "num_tokens": 158602480.0, + "step": 6123 + }, + { + "epoch": 0.6725236108060619, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.034588098526001, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7057656049728394, + "num_tokens": 158631339.0, + "step": 6124 + }, + { + "epoch": 0.6726334285086756, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.492886543273926, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7102799415588379, + "num_tokens": 158651126.0, + "step": 6125 + }, + { + "epoch": 0.6727432462112892, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.317018747329712, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7196966409683228, + "num_tokens": 158674262.0, + "step": 6126 + }, + { + "epoch": 0.6728530639139029, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.432772159576416, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.6962844133377075, + "num_tokens": 158695067.0, + "step": 6127 + }, + { + "epoch": 0.6729628816165166, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.4662230014801025, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7315429449081421, + "num_tokens": 158716763.0, + "step": 6128 + }, + { + "epoch": 0.6730726993191303, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1097347736358643, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7043857574462891, + "num_tokens": 158743441.0, + "step": 6129 + }, + { + "epoch": 0.6731825170217439, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.201138496398926, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7121260166168213, + "num_tokens": 158768551.0, + "step": 6130 + }, + { + "epoch": 0.6732923347243576, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.26802921295166, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.703163743019104, + "num_tokens": 158793558.0, + "step": 6131 + }, + { + "epoch": 0.6734021524269712, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.23333740234375, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7308810353279114, + "num_tokens": 158818430.0, + "step": 6132 + }, + { + "epoch": 0.6735119701295849, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.3086371421813965, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6944183707237244, + "num_tokens": 158843138.0, + "step": 6133 + }, + { + "epoch": 0.6736217878321985, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.070340394973755, + "learning_rate": 1e-06, + "loss": 1.0133, + "mean_token_accuracy": 0.7037621736526489, + "num_tokens": 158874477.0, + "step": 6134 + }, + { + "epoch": 0.6737316055348123, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.2207250595092773, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7117997407913208, + "num_tokens": 158899004.0, + "step": 6135 + }, + { + "epoch": 0.6738414232374259, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1857707500457764, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7052074670791626, + "num_tokens": 158924939.0, + "step": 6136 + }, + { + "epoch": 0.6739512409400396, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.178006410598755, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7191106677055359, + "num_tokens": 158950178.0, + "step": 6137 + }, + { + "epoch": 0.6740610586426532, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 1.9991528987884521, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7003663778305054, + "num_tokens": 158979755.0, + "step": 6138 + }, + { + "epoch": 0.6741708763452668, + "ewc_loss": 3.1888484954833984e-06, + "grad_norm": 2.1219677925109863, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7089434266090393, + "num_tokens": 159007944.0, + "step": 6139 + }, + { + "epoch": 0.6742806940478805, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.9750434160232544, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7195500135421753, + "num_tokens": 159037619.0, + "step": 6140 + }, + { + "epoch": 0.6743905117504941, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.118385076522827, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.715003252029419, + "num_tokens": 159065207.0, + "step": 6141 + }, + { + "epoch": 0.6745003294531079, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1713480949401855, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.7001955509185791, + "num_tokens": 159091637.0, + "step": 6142 + }, + { + "epoch": 0.6746101471557215, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.386268138885498, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7096209526062012, + "num_tokens": 159114372.0, + "step": 6143 + }, + { + "epoch": 0.6747199648583352, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1589126586914062, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6950167417526245, + "num_tokens": 159141680.0, + "step": 6144 + }, + { + "epoch": 0.6748297825609488, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.4666664600372314, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7157779932022095, + "num_tokens": 159162816.0, + "step": 6145 + }, + { + "epoch": 0.6749396002635625, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1608335971832275, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.6977254748344421, + "num_tokens": 159188403.0, + "step": 6146 + }, + { + "epoch": 0.6750494179661761, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.030672788619995, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.6964359283447266, + "num_tokens": 159218711.0, + "step": 6147 + }, + { + "epoch": 0.6751592356687898, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1438682079315186, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6875730752944946, + "num_tokens": 159246877.0, + "step": 6148 + }, + { + "epoch": 0.6752690533714034, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.0692250728607178, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7103504538536072, + "num_tokens": 159275042.0, + "step": 6149 + }, + { + "epoch": 0.6753788710740172, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.8648629188537598, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7150004506111145, + "num_tokens": 159307958.0, + "step": 6150 + }, + { + "epoch": 0.6754886887766308, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.4189751148223877, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7065447568893433, + "num_tokens": 159330280.0, + "step": 6151 + }, + { + "epoch": 0.6755985064792445, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.0690107345581055, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7017244100570679, + "num_tokens": 159358692.0, + "step": 6152 + }, + { + "epoch": 0.6757083241818581, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.0212326049804688, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.709208607673645, + "num_tokens": 159387486.0, + "step": 6153 + }, + { + "epoch": 0.6758181418844718, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2544872760772705, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7104603052139282, + "num_tokens": 159411417.0, + "step": 6154 + }, + { + "epoch": 0.6759279595870854, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.9996042251586914, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7113237380981445, + "num_tokens": 159442444.0, + "step": 6155 + }, + { + "epoch": 0.676037777289699, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.203613519668579, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.6992999911308289, + "num_tokens": 159466545.0, + "step": 6156 + }, + { + "epoch": 0.6761475949923128, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2496354579925537, + "learning_rate": 1e-06, + "loss": 1.0136, + "mean_token_accuracy": 0.6994190216064453, + "num_tokens": 159491951.0, + "step": 6157 + }, + { + "epoch": 0.6762574126949265, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.9491472244262695, + "learning_rate": 1e-06, + "loss": 1.1061, + "mean_token_accuracy": 0.6745033264160156, + "num_tokens": 159524107.0, + "step": 6158 + }, + { + "epoch": 0.6763672303975401, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.7664918899536133, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7131342887878418, + "num_tokens": 159560268.0, + "step": 6159 + }, + { + "epoch": 0.6764770481001537, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2614150047302246, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7405182123184204, + "num_tokens": 159582247.0, + "step": 6160 + }, + { + "epoch": 0.6765868658027674, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.205091714859009, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7142289876937866, + "num_tokens": 159608502.0, + "step": 6161 + }, + { + "epoch": 0.676696683505381, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.4192850589752197, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7230085134506226, + "num_tokens": 159629373.0, + "step": 6162 + }, + { + "epoch": 0.6768065012079947, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2630889415740967, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6922385096549988, + "num_tokens": 159654888.0, + "step": 6163 + }, + { + "epoch": 0.6769163189106084, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2144718170166016, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7149778604507446, + "num_tokens": 159678909.0, + "step": 6164 + }, + { + "epoch": 0.6770261366132221, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.5772881507873535, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7167901992797852, + "num_tokens": 159699296.0, + "step": 6165 + }, + { + "epoch": 0.6771359543158357, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.0535409450531006, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6927350163459778, + "num_tokens": 159728051.0, + "step": 6166 + }, + { + "epoch": 0.6772457720184494, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.225064992904663, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7167693376541138, + "num_tokens": 159752609.0, + "step": 6167 + }, + { + "epoch": 0.677355589721063, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.4797372817993164, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.698801577091217, + "num_tokens": 159775222.0, + "step": 6168 + }, + { + "epoch": 0.6774654074236767, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.4486961364746094, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6970803141593933, + "num_tokens": 159795594.0, + "step": 6169 + }, + { + "epoch": 0.6775752251262903, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2012763023376465, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.7008596062660217, + "num_tokens": 159821063.0, + "step": 6170 + }, + { + "epoch": 0.6776850428289041, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.961320400238037, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7117631435394287, + "num_tokens": 159850148.0, + "step": 6171 + }, + { + "epoch": 0.6777948605315177, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2582430839538574, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7143374681472778, + "num_tokens": 159875384.0, + "step": 6172 + }, + { + "epoch": 0.6779046782341314, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.207953453063965, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7099276781082153, + "num_tokens": 159898946.0, + "step": 6173 + }, + { + "epoch": 0.678014495936745, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.259958028793335, + "learning_rate": 1e-06, + "loss": 1.0512, + "mean_token_accuracy": 0.6823136210441589, + "num_tokens": 159925065.0, + "step": 6174 + }, + { + "epoch": 0.6781243136393587, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1874542236328125, + "learning_rate": 1e-06, + "loss": 1.0398, + "mean_token_accuracy": 0.6834568977355957, + "num_tokens": 159950968.0, + "step": 6175 + }, + { + "epoch": 0.6782341313419723, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.397233009338379, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6955662965774536, + "num_tokens": 159972968.0, + "step": 6176 + }, + { + "epoch": 0.678343949044586, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.9972004890441895, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7417862415313721, + "num_tokens": 160000824.0, + "step": 6177 + }, + { + "epoch": 0.6784537667471996, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.302658796310425, + "learning_rate": 1e-06, + "loss": 1.0415, + "mean_token_accuracy": 0.6838384866714478, + "num_tokens": 160026215.0, + "step": 6178 + }, + { + "epoch": 0.6785635844498134, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.2881715297698975, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7041142582893372, + "num_tokens": 160048907.0, + "step": 6179 + }, + { + "epoch": 0.678673402152427, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.0557990074157715, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6957746744155884, + "num_tokens": 160078943.0, + "step": 6180 + }, + { + "epoch": 0.6787832198550406, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.1762430667877197, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7110762596130371, + "num_tokens": 160105582.0, + "step": 6181 + }, + { + "epoch": 0.6788930375576543, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.274245500564575, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7072733640670776, + "num_tokens": 160128859.0, + "step": 6182 + }, + { + "epoch": 0.6790028552602679, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 1.9617756605148315, + "learning_rate": 1e-06, + "loss": 1.102, + "mean_token_accuracy": 0.6710343360900879, + "num_tokens": 160161771.0, + "step": 6183 + }, + { + "epoch": 0.6791126729628816, + "ewc_loss": 3.203749656677246e-06, + "grad_norm": 2.3766708374023438, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7116666436195374, + "num_tokens": 160185315.0, + "step": 6184 + }, + { + "epoch": 0.6792224906654952, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.1151716709136963, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.734265923500061, + "num_tokens": 160210093.0, + "step": 6185 + }, + { + "epoch": 0.679332308368109, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.2934062480926514, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6995235681533813, + "num_tokens": 160235947.0, + "step": 6186 + }, + { + "epoch": 0.6794421260707226, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.084465265274048, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7153356075286865, + "num_tokens": 160262839.0, + "step": 6187 + }, + { + "epoch": 0.6795519437733363, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.1059458255767822, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.7031964063644409, + "num_tokens": 160291066.0, + "step": 6188 + }, + { + "epoch": 0.6796617614759499, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.2227587699890137, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.6971691846847534, + "num_tokens": 160316101.0, + "step": 6189 + }, + { + "epoch": 0.6797715791785636, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.055283308029175, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7202628254890442, + "num_tokens": 160342531.0, + "step": 6190 + }, + { + "epoch": 0.6798813968811772, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.5342657566070557, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.709001362323761, + "num_tokens": 160361907.0, + "step": 6191 + }, + { + "epoch": 0.6799912145837909, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.2136476039886475, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7289952039718628, + "num_tokens": 160385012.0, + "step": 6192 + }, + { + "epoch": 0.6801010322864046, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 1.960170865058899, + "learning_rate": 1e-06, + "loss": 1.0772, + "mean_token_accuracy": 0.6773825883865356, + "num_tokens": 160418417.0, + "step": 6193 + }, + { + "epoch": 0.6802108499890183, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.6072492599487305, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7207328081130981, + "num_tokens": 160436585.0, + "step": 6194 + }, + { + "epoch": 0.6803206676916319, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0280253887176514, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6975119709968567, + "num_tokens": 160465543.0, + "step": 6195 + }, + { + "epoch": 0.6804304853942456, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.169696092605591, + "learning_rate": 1e-06, + "loss": 1.0028, + "mean_token_accuracy": 0.690991222858429, + "num_tokens": 160494434.0, + "step": 6196 + }, + { + "epoch": 0.6805403030968592, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.002516984939575, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.6990790367126465, + "num_tokens": 160522554.0, + "step": 6197 + }, + { + "epoch": 0.6806501207994728, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.363966464996338, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7034538388252258, + "num_tokens": 160545157.0, + "step": 6198 + }, + { + "epoch": 0.6807599385020865, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0084266662597656, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6959409713745117, + "num_tokens": 160573827.0, + "step": 6199 + }, + { + "epoch": 0.6808697562047002, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.594860076904297, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7231937050819397, + "num_tokens": 160592697.0, + "step": 6200 + }, + { + "epoch": 0.6809795739073139, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0451529026031494, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7070265412330627, + "num_tokens": 160621152.0, + "step": 6201 + }, + { + "epoch": 0.6810893916099275, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.038085460662842, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7066752910614014, + "num_tokens": 160649670.0, + "step": 6202 + }, + { + "epoch": 0.6811992093125412, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.4425535202026367, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7044948935508728, + "num_tokens": 160672004.0, + "step": 6203 + }, + { + "epoch": 0.6813090270151548, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0143818855285645, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.692022442817688, + "num_tokens": 160701671.0, + "step": 6204 + }, + { + "epoch": 0.6814188447177685, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.3957667350769043, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6921660900115967, + "num_tokens": 160724240.0, + "step": 6205 + }, + { + "epoch": 0.6815286624203821, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.176682710647583, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6949195265769958, + "num_tokens": 160750043.0, + "step": 6206 + }, + { + "epoch": 0.6816384801229959, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.28084135055542, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.6978617310523987, + "num_tokens": 160775094.0, + "step": 6207 + }, + { + "epoch": 0.6817482978256095, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0011706352233887, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7329812049865723, + "num_tokens": 160804642.0, + "step": 6208 + }, + { + "epoch": 0.6818581155282232, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.2721545696258545, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7140275239944458, + "num_tokens": 160828357.0, + "step": 6209 + }, + { + "epoch": 0.6819679332308368, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.2499001026153564, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7292968034744263, + "num_tokens": 160852946.0, + "step": 6210 + }, + { + "epoch": 0.6820777509334505, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.1860709190368652, + "learning_rate": 1e-06, + "loss": 1.044, + "mean_token_accuracy": 0.6876431107521057, + "num_tokens": 160880009.0, + "step": 6211 + }, + { + "epoch": 0.6821875686360641, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 1.9040755033493042, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7093722224235535, + "num_tokens": 160910735.0, + "step": 6212 + }, + { + "epoch": 0.6822973863386778, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.084901809692383, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6791408658027649, + "num_tokens": 160940187.0, + "step": 6213 + }, + { + "epoch": 0.6824072040412914, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0434505939483643, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7174085974693298, + "num_tokens": 160967733.0, + "step": 6214 + }, + { + "epoch": 0.6825170217439052, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.3080670833587646, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.715997040271759, + "num_tokens": 160991485.0, + "step": 6215 + }, + { + "epoch": 0.6826268394465188, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.08498215675354, + "learning_rate": 1e-06, + "loss": 1.0452, + "mean_token_accuracy": 0.6957954168319702, + "num_tokens": 161022484.0, + "step": 6216 + }, + { + "epoch": 0.6827366571491325, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.642900228500366, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7098771929740906, + "num_tokens": 161041677.0, + "step": 6217 + }, + { + "epoch": 0.6828464748517461, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 1.8967818021774292, + "learning_rate": 1e-06, + "loss": 1.0953, + "mean_token_accuracy": 0.6806614398956299, + "num_tokens": 161078083.0, + "step": 6218 + }, + { + "epoch": 0.6829562925543597, + "ewc_loss": 3.2186508178710938e-06, + "grad_norm": 2.0762462615966797, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7091119289398193, + "num_tokens": 161105164.0, + "step": 6219 + }, + { + "epoch": 0.6830661102569734, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.177420139312744, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.7003501653671265, + "num_tokens": 161134797.0, + "step": 6220 + }, + { + "epoch": 0.683175927959587, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 1.965225338935852, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7138295769691467, + "num_tokens": 161163605.0, + "step": 6221 + }, + { + "epoch": 0.6832857456622008, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.1915535926818848, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7086454629898071, + "num_tokens": 161189308.0, + "step": 6222 + }, + { + "epoch": 0.6833955633648144, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.3346259593963623, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7069015502929688, + "num_tokens": 161212337.0, + "step": 6223 + }, + { + "epoch": 0.6835053810674281, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.336496353149414, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.710071325302124, + "num_tokens": 161234858.0, + "step": 6224 + }, + { + "epoch": 0.6836151987700417, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.093132734298706, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7111285924911499, + "num_tokens": 161261948.0, + "step": 6225 + }, + { + "epoch": 0.6837250164726554, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.2035319805145264, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7110376358032227, + "num_tokens": 161285655.0, + "step": 6226 + }, + { + "epoch": 0.683834834175269, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.073460578918457, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7076624631881714, + "num_tokens": 161313042.0, + "step": 6227 + }, + { + "epoch": 0.6839446518778827, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.303086042404175, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.6997104287147522, + "num_tokens": 161339421.0, + "step": 6228 + }, + { + "epoch": 0.6840544695804964, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 1.913567304611206, + "learning_rate": 1e-06, + "loss": 1.0606, + "mean_token_accuracy": 0.6871569156646729, + "num_tokens": 161374361.0, + "step": 6229 + }, + { + "epoch": 0.6841642872831101, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.131654739379883, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7059599161148071, + "num_tokens": 161400878.0, + "step": 6230 + }, + { + "epoch": 0.6842741049857237, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.1094534397125244, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.7076614499092102, + "num_tokens": 161428464.0, + "step": 6231 + }, + { + "epoch": 0.6843839226883374, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.3408641815185547, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7321023344993591, + "num_tokens": 161450688.0, + "step": 6232 + }, + { + "epoch": 0.684493740390951, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.3406362533569336, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7055053114891052, + "num_tokens": 161472646.0, + "step": 6233 + }, + { + "epoch": 0.6846035580935647, + "ewc_loss": 3.2335519790649414e-06, + "grad_norm": 2.3786604404449463, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6977355480194092, + "num_tokens": 161495999.0, + "step": 6234 + }, + { + "epoch": 0.6847133757961783, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.2397384643554688, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7204858064651489, + "num_tokens": 161518854.0, + "step": 6235 + }, + { + "epoch": 0.6848231934987921, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.3876516819000244, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7012390494346619, + "num_tokens": 161541927.0, + "step": 6236 + }, + { + "epoch": 0.6849330112014057, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.1776955127716064, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7070021629333496, + "num_tokens": 161568560.0, + "step": 6237 + }, + { + "epoch": 0.6850428289040194, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.384084701538086, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.705803394317627, + "num_tokens": 161592384.0, + "step": 6238 + }, + { + "epoch": 0.685152646606633, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.1224865913391113, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6913281679153442, + "num_tokens": 161620897.0, + "step": 6239 + }, + { + "epoch": 0.6852624643092466, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.307922124862671, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7181397676467896, + "num_tokens": 161644189.0, + "step": 6240 + }, + { + "epoch": 0.6853722820118603, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.203505754470825, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6925262212753296, + "num_tokens": 161668350.0, + "step": 6241 + }, + { + "epoch": 0.6854820997144739, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.3884100914001465, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7219083309173584, + "num_tokens": 161691163.0, + "step": 6242 + }, + { + "epoch": 0.6855919174170876, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.200350046157837, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7194942831993103, + "num_tokens": 161715801.0, + "step": 6243 + }, + { + "epoch": 0.6857017351197013, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.254835605621338, + "learning_rate": 1e-06, + "loss": 1.025, + "mean_token_accuracy": 0.6922776103019714, + "num_tokens": 161743236.0, + "step": 6244 + }, + { + "epoch": 0.685811552822315, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.0198397636413574, + "learning_rate": 1e-06, + "loss": 1.0885, + "mean_token_accuracy": 0.6763390898704529, + "num_tokens": 161772576.0, + "step": 6245 + }, + { + "epoch": 0.6859213705249286, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.170398712158203, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7089306712150574, + "num_tokens": 161798925.0, + "step": 6246 + }, + { + "epoch": 0.6860311882275423, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.493131637573242, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7119131088256836, + "num_tokens": 161820941.0, + "step": 6247 + }, + { + "epoch": 0.6861410059301559, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.228950262069702, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.6969942450523376, + "num_tokens": 161846081.0, + "step": 6248 + }, + { + "epoch": 0.6862508236327696, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.103698492050171, + "learning_rate": 1e-06, + "loss": 1.1013, + "mean_token_accuracy": 0.6718341112136841, + "num_tokens": 161875326.0, + "step": 6249 + }, + { + "epoch": 0.6863606413353832, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.201880931854248, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.709128737449646, + "num_tokens": 161902576.0, + "step": 6250 + }, + { + "epoch": 0.686470459037997, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.7027194499969482, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.718887984752655, + "num_tokens": 161920422.0, + "step": 6251 + }, + { + "epoch": 0.6865802767406106, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.242483615875244, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7173328399658203, + "num_tokens": 161943729.0, + "step": 6252 + }, + { + "epoch": 0.6866900944432243, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.1178858280181885, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6964018940925598, + "num_tokens": 161970782.0, + "step": 6253 + }, + { + "epoch": 0.6867999121458379, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 1.8211982250213623, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6900076866149902, + "num_tokens": 162007033.0, + "step": 6254 + }, + { + "epoch": 0.6869097298484516, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.471470832824707, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7126759886741638, + "num_tokens": 162027881.0, + "step": 6255 + }, + { + "epoch": 0.6870195475510652, + "ewc_loss": 3.248453140258789e-06, + "grad_norm": 2.017678737640381, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.7006371021270752, + "num_tokens": 162058289.0, + "step": 6256 + }, + { + "epoch": 0.6871293652536788, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.413306474685669, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7268964052200317, + "num_tokens": 162079047.0, + "step": 6257 + }, + { + "epoch": 0.6872391829562926, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.379221200942993, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.704923152923584, + "num_tokens": 162100921.0, + "step": 6258 + }, + { + "epoch": 0.6873490006589063, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.248741626739502, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6992913484573364, + "num_tokens": 162125491.0, + "step": 6259 + }, + { + "epoch": 0.6874588183615199, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.192783832550049, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6974554657936096, + "num_tokens": 162150785.0, + "step": 6260 + }, + { + "epoch": 0.6875686360641335, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.3030805587768555, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6951929330825806, + "num_tokens": 162173188.0, + "step": 6261 + }, + { + "epoch": 0.6876784537667472, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 1.9269157648086548, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7231008410453796, + "num_tokens": 162203116.0, + "step": 6262 + }, + { + "epoch": 0.6877882714693608, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.062554121017456, + "learning_rate": 1e-06, + "loss": 1.004, + "mean_token_accuracy": 0.6998752951622009, + "num_tokens": 162230643.0, + "step": 6263 + }, + { + "epoch": 0.6878980891719745, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.1376681327819824, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.6935072541236877, + "num_tokens": 162259943.0, + "step": 6264 + }, + { + "epoch": 0.6880079068745882, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.2243082523345947, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7058076858520508, + "num_tokens": 162284553.0, + "step": 6265 + }, + { + "epoch": 0.6881177245772019, + "ewc_loss": 3.2782554626464844e-06, + "grad_norm": 2.311260938644409, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6960427761077881, + "num_tokens": 162309858.0, + "step": 6266 + }, + { + "epoch": 0.6882275422798155, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.1002371311187744, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7036772966384888, + "num_tokens": 162337622.0, + "step": 6267 + }, + { + "epoch": 0.6883373599824292, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.0516557693481445, + "learning_rate": 1e-06, + "loss": 1.0365, + "mean_token_accuracy": 0.6970582008361816, + "num_tokens": 162366547.0, + "step": 6268 + }, + { + "epoch": 0.6884471776850428, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.003587007522583, + "learning_rate": 1e-06, + "loss": 1.0452, + "mean_token_accuracy": 0.681804358959198, + "num_tokens": 162396141.0, + "step": 6269 + }, + { + "epoch": 0.6885569953876565, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.2995831966400146, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7185471057891846, + "num_tokens": 162420169.0, + "step": 6270 + }, + { + "epoch": 0.6886668130902701, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.2224526405334473, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6935659646987915, + "num_tokens": 162445872.0, + "step": 6271 + }, + { + "epoch": 0.6887766307928838, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.0267763137817383, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7045558094978333, + "num_tokens": 162473104.0, + "step": 6272 + }, + { + "epoch": 0.6888864484954975, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.1533637046813965, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7216140031814575, + "num_tokens": 162498431.0, + "step": 6273 + }, + { + "epoch": 0.6889962661981112, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 1.9822502136230469, + "learning_rate": 1e-06, + "loss": 1.0649, + "mean_token_accuracy": 0.6806003451347351, + "num_tokens": 162528782.0, + "step": 6274 + }, + { + "epoch": 0.6891060839007248, + "ewc_loss": 3.293156623840332e-06, + "grad_norm": 2.208204746246338, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7016507983207703, + "num_tokens": 162553110.0, + "step": 6275 + }, + { + "epoch": 0.6892159016033385, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.0832910537719727, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6903698444366455, + "num_tokens": 162581838.0, + "step": 6276 + }, + { + "epoch": 0.6893257193059521, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2801713943481445, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6905006170272827, + "num_tokens": 162604965.0, + "step": 6277 + }, + { + "epoch": 0.6894355370085657, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.039716958999634, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7122256755828857, + "num_tokens": 162634386.0, + "step": 6278 + }, + { + "epoch": 0.6895453547111794, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.297261953353882, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.6826635599136353, + "num_tokens": 162660520.0, + "step": 6279 + }, + { + "epoch": 0.6896551724137931, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.3523097038269043, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7213937044143677, + "num_tokens": 162683482.0, + "step": 6280 + }, + { + "epoch": 0.6897649901164068, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.0626182556152344, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7082730531692505, + "num_tokens": 162712341.0, + "step": 6281 + }, + { + "epoch": 0.6898748078190204, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2286126613616943, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7199890613555908, + "num_tokens": 162735854.0, + "step": 6282 + }, + { + "epoch": 0.6899846255216341, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 1.9810622930526733, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7040917277336121, + "num_tokens": 162767452.0, + "step": 6283 + }, + { + "epoch": 0.6900944432242477, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.4559295177459717, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7215338945388794, + "num_tokens": 162788127.0, + "step": 6284 + }, + { + "epoch": 0.6902042609268614, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.0380234718322754, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7075960636138916, + "num_tokens": 162817399.0, + "step": 6285 + }, + { + "epoch": 0.690314078629475, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2919554710388184, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6855930089950562, + "num_tokens": 162842073.0, + "step": 6286 + }, + { + "epoch": 0.6904238963320888, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.217907190322876, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.732594907283783, + "num_tokens": 162865066.0, + "step": 6287 + }, + { + "epoch": 0.6905337140347024, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2573866844177246, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7111573815345764, + "num_tokens": 162889656.0, + "step": 6288 + }, + { + "epoch": 0.6906435317373161, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2182059288024902, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7131377458572388, + "num_tokens": 162913759.0, + "step": 6289 + }, + { + "epoch": 0.6907533494399297, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.162733316421509, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7124412059783936, + "num_tokens": 162939871.0, + "step": 6290 + }, + { + "epoch": 0.6908631671425434, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.11624813079834, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.7043336629867554, + "num_tokens": 162968388.0, + "step": 6291 + }, + { + "epoch": 0.690972984845157, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.3503000736236572, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7248902320861816, + "num_tokens": 162991016.0, + "step": 6292 + }, + { + "epoch": 0.6910828025477707, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.2234296798706055, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7361992597579956, + "num_tokens": 163014896.0, + "step": 6293 + }, + { + "epoch": 0.6911926202503844, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.4928791522979736, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7216593027114868, + "num_tokens": 163035705.0, + "step": 6294 + }, + { + "epoch": 0.6913024379529981, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.1243841648101807, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6820423603057861, + "num_tokens": 163064708.0, + "step": 6295 + }, + { + "epoch": 0.6914122556556117, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.104034185409546, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.67976975440979, + "num_tokens": 163092567.0, + "step": 6296 + }, + { + "epoch": 0.6915220733582254, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.1727817058563232, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7170435190200806, + "num_tokens": 163117372.0, + "step": 6297 + }, + { + "epoch": 0.691631891060839, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.1407389640808105, + "learning_rate": 1e-06, + "loss": 1.0763, + "mean_token_accuracy": 0.6769906282424927, + "num_tokens": 163142418.0, + "step": 6298 + }, + { + "epoch": 0.6917417087634526, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.0309510231018066, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6939103603363037, + "num_tokens": 163172080.0, + "step": 6299 + }, + { + "epoch": 0.6918515264660663, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.1379599571228027, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7006977200508118, + "num_tokens": 163197984.0, + "step": 6300 + }, + { + "epoch": 0.6919613441686799, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.105225086212158, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6963487863540649, + "num_tokens": 163226464.0, + "step": 6301 + }, + { + "epoch": 0.6920711618712937, + "ewc_loss": 3.3229589462280273e-06, + "grad_norm": 2.0905916690826416, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7113894820213318, + "num_tokens": 163256158.0, + "step": 6302 + }, + { + "epoch": 0.6921809795739073, + "ewc_loss": 3.3229589462280273e-06, + "grad_norm": 2.4259705543518066, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7000910639762878, + "num_tokens": 163277834.0, + "step": 6303 + }, + { + "epoch": 0.692290797276521, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 2.3960511684417725, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.715251088142395, + "num_tokens": 163299028.0, + "step": 6304 + }, + { + "epoch": 0.6924006149791346, + "ewc_loss": 3.3080577850341797e-06, + "grad_norm": 1.9535608291625977, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6882153749465942, + "num_tokens": 163329095.0, + "step": 6305 + }, + { + "epoch": 0.6925104326817483, + "ewc_loss": 3.3229589462280273e-06, + "grad_norm": 2.099194049835205, + "learning_rate": 1e-06, + "loss": 1.0593, + "mean_token_accuracy": 0.6817069053649902, + "num_tokens": 163357278.0, + "step": 6306 + }, + { + "epoch": 0.6926202503843619, + "ewc_loss": 3.3229589462280273e-06, + "grad_norm": 2.218426465988159, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7106719017028809, + "num_tokens": 163381779.0, + "step": 6307 + }, + { + "epoch": 0.6927300680869756, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1523656845092773, + "learning_rate": 1e-06, + "loss": 1.0884, + "mean_token_accuracy": 0.6770423650741577, + "num_tokens": 163415146.0, + "step": 6308 + }, + { + "epoch": 0.6928398857895893, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.222111701965332, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6951402425765991, + "num_tokens": 163440947.0, + "step": 6309 + }, + { + "epoch": 0.692949703492203, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.4336326122283936, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7169232368469238, + "num_tokens": 163460802.0, + "step": 6310 + }, + { + "epoch": 0.6930595211948166, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.8415838479995728, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7011046409606934, + "num_tokens": 163496352.0, + "step": 6311 + }, + { + "epoch": 0.6931693388974303, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9631617069244385, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.7094061374664307, + "num_tokens": 163526044.0, + "step": 6312 + }, + { + "epoch": 0.6932791566000439, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.397148370742798, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6914901733398438, + "num_tokens": 163550244.0, + "step": 6313 + }, + { + "epoch": 0.6933889743026576, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2784271240234375, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6914281845092773, + "num_tokens": 163576341.0, + "step": 6314 + }, + { + "epoch": 0.6934987920052712, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0903501510620117, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7039104700088501, + "num_tokens": 163604671.0, + "step": 6315 + }, + { + "epoch": 0.693608609707885, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.301438331604004, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.6981537342071533, + "num_tokens": 163629666.0, + "step": 6316 + }, + { + "epoch": 0.6937184274104986, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.975371241569519, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.710842490196228, + "num_tokens": 163658034.0, + "step": 6317 + }, + { + "epoch": 0.6938282451131123, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.213427782058716, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7111266851425171, + "num_tokens": 163683740.0, + "step": 6318 + }, + { + "epoch": 0.6939380628157259, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2362163066864014, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7282812595367432, + "num_tokens": 163706240.0, + "step": 6319 + }, + { + "epoch": 0.6940478805183395, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.4761815071105957, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7208988666534424, + "num_tokens": 163726391.0, + "step": 6320 + }, + { + "epoch": 0.6941576982209532, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1083667278289795, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7151235938072205, + "num_tokens": 163754046.0, + "step": 6321 + }, + { + "epoch": 0.6942675159235668, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.291029214859009, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7061929106712341, + "num_tokens": 163779277.0, + "step": 6322 + }, + { + "epoch": 0.6943773336261806, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0213489532470703, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6881564259529114, + "num_tokens": 163811487.0, + "step": 6323 + }, + { + "epoch": 0.6944871513287942, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.114586353302002, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.6980403065681458, + "num_tokens": 163837897.0, + "step": 6324 + }, + { + "epoch": 0.6945969690314079, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.083425998687744, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6922264695167542, + "num_tokens": 163866188.0, + "step": 6325 + }, + { + "epoch": 0.6947067867340215, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.034482002258301, + "learning_rate": 1e-06, + "loss": 1.0897, + "mean_token_accuracy": 0.6818209290504456, + "num_tokens": 163897427.0, + "step": 6326 + }, + { + "epoch": 0.6948166044366352, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.034179210662842, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7189022898674011, + "num_tokens": 163926436.0, + "step": 6327 + }, + { + "epoch": 0.6949264221392488, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.070103883743286, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6893720030784607, + "num_tokens": 163956225.0, + "step": 6328 + }, + { + "epoch": 0.6950362398418625, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.214505672454834, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.7076225280761719, + "num_tokens": 163982151.0, + "step": 6329 + }, + { + "epoch": 0.6951460575444761, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2350544929504395, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7011682987213135, + "num_tokens": 164007634.0, + "step": 6330 + }, + { + "epoch": 0.6952558752470899, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9704667329788208, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7127166986465454, + "num_tokens": 164035131.0, + "step": 6331 + }, + { + "epoch": 0.6953656929497035, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9980922937393188, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7006500959396362, + "num_tokens": 164063187.0, + "step": 6332 + }, + { + "epoch": 0.6954755106523172, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.381838321685791, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7214246988296509, + "num_tokens": 164085148.0, + "step": 6333 + }, + { + "epoch": 0.6955853283549308, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2940120697021484, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6903765201568604, + "num_tokens": 164108696.0, + "step": 6334 + }, + { + "epoch": 0.6956951460575445, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.3024837970733643, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7164266109466553, + "num_tokens": 164132496.0, + "step": 6335 + }, + { + "epoch": 0.6958049637601581, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.196425199508667, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7071410417556763, + "num_tokens": 164158020.0, + "step": 6336 + }, + { + "epoch": 0.6959147814627717, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.4375240802764893, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7182384729385376, + "num_tokens": 164179998.0, + "step": 6337 + }, + { + "epoch": 0.6960245991653855, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.767420530319214, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7274877429008484, + "num_tokens": 164197247.0, + "step": 6338 + }, + { + "epoch": 0.6961344168679992, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1579134464263916, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7065709829330444, + "num_tokens": 164222013.0, + "step": 6339 + }, + { + "epoch": 0.6962442345706128, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.255664110183716, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.721076488494873, + "num_tokens": 164244311.0, + "step": 6340 + }, + { + "epoch": 0.6963540522732264, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0270421504974365, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7043509483337402, + "num_tokens": 164272798.0, + "step": 6341 + }, + { + "epoch": 0.6964638699758401, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.151183843612671, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7076510190963745, + "num_tokens": 164298367.0, + "step": 6342 + }, + { + "epoch": 0.6965736876784537, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0949485301971436, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6848818063735962, + "num_tokens": 164328980.0, + "step": 6343 + }, + { + "epoch": 0.6966835053810674, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2470836639404297, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7029148936271667, + "num_tokens": 164354971.0, + "step": 6344 + }, + { + "epoch": 0.6967933230836811, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9939990043640137, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7109465599060059, + "num_tokens": 164382484.0, + "step": 6345 + }, + { + "epoch": 0.6969031407862948, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.3485312461853027, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7134182453155518, + "num_tokens": 164406329.0, + "step": 6346 + }, + { + "epoch": 0.6970129584889084, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9301700592041016, + "learning_rate": 1e-06, + "loss": 1.0983, + "mean_token_accuracy": 0.6788079142570496, + "num_tokens": 164438435.0, + "step": 6347 + }, + { + "epoch": 0.6971227761915221, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.495230197906494, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7015088200569153, + "num_tokens": 164458132.0, + "step": 6348 + }, + { + "epoch": 0.6972325938941357, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2063939571380615, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7182118892669678, + "num_tokens": 164483148.0, + "step": 6349 + }, + { + "epoch": 0.6973424115967494, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.206129550933838, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7196405529975891, + "num_tokens": 164507414.0, + "step": 6350 + }, + { + "epoch": 0.697452229299363, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0475003719329834, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7075501680374146, + "num_tokens": 164535855.0, + "step": 6351 + }, + { + "epoch": 0.6975620470019768, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2954695224761963, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.714092493057251, + "num_tokens": 164558183.0, + "step": 6352 + }, + { + "epoch": 0.6976718647045904, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0718905925750732, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7092614769935608, + "num_tokens": 164584290.0, + "step": 6353 + }, + { + "epoch": 0.6977816824072041, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.6531882286071777, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7158000469207764, + "num_tokens": 164601322.0, + "step": 6354 + }, + { + "epoch": 0.6978915001098177, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.737013816833496, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7318864464759827, + "num_tokens": 164618813.0, + "step": 6355 + }, + { + "epoch": 0.6980013178124314, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1954686641693115, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7163584232330322, + "num_tokens": 164642036.0, + "step": 6356 + }, + { + "epoch": 0.698111135515045, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.91240394115448, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7114841341972351, + "num_tokens": 164674367.0, + "step": 6357 + }, + { + "epoch": 0.6982209532176586, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2735588550567627, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.713788628578186, + "num_tokens": 164696694.0, + "step": 6358 + }, + { + "epoch": 0.6983307709202724, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1221368312835693, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7302302122116089, + "num_tokens": 164722792.0, + "step": 6359 + }, + { + "epoch": 0.698440588622886, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2049753665924072, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7115459442138672, + "num_tokens": 164748014.0, + "step": 6360 + }, + { + "epoch": 0.6985504063254997, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2256510257720947, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6963237524032593, + "num_tokens": 164772019.0, + "step": 6361 + }, + { + "epoch": 0.6986602240281133, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0563979148864746, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6978683471679688, + "num_tokens": 164800535.0, + "step": 6362 + }, + { + "epoch": 0.698770041730727, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1772403717041016, + "learning_rate": 1e-06, + "loss": 1.0292, + "mean_token_accuracy": 0.6864675879478455, + "num_tokens": 164827307.0, + "step": 6363 + }, + { + "epoch": 0.6988798594333406, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.4081249237060547, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7177677154541016, + "num_tokens": 164850156.0, + "step": 6364 + }, + { + "epoch": 0.6989896771359543, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.181546211242676, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7094815969467163, + "num_tokens": 164876107.0, + "step": 6365 + }, + { + "epoch": 0.6990994948385679, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9962553977966309, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7018632888793945, + "num_tokens": 164905536.0, + "step": 6366 + }, + { + "epoch": 0.6992093125411817, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.238302230834961, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7194368243217468, + "num_tokens": 164930711.0, + "step": 6367 + }, + { + "epoch": 0.6993191302437953, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.91591477394104, + "learning_rate": 1e-06, + "loss": 1.0697, + "mean_token_accuracy": 0.6784393787384033, + "num_tokens": 164965215.0, + "step": 6368 + }, + { + "epoch": 0.699428947946409, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.400350332260132, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7031266093254089, + "num_tokens": 164987382.0, + "step": 6369 + }, + { + "epoch": 0.6995387656490226, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1734519004821777, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.689162015914917, + "num_tokens": 165016576.0, + "step": 6370 + }, + { + "epoch": 0.6996485833516363, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.1687514781951904, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6973406076431274, + "num_tokens": 165043318.0, + "step": 6371 + }, + { + "epoch": 0.6997584010542499, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.514530658721924, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7034859657287598, + "num_tokens": 165063912.0, + "step": 6372 + }, + { + "epoch": 0.6998682187568636, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.088740348815918, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.7006661891937256, + "num_tokens": 165091183.0, + "step": 6373 + }, + { + "epoch": 0.6999780364594773, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.081453323364258, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7222064733505249, + "num_tokens": 165117875.0, + "step": 6374 + }, + { + "epoch": 0.700087854162091, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.4490389823913574, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7218687534332275, + "num_tokens": 165139314.0, + "step": 6375 + }, + { + "epoch": 0.7001976718647046, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2630884647369385, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7166162729263306, + "num_tokens": 165164950.0, + "step": 6376 + }, + { + "epoch": 0.7003074895673183, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.220315933227539, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7102572321891785, + "num_tokens": 165190399.0, + "step": 6377 + }, + { + "epoch": 0.7004173072699319, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.491058826446533, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.7071062326431274, + "num_tokens": 165213424.0, + "step": 6378 + }, + { + "epoch": 0.7005271249725455, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.159543514251709, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.713373601436615, + "num_tokens": 165239322.0, + "step": 6379 + }, + { + "epoch": 0.7006369426751592, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.809256911277771, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6783485412597656, + "num_tokens": 165276415.0, + "step": 6380 + }, + { + "epoch": 0.700746760377773, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.144784450531006, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6877735257148743, + "num_tokens": 165304440.0, + "step": 6381 + }, + { + "epoch": 0.7008565780803866, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.231311559677124, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7169091701507568, + "num_tokens": 165329692.0, + "step": 6382 + }, + { + "epoch": 0.7009663957830002, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 1.9230155944824219, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7093837261199951, + "num_tokens": 165361978.0, + "step": 6383 + }, + { + "epoch": 0.7010762134856139, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0287506580352783, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7152820825576782, + "num_tokens": 165391661.0, + "step": 6384 + }, + { + "epoch": 0.7011860311882275, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.5013465881347656, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7067524194717407, + "num_tokens": 165412857.0, + "step": 6385 + }, + { + "epoch": 0.7012958488908412, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.440171718597412, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7275553941726685, + "num_tokens": 165434553.0, + "step": 6386 + }, + { + "epoch": 0.7014056665934548, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2075159549713135, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.722550630569458, + "num_tokens": 165458779.0, + "step": 6387 + }, + { + "epoch": 0.7015154842960686, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.0154571533203125, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7284016609191895, + "num_tokens": 165484387.0, + "step": 6388 + }, + { + "epoch": 0.7016253019986822, + "ewc_loss": 3.337860107421875e-06, + "grad_norm": 2.2304575443267822, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7107083797454834, + "num_tokens": 165508058.0, + "step": 6389 + }, + { + "epoch": 0.7017351197012959, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.141721487045288, + "learning_rate": 1e-06, + "loss": 1.0041, + "mean_token_accuracy": 0.6964857578277588, + "num_tokens": 165534481.0, + "step": 6390 + }, + { + "epoch": 0.7018449374039095, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.038450002670288, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7057368755340576, + "num_tokens": 165562167.0, + "step": 6391 + }, + { + "epoch": 0.7019547551065232, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.6264586448669434, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7282613515853882, + "num_tokens": 165580369.0, + "step": 6392 + }, + { + "epoch": 0.7020645728091368, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.15370512008667, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7014448642730713, + "num_tokens": 165606836.0, + "step": 6393 + }, + { + "epoch": 0.7021743905117505, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.411459445953369, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7032574415206909, + "num_tokens": 165629360.0, + "step": 6394 + }, + { + "epoch": 0.7022842082143641, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.267690420150757, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6940830945968628, + "num_tokens": 165653418.0, + "step": 6395 + }, + { + "epoch": 0.7023940259169779, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1194660663604736, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7322542071342468, + "num_tokens": 165678775.0, + "step": 6396 + }, + { + "epoch": 0.7025038436195915, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.15374493598938, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7115209102630615, + "num_tokens": 165704460.0, + "step": 6397 + }, + { + "epoch": 0.7026136613222052, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.163410186767578, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7188733816146851, + "num_tokens": 165731850.0, + "step": 6398 + }, + { + "epoch": 0.7027234790248188, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.192462921142578, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6996462345123291, + "num_tokens": 165757577.0, + "step": 6399 + }, + { + "epoch": 0.7028332967274324, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.076497793197632, + "learning_rate": 1e-06, + "loss": 1.0762, + "mean_token_accuracy": 0.6773558855056763, + "num_tokens": 165787726.0, + "step": 6400 + }, + { + "epoch": 0.7029431144300461, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.4133687019348145, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7066848278045654, + "num_tokens": 165810931.0, + "step": 6401 + }, + { + "epoch": 0.7030529321326597, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.241591215133667, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6929563283920288, + "num_tokens": 165837238.0, + "step": 6402 + }, + { + "epoch": 0.7031627498352735, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 1.9836483001708984, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7080744504928589, + "num_tokens": 165867065.0, + "step": 6403 + }, + { + "epoch": 0.7032725675378871, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1137571334838867, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.694356381893158, + "num_tokens": 165895370.0, + "step": 6404 + }, + { + "epoch": 0.7033823852405008, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 1.9785652160644531, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6937012076377869, + "num_tokens": 165928520.0, + "step": 6405 + }, + { + "epoch": 0.7034922029431144, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.3042964935302734, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.69096839427948, + "num_tokens": 165952076.0, + "step": 6406 + }, + { + "epoch": 0.7036020206457281, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1140928268432617, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7065955400466919, + "num_tokens": 165978496.0, + "step": 6407 + }, + { + "epoch": 0.7037118383483417, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.27237868309021, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7047728896141052, + "num_tokens": 166000883.0, + "step": 6408 + }, + { + "epoch": 0.7038216560509554, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.0656368732452393, + "learning_rate": 1e-06, + "loss": 1.059, + "mean_token_accuracy": 0.6828435659408569, + "num_tokens": 166029905.0, + "step": 6409 + }, + { + "epoch": 0.7039314737535691, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.2620742321014404, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7041018605232239, + "num_tokens": 166052445.0, + "step": 6410 + }, + { + "epoch": 0.7040412914561828, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1611509323120117, + "learning_rate": 1e-06, + "loss": 0.995, + "mean_token_accuracy": 0.701494038105011, + "num_tokens": 166080139.0, + "step": 6411 + }, + { + "epoch": 0.7041511091587964, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.159878969192505, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.7061800956726074, + "num_tokens": 166109391.0, + "step": 6412 + }, + { + "epoch": 0.7042609268614101, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.07924747467041, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6795291900634766, + "num_tokens": 166138729.0, + "step": 6413 + }, + { + "epoch": 0.7043707445640237, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.3244316577911377, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6858923435211182, + "num_tokens": 166164711.0, + "step": 6414 + }, + { + "epoch": 0.7044805622666374, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.6988251209259033, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7208089828491211, + "num_tokens": 166182875.0, + "step": 6415 + }, + { + "epoch": 0.704590379969251, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.219203472137451, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.7053410410881042, + "num_tokens": 166209426.0, + "step": 6416 + }, + { + "epoch": 0.7047001976718648, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.253770351409912, + "learning_rate": 1e-06, + "loss": 1.0237, + "mean_token_accuracy": 0.6825309991836548, + "num_tokens": 166234077.0, + "step": 6417 + }, + { + "epoch": 0.7048100153744784, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.0520427227020264, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7034042477607727, + "num_tokens": 166263591.0, + "step": 6418 + }, + { + "epoch": 0.704919833077092, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1969876289367676, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7093054056167603, + "num_tokens": 166287278.0, + "step": 6419 + }, + { + "epoch": 0.7050296507797057, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.2789223194122314, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.734094500541687, + "num_tokens": 166309647.0, + "step": 6420 + }, + { + "epoch": 0.7051394684823193, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.1268680095672607, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.6870129108428955, + "num_tokens": 166336724.0, + "step": 6421 + }, + { + "epoch": 0.705249286184933, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.374190330505371, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7061818242073059, + "num_tokens": 166361211.0, + "step": 6422 + }, + { + "epoch": 0.7053591038875466, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.076796293258667, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7273387908935547, + "num_tokens": 166385822.0, + "step": 6423 + }, + { + "epoch": 0.7054689215901603, + "ewc_loss": 3.3676624298095703e-06, + "grad_norm": 2.27901029586792, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7450933456420898, + "num_tokens": 166407055.0, + "step": 6424 + }, + { + "epoch": 0.705578739292774, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.258568048477173, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6985747814178467, + "num_tokens": 166430708.0, + "step": 6425 + }, + { + "epoch": 0.7056885569953877, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.13128399848938, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7040482759475708, + "num_tokens": 166456779.0, + "step": 6426 + }, + { + "epoch": 0.7057983746980013, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.1465301513671875, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.711793065071106, + "num_tokens": 166481233.0, + "step": 6427 + }, + { + "epoch": 0.705908192400615, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.516418933868408, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7214368581771851, + "num_tokens": 166501202.0, + "step": 6428 + }, + { + "epoch": 0.7060180101032286, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 1.942837119102478, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7027179002761841, + "num_tokens": 166531152.0, + "step": 6429 + }, + { + "epoch": 0.7061278278058423, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.32549786567688, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7180365324020386, + "num_tokens": 166553589.0, + "step": 6430 + }, + { + "epoch": 0.7062376455084559, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.200578451156616, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7126134634017944, + "num_tokens": 166578512.0, + "step": 6431 + }, + { + "epoch": 0.7063474632110697, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.4603512287139893, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7139723300933838, + "num_tokens": 166599150.0, + "step": 6432 + }, + { + "epoch": 0.7064572809136833, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.1832547187805176, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6885073184967041, + "num_tokens": 166626449.0, + "step": 6433 + }, + { + "epoch": 0.706567098616297, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.4132134914398193, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7241692543029785, + "num_tokens": 166648587.0, + "step": 6434 + }, + { + "epoch": 0.7066769163189106, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.344477653503418, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7222212553024292, + "num_tokens": 166670674.0, + "step": 6435 + }, + { + "epoch": 0.7067867340215243, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.3284075260162354, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.706927478313446, + "num_tokens": 166695813.0, + "step": 6436 + }, + { + "epoch": 0.7068965517241379, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.462827205657959, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7157415151596069, + "num_tokens": 166716119.0, + "step": 6437 + }, + { + "epoch": 0.7070063694267515, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 1.9371421337127686, + "learning_rate": 1e-06, + "loss": 1.0871, + "mean_token_accuracy": 0.6733366250991821, + "num_tokens": 166750666.0, + "step": 6438 + }, + { + "epoch": 0.7071161871293653, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.4010374546051025, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.708274245262146, + "num_tokens": 166772775.0, + "step": 6439 + }, + { + "epoch": 0.707226004831979, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.4031856060028076, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7239794731140137, + "num_tokens": 166792522.0, + "step": 6440 + }, + { + "epoch": 0.7073358225345926, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.080170154571533, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7027143239974976, + "num_tokens": 166821634.0, + "step": 6441 + }, + { + "epoch": 0.7074456402372062, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.28078293800354, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6938043832778931, + "num_tokens": 166847977.0, + "step": 6442 + }, + { + "epoch": 0.7075554579398199, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.289508581161499, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.7016822695732117, + "num_tokens": 166871906.0, + "step": 6443 + }, + { + "epoch": 0.7076652756424335, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 1.8695766925811768, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7050879001617432, + "num_tokens": 166903174.0, + "step": 6444 + }, + { + "epoch": 0.7077750933450472, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.128452777862549, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7270885705947876, + "num_tokens": 166927652.0, + "step": 6445 + }, + { + "epoch": 0.7078849110476609, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.230656623840332, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7089956998825073, + "num_tokens": 166951923.0, + "step": 6446 + }, + { + "epoch": 0.7079947287502746, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.4699010848999023, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7117543816566467, + "num_tokens": 166973547.0, + "step": 6447 + }, + { + "epoch": 0.7081045464528882, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.5157203674316406, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7293599843978882, + "num_tokens": 166992928.0, + "step": 6448 + }, + { + "epoch": 0.7082143641555019, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.319783926010132, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7183289527893066, + "num_tokens": 167014443.0, + "step": 6449 + }, + { + "epoch": 0.7083241818581155, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.091186761856079, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7018330693244934, + "num_tokens": 167040164.0, + "step": 6450 + }, + { + "epoch": 0.7084339995607292, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.1013240814208984, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6999175548553467, + "num_tokens": 167065933.0, + "step": 6451 + }, + { + "epoch": 0.7085438172633428, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.70737624168396, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7136434316635132, + "num_tokens": 167083429.0, + "step": 6452 + }, + { + "epoch": 0.7086536349659565, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.200716972351074, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7108596563339233, + "num_tokens": 167108697.0, + "step": 6453 + }, + { + "epoch": 0.7087634526685702, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 1.964725375175476, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7177263498306274, + "num_tokens": 167138333.0, + "step": 6454 + }, + { + "epoch": 0.7088732703711839, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.0725231170654297, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7027126550674438, + "num_tokens": 167166755.0, + "step": 6455 + }, + { + "epoch": 0.7089830880737975, + "ewc_loss": 3.3974647521972656e-06, + "grad_norm": 2.381223678588867, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7170522212982178, + "num_tokens": 167189785.0, + "step": 6456 + }, + { + "epoch": 0.7090929057764112, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.533792734146118, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.720233678817749, + "num_tokens": 167211804.0, + "step": 6457 + }, + { + "epoch": 0.7092027234790248, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.399996280670166, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.6995166540145874, + "num_tokens": 167233908.0, + "step": 6458 + }, + { + "epoch": 0.7093125411816384, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.26629900932312, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.709871232509613, + "num_tokens": 167258245.0, + "step": 6459 + }, + { + "epoch": 0.7094223588842521, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.135967969894409, + "learning_rate": 1e-06, + "loss": 1.0083, + "mean_token_accuracy": 0.6987786889076233, + "num_tokens": 167284185.0, + "step": 6460 + }, + { + "epoch": 0.7095321765868658, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.2233290672302246, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7114871740341187, + "num_tokens": 167309018.0, + "step": 6461 + }, + { + "epoch": 0.7096419942894795, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 1.972264051437378, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7122554779052734, + "num_tokens": 167337963.0, + "step": 6462 + }, + { + "epoch": 0.7097518119920931, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 1.9477306604385376, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6854184865951538, + "num_tokens": 167368817.0, + "step": 6463 + }, + { + "epoch": 0.7098616296947068, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.1076390743255615, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.7062196731567383, + "num_tokens": 167395048.0, + "step": 6464 + }, + { + "epoch": 0.7099714473973204, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 1.936604380607605, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7108215093612671, + "num_tokens": 167424637.0, + "step": 6465 + }, + { + "epoch": 0.7100812650999341, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 1.8408551216125488, + "learning_rate": 1e-06, + "loss": 1.0366, + "mean_token_accuracy": 0.6839815974235535, + "num_tokens": 167459751.0, + "step": 6466 + }, + { + "epoch": 0.7101910828025477, + "ewc_loss": 3.4123659133911133e-06, + "grad_norm": 2.1234471797943115, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7193784117698669, + "num_tokens": 167484847.0, + "step": 6467 + }, + { + "epoch": 0.7103009005051615, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.2378089427948, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7143946290016174, + "num_tokens": 167507780.0, + "step": 6468 + }, + { + "epoch": 0.7104107182077751, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.281764030456543, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7213671207427979, + "num_tokens": 167530304.0, + "step": 6469 + }, + { + "epoch": 0.7105205359103888, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.4268126487731934, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6968017816543579, + "num_tokens": 167551363.0, + "step": 6470 + }, + { + "epoch": 0.7106303536130024, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.028928279876709, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6844121813774109, + "num_tokens": 167582669.0, + "step": 6471 + }, + { + "epoch": 0.7107401713156161, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 1.9061256647109985, + "learning_rate": 1e-06, + "loss": 1.0315, + "mean_token_accuracy": 0.6898488402366638, + "num_tokens": 167613850.0, + "step": 6472 + }, + { + "epoch": 0.7108499890182297, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 1.897716999053955, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7057520151138306, + "num_tokens": 167646709.0, + "step": 6473 + }, + { + "epoch": 0.7109598067208434, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.0285933017730713, + "learning_rate": 1e-06, + "loss": 1.0174, + "mean_token_accuracy": 0.7123205661773682, + "num_tokens": 167675359.0, + "step": 6474 + }, + { + "epoch": 0.7110696244234571, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.3070523738861084, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7024015188217163, + "num_tokens": 167698157.0, + "step": 6475 + }, + { + "epoch": 0.7111794421260708, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.104625701904297, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6981927156448364, + "num_tokens": 167725839.0, + "step": 6476 + }, + { + "epoch": 0.7112892598286844, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.2245521545410156, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7259497046470642, + "num_tokens": 167749767.0, + "step": 6477 + }, + { + "epoch": 0.711399077531298, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.075732707977295, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6957802772521973, + "num_tokens": 167777055.0, + "step": 6478 + }, + { + "epoch": 0.7115088952339117, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.2963943481445312, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7281513214111328, + "num_tokens": 167799343.0, + "step": 6479 + }, + { + "epoch": 0.7116187129365253, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.2587509155273438, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7224167585372925, + "num_tokens": 167822339.0, + "step": 6480 + }, + { + "epoch": 0.711728530639139, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.138544797897339, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7279999256134033, + "num_tokens": 167847874.0, + "step": 6481 + }, + { + "epoch": 0.7118383483417526, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.280820608139038, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6956489086151123, + "num_tokens": 167871087.0, + "step": 6482 + }, + { + "epoch": 0.7119481660443664, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.115145683288574, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6928021907806396, + "num_tokens": 167898170.0, + "step": 6483 + }, + { + "epoch": 0.71205798374698, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.133589744567871, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7084488868713379, + "num_tokens": 167923656.0, + "step": 6484 + }, + { + "epoch": 0.7121678014495937, + "ewc_loss": 3.427267074584961e-06, + "grad_norm": 2.1436519622802734, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6875408291816711, + "num_tokens": 167949266.0, + "step": 6485 + }, + { + "epoch": 0.7122776191522073, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2046852111816406, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7101797461509705, + "num_tokens": 167972994.0, + "step": 6486 + }, + { + "epoch": 0.712387436854821, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2805140018463135, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7030830383300781, + "num_tokens": 167997392.0, + "step": 6487 + }, + { + "epoch": 0.7124972545574346, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.33152174949646, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6967233419418335, + "num_tokens": 168022751.0, + "step": 6488 + }, + { + "epoch": 0.7126070722600483, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.109036684036255, + "learning_rate": 1e-06, + "loss": 1.0634, + "mean_token_accuracy": 0.6863765716552734, + "num_tokens": 168052042.0, + "step": 6489 + }, + { + "epoch": 0.712716889962662, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.605264663696289, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7056339979171753, + "num_tokens": 168071391.0, + "step": 6490 + }, + { + "epoch": 0.7128267076652757, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.109544515609741, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6988139748573303, + "num_tokens": 168102465.0, + "step": 6491 + }, + { + "epoch": 0.7129365253678893, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.644129991531372, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.6996889710426331, + "num_tokens": 168122278.0, + "step": 6492 + }, + { + "epoch": 0.713046343070503, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.050945997238159, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7094606161117554, + "num_tokens": 168151102.0, + "step": 6493 + }, + { + "epoch": 0.7131561607731166, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2466681003570557, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7342103719711304, + "num_tokens": 168175128.0, + "step": 6494 + }, + { + "epoch": 0.7132659784757303, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 1.992087960243225, + "learning_rate": 1e-06, + "loss": 1.055, + "mean_token_accuracy": 0.6769928932189941, + "num_tokens": 168207639.0, + "step": 6495 + }, + { + "epoch": 0.7133757961783439, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2324483394622803, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.706298291683197, + "num_tokens": 168233243.0, + "step": 6496 + }, + { + "epoch": 0.7134856138809577, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2371342182159424, + "learning_rate": 1e-06, + "loss": 1.0774, + "mean_token_accuracy": 0.6788261532783508, + "num_tokens": 168260065.0, + "step": 6497 + }, + { + "epoch": 0.7135954315835713, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.26106595993042, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7227709889411926, + "num_tokens": 168284638.0, + "step": 6498 + }, + { + "epoch": 0.713705249286185, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2704923152923584, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7065243124961853, + "num_tokens": 168310957.0, + "step": 6499 + }, + { + "epoch": 0.7138150669887986, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.2930407524108887, + "learning_rate": 1e-06, + "loss": 1.0523, + "mean_token_accuracy": 0.6846673488616943, + "num_tokens": 168336000.0, + "step": 6500 + }, + { + "epoch": 0.7139248846914122, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.4329891204833984, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7111260294914246, + "num_tokens": 168358893.0, + "step": 6501 + }, + { + "epoch": 0.7140347023940259, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.19502329826355, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7279353737831116, + "num_tokens": 168382930.0, + "step": 6502 + }, + { + "epoch": 0.7141445200966395, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.1262717247009277, + "learning_rate": 1e-06, + "loss": 1.0244, + "mean_token_accuracy": 0.6924997568130493, + "num_tokens": 168410606.0, + "step": 6503 + }, + { + "epoch": 0.7142543377992533, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.041952610015869, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7078695297241211, + "num_tokens": 168441184.0, + "step": 6504 + }, + { + "epoch": 0.7143641555018669, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.4305315017700195, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7179256677627563, + "num_tokens": 168464355.0, + "step": 6505 + }, + { + "epoch": 0.7144739732044806, + "ewc_loss": 3.4421682357788086e-06, + "grad_norm": 2.0413968563079834, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7106114029884338, + "num_tokens": 168491266.0, + "step": 6506 + }, + { + "epoch": 0.7145837909070942, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.3736374378204346, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6945263147354126, + "num_tokens": 168516585.0, + "step": 6507 + }, + { + "epoch": 0.7146936086097079, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.7815651893615723, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6890019774436951, + "num_tokens": 168535196.0, + "step": 6508 + }, + { + "epoch": 0.7148034263123215, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.113166093826294, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7106198072433472, + "num_tokens": 168561111.0, + "step": 6509 + }, + { + "epoch": 0.7149132440149352, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.7370572090148926, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7267355918884277, + "num_tokens": 168578428.0, + "step": 6510 + }, + { + "epoch": 0.7150230617175488, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.0193350315093994, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7086694240570068, + "num_tokens": 168607648.0, + "step": 6511 + }, + { + "epoch": 0.7151328794201626, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.0685513019561768, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.695870041847229, + "num_tokens": 168636005.0, + "step": 6512 + }, + { + "epoch": 0.7152426971227762, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.0855164527893066, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6961114406585693, + "num_tokens": 168664419.0, + "step": 6513 + }, + { + "epoch": 0.7153525148253899, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.280672788619995, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.728123128414154, + "num_tokens": 168687639.0, + "step": 6514 + }, + { + "epoch": 0.7154623325280035, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.273359775543213, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7066099643707275, + "num_tokens": 168712456.0, + "step": 6515 + }, + { + "epoch": 0.7155721502306172, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.037750482559204, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7214369773864746, + "num_tokens": 168740946.0, + "step": 6516 + }, + { + "epoch": 0.7156819679332308, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.0644044876098633, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6889550685882568, + "num_tokens": 168769603.0, + "step": 6517 + }, + { + "epoch": 0.7157917856358444, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.149409532546997, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7263593673706055, + "num_tokens": 168795783.0, + "step": 6518 + }, + { + "epoch": 0.7159016033384582, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.743398666381836, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7326305508613586, + "num_tokens": 168813454.0, + "step": 6519 + }, + { + "epoch": 0.7160114210410718, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.4441421031951904, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7203879952430725, + "num_tokens": 168834759.0, + "step": 6520 + }, + { + "epoch": 0.7161212387436855, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.1749794483184814, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.688675045967102, + "num_tokens": 168861563.0, + "step": 6521 + }, + { + "epoch": 0.7162310564462991, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.0958662033081055, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7066129446029663, + "num_tokens": 168888884.0, + "step": 6522 + }, + { + "epoch": 0.7163408741489128, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.021865129470825, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7077795267105103, + "num_tokens": 168917774.0, + "step": 6523 + }, + { + "epoch": 0.7164506918515264, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.510258197784424, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7142957448959351, + "num_tokens": 168936506.0, + "step": 6524 + }, + { + "epoch": 0.7165605095541401, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.1323680877685547, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.697784423828125, + "num_tokens": 168963308.0, + "step": 6525 + }, + { + "epoch": 0.7166703272567538, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.474215269088745, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7280452847480774, + "num_tokens": 168985502.0, + "step": 6526 + }, + { + "epoch": 0.7167801449593675, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.345463752746582, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.738213300704956, + "num_tokens": 169006486.0, + "step": 6527 + }, + { + "epoch": 0.7168899626619811, + "ewc_loss": 3.471970558166504e-06, + "grad_norm": 2.2383692264556885, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7014281749725342, + "num_tokens": 169030895.0, + "step": 6528 + }, + { + "epoch": 0.7169997803645948, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.341683864593506, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.70335853099823, + "num_tokens": 169054078.0, + "step": 6529 + }, + { + "epoch": 0.7171095980672084, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.5196831226348877, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7195552587509155, + "num_tokens": 169072668.0, + "step": 6530 + }, + { + "epoch": 0.7172194157698221, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.1663100719451904, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6954593062400818, + "num_tokens": 169099329.0, + "step": 6531 + }, + { + "epoch": 0.7173292334724357, + "ewc_loss": 3.4868717193603516e-06, + "grad_norm": 2.4561336040496826, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7134895324707031, + "num_tokens": 169120903.0, + "step": 6532 + }, + { + "epoch": 0.7174390511750495, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.0920913219451904, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6890102028846741, + "num_tokens": 169149640.0, + "step": 6533 + }, + { + "epoch": 0.7175488688776631, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.389056921005249, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7130710482597351, + "num_tokens": 169170780.0, + "step": 6534 + }, + { + "epoch": 0.7176586865802768, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.0656750202178955, + "learning_rate": 1e-06, + "loss": 1.0456, + "mean_token_accuracy": 0.693941593170166, + "num_tokens": 169200303.0, + "step": 6535 + }, + { + "epoch": 0.7177685042828904, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 1.9594523906707764, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7253421545028687, + "num_tokens": 169229281.0, + "step": 6536 + }, + { + "epoch": 0.717878321985504, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.417642831802368, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7251190543174744, + "num_tokens": 169252296.0, + "step": 6537 + }, + { + "epoch": 0.7179881396881177, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.2485225200653076, + "learning_rate": 1e-06, + "loss": 1.0395, + "mean_token_accuracy": 0.6902482509613037, + "num_tokens": 169279224.0, + "step": 6538 + }, + { + "epoch": 0.7180979573907313, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.122917652130127, + "learning_rate": 1e-06, + "loss": 1.0151, + "mean_token_accuracy": 0.6994173526763916, + "num_tokens": 169307045.0, + "step": 6539 + }, + { + "epoch": 0.7182077750933451, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.1750423908233643, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7047396898269653, + "num_tokens": 169333309.0, + "step": 6540 + }, + { + "epoch": 0.7183175927959587, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.143242835998535, + "learning_rate": 1e-06, + "loss": 1.0003, + "mean_token_accuracy": 0.6911393404006958, + "num_tokens": 169362714.0, + "step": 6541 + }, + { + "epoch": 0.7184274104985724, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.218043804168701, + "learning_rate": 1e-06, + "loss": 1.0285, + "mean_token_accuracy": 0.6950489282608032, + "num_tokens": 169390450.0, + "step": 6542 + }, + { + "epoch": 0.718537228201186, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.0563666820526123, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.6959947347640991, + "num_tokens": 169419864.0, + "step": 6543 + }, + { + "epoch": 0.7186470459037997, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.2999231815338135, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7375261783599854, + "num_tokens": 169441594.0, + "step": 6544 + }, + { + "epoch": 0.7187568636064133, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.155181646347046, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.704466700553894, + "num_tokens": 169467608.0, + "step": 6545 + }, + { + "epoch": 0.718866681309027, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.172494888305664, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.6987656354904175, + "num_tokens": 169494768.0, + "step": 6546 + }, + { + "epoch": 0.7189764990116406, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.0547027587890625, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7023616433143616, + "num_tokens": 169521616.0, + "step": 6547 + }, + { + "epoch": 0.7190863167142544, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.4606192111968994, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7210488319396973, + "num_tokens": 169542364.0, + "step": 6548 + }, + { + "epoch": 0.719196134416868, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.0511810779571533, + "learning_rate": 1e-06, + "loss": 1.0623, + "mean_token_accuracy": 0.6786943674087524, + "num_tokens": 169574336.0, + "step": 6549 + }, + { + "epoch": 0.7193059521194817, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.956738829612732, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7027873396873474, + "num_tokens": 169602383.0, + "step": 6550 + }, + { + "epoch": 0.7194157698220953, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.2591044902801514, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7051258683204651, + "num_tokens": 169625680.0, + "step": 6551 + }, + { + "epoch": 0.719525587524709, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.2208921909332275, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7145825028419495, + "num_tokens": 169649483.0, + "step": 6552 + }, + { + "epoch": 0.7196354052273226, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.1300606727600098, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7172919511795044, + "num_tokens": 169676176.0, + "step": 6553 + }, + { + "epoch": 0.7197452229299363, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.3596131801605225, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7049250602722168, + "num_tokens": 169700230.0, + "step": 6554 + }, + { + "epoch": 0.71985504063255, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.2645857334136963, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6989660263061523, + "num_tokens": 169724249.0, + "step": 6555 + }, + { + "epoch": 0.7199648583351637, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.1874465942382812, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.7000939846038818, + "num_tokens": 169750728.0, + "step": 6556 + }, + { + "epoch": 0.7200746760377773, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.956993818283081, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.7018072605133057, + "num_tokens": 169782386.0, + "step": 6557 + }, + { + "epoch": 0.720184493740391, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.0178580284118652, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.721403956413269, + "num_tokens": 169809532.0, + "step": 6558 + }, + { + "epoch": 0.7202943114430046, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.0732619762420654, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6986322999000549, + "num_tokens": 169838593.0, + "step": 6559 + }, + { + "epoch": 0.7204041291456182, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.11842942237854, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7127223014831543, + "num_tokens": 169865464.0, + "step": 6560 + }, + { + "epoch": 0.7205139468482319, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.9880154132843018, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7250096797943115, + "num_tokens": 169894499.0, + "step": 6561 + }, + { + "epoch": 0.7206237645508456, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.095317840576172, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6951466798782349, + "num_tokens": 169922616.0, + "step": 6562 + }, + { + "epoch": 0.7207335822534593, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.1036922931671143, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6867398023605347, + "num_tokens": 169950672.0, + "step": 6563 + }, + { + "epoch": 0.7208433999560729, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.239264726638794, + "learning_rate": 1e-06, + "loss": 1.037, + "mean_token_accuracy": 0.6986737847328186, + "num_tokens": 169974535.0, + "step": 6564 + }, + { + "epoch": 0.7209532176586866, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.111614227294922, + "learning_rate": 1e-06, + "loss": 1.08, + "mean_token_accuracy": 0.6808591485023499, + "num_tokens": 170002421.0, + "step": 6565 + }, + { + "epoch": 0.7210630353613002, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.9346847534179688, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7182290554046631, + "num_tokens": 170032139.0, + "step": 6566 + }, + { + "epoch": 0.7211728530639139, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.4616425037384033, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7049461007118225, + "num_tokens": 170051814.0, + "step": 6567 + }, + { + "epoch": 0.7212826707665275, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.9710444211959839, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7137048244476318, + "num_tokens": 170080831.0, + "step": 6568 + }, + { + "epoch": 0.7213924884691413, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.0213873386383057, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.701984703540802, + "num_tokens": 170108149.0, + "step": 6569 + }, + { + "epoch": 0.7215023061717549, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.096691131591797, + "learning_rate": 1e-06, + "loss": 1.0523, + "mean_token_accuracy": 0.6863089799880981, + "num_tokens": 170137356.0, + "step": 6570 + }, + { + "epoch": 0.7216121238743686, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.280015707015991, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7104960083961487, + "num_tokens": 170161300.0, + "step": 6571 + }, + { + "epoch": 0.7217219415769822, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.232297897338867, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7264506816864014, + "num_tokens": 170184667.0, + "step": 6572 + }, + { + "epoch": 0.7218317592795959, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.9538239240646362, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.693295419216156, + "num_tokens": 170215863.0, + "step": 6573 + }, + { + "epoch": 0.7219415769822095, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.2690021991729736, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7139829397201538, + "num_tokens": 170239781.0, + "step": 6574 + }, + { + "epoch": 0.7220513946848232, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.092318058013916, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6935724020004272, + "num_tokens": 170266272.0, + "step": 6575 + }, + { + "epoch": 0.7221612123874368, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.299179792404175, + "learning_rate": 1e-06, + "loss": 1.074, + "mean_token_accuracy": 0.6751351356506348, + "num_tokens": 170293997.0, + "step": 6576 + }, + { + "epoch": 0.7222710300900506, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.243501663208008, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.7366417646408081, + "num_tokens": 170316894.0, + "step": 6577 + }, + { + "epoch": 0.7223808477926642, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.112691640853882, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7249242067337036, + "num_tokens": 170343035.0, + "step": 6578 + }, + { + "epoch": 0.7224906654952779, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.262486457824707, + "learning_rate": 1e-06, + "loss": 1.0368, + "mean_token_accuracy": 0.6986246109008789, + "num_tokens": 170368423.0, + "step": 6579 + }, + { + "epoch": 0.7226004831978915, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.158219337463379, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.724614679813385, + "num_tokens": 170393599.0, + "step": 6580 + }, + { + "epoch": 0.7227103009005051, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.115419387817383, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.727826714515686, + "num_tokens": 170419722.0, + "step": 6581 + }, + { + "epoch": 0.7228201186031188, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 2.2915971279144287, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7120702862739563, + "num_tokens": 170443024.0, + "step": 6582 + }, + { + "epoch": 0.7229299363057324, + "ewc_loss": 3.516674041748047e-06, + "grad_norm": 1.9919883012771606, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7104759812355042, + "num_tokens": 170471462.0, + "step": 6583 + }, + { + "epoch": 0.7230397540083462, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.047821283340454, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6844964027404785, + "num_tokens": 170502437.0, + "step": 6584 + }, + { + "epoch": 0.7231495717109598, + "ewc_loss": 3.5017728805541992e-06, + "grad_norm": 2.2794582843780518, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7191887497901917, + "num_tokens": 170525603.0, + "step": 6585 + }, + { + "epoch": 0.7232593894135735, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.048492193222046, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7033072710037231, + "num_tokens": 170555958.0, + "step": 6586 + }, + { + "epoch": 0.7233692071161871, + "ewc_loss": 3.5315752029418945e-06, + "grad_norm": 2.232921838760376, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7035685181617737, + "num_tokens": 170579904.0, + "step": 6587 + }, + { + "epoch": 0.7234790248188008, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.5956790447235107, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.718070924282074, + "num_tokens": 170599073.0, + "step": 6588 + }, + { + "epoch": 0.7235888425214144, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.0349538326263428, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7294265031814575, + "num_tokens": 170624812.0, + "step": 6589 + }, + { + "epoch": 0.7236986602240281, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.289152145385742, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6979732513427734, + "num_tokens": 170649253.0, + "step": 6590 + }, + { + "epoch": 0.7238084779266418, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.3064768314361572, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6945370435714722, + "num_tokens": 170672443.0, + "step": 6591 + }, + { + "epoch": 0.7239182956292555, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.9907336235046387, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6940999031066895, + "num_tokens": 170701394.0, + "step": 6592 + }, + { + "epoch": 0.7240281133318691, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.3495755195617676, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7096571922302246, + "num_tokens": 170725227.0, + "step": 6593 + }, + { + "epoch": 0.7241379310344828, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.426072120666504, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7088888883590698, + "num_tokens": 170747080.0, + "step": 6594 + }, + { + "epoch": 0.7242477487370964, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.1479132175445557, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7239933013916016, + "num_tokens": 170769740.0, + "step": 6595 + }, + { + "epoch": 0.72435756643971, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.2495548725128174, + "learning_rate": 1e-06, + "loss": 1.0253, + "mean_token_accuracy": 0.6916552782058716, + "num_tokens": 170795664.0, + "step": 6596 + }, + { + "epoch": 0.7244673841423237, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.9663571119308472, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7170891761779785, + "num_tokens": 170824411.0, + "step": 6597 + }, + { + "epoch": 0.7245772018449375, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.2453556060791016, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6883317828178406, + "num_tokens": 170850123.0, + "step": 6598 + }, + { + "epoch": 0.7246870195475511, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.362003803253174, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7033985257148743, + "num_tokens": 170873265.0, + "step": 6599 + }, + { + "epoch": 0.7247968372501647, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.1287477016448975, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.6996122002601624, + "num_tokens": 170899152.0, + "step": 6600 + }, + { + "epoch": 0.7249066549527784, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.054187536239624, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6981039643287659, + "num_tokens": 170929558.0, + "step": 6601 + }, + { + "epoch": 0.725016472655392, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.165903091430664, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7199625968933105, + "num_tokens": 170955891.0, + "step": 6602 + }, + { + "epoch": 0.7251262903580057, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.9992899894714355, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.7148827314376831, + "num_tokens": 170984853.0, + "step": 6603 + }, + { + "epoch": 0.7252361080606193, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.8987171649932861, + "learning_rate": 1e-06, + "loss": 1.0899, + "mean_token_accuracy": 0.6775668859481812, + "num_tokens": 171017608.0, + "step": 6604 + }, + { + "epoch": 0.725345925763233, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.29960560798645, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.6995694637298584, + "num_tokens": 171043941.0, + "step": 6605 + }, + { + "epoch": 0.7254557434658467, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.336080551147461, + "learning_rate": 1e-06, + "loss": 0.9994, + "mean_token_accuracy": 0.6946197748184204, + "num_tokens": 171067231.0, + "step": 6606 + }, + { + "epoch": 0.7255655611684604, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.022758960723877, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6994920372962952, + "num_tokens": 171094751.0, + "step": 6607 + }, + { + "epoch": 0.725675378871074, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.216052293777466, + "learning_rate": 1e-06, + "loss": 1.0406, + "mean_token_accuracy": 0.6922945976257324, + "num_tokens": 171121793.0, + "step": 6608 + }, + { + "epoch": 0.7257851965736877, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.197131633758545, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.6970380544662476, + "num_tokens": 171147151.0, + "step": 6609 + }, + { + "epoch": 0.7258950142763013, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.1163599491119385, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.7020143270492554, + "num_tokens": 171173165.0, + "step": 6610 + }, + { + "epoch": 0.726004831978915, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.2510948181152344, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7235720157623291, + "num_tokens": 171195083.0, + "step": 6611 + }, + { + "epoch": 0.7261146496815286, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.9144525527954102, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.688444972038269, + "num_tokens": 171227814.0, + "step": 6612 + }, + { + "epoch": 0.7262244673841424, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.1857426166534424, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6927098631858826, + "num_tokens": 171255157.0, + "step": 6613 + }, + { + "epoch": 0.726334285086756, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.142224073410034, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7097252011299133, + "num_tokens": 171279921.0, + "step": 6614 + }, + { + "epoch": 0.7264441027893697, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.268145799636841, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.6913673281669617, + "num_tokens": 171306367.0, + "step": 6615 + }, + { + "epoch": 0.7265539204919833, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.0783870220184326, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6909312605857849, + "num_tokens": 171335470.0, + "step": 6616 + }, + { + "epoch": 0.726663738194597, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.2842636108398438, + "learning_rate": 1e-06, + "loss": 0.9925, + "mean_token_accuracy": 0.6957228183746338, + "num_tokens": 171360291.0, + "step": 6617 + }, + { + "epoch": 0.7267735558972106, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.2136423587799072, + "learning_rate": 1e-06, + "loss": 1.0063, + "mean_token_accuracy": 0.7092711925506592, + "num_tokens": 171386241.0, + "step": 6618 + }, + { + "epoch": 0.7268833735998242, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.1646223068237305, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6898028254508972, + "num_tokens": 171412587.0, + "step": 6619 + }, + { + "epoch": 0.726993191302438, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.4087202548980713, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7048288583755493, + "num_tokens": 171434516.0, + "step": 6620 + }, + { + "epoch": 0.7271030090050516, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.0374011993408203, + "learning_rate": 1e-06, + "loss": 1.0352, + "mean_token_accuracy": 0.6874969601631165, + "num_tokens": 171464916.0, + "step": 6621 + }, + { + "epoch": 0.7272128267076653, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.272435188293457, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6969691514968872, + "num_tokens": 171489467.0, + "step": 6622 + }, + { + "epoch": 0.7273226444102789, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.627270460128784, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7393849492073059, + "num_tokens": 171506265.0, + "step": 6623 + }, + { + "epoch": 0.7274324621128926, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.3208072185516357, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7303544282913208, + "num_tokens": 171527160.0, + "step": 6624 + }, + { + "epoch": 0.7275422798155062, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.131317615509033, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7126086354255676, + "num_tokens": 171552198.0, + "step": 6625 + }, + { + "epoch": 0.7276520975181199, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.457749843597412, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7095129489898682, + "num_tokens": 171573001.0, + "step": 6626 + }, + { + "epoch": 0.7277619152207336, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 1.9621388912200928, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.69097900390625, + "num_tokens": 171602431.0, + "step": 6627 + }, + { + "epoch": 0.7278717329233473, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.077775001525879, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.7023894190788269, + "num_tokens": 171631809.0, + "step": 6628 + }, + { + "epoch": 0.7279815506259609, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.282158136367798, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7032228708267212, + "num_tokens": 171655463.0, + "step": 6629 + }, + { + "epoch": 0.7280913683285746, + "ewc_loss": 3.546476364135742e-06, + "grad_norm": 2.250609874725342, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.721106767654419, + "num_tokens": 171679233.0, + "step": 6630 + }, + { + "epoch": 0.7282011860311882, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2435379028320312, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7136889696121216, + "num_tokens": 171704875.0, + "step": 6631 + }, + { + "epoch": 0.7283110037338019, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.8682187795639038, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7150408029556274, + "num_tokens": 171736957.0, + "step": 6632 + }, + { + "epoch": 0.7284208214364155, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.07564377784729, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6830358505249023, + "num_tokens": 171765231.0, + "step": 6633 + }, + { + "epoch": 0.7285306391390292, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.346198797225952, + "learning_rate": 1e-06, + "loss": 1.0467, + "mean_token_accuracy": 0.6893301010131836, + "num_tokens": 171790280.0, + "step": 6634 + }, + { + "epoch": 0.7286404568416429, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.0204954147338867, + "learning_rate": 1e-06, + "loss": 1.0719, + "mean_token_accuracy": 0.6858872771263123, + "num_tokens": 171823396.0, + "step": 6635 + }, + { + "epoch": 0.7287502745442566, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.963579773902893, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7158183455467224, + "num_tokens": 171852932.0, + "step": 6636 + }, + { + "epoch": 0.7288600922468702, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.0242183208465576, + "learning_rate": 1e-06, + "loss": 1.0736, + "mean_token_accuracy": 0.676767110824585, + "num_tokens": 171882645.0, + "step": 6637 + }, + { + "epoch": 0.7289699099494839, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2563140392303467, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.689082682132721, + "num_tokens": 171907999.0, + "step": 6638 + }, + { + "epoch": 0.7290797276520975, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.462946653366089, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7197126746177673, + "num_tokens": 171929437.0, + "step": 6639 + }, + { + "epoch": 0.7291895453547111, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2554452419281006, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7203430533409119, + "num_tokens": 171952768.0, + "step": 6640 + }, + { + "epoch": 0.7292993630573248, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9954984188079834, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6977028250694275, + "num_tokens": 171982817.0, + "step": 6641 + }, + { + "epoch": 0.7294091807599385, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1868772506713867, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7055463790893555, + "num_tokens": 172008264.0, + "step": 6642 + }, + { + "epoch": 0.7295189984625522, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.998940110206604, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7009992003440857, + "num_tokens": 172037187.0, + "step": 6643 + }, + { + "epoch": 0.7296288161651658, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2968735694885254, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.7014549374580383, + "num_tokens": 172061265.0, + "step": 6644 + }, + { + "epoch": 0.7297386338677795, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2128801345825195, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7127302885055542, + "num_tokens": 172088106.0, + "step": 6645 + }, + { + "epoch": 0.7298484515703931, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2524776458740234, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.7003191709518433, + "num_tokens": 172111278.0, + "step": 6646 + }, + { + "epoch": 0.7299582692730068, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.0964012145996094, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6950314044952393, + "num_tokens": 172139161.0, + "step": 6647 + }, + { + "epoch": 0.7300680869756204, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.211738348007202, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6866347789764404, + "num_tokens": 172164497.0, + "step": 6648 + }, + { + "epoch": 0.7301779046782342, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9196308851242065, + "learning_rate": 1e-06, + "loss": 1.0284, + "mean_token_accuracy": 0.6896335482597351, + "num_tokens": 172195776.0, + "step": 6649 + }, + { + "epoch": 0.7302877223808478, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.105391025543213, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7147064805030823, + "num_tokens": 172222583.0, + "step": 6650 + }, + { + "epoch": 0.7303975400834615, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2388269901275635, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6950640082359314, + "num_tokens": 172247462.0, + "step": 6651 + }, + { + "epoch": 0.7305073577860751, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.084493637084961, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6937699913978577, + "num_tokens": 172275377.0, + "step": 6652 + }, + { + "epoch": 0.7306171754886888, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2639241218566895, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7262425422668457, + "num_tokens": 172297610.0, + "step": 6653 + }, + { + "epoch": 0.7307269931913024, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1277904510498047, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7134476900100708, + "num_tokens": 172322826.0, + "step": 6654 + }, + { + "epoch": 0.7308368108939161, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2831079959869385, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.699246883392334, + "num_tokens": 172347222.0, + "step": 6655 + }, + { + "epoch": 0.7309466285965298, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9822019338607788, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7271228432655334, + "num_tokens": 172373660.0, + "step": 6656 + }, + { + "epoch": 0.7310564462991435, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.985516905784607, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6951002478599548, + "num_tokens": 172402897.0, + "step": 6657 + }, + { + "epoch": 0.7311662640017571, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.3303229808807373, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7145644426345825, + "num_tokens": 172424908.0, + "step": 6658 + }, + { + "epoch": 0.7312760817043708, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.140108823776245, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7025060057640076, + "num_tokens": 172452343.0, + "step": 6659 + }, + { + "epoch": 0.7313858994069844, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.182499885559082, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7331180572509766, + "num_tokens": 172474687.0, + "step": 6660 + }, + { + "epoch": 0.731495717109598, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.0537588596343994, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7165136933326721, + "num_tokens": 172501083.0, + "step": 6661 + }, + { + "epoch": 0.7316055348122117, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.0494415760040283, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6949558258056641, + "num_tokens": 172529837.0, + "step": 6662 + }, + { + "epoch": 0.7317153525148253, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.8802319765090942, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7013369798660278, + "num_tokens": 172565242.0, + "step": 6663 + }, + { + "epoch": 0.7318251702174391, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1691293716430664, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7238979935646057, + "num_tokens": 172590053.0, + "step": 6664 + }, + { + "epoch": 0.7319349879200527, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9527297019958496, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7169046401977539, + "num_tokens": 172617737.0, + "step": 6665 + }, + { + "epoch": 0.7320448056226664, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.110963821411133, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7145472764968872, + "num_tokens": 172643719.0, + "step": 6666 + }, + { + "epoch": 0.73215462332528, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.067042589187622, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7120644450187683, + "num_tokens": 172670802.0, + "step": 6667 + }, + { + "epoch": 0.7322644410278937, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9361748695373535, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.7021752595901489, + "num_tokens": 172701329.0, + "step": 6668 + }, + { + "epoch": 0.7323742587305073, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.3573081493377686, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7011349201202393, + "num_tokens": 172723687.0, + "step": 6669 + }, + { + "epoch": 0.732484076433121, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 1.9815725088119507, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7015738487243652, + "num_tokens": 172752354.0, + "step": 6670 + }, + { + "epoch": 0.7325938941357347, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.305345296859741, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6998733878135681, + "num_tokens": 172776323.0, + "step": 6671 + }, + { + "epoch": 0.7327037118383484, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.243767261505127, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7164424061775208, + "num_tokens": 172800904.0, + "step": 6672 + }, + { + "epoch": 0.732813529540962, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.283885955810547, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7143186926841736, + "num_tokens": 172823370.0, + "step": 6673 + }, + { + "epoch": 0.7329233472435757, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.389965295791626, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.714207649230957, + "num_tokens": 172846258.0, + "step": 6674 + }, + { + "epoch": 0.7330331649461893, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.4712889194488525, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7051013112068176, + "num_tokens": 172867940.0, + "step": 6675 + }, + { + "epoch": 0.733142982648803, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.643544912338257, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.70464688539505, + "num_tokens": 172886700.0, + "step": 6676 + }, + { + "epoch": 0.7332528003514166, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1764683723449707, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7043493986129761, + "num_tokens": 172914941.0, + "step": 6677 + }, + { + "epoch": 0.7333626180540304, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1495044231414795, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7184666395187378, + "num_tokens": 172939971.0, + "step": 6678 + }, + { + "epoch": 0.733472435756644, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1490564346313477, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7109225392341614, + "num_tokens": 172968136.0, + "step": 6679 + }, + { + "epoch": 0.7335822534592576, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.1803348064422607, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7052791714668274, + "num_tokens": 172993683.0, + "step": 6680 + }, + { + "epoch": 0.7336920711618713, + "ewc_loss": 3.56137752532959e-06, + "grad_norm": 2.2884528636932373, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7012133598327637, + "num_tokens": 173018448.0, + "step": 6681 + }, + { + "epoch": 0.7338018888644849, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.172797918319702, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7026620507240295, + "num_tokens": 173044670.0, + "step": 6682 + }, + { + "epoch": 0.7339117065670986, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1965324878692627, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6943668127059937, + "num_tokens": 173072158.0, + "step": 6683 + }, + { + "epoch": 0.7340215242697122, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9696794748306274, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7109718918800354, + "num_tokens": 173101918.0, + "step": 6684 + }, + { + "epoch": 0.734131341972326, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.216948986053467, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6990799903869629, + "num_tokens": 173127160.0, + "step": 6685 + }, + { + "epoch": 0.7342411596749396, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.098693370819092, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6924149394035339, + "num_tokens": 173153811.0, + "step": 6686 + }, + { + "epoch": 0.7343509773775533, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.251330852508545, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.6870427131652832, + "num_tokens": 173178814.0, + "step": 6687 + }, + { + "epoch": 0.7344607950801669, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.391601324081421, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.693027675151825, + "num_tokens": 173200290.0, + "step": 6688 + }, + { + "epoch": 0.7345706127827806, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0815327167510986, + "learning_rate": 1e-06, + "loss": 1.0421, + "mean_token_accuracy": 0.693927526473999, + "num_tokens": 173226366.0, + "step": 6689 + }, + { + "epoch": 0.7346804304853942, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.262072801589966, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6978061199188232, + "num_tokens": 173253655.0, + "step": 6690 + }, + { + "epoch": 0.7347902481880079, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.52553391456604, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7206172943115234, + "num_tokens": 173275181.0, + "step": 6691 + }, + { + "epoch": 0.7349000658906216, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1364545822143555, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7104088068008423, + "num_tokens": 173302924.0, + "step": 6692 + }, + { + "epoch": 0.7350098835932353, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.208843469619751, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7004896402359009, + "num_tokens": 173327534.0, + "step": 6693 + }, + { + "epoch": 0.7351197012958489, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2124643325805664, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7210831642150879, + "num_tokens": 173352614.0, + "step": 6694 + }, + { + "epoch": 0.7352295189984626, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3903090953826904, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.706196129322052, + "num_tokens": 173374986.0, + "step": 6695 + }, + { + "epoch": 0.7353393367010762, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.642334222793579, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7053954601287842, + "num_tokens": 173394416.0, + "step": 6696 + }, + { + "epoch": 0.7354491544036899, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2861926555633545, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.719309389591217, + "num_tokens": 173417035.0, + "step": 6697 + }, + { + "epoch": 0.7355589721063035, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9355460405349731, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7167949080467224, + "num_tokens": 173447210.0, + "step": 6698 + }, + { + "epoch": 0.7356687898089171, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.163633346557617, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7043033838272095, + "num_tokens": 173473313.0, + "step": 6699 + }, + { + "epoch": 0.7357786075115309, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2594566345214844, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7390990257263184, + "num_tokens": 173495252.0, + "step": 6700 + }, + { + "epoch": 0.7358884252141445, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1836822032928467, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7263612151145935, + "num_tokens": 173521442.0, + "step": 6701 + }, + { + "epoch": 0.7359982429167582, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1695034503936768, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7233657836914062, + "num_tokens": 173547300.0, + "step": 6702 + }, + { + "epoch": 0.7361080606193718, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.117283344268799, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7141828536987305, + "num_tokens": 173576514.0, + "step": 6703 + }, + { + "epoch": 0.7362178783219855, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3241047859191895, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.7016176581382751, + "num_tokens": 173602159.0, + "step": 6704 + }, + { + "epoch": 0.7363276960245991, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1581292152404785, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7113074064254761, + "num_tokens": 173627353.0, + "step": 6705 + }, + { + "epoch": 0.7364375137272128, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.061285972595215, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7252621650695801, + "num_tokens": 173655914.0, + "step": 6706 + }, + { + "epoch": 0.7365473314298265, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1814441680908203, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7004546523094177, + "num_tokens": 173682089.0, + "step": 6707 + }, + { + "epoch": 0.7366571491324402, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.062767505645752, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7217282652854919, + "num_tokens": 173710899.0, + "step": 6708 + }, + { + "epoch": 0.7367669668350538, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3731906414031982, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.6977503299713135, + "num_tokens": 173734020.0, + "step": 6709 + }, + { + "epoch": 0.7368767845376675, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9836064577102661, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7174210548400879, + "num_tokens": 173763730.0, + "step": 6710 + }, + { + "epoch": 0.7369866022402811, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3181583881378174, + "learning_rate": 1e-06, + "loss": 1.0566, + "mean_token_accuracy": 0.6735620498657227, + "num_tokens": 173789322.0, + "step": 6711 + }, + { + "epoch": 0.7370964199428948, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9086905717849731, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6910442113876343, + "num_tokens": 173821838.0, + "step": 6712 + }, + { + "epoch": 0.7372062376455084, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1265063285827637, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7194605469703674, + "num_tokens": 173848132.0, + "step": 6713 + }, + { + "epoch": 0.7373160553481222, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.203113317489624, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7302735447883606, + "num_tokens": 173872793.0, + "step": 6714 + }, + { + "epoch": 0.7374258730507358, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.182209014892578, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7178057432174683, + "num_tokens": 173897476.0, + "step": 6715 + }, + { + "epoch": 0.7375356907533495, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2726476192474365, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7046856880187988, + "num_tokens": 173921544.0, + "step": 6716 + }, + { + "epoch": 0.7376455084559631, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.4502556324005127, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7191983461380005, + "num_tokens": 173942604.0, + "step": 6717 + }, + { + "epoch": 0.7377553261585768, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.214841842651367, + "learning_rate": 1e-06, + "loss": 0.9729, + "mean_token_accuracy": 0.707871675491333, + "num_tokens": 173967385.0, + "step": 6718 + }, + { + "epoch": 0.7378651438611904, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3157145977020264, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7161577939987183, + "num_tokens": 173989018.0, + "step": 6719 + }, + { + "epoch": 0.737974961563804, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2087249755859375, + "learning_rate": 1e-06, + "loss": 1.0583, + "mean_token_accuracy": 0.6890826225280762, + "num_tokens": 174014756.0, + "step": 6720 + }, + { + "epoch": 0.7380847792664178, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.51613712310791, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6954301595687866, + "num_tokens": 174036322.0, + "step": 6721 + }, + { + "epoch": 0.7381945969690314, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.135138511657715, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.731863260269165, + "num_tokens": 174061966.0, + "step": 6722 + }, + { + "epoch": 0.7383044146716451, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.5087318420410156, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7005873918533325, + "num_tokens": 174082133.0, + "step": 6723 + }, + { + "epoch": 0.7384142323742587, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9275788068771362, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7029892802238464, + "num_tokens": 174114587.0, + "step": 6724 + }, + { + "epoch": 0.7385240500768724, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.004141092300415, + "learning_rate": 1e-06, + "loss": 1.0424, + "mean_token_accuracy": 0.685550332069397, + "num_tokens": 174144910.0, + "step": 6725 + }, + { + "epoch": 0.738633867779486, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1922049522399902, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7071496844291687, + "num_tokens": 174170044.0, + "step": 6726 + }, + { + "epoch": 0.7387436854820997, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.163681983947754, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7303799390792847, + "num_tokens": 174194128.0, + "step": 6727 + }, + { + "epoch": 0.7388535031847133, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.404088020324707, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7310366630554199, + "num_tokens": 174213613.0, + "step": 6728 + }, + { + "epoch": 0.7389633208873271, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2038052082061768, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7202804684638977, + "num_tokens": 174239072.0, + "step": 6729 + }, + { + "epoch": 0.7390731385899407, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0762531757354736, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6944119930267334, + "num_tokens": 174266710.0, + "step": 6730 + }, + { + "epoch": 0.7391829562925544, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3149030208587646, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7082110643386841, + "num_tokens": 174289578.0, + "step": 6731 + }, + { + "epoch": 0.739292773995168, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2975356578826904, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7084265351295471, + "num_tokens": 174312357.0, + "step": 6732 + }, + { + "epoch": 0.7394025916977817, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.4812965393066406, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7114945650100708, + "num_tokens": 174333509.0, + "step": 6733 + }, + { + "epoch": 0.7395124094003953, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3293955326080322, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7261162400245667, + "num_tokens": 174354178.0, + "step": 6734 + }, + { + "epoch": 0.739622227103009, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3724138736724854, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7037365436553955, + "num_tokens": 174377254.0, + "step": 6735 + }, + { + "epoch": 0.7397320448056227, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2836132049560547, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.71478271484375, + "num_tokens": 174400581.0, + "step": 6736 + }, + { + "epoch": 0.7398418625082364, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9239907264709473, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7000908851623535, + "num_tokens": 174431436.0, + "step": 6737 + }, + { + "epoch": 0.73995168021085, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.312690496444702, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.7059328556060791, + "num_tokens": 174456476.0, + "step": 6738 + }, + { + "epoch": 0.7400614979134637, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3831231594085693, + "learning_rate": 1e-06, + "loss": 1.0305, + "mean_token_accuracy": 0.6895403861999512, + "num_tokens": 174479634.0, + "step": 6739 + }, + { + "epoch": 0.7401713156160773, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2276082038879395, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7063906192779541, + "num_tokens": 174504223.0, + "step": 6740 + }, + { + "epoch": 0.7402811333186909, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.012394905090332, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7263948321342468, + "num_tokens": 174532251.0, + "step": 6741 + }, + { + "epoch": 0.7403909510213046, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.5473170280456543, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7214912176132202, + "num_tokens": 174551363.0, + "step": 6742 + }, + { + "epoch": 0.7405007687239183, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.8722376823425293, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7127535343170166, + "num_tokens": 174585552.0, + "step": 6743 + }, + { + "epoch": 0.740610586426532, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2417590618133545, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6947143077850342, + "num_tokens": 174612006.0, + "step": 6744 + }, + { + "epoch": 0.7407204041291456, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.346862554550171, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7000160217285156, + "num_tokens": 174635438.0, + "step": 6745 + }, + { + "epoch": 0.7408302218317593, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0070831775665283, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.6999484300613403, + "num_tokens": 174663764.0, + "step": 6746 + }, + { + "epoch": 0.7409400395343729, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3105835914611816, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7112849950790405, + "num_tokens": 174686855.0, + "step": 6747 + }, + { + "epoch": 0.7410498572369866, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.264500856399536, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6948211193084717, + "num_tokens": 174712408.0, + "step": 6748 + }, + { + "epoch": 0.7411596749396002, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9644241333007812, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.6996797323226929, + "num_tokens": 174746301.0, + "step": 6749 + }, + { + "epoch": 0.741269492642214, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1437106132507324, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.70614093542099, + "num_tokens": 174772707.0, + "step": 6750 + }, + { + "epoch": 0.7413793103448276, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 1.9972480535507202, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7022062540054321, + "num_tokens": 174800887.0, + "step": 6751 + }, + { + "epoch": 0.7414891280474413, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.17330265045166, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7228084802627563, + "num_tokens": 174825484.0, + "step": 6752 + }, + { + "epoch": 0.7415989457500549, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0046372413635254, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6985687017440796, + "num_tokens": 174856788.0, + "step": 6753 + }, + { + "epoch": 0.7417087634526686, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.075665235519409, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7190136909484863, + "num_tokens": 174883870.0, + "step": 6754 + }, + { + "epoch": 0.7418185811552822, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1184024810791016, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7145661115646362, + "num_tokens": 174911277.0, + "step": 6755 + }, + { + "epoch": 0.7419283988578959, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1547019481658936, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7190876007080078, + "num_tokens": 174937395.0, + "step": 6756 + }, + { + "epoch": 0.7420382165605095, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2298214435577393, + "learning_rate": 1e-06, + "loss": 1.0349, + "mean_token_accuracy": 0.6875171661376953, + "num_tokens": 174962830.0, + "step": 6757 + }, + { + "epoch": 0.7421480342631233, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.117436408996582, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7009727954864502, + "num_tokens": 174989345.0, + "step": 6758 + }, + { + "epoch": 0.7422578519657369, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2908670902252197, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7208263874053955, + "num_tokens": 175012986.0, + "step": 6759 + }, + { + "epoch": 0.7423676696683505, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.579282760620117, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7106907963752747, + "num_tokens": 175032096.0, + "step": 6760 + }, + { + "epoch": 0.7424774873709642, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.53108286857605, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7325050830841064, + "num_tokens": 175051955.0, + "step": 6761 + }, + { + "epoch": 0.7425873050735778, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2045652866363525, + "learning_rate": 1e-06, + "loss": 1.071, + "mean_token_accuracy": 0.6804320812225342, + "num_tokens": 175078938.0, + "step": 6762 + }, + { + "epoch": 0.7426971227761915, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.123495101928711, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.722158670425415, + "num_tokens": 175104952.0, + "step": 6763 + }, + { + "epoch": 0.7428069404788051, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2473812103271484, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7006341218948364, + "num_tokens": 175129499.0, + "step": 6764 + }, + { + "epoch": 0.7429167581814189, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.295705795288086, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.6782705783843994, + "num_tokens": 175155643.0, + "step": 6765 + }, + { + "epoch": 0.7430265758840325, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.356454610824585, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6997774839401245, + "num_tokens": 175179215.0, + "step": 6766 + }, + { + "epoch": 0.7431363935866462, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.056600332260132, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.6988064646720886, + "num_tokens": 175208654.0, + "step": 6767 + }, + { + "epoch": 0.7432462112892598, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1398656368255615, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.7036667466163635, + "num_tokens": 175235570.0, + "step": 6768 + }, + { + "epoch": 0.7433560289918735, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1665687561035156, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6967589259147644, + "num_tokens": 175260521.0, + "step": 6769 + }, + { + "epoch": 0.7434658466944871, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.126436710357666, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7215372323989868, + "num_tokens": 175287083.0, + "step": 6770 + }, + { + "epoch": 0.7435756643971008, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3837156295776367, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7180556058883667, + "num_tokens": 175309073.0, + "step": 6771 + }, + { + "epoch": 0.7436854820997145, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.57930850982666, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7109988927841187, + "num_tokens": 175327506.0, + "step": 6772 + }, + { + "epoch": 0.7437952998023282, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.4710607528686523, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7101231217384338, + "num_tokens": 175350031.0, + "step": 6773 + }, + { + "epoch": 0.7439051175049418, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3043179512023926, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7131292819976807, + "num_tokens": 175371750.0, + "step": 6774 + }, + { + "epoch": 0.7440149352075555, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0650126934051514, + "learning_rate": 1e-06, + "loss": 1.05, + "mean_token_accuracy": 0.6804267168045044, + "num_tokens": 175398988.0, + "step": 6775 + }, + { + "epoch": 0.7441247529101691, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0675525665283203, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7101227045059204, + "num_tokens": 175427102.0, + "step": 6776 + }, + { + "epoch": 0.7442345706127828, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.135446548461914, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7076521515846252, + "num_tokens": 175454098.0, + "step": 6777 + }, + { + "epoch": 0.7443443883153964, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.063958168029785, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6914668679237366, + "num_tokens": 175481485.0, + "step": 6778 + }, + { + "epoch": 0.7444542060180102, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2720770835876465, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7074804306030273, + "num_tokens": 175505092.0, + "step": 6779 + }, + { + "epoch": 0.7445640237206238, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.2241063117980957, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.6992955803871155, + "num_tokens": 175529187.0, + "step": 6780 + }, + { + "epoch": 0.7446738414232374, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3386125564575195, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6893730759620667, + "num_tokens": 175554714.0, + "step": 6781 + }, + { + "epoch": 0.7447836591258511, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.1150152683258057, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.6989763975143433, + "num_tokens": 175582796.0, + "step": 6782 + }, + { + "epoch": 0.7448934768284647, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.310004234313965, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7043383121490479, + "num_tokens": 175606081.0, + "step": 6783 + }, + { + "epoch": 0.7450032945310784, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.135831594467163, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7223575115203857, + "num_tokens": 175632000.0, + "step": 6784 + }, + { + "epoch": 0.745113112233692, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0787909030914307, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7178588509559631, + "num_tokens": 175658364.0, + "step": 6785 + }, + { + "epoch": 0.7452229299363057, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3150603771209717, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7120568752288818, + "num_tokens": 175683050.0, + "step": 6786 + }, + { + "epoch": 0.7453327476389194, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.263723134994507, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.717704176902771, + "num_tokens": 175706060.0, + "step": 6787 + }, + { + "epoch": 0.7454425653415331, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.0331125259399414, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.716233491897583, + "num_tokens": 175734301.0, + "step": 6788 + }, + { + "epoch": 0.7455523830441467, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.263272762298584, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7232204675674438, + "num_tokens": 175757178.0, + "step": 6789 + }, + { + "epoch": 0.7456622007467604, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.493779420852661, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6904520988464355, + "num_tokens": 175778863.0, + "step": 6790 + }, + { + "epoch": 0.745772018449374, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.075779914855957, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7115943431854248, + "num_tokens": 175806023.0, + "step": 6791 + }, + { + "epoch": 0.7458818361519877, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.3830997943878174, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7056235074996948, + "num_tokens": 175828049.0, + "step": 6792 + }, + { + "epoch": 0.7459916538546013, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.541086196899414, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7110649347305298, + "num_tokens": 175848135.0, + "step": 6793 + }, + { + "epoch": 0.7461014715572151, + "ewc_loss": 3.5762786865234375e-06, + "grad_norm": 2.081026554107666, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7038600444793701, + "num_tokens": 175876498.0, + "step": 6794 + }, + { + "epoch": 0.7462112892598287, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.1981613636016846, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7007684707641602, + "num_tokens": 175902733.0, + "step": 6795 + }, + { + "epoch": 0.7463211069624424, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.392099380493164, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7109799385070801, + "num_tokens": 175925566.0, + "step": 6796 + }, + { + "epoch": 0.746430924665056, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.0507826805114746, + "learning_rate": 1e-06, + "loss": 1.027, + "mean_token_accuracy": 0.6884394288063049, + "num_tokens": 175955080.0, + "step": 6797 + }, + { + "epoch": 0.7465407423676697, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.3207435607910156, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7213284969329834, + "num_tokens": 175977085.0, + "step": 6798 + }, + { + "epoch": 0.7466505600702833, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.2493162155151367, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7109980583190918, + "num_tokens": 176003377.0, + "step": 6799 + }, + { + "epoch": 0.7467603777728969, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 1.9920753240585327, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7084691524505615, + "num_tokens": 176030427.0, + "step": 6800 + }, + { + "epoch": 0.7468701954755107, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.040923595428467, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6877338886260986, + "num_tokens": 176058425.0, + "step": 6801 + }, + { + "epoch": 0.7469800131781243, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.186343193054199, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7338141202926636, + "num_tokens": 176081969.0, + "step": 6802 + }, + { + "epoch": 0.747089830880738, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.185126781463623, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7200940251350403, + "num_tokens": 176106404.0, + "step": 6803 + }, + { + "epoch": 0.7471996485833516, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.050407648086548, + "learning_rate": 1e-06, + "loss": 1.0654, + "mean_token_accuracy": 0.6722183227539062, + "num_tokens": 176137354.0, + "step": 6804 + }, + { + "epoch": 0.7473094662859653, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.212766170501709, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7030978202819824, + "num_tokens": 176162288.0, + "step": 6805 + }, + { + "epoch": 0.7474192839885789, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.174147367477417, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.7019649147987366, + "num_tokens": 176189523.0, + "step": 6806 + }, + { + "epoch": 0.7475291016911926, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.3070740699768066, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7076314091682434, + "num_tokens": 176213616.0, + "step": 6807 + }, + { + "epoch": 0.7476389193938063, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.19506573677063, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7233723402023315, + "num_tokens": 176236767.0, + "step": 6808 + }, + { + "epoch": 0.74774873709642, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.0764029026031494, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7260345816612244, + "num_tokens": 176263344.0, + "step": 6809 + }, + { + "epoch": 0.7478585547990336, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.2114076614379883, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7059308290481567, + "num_tokens": 176288047.0, + "step": 6810 + }, + { + "epoch": 0.7479683725016473, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.0258171558380127, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7012460231781006, + "num_tokens": 176317781.0, + "step": 6811 + }, + { + "epoch": 0.7480781902042609, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.15757417678833, + "learning_rate": 1e-06, + "loss": 1.0534, + "mean_token_accuracy": 0.6826533079147339, + "num_tokens": 176347624.0, + "step": 6812 + }, + { + "epoch": 0.7481880079068746, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.558283567428589, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7173146605491638, + "num_tokens": 176367772.0, + "step": 6813 + }, + { + "epoch": 0.7482978256094882, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.0770676136016846, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7122521996498108, + "num_tokens": 176400767.0, + "step": 6814 + }, + { + "epoch": 0.7484076433121019, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.1921823024749756, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.687621533870697, + "num_tokens": 176425357.0, + "step": 6815 + }, + { + "epoch": 0.7485174610147156, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.4876108169555664, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6947892904281616, + "num_tokens": 176445214.0, + "step": 6816 + }, + { + "epoch": 0.7486272787173293, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.1731455326080322, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7069765329360962, + "num_tokens": 176472967.0, + "step": 6817 + }, + { + "epoch": 0.7487370964199429, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 1.967923879623413, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7203419804573059, + "num_tokens": 176502370.0, + "step": 6818 + }, + { + "epoch": 0.7488469141225566, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.2612454891204834, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7056859731674194, + "num_tokens": 176525646.0, + "step": 6819 + }, + { + "epoch": 0.7489567318251702, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.1012680530548096, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.695408046245575, + "num_tokens": 176552922.0, + "step": 6820 + }, + { + "epoch": 0.7490665495277838, + "ewc_loss": 3.591179847717285e-06, + "grad_norm": 2.1121981143951416, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.709691047668457, + "num_tokens": 176580090.0, + "step": 6821 + }, + { + "epoch": 0.7491763672303975, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0363223552703857, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7081021070480347, + "num_tokens": 176607392.0, + "step": 6822 + }, + { + "epoch": 0.7492861849330112, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.382066249847412, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.720835268497467, + "num_tokens": 176631020.0, + "step": 6823 + }, + { + "epoch": 0.7493960026356249, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0776779651641846, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.726202130317688, + "num_tokens": 176659765.0, + "step": 6824 + }, + { + "epoch": 0.7495058203382385, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0631580352783203, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7266340255737305, + "num_tokens": 176688642.0, + "step": 6825 + }, + { + "epoch": 0.7496156380408522, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.304839611053467, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7337433099746704, + "num_tokens": 176712306.0, + "step": 6826 + }, + { + "epoch": 0.7497254557434658, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.4007160663604736, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7302741408348083, + "num_tokens": 176731830.0, + "step": 6827 + }, + { + "epoch": 0.7498352734460795, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0167713165283203, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.7011716365814209, + "num_tokens": 176762912.0, + "step": 6828 + }, + { + "epoch": 0.7499450911486931, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.191983222961426, + "learning_rate": 1e-06, + "loss": 0.9877, + "mean_token_accuracy": 0.7026209831237793, + "num_tokens": 176787808.0, + "step": 6829 + }, + { + "epoch": 0.7500549088513069, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.381967782974243, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7147844433784485, + "num_tokens": 176809620.0, + "step": 6830 + }, + { + "epoch": 0.7501647265539205, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 1.9705009460449219, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7010349035263062, + "num_tokens": 176838777.0, + "step": 6831 + }, + { + "epoch": 0.7502745442565342, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.438735246658325, + "learning_rate": 1e-06, + "loss": 1.0661, + "mean_token_accuracy": 0.6837736964225769, + "num_tokens": 176863801.0, + "step": 6832 + }, + { + "epoch": 0.7503843619591478, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0941097736358643, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7114152908325195, + "num_tokens": 176892098.0, + "step": 6833 + }, + { + "epoch": 0.7504941796617615, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.044257164001465, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7203400731086731, + "num_tokens": 176919806.0, + "step": 6834 + }, + { + "epoch": 0.7506039973643751, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2017054557800293, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6859781742095947, + "num_tokens": 176944799.0, + "step": 6835 + }, + { + "epoch": 0.7507138150669888, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3907134532928467, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.6970686912536621, + "num_tokens": 176968499.0, + "step": 6836 + }, + { + "epoch": 0.7508236327696025, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2285749912261963, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7144988775253296, + "num_tokens": 176992760.0, + "step": 6837 + }, + { + "epoch": 0.7509334504722162, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0706169605255127, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.695903480052948, + "num_tokens": 177019679.0, + "step": 6838 + }, + { + "epoch": 0.7510432681748298, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.207808256149292, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6955996751785278, + "num_tokens": 177048441.0, + "step": 6839 + }, + { + "epoch": 0.7511530858774434, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0771291255950928, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7062536478042603, + "num_tokens": 177076460.0, + "step": 6840 + }, + { + "epoch": 0.7512629035800571, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3400466442108154, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7041587829589844, + "num_tokens": 177099730.0, + "step": 6841 + }, + { + "epoch": 0.7513727212826707, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3431942462921143, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7055731415748596, + "num_tokens": 177122154.0, + "step": 6842 + }, + { + "epoch": 0.7514825389852844, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.423562526702881, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7153431177139282, + "num_tokens": 177142945.0, + "step": 6843 + }, + { + "epoch": 0.7515923566878981, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.106637716293335, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7033862471580505, + "num_tokens": 177170514.0, + "step": 6844 + }, + { + "epoch": 0.7517021743905118, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.229261636734009, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7002962827682495, + "num_tokens": 177196408.0, + "step": 6845 + }, + { + "epoch": 0.7518119920931254, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3484153747558594, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7232881188392639, + "num_tokens": 177220300.0, + "step": 6846 + }, + { + "epoch": 0.7519218097957391, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0033462047576904, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.6961897611618042, + "num_tokens": 177251674.0, + "step": 6847 + }, + { + "epoch": 0.7520316274983527, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 1.9419621229171753, + "learning_rate": 1e-06, + "loss": 0.9989, + "mean_token_accuracy": 0.6972578763961792, + "num_tokens": 177284386.0, + "step": 6848 + }, + { + "epoch": 0.7521414452009664, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 1.9935303926467896, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.6989494562149048, + "num_tokens": 177314575.0, + "step": 6849 + }, + { + "epoch": 0.75225126290358, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.1728768348693848, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6935542821884155, + "num_tokens": 177340530.0, + "step": 6850 + }, + { + "epoch": 0.7523610806061937, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3127448558807373, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7022308111190796, + "num_tokens": 177363114.0, + "step": 6851 + }, + { + "epoch": 0.7524708983088074, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0166985988616943, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7047134637832642, + "num_tokens": 177392813.0, + "step": 6852 + }, + { + "epoch": 0.7525807160114211, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.376162052154541, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7087218165397644, + "num_tokens": 177416349.0, + "step": 6853 + }, + { + "epoch": 0.7526905337140347, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.293985605239868, + "learning_rate": 1e-06, + "loss": 1.0237, + "mean_token_accuracy": 0.6879147291183472, + "num_tokens": 177441313.0, + "step": 6854 + }, + { + "epoch": 0.7528003514166484, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.398595094680786, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7452102899551392, + "num_tokens": 177462066.0, + "step": 6855 + }, + { + "epoch": 0.752910169119262, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2550265789031982, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.707420289516449, + "num_tokens": 177488623.0, + "step": 6856 + }, + { + "epoch": 0.7530199868218757, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.054785966873169, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7094636559486389, + "num_tokens": 177517433.0, + "step": 6857 + }, + { + "epoch": 0.7531298045244893, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.1052908897399902, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.7032818794250488, + "num_tokens": 177546971.0, + "step": 6858 + }, + { + "epoch": 0.7532396222271031, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3598740100860596, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.6900995969772339, + "num_tokens": 177568348.0, + "step": 6859 + }, + { + "epoch": 0.7533494399297167, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.200932741165161, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7106618881225586, + "num_tokens": 177595015.0, + "step": 6860 + }, + { + "epoch": 0.7534592576323303, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.1762938499450684, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7112616300582886, + "num_tokens": 177621428.0, + "step": 6861 + }, + { + "epoch": 0.753569075334944, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.1288726329803467, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7110636234283447, + "num_tokens": 177648253.0, + "step": 6862 + }, + { + "epoch": 0.7536788930375576, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.6710662841796875, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7358891367912292, + "num_tokens": 177666539.0, + "step": 6863 + }, + { + "epoch": 0.7537887107401713, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.3224594593048096, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7187500596046448, + "num_tokens": 177691319.0, + "step": 6864 + }, + { + "epoch": 0.7538985284427849, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.139503240585327, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7014251351356506, + "num_tokens": 177721356.0, + "step": 6865 + }, + { + "epoch": 0.7540083461453987, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 1.9988042116165161, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.6842158436775208, + "num_tokens": 177753114.0, + "step": 6866 + }, + { + "epoch": 0.7541181638480123, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.242403507232666, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.725063681602478, + "num_tokens": 177777685.0, + "step": 6867 + }, + { + "epoch": 0.754227981550626, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.061535596847534, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7079517841339111, + "num_tokens": 177805844.0, + "step": 6868 + }, + { + "epoch": 0.7543377992532396, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.107170581817627, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.6966961026191711, + "num_tokens": 177835153.0, + "step": 6869 + }, + { + "epoch": 0.7544476169558533, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.098055601119995, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7128631472587585, + "num_tokens": 177861316.0, + "step": 6870 + }, + { + "epoch": 0.7545574346584669, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.191710948944092, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7113750576972961, + "num_tokens": 177888119.0, + "step": 6871 + }, + { + "epoch": 0.7546672523610806, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.032336711883545, + "learning_rate": 1e-06, + "loss": 1.0404, + "mean_token_accuracy": 0.6836885809898376, + "num_tokens": 177920001.0, + "step": 6872 + }, + { + "epoch": 0.7547770700636943, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.6480655670166016, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7117235064506531, + "num_tokens": 177938096.0, + "step": 6873 + }, + { + "epoch": 0.754886887766308, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2264344692230225, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7012177109718323, + "num_tokens": 177961331.0, + "step": 6874 + }, + { + "epoch": 0.7549967054689216, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.1788322925567627, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7287098169326782, + "num_tokens": 177985902.0, + "step": 6875 + }, + { + "epoch": 0.7551065231715353, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2968783378601074, + "learning_rate": 1e-06, + "loss": 1.015, + "mean_token_accuracy": 0.6882550716400146, + "num_tokens": 178011077.0, + "step": 6876 + }, + { + "epoch": 0.7552163408741489, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.396388053894043, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7195667028427124, + "num_tokens": 178033075.0, + "step": 6877 + }, + { + "epoch": 0.7553261585767626, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0333809852600098, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7050951719284058, + "num_tokens": 178060752.0, + "step": 6878 + }, + { + "epoch": 0.7554359762793762, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.0104117393493652, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7267594337463379, + "num_tokens": 178089622.0, + "step": 6879 + }, + { + "epoch": 0.7555457939819898, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.002253770828247, + "learning_rate": 1e-06, + "loss": 1.0514, + "mean_token_accuracy": 0.6794667840003967, + "num_tokens": 178120948.0, + "step": 6880 + }, + { + "epoch": 0.7556556116846036, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.666080951690674, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6927711963653564, + "num_tokens": 178143058.0, + "step": 6881 + }, + { + "epoch": 0.7557654293872172, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.2431554794311523, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7048768401145935, + "num_tokens": 178166367.0, + "step": 6882 + }, + { + "epoch": 0.7558752470898309, + "ewc_loss": 3.606081008911133e-06, + "grad_norm": 2.5370917320251465, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7154419422149658, + "num_tokens": 178185632.0, + "step": 6883 + }, + { + "epoch": 0.7559850647924445, + "ewc_loss": 3.6209821701049805e-06, + "grad_norm": 2.1859748363494873, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6951945424079895, + "num_tokens": 178210509.0, + "step": 6884 + }, + { + "epoch": 0.7560948824950582, + "ewc_loss": 3.6209821701049805e-06, + "grad_norm": 1.8869884014129639, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6944369673728943, + "num_tokens": 178246098.0, + "step": 6885 + }, + { + "epoch": 0.7562047001976718, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2447004318237305, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.7033991813659668, + "num_tokens": 178269653.0, + "step": 6886 + }, + { + "epoch": 0.7563145179002855, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.1650335788726807, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7059970498085022, + "num_tokens": 178297507.0, + "step": 6887 + }, + { + "epoch": 0.7564243356028992, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2356646060943604, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7131128311157227, + "num_tokens": 178322444.0, + "step": 6888 + }, + { + "epoch": 0.7565341533055129, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.4082326889038086, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6906230449676514, + "num_tokens": 178345073.0, + "step": 6889 + }, + { + "epoch": 0.7566439710081265, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.027836322784424, + "learning_rate": 1e-06, + "loss": 1.0203, + "mean_token_accuracy": 0.6948469877243042, + "num_tokens": 178373921.0, + "step": 6890 + }, + { + "epoch": 0.7567537887107402, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.050703763961792, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7076411247253418, + "num_tokens": 178401628.0, + "step": 6891 + }, + { + "epoch": 0.7568636064133538, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2821500301361084, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7326511740684509, + "num_tokens": 178424561.0, + "step": 6892 + }, + { + "epoch": 0.7569734241159675, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.146151542663574, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7311372756958008, + "num_tokens": 178447980.0, + "step": 6893 + }, + { + "epoch": 0.7570832418185811, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.1546812057495117, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7177397012710571, + "num_tokens": 178471718.0, + "step": 6894 + }, + { + "epoch": 0.7571930595211949, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2655725479125977, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7045291662216187, + "num_tokens": 178496847.0, + "step": 6895 + }, + { + "epoch": 0.7573028772238085, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0144155025482178, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6891389489173889, + "num_tokens": 178527000.0, + "step": 6896 + }, + { + "epoch": 0.7574126949264222, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.3490054607391357, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.7057054042816162, + "num_tokens": 178550133.0, + "step": 6897 + }, + { + "epoch": 0.7575225126290358, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2972843647003174, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7013932466506958, + "num_tokens": 178577050.0, + "step": 6898 + }, + { + "epoch": 0.7576323303316495, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.1491782665252686, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7241238355636597, + "num_tokens": 178602725.0, + "step": 6899 + }, + { + "epoch": 0.7577421480342631, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 1.9529166221618652, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7140854597091675, + "num_tokens": 178633785.0, + "step": 6900 + }, + { + "epoch": 0.7578519657368767, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.515690565109253, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7056597471237183, + "num_tokens": 178654144.0, + "step": 6901 + }, + { + "epoch": 0.7579617834394905, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.1600897312164307, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7022234797477722, + "num_tokens": 178680025.0, + "step": 6902 + }, + { + "epoch": 0.7580716011421041, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.3925702571868896, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7049867510795593, + "num_tokens": 178702119.0, + "step": 6903 + }, + { + "epoch": 0.7581814188447178, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 1.9942066669464111, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6973645091056824, + "num_tokens": 178731672.0, + "step": 6904 + }, + { + "epoch": 0.7582912365473314, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.07938551902771, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7258949279785156, + "num_tokens": 178758582.0, + "step": 6905 + }, + { + "epoch": 0.7584010542499451, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 1.935439944267273, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6896989345550537, + "num_tokens": 178787672.0, + "step": 6906 + }, + { + "epoch": 0.7585108719525587, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.132910966873169, + "learning_rate": 1e-06, + "loss": 1.0271, + "mean_token_accuracy": 0.6877466440200806, + "num_tokens": 178814095.0, + "step": 6907 + }, + { + "epoch": 0.7586206896551724, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0298428535461426, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.6943714618682861, + "num_tokens": 178843461.0, + "step": 6908 + }, + { + "epoch": 0.758730507357786, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0912914276123047, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.7041033506393433, + "num_tokens": 178870250.0, + "step": 6909 + }, + { + "epoch": 0.7588403250603998, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.225417375564575, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7190799713134766, + "num_tokens": 178893417.0, + "step": 6910 + }, + { + "epoch": 0.7589501427630134, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.3178892135620117, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7037596702575684, + "num_tokens": 178915244.0, + "step": 6911 + }, + { + "epoch": 0.7590599604656271, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2996163368225098, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7016246914863586, + "num_tokens": 178938120.0, + "step": 6912 + }, + { + "epoch": 0.7591697781682407, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.1211249828338623, + "learning_rate": 1e-06, + "loss": 1.0554, + "mean_token_accuracy": 0.684627890586853, + "num_tokens": 178965558.0, + "step": 6913 + }, + { + "epoch": 0.7592795958708544, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0666346549987793, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.717715859413147, + "num_tokens": 178994111.0, + "step": 6914 + }, + { + "epoch": 0.759389413573468, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.3572990894317627, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.706406831741333, + "num_tokens": 179017532.0, + "step": 6915 + }, + { + "epoch": 0.7594992312760817, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.648594856262207, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7040972709655762, + "num_tokens": 179036660.0, + "step": 6916 + }, + { + "epoch": 0.7596090489786954, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.034048318862915, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7060227394104004, + "num_tokens": 179064753.0, + "step": 6917 + }, + { + "epoch": 0.7597188666813091, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2618768215179443, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7191087007522583, + "num_tokens": 179087979.0, + "step": 6918 + }, + { + "epoch": 0.7598286843839227, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.3456788063049316, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.7007763981819153, + "num_tokens": 179113525.0, + "step": 6919 + }, + { + "epoch": 0.7599385020865363, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0891332626342773, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.717049241065979, + "num_tokens": 179140003.0, + "step": 6920 + }, + { + "epoch": 0.76004831978915, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.2087507247924805, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7223268747329712, + "num_tokens": 179164191.0, + "step": 6921 + }, + { + "epoch": 0.7601581374917636, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.0285401344299316, + "learning_rate": 1e-06, + "loss": 1.0411, + "mean_token_accuracy": 0.6851433515548706, + "num_tokens": 179194456.0, + "step": 6922 + }, + { + "epoch": 0.7602679551943773, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.144404172897339, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7310577034950256, + "num_tokens": 179220806.0, + "step": 6923 + }, + { + "epoch": 0.760377772896991, + "ewc_loss": 3.635883331298828e-06, + "grad_norm": 2.382603168487549, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.705050528049469, + "num_tokens": 179243326.0, + "step": 6924 + }, + { + "epoch": 0.7604875905996047, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.1077797412872314, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7099502086639404, + "num_tokens": 179270466.0, + "step": 6925 + }, + { + "epoch": 0.7605974083022183, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.001357078552246, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7193963527679443, + "num_tokens": 179300590.0, + "step": 6926 + }, + { + "epoch": 0.760707226004832, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.2335712909698486, + "learning_rate": 1e-06, + "loss": 1.0095, + "mean_token_accuracy": 0.6893738508224487, + "num_tokens": 179324837.0, + "step": 6927 + }, + { + "epoch": 0.7608170437074456, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 1.9805152416229248, + "learning_rate": 1e-06, + "loss": 1.0244, + "mean_token_accuracy": 0.6997553110122681, + "num_tokens": 179356085.0, + "step": 6928 + }, + { + "epoch": 0.7609268614100593, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.259335517883301, + "learning_rate": 1e-06, + "loss": 1.0637, + "mean_token_accuracy": 0.6809096336364746, + "num_tokens": 179380943.0, + "step": 6929 + }, + { + "epoch": 0.7610366791126729, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.143819570541382, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6999393701553345, + "num_tokens": 179406636.0, + "step": 6930 + }, + { + "epoch": 0.7611464968152867, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.1635591983795166, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.6980452537536621, + "num_tokens": 179432128.0, + "step": 6931 + }, + { + "epoch": 0.7612563145179003, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.240800380706787, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.7027598023414612, + "num_tokens": 179457591.0, + "step": 6932 + }, + { + "epoch": 0.761366132220514, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.261631727218628, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7123352289199829, + "num_tokens": 179483646.0, + "step": 6933 + }, + { + "epoch": 0.7614759499231276, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.0617125034332275, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7123137712478638, + "num_tokens": 179511204.0, + "step": 6934 + }, + { + "epoch": 0.7615857676257413, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.148118495941162, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7063552737236023, + "num_tokens": 179536067.0, + "step": 6935 + }, + { + "epoch": 0.7616955853283549, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.0148959159851074, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7374239563941956, + "num_tokens": 179563707.0, + "step": 6936 + }, + { + "epoch": 0.7618054030309686, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.1979429721832275, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7245094776153564, + "num_tokens": 179587191.0, + "step": 6937 + }, + { + "epoch": 0.7619152207335822, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.175647735595703, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7326290607452393, + "num_tokens": 179610040.0, + "step": 6938 + }, + { + "epoch": 0.762025038436196, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.2424204349517822, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7032884955406189, + "num_tokens": 179635594.0, + "step": 6939 + }, + { + "epoch": 0.7621348561388096, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.324854850769043, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.713860273361206, + "num_tokens": 179657892.0, + "step": 6940 + }, + { + "epoch": 0.7622446738414232, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.2674760818481445, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7289865016937256, + "num_tokens": 179683394.0, + "step": 6941 + }, + { + "epoch": 0.7623544915440369, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.040213108062744, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.6855252385139465, + "num_tokens": 179711403.0, + "step": 6942 + }, + { + "epoch": 0.7624643092466505, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.5902390480041504, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7203377485275269, + "num_tokens": 179730406.0, + "step": 6943 + }, + { + "epoch": 0.7625741269492642, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.0533835887908936, + "learning_rate": 1e-06, + "loss": 1.0129, + "mean_token_accuracy": 0.6927498579025269, + "num_tokens": 179761665.0, + "step": 6944 + }, + { + "epoch": 0.7626839446518778, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.236785888671875, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.709945797920227, + "num_tokens": 179784254.0, + "step": 6945 + }, + { + "epoch": 0.7627937623544916, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.4936766624450684, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.705091118812561, + "num_tokens": 179807077.0, + "step": 6946 + }, + { + "epoch": 0.7629035800571052, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 1.8638216257095337, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7169650793075562, + "num_tokens": 179838235.0, + "step": 6947 + }, + { + "epoch": 0.7630133977597189, + "ewc_loss": 3.6507844924926758e-06, + "grad_norm": 2.082509994506836, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7059988975524902, + "num_tokens": 179866103.0, + "step": 6948 + }, + { + "epoch": 0.7631232154623325, + "ewc_loss": 3.680586814880371e-06, + "grad_norm": 2.366227626800537, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.721426248550415, + "num_tokens": 179888382.0, + "step": 6949 + }, + { + "epoch": 0.7632330331649462, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.4061005115509033, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7118304371833801, + "num_tokens": 179908263.0, + "step": 6950 + }, + { + "epoch": 0.7633428508675598, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.146148681640625, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7163552045822144, + "num_tokens": 179935433.0, + "step": 6951 + }, + { + "epoch": 0.7634526685701735, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.1565744876861572, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7009992003440857, + "num_tokens": 179961446.0, + "step": 6952 + }, + { + "epoch": 0.7635624862727872, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.095378875732422, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.718949556350708, + "num_tokens": 179987906.0, + "step": 6953 + }, + { + "epoch": 0.7636723039754009, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.2716598510742188, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.6759186387062073, + "num_tokens": 180011551.0, + "step": 6954 + }, + { + "epoch": 0.7637821216780145, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.293452501296997, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7150466442108154, + "num_tokens": 180037809.0, + "step": 6955 + }, + { + "epoch": 0.7638919393806282, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.168855667114258, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.713213324546814, + "num_tokens": 180063827.0, + "step": 6956 + }, + { + "epoch": 0.7640017570832418, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.158283233642578, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.6945828795433044, + "num_tokens": 180089028.0, + "step": 6957 + }, + { + "epoch": 0.7641115747858555, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.30971622467041, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7115042805671692, + "num_tokens": 180113074.0, + "step": 6958 + }, + { + "epoch": 0.7642213924884691, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.160858631134033, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6911657452583313, + "num_tokens": 180140288.0, + "step": 6959 + }, + { + "epoch": 0.7643312101910829, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.1068034172058105, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.6937823295593262, + "num_tokens": 180165848.0, + "step": 6960 + }, + { + "epoch": 0.7644410278936965, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.0652663707733154, + "learning_rate": 1e-06, + "loss": 1.0627, + "mean_token_accuracy": 0.6842583417892456, + "num_tokens": 180193944.0, + "step": 6961 + }, + { + "epoch": 0.7645508455963101, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.335984468460083, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7041640281677246, + "num_tokens": 180217722.0, + "step": 6962 + }, + { + "epoch": 0.7646606632989238, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.1512176990509033, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7267886400222778, + "num_tokens": 180242650.0, + "step": 6963 + }, + { + "epoch": 0.7647704810015374, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 1.918264627456665, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7358279228210449, + "num_tokens": 180272453.0, + "step": 6964 + }, + { + "epoch": 0.7648802987041511, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.0208146572113037, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7176987528800964, + "num_tokens": 180300422.0, + "step": 6965 + }, + { + "epoch": 0.7649901164067647, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.386274814605713, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7307679653167725, + "num_tokens": 180322956.0, + "step": 6966 + }, + { + "epoch": 0.7650999341093784, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.1946864128112793, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7025774717330933, + "num_tokens": 180348589.0, + "step": 6967 + }, + { + "epoch": 0.7652097518119921, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.4812870025634766, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7064169645309448, + "num_tokens": 180369574.0, + "step": 6968 + }, + { + "epoch": 0.7653195695146058, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.323773145675659, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.7042971849441528, + "num_tokens": 180391533.0, + "step": 6969 + }, + { + "epoch": 0.7654293872172194, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.080291271209717, + "learning_rate": 1e-06, + "loss": 1.0157, + "mean_token_accuracy": 0.6875351667404175, + "num_tokens": 180419677.0, + "step": 6970 + }, + { + "epoch": 0.7655392049198331, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.0923831462860107, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6877075433731079, + "num_tokens": 180449792.0, + "step": 6971 + }, + { + "epoch": 0.7656490226224467, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.5048301219940186, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7069388031959534, + "num_tokens": 180470960.0, + "step": 6972 + }, + { + "epoch": 0.7657588403250604, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.015291213989258, + "learning_rate": 1e-06, + "loss": 1.0414, + "mean_token_accuracy": 0.6816650629043579, + "num_tokens": 180501704.0, + "step": 6973 + }, + { + "epoch": 0.765868658027674, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.452981948852539, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7049695253372192, + "num_tokens": 180522239.0, + "step": 6974 + }, + { + "epoch": 0.7659784757302878, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.3135600090026855, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7068287134170532, + "num_tokens": 180545659.0, + "step": 6975 + }, + { + "epoch": 0.7660882934329014, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.1090807914733887, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7253533005714417, + "num_tokens": 180570244.0, + "step": 6976 + }, + { + "epoch": 0.7661981111355151, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.005758762359619, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7223086357116699, + "num_tokens": 180597361.0, + "step": 6977 + }, + { + "epoch": 0.7663079288381287, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.2313907146453857, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.701274037361145, + "num_tokens": 180622400.0, + "step": 6978 + }, + { + "epoch": 0.7664177465407424, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.0716328620910645, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7091710567474365, + "num_tokens": 180649632.0, + "step": 6979 + }, + { + "epoch": 0.766527564243356, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.047553300857544, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7163674235343933, + "num_tokens": 180678695.0, + "step": 6980 + }, + { + "epoch": 0.7666373819459696, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.4331884384155273, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7226173877716064, + "num_tokens": 180700718.0, + "step": 6981 + }, + { + "epoch": 0.7667471996485834, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.2312369346618652, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.731918215751648, + "num_tokens": 180723553.0, + "step": 6982 + }, + { + "epoch": 0.766857017351197, + "ewc_loss": 3.6954879760742188e-06, + "grad_norm": 2.2448039054870605, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7146493196487427, + "num_tokens": 180748214.0, + "step": 6983 + }, + { + "epoch": 0.7669668350538107, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1964080333709717, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7061952352523804, + "num_tokens": 180773817.0, + "step": 6984 + }, + { + "epoch": 0.7670766527564243, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.354482650756836, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7108629941940308, + "num_tokens": 180794784.0, + "step": 6985 + }, + { + "epoch": 0.767186470459038, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.445568323135376, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7025265693664551, + "num_tokens": 180815988.0, + "step": 6986 + }, + { + "epoch": 0.7672962881616516, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.0835065841674805, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6993340849876404, + "num_tokens": 180842777.0, + "step": 6987 + }, + { + "epoch": 0.7674061058642653, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.2004218101501465, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7158190011978149, + "num_tokens": 180868011.0, + "step": 6988 + }, + { + "epoch": 0.767515923566879, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.0049805641174316, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6918494701385498, + "num_tokens": 180899163.0, + "step": 6989 + }, + { + "epoch": 0.7676257412694927, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1745190620422363, + "learning_rate": 1e-06, + "loss": 1.0842, + "mean_token_accuracy": 0.6737939119338989, + "num_tokens": 180926392.0, + "step": 6990 + }, + { + "epoch": 0.7677355589721063, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.176729440689087, + "learning_rate": 1e-06, + "loss": 1.0169, + "mean_token_accuracy": 0.6971917152404785, + "num_tokens": 180951585.0, + "step": 6991 + }, + { + "epoch": 0.76784537667472, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.273951768875122, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7066617012023926, + "num_tokens": 180975363.0, + "step": 6992 + }, + { + "epoch": 0.7679551943773336, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1552047729492188, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6916040182113647, + "num_tokens": 181003116.0, + "step": 6993 + }, + { + "epoch": 0.7680650120799473, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.170332431793213, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6922771334648132, + "num_tokens": 181029285.0, + "step": 6994 + }, + { + "epoch": 0.7681748297825609, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.192857265472412, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.712155818939209, + "num_tokens": 181053653.0, + "step": 6995 + }, + { + "epoch": 0.7682846474851747, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1618640422821045, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7061906456947327, + "num_tokens": 181079899.0, + "step": 6996 + }, + { + "epoch": 0.7683944651877883, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.2317779064178467, + "learning_rate": 1e-06, + "loss": 1.0369, + "mean_token_accuracy": 0.6919562816619873, + "num_tokens": 181105803.0, + "step": 6997 + }, + { + "epoch": 0.768504282890402, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.159353494644165, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7077649831771851, + "num_tokens": 181132456.0, + "step": 6998 + }, + { + "epoch": 0.7686141005930156, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.3863136768341064, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.6949259042739868, + "num_tokens": 181155100.0, + "step": 6999 + }, + { + "epoch": 0.7687239182956292, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.3982672691345215, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7076488733291626, + "num_tokens": 181176842.0, + "step": 7000 + }, + { + "epoch": 0.7688337359982429, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.2704360485076904, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7023431658744812, + "num_tokens": 181200229.0, + "step": 7001 + }, + { + "epoch": 0.7689435537008565, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.0684289932250977, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7042896747589111, + "num_tokens": 181229968.0, + "step": 7002 + }, + { + "epoch": 0.7690533714034702, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1672444343566895, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.697824239730835, + "num_tokens": 181256302.0, + "step": 7003 + }, + { + "epoch": 0.7691631891060839, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 1.9160128831863403, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7325870394706726, + "num_tokens": 181286191.0, + "step": 7004 + }, + { + "epoch": 0.7692730068086976, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.4238855838775635, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.704740047454834, + "num_tokens": 181307029.0, + "step": 7005 + }, + { + "epoch": 0.7693828245113112, + "ewc_loss": 3.7103891372680664e-06, + "grad_norm": 2.1638870239257812, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7076605558395386, + "num_tokens": 181332716.0, + "step": 7006 + }, + { + "epoch": 0.7694926422139249, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0048577785491943, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7082713842391968, + "num_tokens": 181361598.0, + "step": 7007 + }, + { + "epoch": 0.7696024599165385, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.251464605331421, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6918395757675171, + "num_tokens": 181385434.0, + "step": 7008 + }, + { + "epoch": 0.7697122776191522, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.218181610107422, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6968870162963867, + "num_tokens": 181411830.0, + "step": 7009 + }, + { + "epoch": 0.7698220953217658, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.9585559368133545, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7022348642349243, + "num_tokens": 181442415.0, + "step": 7010 + }, + { + "epoch": 0.7699319130243796, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3052115440368652, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7036476135253906, + "num_tokens": 181466967.0, + "step": 7011 + }, + { + "epoch": 0.7700417307269932, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.5685317516326904, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7215890288352966, + "num_tokens": 181487298.0, + "step": 7012 + }, + { + "epoch": 0.7701515484296069, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.440549373626709, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.739567756652832, + "num_tokens": 181507506.0, + "step": 7013 + }, + { + "epoch": 0.7702613661322205, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0307865142822266, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7058155536651611, + "num_tokens": 181536470.0, + "step": 7014 + }, + { + "epoch": 0.7703711838348342, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.204639196395874, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7078430652618408, + "num_tokens": 181562596.0, + "step": 7015 + }, + { + "epoch": 0.7704810015374478, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1541943550109863, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6900262832641602, + "num_tokens": 181591441.0, + "step": 7016 + }, + { + "epoch": 0.7705908192400615, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1552488803863525, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6983464956283569, + "num_tokens": 181617441.0, + "step": 7017 + }, + { + "epoch": 0.7707006369426752, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.185387134552002, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7130839824676514, + "num_tokens": 181644105.0, + "step": 7018 + }, + { + "epoch": 0.7708104546452889, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1340818405151367, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7155860662460327, + "num_tokens": 181669368.0, + "step": 7019 + }, + { + "epoch": 0.7709202723479025, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.4054040908813477, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7047470808029175, + "num_tokens": 181690185.0, + "step": 7020 + }, + { + "epoch": 0.7710300900505161, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.050957202911377, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7170580625534058, + "num_tokens": 181717010.0, + "step": 7021 + }, + { + "epoch": 0.7711399077531298, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.151357650756836, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6922871470451355, + "num_tokens": 181742561.0, + "step": 7022 + }, + { + "epoch": 0.7712497254557434, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3159847259521484, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6906494498252869, + "num_tokens": 181766667.0, + "step": 7023 + }, + { + "epoch": 0.7713595431583571, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2209410667419434, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7308177351951599, + "num_tokens": 181792346.0, + "step": 7024 + }, + { + "epoch": 0.7714693608609708, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2303340435028076, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6836440563201904, + "num_tokens": 181817478.0, + "step": 7025 + }, + { + "epoch": 0.7715791785635845, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2736306190490723, + "learning_rate": 1e-06, + "loss": 1.0382, + "mean_token_accuracy": 0.6968450546264648, + "num_tokens": 181841823.0, + "step": 7026 + }, + { + "epoch": 0.7716889962661981, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.348205089569092, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.6991679072380066, + "num_tokens": 181863086.0, + "step": 7027 + }, + { + "epoch": 0.7717988139688118, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.21647310256958, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6841583251953125, + "num_tokens": 181888119.0, + "step": 7028 + }, + { + "epoch": 0.7719086316714254, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0670642852783203, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6934127807617188, + "num_tokens": 181914649.0, + "step": 7029 + }, + { + "epoch": 0.7720184493740391, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.567333936691284, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6980751752853394, + "num_tokens": 181940137.0, + "step": 7030 + }, + { + "epoch": 0.7721282670766527, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2514212131500244, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7268816828727722, + "num_tokens": 181963552.0, + "step": 7031 + }, + { + "epoch": 0.7722380847792664, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.444195508956909, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7394278049468994, + "num_tokens": 181983514.0, + "step": 7032 + }, + { + "epoch": 0.7723479024818801, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0366404056549072, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.709357738494873, + "num_tokens": 182012525.0, + "step": 7033 + }, + { + "epoch": 0.7724577201844938, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.264112710952759, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6907839775085449, + "num_tokens": 182036340.0, + "step": 7034 + }, + { + "epoch": 0.7725675378871074, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.5518999099731445, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.716384768486023, + "num_tokens": 182055692.0, + "step": 7035 + }, + { + "epoch": 0.7726773555897211, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.028864860534668, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.710311233997345, + "num_tokens": 182084359.0, + "step": 7036 + }, + { + "epoch": 0.7727871732923347, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3624398708343506, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6894339323043823, + "num_tokens": 182108290.0, + "step": 7037 + }, + { + "epoch": 0.7728969909949484, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.064701557159424, + "learning_rate": 1e-06, + "loss": 1.025, + "mean_token_accuracy": 0.6887979507446289, + "num_tokens": 182137242.0, + "step": 7038 + }, + { + "epoch": 0.773006808697562, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.038985252380371, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6939010620117188, + "num_tokens": 182166149.0, + "step": 7039 + }, + { + "epoch": 0.7731166264001758, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1710562705993652, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7155579328536987, + "num_tokens": 182189087.0, + "step": 7040 + }, + { + "epoch": 0.7732264441027894, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3132317066192627, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6950007677078247, + "num_tokens": 182213522.0, + "step": 7041 + }, + { + "epoch": 0.773336261805403, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.35335636138916, + "learning_rate": 1e-06, + "loss": 1.0314, + "mean_token_accuracy": 0.6874169111251831, + "num_tokens": 182236627.0, + "step": 7042 + }, + { + "epoch": 0.7734460795080167, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.440047264099121, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7314491271972656, + "num_tokens": 182259748.0, + "step": 7043 + }, + { + "epoch": 0.7735558972106303, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1169443130493164, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7055642008781433, + "num_tokens": 182288135.0, + "step": 7044 + }, + { + "epoch": 0.773665714913244, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.098007917404175, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.695570707321167, + "num_tokens": 182313404.0, + "step": 7045 + }, + { + "epoch": 0.7737755326158576, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0072124004364014, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7124350666999817, + "num_tokens": 182342822.0, + "step": 7046 + }, + { + "epoch": 0.7738853503184714, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.5993688106536865, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7287775874137878, + "num_tokens": 182360758.0, + "step": 7047 + }, + { + "epoch": 0.773995168021085, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.8970402479171753, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6932950019836426, + "num_tokens": 182393280.0, + "step": 7048 + }, + { + "epoch": 0.7741049857236987, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2031121253967285, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7056581377983093, + "num_tokens": 182417912.0, + "step": 7049 + }, + { + "epoch": 0.7742148034263123, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.282015800476074, + "learning_rate": 1e-06, + "loss": 1.0325, + "mean_token_accuracy": 0.6801154613494873, + "num_tokens": 182442234.0, + "step": 7050 + }, + { + "epoch": 0.774324621128926, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2098236083984375, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7178007960319519, + "num_tokens": 182466737.0, + "step": 7051 + }, + { + "epoch": 0.7744344388315396, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.051891326904297, + "learning_rate": 1e-06, + "loss": 1.0413, + "mean_token_accuracy": 0.6896252632141113, + "num_tokens": 182494786.0, + "step": 7052 + }, + { + "epoch": 0.7745442565341533, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3311400413513184, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7156041264533997, + "num_tokens": 182517728.0, + "step": 7053 + }, + { + "epoch": 0.774654074236767, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3602964878082275, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.725395679473877, + "num_tokens": 182539508.0, + "step": 7054 + }, + { + "epoch": 0.7747638919393807, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.122042655944824, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6929627656936646, + "num_tokens": 182565504.0, + "step": 7055 + }, + { + "epoch": 0.7748737096419943, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.081716775894165, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6950690746307373, + "num_tokens": 182593815.0, + "step": 7056 + }, + { + "epoch": 0.774983527344608, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.051239490509033, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7018215656280518, + "num_tokens": 182621249.0, + "step": 7057 + }, + { + "epoch": 0.7750933450472216, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1864712238311768, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7238648533821106, + "num_tokens": 182645287.0, + "step": 7058 + }, + { + "epoch": 0.7752031627498353, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.150817632675171, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6986321806907654, + "num_tokens": 182670630.0, + "step": 7059 + }, + { + "epoch": 0.7753129804524489, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2230136394500732, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7284274101257324, + "num_tokens": 182693031.0, + "step": 7060 + }, + { + "epoch": 0.7754227981550625, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1182830333709717, + "learning_rate": 1e-06, + "loss": 1.0675, + "mean_token_accuracy": 0.6791298389434814, + "num_tokens": 182720336.0, + "step": 7061 + }, + { + "epoch": 0.7755326158576763, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.210864782333374, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7059915065765381, + "num_tokens": 182744926.0, + "step": 7062 + }, + { + "epoch": 0.77564243356029, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.199869155883789, + "learning_rate": 1e-06, + "loss": 1.0551, + "mean_token_accuracy": 0.6840118169784546, + "num_tokens": 182771478.0, + "step": 7063 + }, + { + "epoch": 0.7757522512629036, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.426830291748047, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7032132148742676, + "num_tokens": 182795271.0, + "step": 7064 + }, + { + "epoch": 0.7758620689655172, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1635289192199707, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6943376660346985, + "num_tokens": 182820752.0, + "step": 7065 + }, + { + "epoch": 0.7759718866681309, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.032353639602661, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6910579800605774, + "num_tokens": 182850623.0, + "step": 7066 + }, + { + "epoch": 0.7760817043707445, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1819119453430176, + "learning_rate": 1e-06, + "loss": 1.0541, + "mean_token_accuracy": 0.6835207343101501, + "num_tokens": 182876812.0, + "step": 7067 + }, + { + "epoch": 0.7761915220733582, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.160637140274048, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7238476872444153, + "num_tokens": 182901447.0, + "step": 7068 + }, + { + "epoch": 0.7763013397759719, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.5799756050109863, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.6981310844421387, + "num_tokens": 182922621.0, + "step": 7069 + }, + { + "epoch": 0.7764111574785856, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2931265830993652, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7219012379646301, + "num_tokens": 182945907.0, + "step": 7070 + }, + { + "epoch": 0.7765209751811992, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.505625009536743, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.7072139382362366, + "num_tokens": 182967961.0, + "step": 7071 + }, + { + "epoch": 0.7766307928838129, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0749995708465576, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7010630965232849, + "num_tokens": 182997028.0, + "step": 7072 + }, + { + "epoch": 0.7767406105864265, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.206970691680908, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7177478075027466, + "num_tokens": 183020275.0, + "step": 7073 + }, + { + "epoch": 0.7768504282890402, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.299496650695801, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7066373825073242, + "num_tokens": 183042996.0, + "step": 7074 + }, + { + "epoch": 0.7769602459916538, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2888174057006836, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7041990160942078, + "num_tokens": 183069469.0, + "step": 7075 + }, + { + "epoch": 0.7770700636942676, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.235490322113037, + "learning_rate": 1e-06, + "loss": 1.0292, + "mean_token_accuracy": 0.7030215263366699, + "num_tokens": 183094259.0, + "step": 7076 + }, + { + "epoch": 0.7771798813968812, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.3449573516845703, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7342338562011719, + "num_tokens": 183114980.0, + "step": 7077 + }, + { + "epoch": 0.7772896990994949, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.253232955932617, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7077221274375916, + "num_tokens": 183140267.0, + "step": 7078 + }, + { + "epoch": 0.7773995168021085, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.2282307147979736, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7202303409576416, + "num_tokens": 183165152.0, + "step": 7079 + }, + { + "epoch": 0.7775093345047221, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0083861351013184, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7142512202262878, + "num_tokens": 183192657.0, + "step": 7080 + }, + { + "epoch": 0.7776191522073358, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.9627140760421753, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7371107339859009, + "num_tokens": 183221786.0, + "step": 7081 + }, + { + "epoch": 0.7777289699099494, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.0438308715820312, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7281363010406494, + "num_tokens": 183248432.0, + "step": 7082 + }, + { + "epoch": 0.7778387876125632, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.110985040664673, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7145036458969116, + "num_tokens": 183276389.0, + "step": 7083 + }, + { + "epoch": 0.7779486053151768, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1367790699005127, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7123599052429199, + "num_tokens": 183302609.0, + "step": 7084 + }, + { + "epoch": 0.7780584230177905, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.9117902517318726, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6974183320999146, + "num_tokens": 183336102.0, + "step": 7085 + }, + { + "epoch": 0.7781682407204041, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.9293501377105713, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.6986659169197083, + "num_tokens": 183365553.0, + "step": 7086 + }, + { + "epoch": 0.7782780584230178, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.1308445930480957, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6965732574462891, + "num_tokens": 183391577.0, + "step": 7087 + }, + { + "epoch": 0.7783878761256314, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.148289203643799, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6968653798103333, + "num_tokens": 183416801.0, + "step": 7088 + }, + { + "epoch": 0.7784976938282451, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 2.5218122005462646, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7363934516906738, + "num_tokens": 183435573.0, + "step": 7089 + }, + { + "epoch": 0.7786075115308587, + "ewc_loss": 3.7401914596557617e-06, + "grad_norm": 1.9813907146453857, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6901006698608398, + "num_tokens": 183467300.0, + "step": 7090 + }, + { + "epoch": 0.7787173292334725, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.3213322162628174, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7080574631690979, + "num_tokens": 183490195.0, + "step": 7091 + }, + { + "epoch": 0.7788271469360861, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.387428045272827, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7032321691513062, + "num_tokens": 183514431.0, + "step": 7092 + }, + { + "epoch": 0.7789369646386998, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2028138637542725, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7423766255378723, + "num_tokens": 183539682.0, + "step": 7093 + }, + { + "epoch": 0.7790467823413134, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.347264528274536, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7181968092918396, + "num_tokens": 183562472.0, + "step": 7094 + }, + { + "epoch": 0.7791566000439271, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.093262195587158, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7079176902770996, + "num_tokens": 183589502.0, + "step": 7095 + }, + { + "epoch": 0.7792664177465407, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.488283157348633, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.700547456741333, + "num_tokens": 183615986.0, + "step": 7096 + }, + { + "epoch": 0.7793762354491544, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.099634885787964, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6938433051109314, + "num_tokens": 183642905.0, + "step": 7097 + }, + { + "epoch": 0.7794860531517681, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.4918179512023926, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7181081771850586, + "num_tokens": 183664664.0, + "step": 7098 + }, + { + "epoch": 0.7795958708543818, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2397043704986572, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7310097217559814, + "num_tokens": 183688895.0, + "step": 7099 + }, + { + "epoch": 0.7797056885569954, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2129464149475098, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6985561847686768, + "num_tokens": 183716026.0, + "step": 7100 + }, + { + "epoch": 0.779815506259609, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 1.9743329286575317, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7075858116149902, + "num_tokens": 183744988.0, + "step": 7101 + }, + { + "epoch": 0.7799253239622227, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.180567502975464, + "learning_rate": 1e-06, + "loss": 1.0609, + "mean_token_accuracy": 0.678463339805603, + "num_tokens": 183771237.0, + "step": 7102 + }, + { + "epoch": 0.7800351416648363, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.1038548946380615, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.700889527797699, + "num_tokens": 183798604.0, + "step": 7103 + }, + { + "epoch": 0.78014495936745, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2361032962799072, + "learning_rate": 1e-06, + "loss": 1.0495, + "mean_token_accuracy": 0.681640625, + "num_tokens": 183824031.0, + "step": 7104 + }, + { + "epoch": 0.7802547770700637, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.330070972442627, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7104858160018921, + "num_tokens": 183847258.0, + "step": 7105 + }, + { + "epoch": 0.7803645947726774, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.174208641052246, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7171773910522461, + "num_tokens": 183872247.0, + "step": 7106 + }, + { + "epoch": 0.780474412475291, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.13271427154541, + "learning_rate": 1e-06, + "loss": 1.0612, + "mean_token_accuracy": 0.6900081038475037, + "num_tokens": 183899771.0, + "step": 7107 + }, + { + "epoch": 0.7805842301779047, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.035163640975952, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7039982080459595, + "num_tokens": 183927184.0, + "step": 7108 + }, + { + "epoch": 0.7806940478805183, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 1.9842824935913086, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7053971290588379, + "num_tokens": 183957129.0, + "step": 7109 + }, + { + "epoch": 0.780803865583132, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.04130482673645, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7053794264793396, + "num_tokens": 183985840.0, + "step": 7110 + }, + { + "epoch": 0.7809136832857456, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.1992897987365723, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6937861442565918, + "num_tokens": 184012366.0, + "step": 7111 + }, + { + "epoch": 0.7810235009883594, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.0563759803771973, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6905862092971802, + "num_tokens": 184040951.0, + "step": 7112 + }, + { + "epoch": 0.781133318690973, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.154352903366089, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7142970561981201, + "num_tokens": 184067588.0, + "step": 7113 + }, + { + "epoch": 0.7812431363935867, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.328200340270996, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7344343662261963, + "num_tokens": 184089857.0, + "step": 7114 + }, + { + "epoch": 0.7813529540962003, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.3402838706970215, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7239487171173096, + "num_tokens": 184113498.0, + "step": 7115 + }, + { + "epoch": 0.781462771798814, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.292861223220825, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6929718852043152, + "num_tokens": 184137404.0, + "step": 7116 + }, + { + "epoch": 0.7815725895014276, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.4667441844940186, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7010206580162048, + "num_tokens": 184159063.0, + "step": 7117 + }, + { + "epoch": 0.7816824072040413, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2314305305480957, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7136613130569458, + "num_tokens": 184181374.0, + "step": 7118 + }, + { + "epoch": 0.7817922249066549, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.2594006061553955, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7189087867736816, + "num_tokens": 184204543.0, + "step": 7119 + }, + { + "epoch": 0.7819020426092687, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.3247320652008057, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7135474681854248, + "num_tokens": 184227436.0, + "step": 7120 + }, + { + "epoch": 0.7820118603118823, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.176621437072754, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7171574234962463, + "num_tokens": 184252789.0, + "step": 7121 + }, + { + "epoch": 0.782121678014496, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.1658523082733154, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7154178619384766, + "num_tokens": 184279574.0, + "step": 7122 + }, + { + "epoch": 0.7822314957171096, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 1.9928741455078125, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6996939182281494, + "num_tokens": 184311071.0, + "step": 7123 + }, + { + "epoch": 0.7823413134197232, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.0617787837982178, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.686366081237793, + "num_tokens": 184339486.0, + "step": 7124 + }, + { + "epoch": 0.7824511311223369, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.3771862983703613, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7276938557624817, + "num_tokens": 184362691.0, + "step": 7125 + }, + { + "epoch": 0.7825609488249505, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 1.9691872596740723, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7043110728263855, + "num_tokens": 184391019.0, + "step": 7126 + }, + { + "epoch": 0.7826707665275643, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.285900831222534, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7198307514190674, + "num_tokens": 184412939.0, + "step": 7127 + }, + { + "epoch": 0.7827805842301779, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.009883403778076, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7159827351570129, + "num_tokens": 184441180.0, + "step": 7128 + }, + { + "epoch": 0.7828904019327916, + "ewc_loss": 3.7550926208496094e-06, + "grad_norm": 2.0137386322021484, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.730069100856781, + "num_tokens": 184471774.0, + "step": 7129 + }, + { + "epoch": 0.7830002196354052, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.32749080657959, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7102845907211304, + "num_tokens": 184494838.0, + "step": 7130 + }, + { + "epoch": 0.7831100373380189, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.0632874965667725, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7242422103881836, + "num_tokens": 184520574.0, + "step": 7131 + }, + { + "epoch": 0.7832198550406325, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.392799139022827, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7127367258071899, + "num_tokens": 184542020.0, + "step": 7132 + }, + { + "epoch": 0.7833296727432462, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.223115921020508, + "learning_rate": 1e-06, + "loss": 1.0114, + "mean_token_accuracy": 0.6884541511535645, + "num_tokens": 184567106.0, + "step": 7133 + }, + { + "epoch": 0.7834394904458599, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.096353769302368, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7141014337539673, + "num_tokens": 184592293.0, + "step": 7134 + }, + { + "epoch": 0.7835493081484736, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.3438334465026855, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6812340021133423, + "num_tokens": 184614675.0, + "step": 7135 + }, + { + "epoch": 0.7836591258510872, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.0347347259521484, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6927061676979065, + "num_tokens": 184643311.0, + "step": 7136 + }, + { + "epoch": 0.7837689435537009, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.078275442123413, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6890018582344055, + "num_tokens": 184671028.0, + "step": 7137 + }, + { + "epoch": 0.7838787612563145, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.0956358909606934, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7045584321022034, + "num_tokens": 184697887.0, + "step": 7138 + }, + { + "epoch": 0.7839885789589282, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 1.9894189834594727, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7133212685585022, + "num_tokens": 184728433.0, + "step": 7139 + }, + { + "epoch": 0.7840983966615418, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.306044340133667, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.709061861038208, + "num_tokens": 184754313.0, + "step": 7140 + }, + { + "epoch": 0.7842082143641556, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 1.9577621221542358, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6973309516906738, + "num_tokens": 184785843.0, + "step": 7141 + }, + { + "epoch": 0.7843180320667692, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.1184661388397217, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7178189754486084, + "num_tokens": 184811996.0, + "step": 7142 + }, + { + "epoch": 0.7844278497693828, + "ewc_loss": 3.769993782043457e-06, + "grad_norm": 2.4162039756774902, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.691907525062561, + "num_tokens": 184834675.0, + "step": 7143 + }, + { + "epoch": 0.7845376674719965, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.016028881072998, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7094588279724121, + "num_tokens": 184863266.0, + "step": 7144 + }, + { + "epoch": 0.7846474851746101, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.299703598022461, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.724455714225769, + "num_tokens": 184886024.0, + "step": 7145 + }, + { + "epoch": 0.7847573028772238, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.3436458110809326, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6986058950424194, + "num_tokens": 184909965.0, + "step": 7146 + }, + { + "epoch": 0.7848671205798374, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.3116848468780518, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7198272943496704, + "num_tokens": 184931186.0, + "step": 7147 + }, + { + "epoch": 0.7849769382824512, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.1537857055664062, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6894012689590454, + "num_tokens": 184959886.0, + "step": 7148 + }, + { + "epoch": 0.7850867559850648, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.1684350967407227, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7348116040229797, + "num_tokens": 184983109.0, + "step": 7149 + }, + { + "epoch": 0.7851965736876785, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.1677494049072266, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.6888619661331177, + "num_tokens": 185012198.0, + "step": 7150 + }, + { + "epoch": 0.7853063913902921, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.2764759063720703, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7382268905639648, + "num_tokens": 185033034.0, + "step": 7151 + }, + { + "epoch": 0.7854162090929058, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.126847505569458, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6995958089828491, + "num_tokens": 185059299.0, + "step": 7152 + }, + { + "epoch": 0.7855260267955194, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.076305389404297, + "learning_rate": 1e-06, + "loss": 1.0817, + "mean_token_accuracy": 0.6673191785812378, + "num_tokens": 185087271.0, + "step": 7153 + }, + { + "epoch": 0.7856358444981331, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.0775859355926514, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.7013558745384216, + "num_tokens": 185113305.0, + "step": 7154 + }, + { + "epoch": 0.7857456622007467, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.442784070968628, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7144727110862732, + "num_tokens": 185134830.0, + "step": 7155 + }, + { + "epoch": 0.7858554799033605, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.408385753631592, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7248237133026123, + "num_tokens": 185154646.0, + "step": 7156 + }, + { + "epoch": 0.7859652976059741, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 1.987910509109497, + "learning_rate": 1e-06, + "loss": 1.0255, + "mean_token_accuracy": 0.6977710723876953, + "num_tokens": 185183577.0, + "step": 7157 + }, + { + "epoch": 0.7860751153085878, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.4687821865081787, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.722296953201294, + "num_tokens": 185203013.0, + "step": 7158 + }, + { + "epoch": 0.7861849330112014, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.043478488922119, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6887112259864807, + "num_tokens": 185229869.0, + "step": 7159 + }, + { + "epoch": 0.786294750713815, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.181243896484375, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.7008619904518127, + "num_tokens": 185255496.0, + "step": 7160 + }, + { + "epoch": 0.7864045684164287, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.0160696506500244, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6941366195678711, + "num_tokens": 185284771.0, + "step": 7161 + }, + { + "epoch": 0.7865143861190423, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.416785955429077, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.72514808177948, + "num_tokens": 185306814.0, + "step": 7162 + }, + { + "epoch": 0.7866242038216561, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.2829627990722656, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6912813782691956, + "num_tokens": 185331860.0, + "step": 7163 + }, + { + "epoch": 0.7867340215242697, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.0537545680999756, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.7018863558769226, + "num_tokens": 185359863.0, + "step": 7164 + }, + { + "epoch": 0.7868438392268834, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.095111131668091, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.6992995142936707, + "num_tokens": 185387084.0, + "step": 7165 + }, + { + "epoch": 0.786953656929497, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.1618034839630127, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7279549837112427, + "num_tokens": 185413133.0, + "step": 7166 + }, + { + "epoch": 0.7870634746321107, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.1348860263824463, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.6952348947525024, + "num_tokens": 185437925.0, + "step": 7167 + }, + { + "epoch": 0.7871732923347243, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.160144567489624, + "learning_rate": 1e-06, + "loss": 1.0513, + "mean_token_accuracy": 0.6806739568710327, + "num_tokens": 185466086.0, + "step": 7168 + }, + { + "epoch": 0.787283110037338, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.073801279067993, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7131528854370117, + "num_tokens": 185493799.0, + "step": 7169 + }, + { + "epoch": 0.7873929277399517, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.113337516784668, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7029033899307251, + "num_tokens": 185522535.0, + "step": 7170 + }, + { + "epoch": 0.7875027454425654, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.03995418548584, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6836884021759033, + "num_tokens": 185556122.0, + "step": 7171 + }, + { + "epoch": 0.787612563145179, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.2420318126678467, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7166488170623779, + "num_tokens": 185582159.0, + "step": 7172 + }, + { + "epoch": 0.7877223808477927, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.254808187484741, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.6810449957847595, + "num_tokens": 185607770.0, + "step": 7173 + }, + { + "epoch": 0.7878321985504063, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.176204204559326, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7174965143203735, + "num_tokens": 185633153.0, + "step": 7174 + }, + { + "epoch": 0.78794201625302, + "ewc_loss": 3.7848949432373047e-06, + "grad_norm": 2.0359103679656982, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7143062353134155, + "num_tokens": 185662012.0, + "step": 7175 + }, + { + "epoch": 0.7880518339556336, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 1.9564703702926636, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6873687505722046, + "num_tokens": 185691755.0, + "step": 7176 + }, + { + "epoch": 0.7881616516582474, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.224651575088501, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7050652503967285, + "num_tokens": 185715651.0, + "step": 7177 + }, + { + "epoch": 0.788271469360861, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.2072575092315674, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6796100735664368, + "num_tokens": 185741311.0, + "step": 7178 + }, + { + "epoch": 0.7883812870634747, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.0694222450256348, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7237530946731567, + "num_tokens": 185768557.0, + "step": 7179 + }, + { + "epoch": 0.7884911047660883, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.1792755126953125, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6915993690490723, + "num_tokens": 185795015.0, + "step": 7180 + }, + { + "epoch": 0.788600922468702, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.2124147415161133, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.6930044889450073, + "num_tokens": 185819116.0, + "step": 7181 + }, + { + "epoch": 0.7887107401713156, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.366719961166382, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7101690769195557, + "num_tokens": 185841426.0, + "step": 7182 + }, + { + "epoch": 0.7888205578739292, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.1770074367523193, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.7043173313140869, + "num_tokens": 185866540.0, + "step": 7183 + }, + { + "epoch": 0.7889303755765429, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.1826610565185547, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7042103409767151, + "num_tokens": 185892827.0, + "step": 7184 + }, + { + "epoch": 0.7890401932791566, + "ewc_loss": 3.7997961044311523e-06, + "grad_norm": 2.4303760528564453, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7295889854431152, + "num_tokens": 185913702.0, + "step": 7185 + }, + { + "epoch": 0.7891500109817703, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.177968740463257, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7143517732620239, + "num_tokens": 185939475.0, + "step": 7186 + }, + { + "epoch": 0.7892598286843839, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.926339864730835, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6885216236114502, + "num_tokens": 185972425.0, + "step": 7187 + }, + { + "epoch": 0.7893696463869976, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.8655754327774048, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6977099776268005, + "num_tokens": 186007050.0, + "step": 7188 + }, + { + "epoch": 0.7894794640896112, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.2552578449249268, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7158370018005371, + "num_tokens": 186030479.0, + "step": 7189 + }, + { + "epoch": 0.7895892817922249, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.9829574823379517, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7019925117492676, + "num_tokens": 186061037.0, + "step": 7190 + }, + { + "epoch": 0.7896990994948385, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.154118776321411, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7057182788848877, + "num_tokens": 186086817.0, + "step": 7191 + }, + { + "epoch": 0.7898089171974523, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.0657901763916016, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.713545024394989, + "num_tokens": 186114481.0, + "step": 7192 + }, + { + "epoch": 0.7899187349000659, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.161238670349121, + "learning_rate": 1e-06, + "loss": 1.0246, + "mean_token_accuracy": 0.6895852088928223, + "num_tokens": 186142585.0, + "step": 7193 + }, + { + "epoch": 0.7900285526026796, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.110222816467285, + "learning_rate": 1e-06, + "loss": 0.9923, + "mean_token_accuracy": 0.7026633620262146, + "num_tokens": 186170466.0, + "step": 7194 + }, + { + "epoch": 0.7901383703052932, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.1540541648864746, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7092490196228027, + "num_tokens": 186197787.0, + "step": 7195 + }, + { + "epoch": 0.7902481880079069, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.3240814208984375, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7188969850540161, + "num_tokens": 186220283.0, + "step": 7196 + }, + { + "epoch": 0.7903580057105205, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.9774560928344727, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7033888101577759, + "num_tokens": 186250044.0, + "step": 7197 + }, + { + "epoch": 0.7904678234131342, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.9994659423828125, + "learning_rate": 1e-06, + "loss": 1.0357, + "mean_token_accuracy": 0.6911382079124451, + "num_tokens": 186279918.0, + "step": 7198 + }, + { + "epoch": 0.7905776411157479, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.2339296340942383, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6980812549591064, + "num_tokens": 186306253.0, + "step": 7199 + }, + { + "epoch": 0.7906874588183616, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.055877923965454, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.711147129535675, + "num_tokens": 186334988.0, + "step": 7200 + }, + { + "epoch": 0.7907972765209752, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.192979335784912, + "learning_rate": 1e-06, + "loss": 1.0602, + "mean_token_accuracy": 0.6766952276229858, + "num_tokens": 186362393.0, + "step": 7201 + }, + { + "epoch": 0.7909070942235888, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.931213140487671, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.725944995880127, + "num_tokens": 186391206.0, + "step": 7202 + }, + { + "epoch": 0.7910169119262025, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.996184229850769, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7053983807563782, + "num_tokens": 186420976.0, + "step": 7203 + }, + { + "epoch": 0.7911267296288161, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.1756181716918945, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7108378410339355, + "num_tokens": 186447364.0, + "step": 7204 + }, + { + "epoch": 0.7912365473314298, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.4997429847717285, + "learning_rate": 1e-06, + "loss": 1.0628, + "mean_token_accuracy": 0.6958997249603271, + "num_tokens": 186470996.0, + "step": 7205 + }, + { + "epoch": 0.7913463650340435, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.148993492126465, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7277880907058716, + "num_tokens": 186496877.0, + "step": 7206 + }, + { + "epoch": 0.7914561827366572, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.2775135040283203, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7201846837997437, + "num_tokens": 186519265.0, + "step": 7207 + }, + { + "epoch": 0.7915660004392708, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.086860179901123, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7011034488677979, + "num_tokens": 186548132.0, + "step": 7208 + }, + { + "epoch": 0.7916758181418845, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.0571463108062744, + "learning_rate": 1e-06, + "loss": 1.1143, + "mean_token_accuracy": 0.6648149490356445, + "num_tokens": 186577678.0, + "step": 7209 + }, + { + "epoch": 0.7917856358444981, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.9174693822860718, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7119513750076294, + "num_tokens": 186608047.0, + "step": 7210 + }, + { + "epoch": 0.7918954535471118, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.87024986743927, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6992309093475342, + "num_tokens": 186642725.0, + "step": 7211 + }, + { + "epoch": 0.7920052712497254, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.911651372909546, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7069154977798462, + "num_tokens": 186673867.0, + "step": 7212 + }, + { + "epoch": 0.7921150889523391, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 1.9356790781021118, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.691717267036438, + "num_tokens": 186704477.0, + "step": 7213 + }, + { + "epoch": 0.7922249066549528, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.1827917098999023, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6985063552856445, + "num_tokens": 186730133.0, + "step": 7214 + }, + { + "epoch": 0.7923347243575665, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.6206068992614746, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7105633020401001, + "num_tokens": 186750994.0, + "step": 7215 + }, + { + "epoch": 0.7924445420601801, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.545407772064209, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7139422297477722, + "num_tokens": 186770534.0, + "step": 7216 + }, + { + "epoch": 0.7925543597627938, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.098824977874756, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.695651113986969, + "num_tokens": 186798520.0, + "step": 7217 + }, + { + "epoch": 0.7926641774654074, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.025771141052246, + "learning_rate": 1e-06, + "loss": 1.0813, + "mean_token_accuracy": 0.6729578971862793, + "num_tokens": 186828948.0, + "step": 7218 + }, + { + "epoch": 0.792773995168021, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.1539559364318848, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.7003659009933472, + "num_tokens": 186857082.0, + "step": 7219 + }, + { + "epoch": 0.7928838128706347, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.4847326278686523, + "learning_rate": 1e-06, + "loss": 1.0244, + "mean_token_accuracy": 0.6913553476333618, + "num_tokens": 186880348.0, + "step": 7220 + }, + { + "epoch": 0.7929936305732485, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.597243547439575, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7083191275596619, + "num_tokens": 186899871.0, + "step": 7221 + }, + { + "epoch": 0.7931034482758621, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.130099058151245, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.6908837556838989, + "num_tokens": 186930246.0, + "step": 7222 + }, + { + "epoch": 0.7932132659784757, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.136822462081909, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7149765491485596, + "num_tokens": 186957223.0, + "step": 7223 + }, + { + "epoch": 0.7933230836810894, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.2969048023223877, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7093908786773682, + "num_tokens": 186980450.0, + "step": 7224 + }, + { + "epoch": 0.793432901383703, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.251251697540283, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7140688896179199, + "num_tokens": 187004492.0, + "step": 7225 + }, + { + "epoch": 0.7935427190863167, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.1508166790008545, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7175981998443604, + "num_tokens": 187030339.0, + "step": 7226 + }, + { + "epoch": 0.7936525367889303, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.2231802940368652, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.690963625907898, + "num_tokens": 187057485.0, + "step": 7227 + }, + { + "epoch": 0.7937623544915441, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.5039167404174805, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7102127075195312, + "num_tokens": 187077206.0, + "step": 7228 + }, + { + "epoch": 0.7938721721941577, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.090891122817993, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7110768556594849, + "num_tokens": 187104941.0, + "step": 7229 + }, + { + "epoch": 0.7939819898967714, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.3934977054595947, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7240055203437805, + "num_tokens": 187127950.0, + "step": 7230 + }, + { + "epoch": 0.794091807599385, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.174097776412964, + "learning_rate": 1e-06, + "loss": 1.0351, + "mean_token_accuracy": 0.6931412220001221, + "num_tokens": 187153998.0, + "step": 7231 + }, + { + "epoch": 0.7942016253019987, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.3229053020477295, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7243071794509888, + "num_tokens": 187174998.0, + "step": 7232 + }, + { + "epoch": 0.7943114430046123, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.104038953781128, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7101758718490601, + "num_tokens": 187201294.0, + "step": 7233 + }, + { + "epoch": 0.794421260707226, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.038513660430908, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.6997913122177124, + "num_tokens": 187229083.0, + "step": 7234 + }, + { + "epoch": 0.7945310784098397, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.3215789794921875, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7135101556777954, + "num_tokens": 187250489.0, + "step": 7235 + }, + { + "epoch": 0.7946408961124534, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.0974974632263184, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7195920348167419, + "num_tokens": 187277302.0, + "step": 7236 + }, + { + "epoch": 0.794750713815067, + "ewc_loss": 3.844499588012695e-06, + "grad_norm": 2.279090166091919, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7125537395477295, + "num_tokens": 187300062.0, + "step": 7237 + }, + { + "epoch": 0.7948605315176807, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.5033860206604004, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7208395600318909, + "num_tokens": 187319118.0, + "step": 7238 + }, + { + "epoch": 0.7949703492202943, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.148986339569092, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7089727520942688, + "num_tokens": 187345374.0, + "step": 7239 + }, + { + "epoch": 0.795080166922908, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.117157220840454, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.728507399559021, + "num_tokens": 187370379.0, + "step": 7240 + }, + { + "epoch": 0.7951899846255216, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.135711669921875, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7201160788536072, + "num_tokens": 187395194.0, + "step": 7241 + }, + { + "epoch": 0.7952998023281352, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3093831539154053, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7036410570144653, + "num_tokens": 187418656.0, + "step": 7242 + }, + { + "epoch": 0.795409620030749, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1989684104919434, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.7013276219367981, + "num_tokens": 187442710.0, + "step": 7243 + }, + { + "epoch": 0.7955194377333626, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.0936405658721924, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7125168442726135, + "num_tokens": 187470929.0, + "step": 7244 + }, + { + "epoch": 0.7956292554359763, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.0145325660705566, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.719140887260437, + "num_tokens": 187499503.0, + "step": 7245 + }, + { + "epoch": 0.7957390731385899, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.242692232131958, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7241604328155518, + "num_tokens": 187523173.0, + "step": 7246 + }, + { + "epoch": 0.7958488908412036, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.016782760620117, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7100281715393066, + "num_tokens": 187551392.0, + "step": 7247 + }, + { + "epoch": 0.7959587085438172, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.872245192527771, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6891096830368042, + "num_tokens": 187583865.0, + "step": 7248 + }, + { + "epoch": 0.7960685262464309, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3555784225463867, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7148556709289551, + "num_tokens": 187605803.0, + "step": 7249 + }, + { + "epoch": 0.7961783439490446, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3774681091308594, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7028511762619019, + "num_tokens": 187628523.0, + "step": 7250 + }, + { + "epoch": 0.7962881616516583, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.5442123413085938, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7036523818969727, + "num_tokens": 187648844.0, + "step": 7251 + }, + { + "epoch": 0.7963979793542719, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.5824148654937744, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.6997804641723633, + "num_tokens": 187668170.0, + "step": 7252 + }, + { + "epoch": 0.7965077970568856, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.143602132797241, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7090162038803101, + "num_tokens": 187695231.0, + "step": 7253 + }, + { + "epoch": 0.7966176147594992, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.392273187637329, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7491204142570496, + "num_tokens": 187716433.0, + "step": 7254 + }, + { + "epoch": 0.7967274324621129, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.390749931335449, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7411531209945679, + "num_tokens": 187736921.0, + "step": 7255 + }, + { + "epoch": 0.7968372501647265, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1236376762390137, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6987228393554688, + "num_tokens": 187761659.0, + "step": 7256 + }, + { + "epoch": 0.7969470678673403, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.9409693479537964, + "learning_rate": 1e-06, + "loss": 1.0675, + "mean_token_accuracy": 0.6784819960594177, + "num_tokens": 187794860.0, + "step": 7257 + }, + { + "epoch": 0.7970568855699539, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1023004055023193, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7042778730392456, + "num_tokens": 187822637.0, + "step": 7258 + }, + { + "epoch": 0.7971667032725676, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.9624096155166626, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.7039682269096375, + "num_tokens": 187853797.0, + "step": 7259 + }, + { + "epoch": 0.7972765209751812, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.44524884223938, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7179051637649536, + "num_tokens": 187874537.0, + "step": 7260 + }, + { + "epoch": 0.7973863386777948, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.9264562129974365, + "learning_rate": 1e-06, + "loss": 1.0806, + "mean_token_accuracy": 0.6722100377082825, + "num_tokens": 187907296.0, + "step": 7261 + }, + { + "epoch": 0.7974961563804085, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1081056594848633, + "learning_rate": 1e-06, + "loss": 1.1133, + "mean_token_accuracy": 0.6667075753211975, + "num_tokens": 187935828.0, + "step": 7262 + }, + { + "epoch": 0.7976059740830221, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.248793840408325, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.727636456489563, + "num_tokens": 187958585.0, + "step": 7263 + }, + { + "epoch": 0.7977157917856359, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3616254329681396, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7188469171524048, + "num_tokens": 187979118.0, + "step": 7264 + }, + { + "epoch": 0.7978256094882495, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.225149631500244, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7088068127632141, + "num_tokens": 188005438.0, + "step": 7265 + }, + { + "epoch": 0.7979354271908632, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3720479011535645, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7083780169487, + "num_tokens": 188029090.0, + "step": 7266 + }, + { + "epoch": 0.7980452448934768, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1984734535217285, + "learning_rate": 1e-06, + "loss": 1.0616, + "mean_token_accuracy": 0.6800404787063599, + "num_tokens": 188053893.0, + "step": 7267 + }, + { + "epoch": 0.7981550625960905, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.176461696624756, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.7002565860748291, + "num_tokens": 188078874.0, + "step": 7268 + }, + { + "epoch": 0.7982648802987041, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.198780059814453, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7030513882637024, + "num_tokens": 188105436.0, + "step": 7269 + }, + { + "epoch": 0.7983746980013178, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.9692203998565674, + "learning_rate": 1e-06, + "loss": 1.0677, + "mean_token_accuracy": 0.6879221200942993, + "num_tokens": 188139579.0, + "step": 7270 + }, + { + "epoch": 0.7984845157039314, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1690680980682373, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7206605672836304, + "num_tokens": 188165248.0, + "step": 7271 + }, + { + "epoch": 0.7985943334065452, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.0671920776367188, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6948370933532715, + "num_tokens": 188192813.0, + "step": 7272 + }, + { + "epoch": 0.7987041511091588, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.500368595123291, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7169030904769897, + "num_tokens": 188212555.0, + "step": 7273 + }, + { + "epoch": 0.7988139688117725, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1044275760650635, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7051304578781128, + "num_tokens": 188240737.0, + "step": 7274 + }, + { + "epoch": 0.7989237865143861, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.003384828567505, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7073982954025269, + "num_tokens": 188268722.0, + "step": 7275 + }, + { + "epoch": 0.7990336042169998, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.2961645126342773, + "learning_rate": 1e-06, + "loss": 1.0394, + "mean_token_accuracy": 0.6856375932693481, + "num_tokens": 188295061.0, + "step": 7276 + }, + { + "epoch": 0.7991434219196134, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1539390087127686, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7191897630691528, + "num_tokens": 188322008.0, + "step": 7277 + }, + { + "epoch": 0.799253239622227, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.0172762870788574, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6904667615890503, + "num_tokens": 188351369.0, + "step": 7278 + }, + { + "epoch": 0.7993630573248408, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.3084511756896973, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7225792407989502, + "num_tokens": 188374342.0, + "step": 7279 + }, + { + "epoch": 0.7994728750274545, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.1002984046936035, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6830666661262512, + "num_tokens": 188403080.0, + "step": 7280 + }, + { + "epoch": 0.7995826927300681, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 1.891159176826477, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6890178322792053, + "num_tokens": 188434808.0, + "step": 7281 + }, + { + "epoch": 0.7996925104326817, + "ewc_loss": 3.874301910400391e-06, + "grad_norm": 2.013033151626587, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.7138136625289917, + "num_tokens": 188465375.0, + "step": 7282 + }, + { + "epoch": 0.7998023281352954, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.9746677875518799, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7085549235343933, + "num_tokens": 188495748.0, + "step": 7283 + }, + { + "epoch": 0.799912145837909, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.4296300411224365, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.707199215888977, + "num_tokens": 188517345.0, + "step": 7284 + }, + { + "epoch": 0.8000219635405227, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1772615909576416, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7101116180419922, + "num_tokens": 188543347.0, + "step": 7285 + }, + { + "epoch": 0.8001317812431364, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.848452091217041, + "learning_rate": 1e-06, + "loss": 1.0856, + "mean_token_accuracy": 0.6787152886390686, + "num_tokens": 188577493.0, + "step": 7286 + }, + { + "epoch": 0.8002415989457501, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0683629512786865, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7152631878852844, + "num_tokens": 188602905.0, + "step": 7287 + }, + { + "epoch": 0.8003514166483637, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.151777982711792, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7087606191635132, + "num_tokens": 188629470.0, + "step": 7288 + }, + { + "epoch": 0.8004612343509774, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1144633293151855, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6959319114685059, + "num_tokens": 188657305.0, + "step": 7289 + }, + { + "epoch": 0.800571052053591, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.17208194732666, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.682249903678894, + "num_tokens": 188685231.0, + "step": 7290 + }, + { + "epoch": 0.8006808697562047, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.9562774896621704, + "learning_rate": 1e-06, + "loss": 1.0592, + "mean_token_accuracy": 0.6842875480651855, + "num_tokens": 188719881.0, + "step": 7291 + }, + { + "epoch": 0.8007906874588183, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.407709836959839, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7356815934181213, + "num_tokens": 188740271.0, + "step": 7292 + }, + { + "epoch": 0.8009005051614321, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.883571982383728, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7198737859725952, + "num_tokens": 188771459.0, + "step": 7293 + }, + { + "epoch": 0.8010103228640457, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.240586519241333, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7210370302200317, + "num_tokens": 188796163.0, + "step": 7294 + }, + { + "epoch": 0.8011201405666594, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.4525604248046875, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7107791900634766, + "num_tokens": 188818117.0, + "step": 7295 + }, + { + "epoch": 0.801229958269273, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3764729499816895, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.713716983795166, + "num_tokens": 188839835.0, + "step": 7296 + }, + { + "epoch": 0.8013397759718867, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.2237746715545654, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7060111165046692, + "num_tokens": 188864548.0, + "step": 7297 + }, + { + "epoch": 0.8014495936745003, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.027517080307007, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6976355314254761, + "num_tokens": 188894610.0, + "step": 7298 + }, + { + "epoch": 0.801559411377114, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.4667980670928955, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7236633896827698, + "num_tokens": 188915161.0, + "step": 7299 + }, + { + "epoch": 0.8016692290797276, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.179643392562866, + "learning_rate": 1e-06, + "loss": 1.0261, + "mean_token_accuracy": 0.6901007890701294, + "num_tokens": 188944775.0, + "step": 7300 + }, + { + "epoch": 0.8017790467823414, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.103945255279541, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7115854024887085, + "num_tokens": 188973964.0, + "step": 7301 + }, + { + "epoch": 0.801888864484955, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.204524040222168, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7132306694984436, + "num_tokens": 189000178.0, + "step": 7302 + }, + { + "epoch": 0.8019986821875686, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0157840251922607, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.7031145095825195, + "num_tokens": 189030193.0, + "step": 7303 + }, + { + "epoch": 0.8021084998901823, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.325779914855957, + "learning_rate": 1e-06, + "loss": 1.0012, + "mean_token_accuracy": 0.6884874701499939, + "num_tokens": 189053274.0, + "step": 7304 + }, + { + "epoch": 0.8022183175927959, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.078737735748291, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.689940333366394, + "num_tokens": 189082363.0, + "step": 7305 + }, + { + "epoch": 0.8023281352954096, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0923171043395996, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.6986618041992188, + "num_tokens": 189109955.0, + "step": 7306 + }, + { + "epoch": 0.8024379529980232, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.110316753387451, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7241949439048767, + "num_tokens": 189134395.0, + "step": 7307 + }, + { + "epoch": 0.802547770700637, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.142674446105957, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7046101689338684, + "num_tokens": 189162058.0, + "step": 7308 + }, + { + "epoch": 0.8026575884032506, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1493239402770996, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7002720832824707, + "num_tokens": 189187227.0, + "step": 7309 + }, + { + "epoch": 0.8027674061058643, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.193021535873413, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7183954119682312, + "num_tokens": 189212907.0, + "step": 7310 + }, + { + "epoch": 0.8028772238084779, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0508601665496826, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.711711049079895, + "num_tokens": 189241018.0, + "step": 7311 + }, + { + "epoch": 0.8029870415110916, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3215794563293457, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.7060679197311401, + "num_tokens": 189263735.0, + "step": 7312 + }, + { + "epoch": 0.8030968592137052, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.9841008186340332, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7243572473526001, + "num_tokens": 189293717.0, + "step": 7313 + }, + { + "epoch": 0.8032066769163189, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3098220825195312, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.7008618712425232, + "num_tokens": 189316183.0, + "step": 7314 + }, + { + "epoch": 0.8033164946189326, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.119554042816162, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7058225870132446, + "num_tokens": 189343617.0, + "step": 7315 + }, + { + "epoch": 0.8034263123215463, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.169745922088623, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6988475322723389, + "num_tokens": 189371355.0, + "step": 7316 + }, + { + "epoch": 0.8035361300241599, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.597695827484131, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7068041563034058, + "num_tokens": 189391749.0, + "step": 7317 + }, + { + "epoch": 0.8036459477267736, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.054525852203369, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6923681497573853, + "num_tokens": 189421161.0, + "step": 7318 + }, + { + "epoch": 0.8037557654293872, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.5146560668945312, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7212481498718262, + "num_tokens": 189440703.0, + "step": 7319 + }, + { + "epoch": 0.8038655831320008, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.467883825302124, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7117474675178528, + "num_tokens": 189460939.0, + "step": 7320 + }, + { + "epoch": 0.8039754008346145, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.018500804901123, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6965351700782776, + "num_tokens": 189492736.0, + "step": 7321 + }, + { + "epoch": 0.8040852185372283, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.251934289932251, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7018455266952515, + "num_tokens": 189517571.0, + "step": 7322 + }, + { + "epoch": 0.8041950362398419, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.080554246902466, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7152543663978577, + "num_tokens": 189544763.0, + "step": 7323 + }, + { + "epoch": 0.8043048539424555, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.091341018676758, + "learning_rate": 1e-06, + "loss": 1.0756, + "mean_token_accuracy": 0.6859669089317322, + "num_tokens": 189571734.0, + "step": 7324 + }, + { + "epoch": 0.8044146716450692, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0503973960876465, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7167659401893616, + "num_tokens": 189599544.0, + "step": 7325 + }, + { + "epoch": 0.8045244893476828, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.473597288131714, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6962300539016724, + "num_tokens": 189620801.0, + "step": 7326 + }, + { + "epoch": 0.8046343070502965, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.309293270111084, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.6957457065582275, + "num_tokens": 189642798.0, + "step": 7327 + }, + { + "epoch": 0.8047441247529101, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3523550033569336, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7139101028442383, + "num_tokens": 189667045.0, + "step": 7328 + }, + { + "epoch": 0.8048539424555239, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.188852310180664, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7184844613075256, + "num_tokens": 189693170.0, + "step": 7329 + }, + { + "epoch": 0.8049637601581375, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.426269769668579, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6962295770645142, + "num_tokens": 189717385.0, + "step": 7330 + }, + { + "epoch": 0.8050735778607512, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.146512269973755, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.707700788974762, + "num_tokens": 189743773.0, + "step": 7331 + }, + { + "epoch": 0.8051833955633648, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0293498039245605, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.6852182745933533, + "num_tokens": 189775152.0, + "step": 7332 + }, + { + "epoch": 0.8052932132659785, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.101242780685425, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7388383150100708, + "num_tokens": 189801396.0, + "step": 7333 + }, + { + "epoch": 0.8054030309685921, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3864500522613525, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6924265623092651, + "num_tokens": 189824889.0, + "step": 7334 + }, + { + "epoch": 0.8055128486712058, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.478440523147583, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7134743928909302, + "num_tokens": 189845178.0, + "step": 7335 + }, + { + "epoch": 0.8056226663738194, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1478939056396484, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7101613283157349, + "num_tokens": 189869934.0, + "step": 7336 + }, + { + "epoch": 0.8057324840764332, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1095778942108154, + "learning_rate": 1e-06, + "loss": 1.0258, + "mean_token_accuracy": 0.6902854442596436, + "num_tokens": 189897957.0, + "step": 7337 + }, + { + "epoch": 0.8058423017790468, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1263670921325684, + "learning_rate": 1e-06, + "loss": 1.0507, + "mean_token_accuracy": 0.6832023859024048, + "num_tokens": 189926967.0, + "step": 7338 + }, + { + "epoch": 0.8059521194816605, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.3369271755218506, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7246427536010742, + "num_tokens": 189948959.0, + "step": 7339 + }, + { + "epoch": 0.8060619371842741, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.9500480890274048, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7008651494979858, + "num_tokens": 189979548.0, + "step": 7340 + }, + { + "epoch": 0.8061717548868877, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.348991632461548, + "learning_rate": 1e-06, + "loss": 1.0147, + "mean_token_accuracy": 0.6928524374961853, + "num_tokens": 190003615.0, + "step": 7341 + }, + { + "epoch": 0.8062815725895014, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1086466312408447, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7305989861488342, + "num_tokens": 190027255.0, + "step": 7342 + }, + { + "epoch": 0.806391390292115, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0798630714416504, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7077538371086121, + "num_tokens": 190054949.0, + "step": 7343 + }, + { + "epoch": 0.8065012079947288, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1478798389434814, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7319599390029907, + "num_tokens": 190077434.0, + "step": 7344 + }, + { + "epoch": 0.8066110256973424, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.1468122005462646, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7096970677375793, + "num_tokens": 190105048.0, + "step": 7345 + }, + { + "epoch": 0.8067208433999561, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 1.9788835048675537, + "learning_rate": 1e-06, + "loss": 1.1267, + "mean_token_accuracy": 0.6678600907325745, + "num_tokens": 190135929.0, + "step": 7346 + }, + { + "epoch": 0.8068306611025697, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.0022976398468018, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.6930810213088989, + "num_tokens": 190164509.0, + "step": 7347 + }, + { + "epoch": 0.8069404788051834, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.2461695671081543, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.722960889339447, + "num_tokens": 190186916.0, + "step": 7348 + }, + { + "epoch": 0.807050296507797, + "ewc_loss": 3.933906555175781e-06, + "grad_norm": 2.113600492477417, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6972591876983643, + "num_tokens": 190215352.0, + "step": 7349 + }, + { + "epoch": 0.8071601142104107, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1331565380096436, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7160417437553406, + "num_tokens": 190240627.0, + "step": 7350 + }, + { + "epoch": 0.8072699319130244, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.171909809112549, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.701321005821228, + "num_tokens": 190266976.0, + "step": 7351 + }, + { + "epoch": 0.8073797496156381, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.322474479675293, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7274929285049438, + "num_tokens": 190290414.0, + "step": 7352 + }, + { + "epoch": 0.8074895673182517, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.4317731857299805, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6932907104492188, + "num_tokens": 190314442.0, + "step": 7353 + }, + { + "epoch": 0.8075993850208654, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.093827724456787, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7281414270401001, + "num_tokens": 190341818.0, + "step": 7354 + }, + { + "epoch": 0.807709202723479, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.10562801361084, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7276711463928223, + "num_tokens": 190368152.0, + "step": 7355 + }, + { + "epoch": 0.8078190204260927, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.41926646232605, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7077957987785339, + "num_tokens": 190387620.0, + "step": 7356 + }, + { + "epoch": 0.8079288381287063, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.227222442626953, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7092957496643066, + "num_tokens": 190411877.0, + "step": 7357 + }, + { + "epoch": 0.8080386558313201, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.2255749702453613, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7168389558792114, + "num_tokens": 190436207.0, + "step": 7358 + }, + { + "epoch": 0.8081484735339337, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.910588026046753, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7051835060119629, + "num_tokens": 190469819.0, + "step": 7359 + }, + { + "epoch": 0.8082582912365474, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0459001064300537, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7139066457748413, + "num_tokens": 190497336.0, + "step": 7360 + }, + { + "epoch": 0.808368108939161, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0724823474884033, + "learning_rate": 1e-06, + "loss": 1.0383, + "mean_token_accuracy": 0.6865267753601074, + "num_tokens": 190526504.0, + "step": 7361 + }, + { + "epoch": 0.8084779266417746, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.9607223272323608, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.698035717010498, + "num_tokens": 190559033.0, + "step": 7362 + }, + { + "epoch": 0.8085877443443883, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.6441009044647217, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7296501994132996, + "num_tokens": 190576453.0, + "step": 7363 + }, + { + "epoch": 0.8086975620470019, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.270197629928589, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6998440027236938, + "num_tokens": 190601617.0, + "step": 7364 + }, + { + "epoch": 0.8088073797496156, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.4624688625335693, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6992594599723816, + "num_tokens": 190622769.0, + "step": 7365 + }, + { + "epoch": 0.8089171974522293, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.9462363719940186, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7051193714141846, + "num_tokens": 190654941.0, + "step": 7366 + }, + { + "epoch": 0.809027015154843, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.2588956356048584, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6929961442947388, + "num_tokens": 190679722.0, + "step": 7367 + }, + { + "epoch": 0.8091368328574566, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.001669406890869, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7012882232666016, + "num_tokens": 190708192.0, + "step": 7368 + }, + { + "epoch": 0.8092466505600703, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.297621250152588, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6984516382217407, + "num_tokens": 190731374.0, + "step": 7369 + }, + { + "epoch": 0.8093564682626839, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1473658084869385, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.6980337500572205, + "num_tokens": 190759008.0, + "step": 7370 + }, + { + "epoch": 0.8094662859652976, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.123737335205078, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.708342432975769, + "num_tokens": 190784306.0, + "step": 7371 + }, + { + "epoch": 0.8095761036679112, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1764888763427734, + "learning_rate": 1e-06, + "loss": 1.0686, + "mean_token_accuracy": 0.6839348673820496, + "num_tokens": 190812971.0, + "step": 7372 + }, + { + "epoch": 0.809685921370525, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.2907609939575195, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7186440825462341, + "num_tokens": 190835592.0, + "step": 7373 + }, + { + "epoch": 0.8097957390731386, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.130253791809082, + "learning_rate": 1e-06, + "loss": 1.0631, + "mean_token_accuracy": 0.682039737701416, + "num_tokens": 190864363.0, + "step": 7374 + }, + { + "epoch": 0.8099055567757523, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.7065303325653076, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7079483866691589, + "num_tokens": 190882458.0, + "step": 7375 + }, + { + "epoch": 0.8100153744783659, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.16031551361084, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6962905526161194, + "num_tokens": 190909073.0, + "step": 7376 + }, + { + "epoch": 0.8101251921809796, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1882779598236084, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7191406488418579, + "num_tokens": 190933418.0, + "step": 7377 + }, + { + "epoch": 0.8102350098835932, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.3167178630828857, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6931136846542358, + "num_tokens": 190957448.0, + "step": 7378 + }, + { + "epoch": 0.8103448275862069, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0516140460968018, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7206673622131348, + "num_tokens": 190984754.0, + "step": 7379 + }, + { + "epoch": 0.8104546452888206, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.398172378540039, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6991782784461975, + "num_tokens": 191007254.0, + "step": 7380 + }, + { + "epoch": 0.8105644629914343, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.2620608806610107, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7081423997879028, + "num_tokens": 191030203.0, + "step": 7381 + }, + { + "epoch": 0.8106742806940479, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.255998373031616, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7030305862426758, + "num_tokens": 191053705.0, + "step": 7382 + }, + { + "epoch": 0.8107840983966615, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0863559246063232, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7104178071022034, + "num_tokens": 191081547.0, + "step": 7383 + }, + { + "epoch": 0.8108939160992752, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.758363723754883, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7260032296180725, + "num_tokens": 191100151.0, + "step": 7384 + }, + { + "epoch": 0.8110037338018888, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0333707332611084, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7170355319976807, + "num_tokens": 191128014.0, + "step": 7385 + }, + { + "epoch": 0.8111135515045025, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.107708692550659, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6849002242088318, + "num_tokens": 191154703.0, + "step": 7386 + }, + { + "epoch": 0.8112233692071162, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1888997554779053, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.706660270690918, + "num_tokens": 191180285.0, + "step": 7387 + }, + { + "epoch": 0.8113331869097299, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.3796591758728027, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.7029997110366821, + "num_tokens": 191201760.0, + "step": 7388 + }, + { + "epoch": 0.8114430046123435, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0175795555114746, + "learning_rate": 1e-06, + "loss": 1.0183, + "mean_token_accuracy": 0.6981415152549744, + "num_tokens": 191231226.0, + "step": 7389 + }, + { + "epoch": 0.8115528223149572, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.3424930572509766, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6845061779022217, + "num_tokens": 191254797.0, + "step": 7390 + }, + { + "epoch": 0.8116626400175708, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.383188009262085, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7247172594070435, + "num_tokens": 191278029.0, + "step": 7391 + }, + { + "epoch": 0.8117724577201845, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.9559733867645264, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7024976015090942, + "num_tokens": 191306205.0, + "step": 7392 + }, + { + "epoch": 0.8118822754227981, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.19577956199646, + "learning_rate": 1e-06, + "loss": 1.0213, + "mean_token_accuracy": 0.6908501982688904, + "num_tokens": 191331158.0, + "step": 7393 + }, + { + "epoch": 0.8119920931254118, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0533812046051025, + "learning_rate": 1e-06, + "loss": 1.04, + "mean_token_accuracy": 0.6945148706436157, + "num_tokens": 191362487.0, + "step": 7394 + }, + { + "epoch": 0.8121019108280255, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1578710079193115, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.7011463642120361, + "num_tokens": 191389257.0, + "step": 7395 + }, + { + "epoch": 0.8122117285306392, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.25821852684021, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7017467021942139, + "num_tokens": 191414774.0, + "step": 7396 + }, + { + "epoch": 0.8123215462332528, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.287367820739746, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7145460844039917, + "num_tokens": 191438055.0, + "step": 7397 + }, + { + "epoch": 0.8124313639358665, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.270124912261963, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7019690275192261, + "num_tokens": 191461414.0, + "step": 7398 + }, + { + "epoch": 0.8125411816384801, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.304032802581787, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7209784388542175, + "num_tokens": 191485426.0, + "step": 7399 + }, + { + "epoch": 0.8126509993410937, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.1323273181915283, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7059231400489807, + "num_tokens": 191513727.0, + "step": 7400 + }, + { + "epoch": 0.8127608170437074, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.214472532272339, + "learning_rate": 1e-06, + "loss": 1.0309, + "mean_token_accuracy": 0.686111330986023, + "num_tokens": 191541044.0, + "step": 7401 + }, + { + "epoch": 0.8128706347463212, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.769726037979126, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7332524061203003, + "num_tokens": 191575243.0, + "step": 7402 + }, + { + "epoch": 0.8129804524489348, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.171631336212158, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7152019143104553, + "num_tokens": 191599817.0, + "step": 7403 + }, + { + "epoch": 0.8130902701515484, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 1.9353692531585693, + "learning_rate": 1e-06, + "loss": 1.0753, + "mean_token_accuracy": 0.6739404201507568, + "num_tokens": 191633627.0, + "step": 7404 + }, + { + "epoch": 0.8132000878541621, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.129716157913208, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6896033883094788, + "num_tokens": 191661966.0, + "step": 7405 + }, + { + "epoch": 0.8133099055567757, + "ewc_loss": 4.023313522338867e-06, + "grad_norm": 2.0976755619049072, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.7012903690338135, + "num_tokens": 191689951.0, + "step": 7406 + }, + { + "epoch": 0.8134197232593894, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1783225536346436, + "learning_rate": 1e-06, + "loss": 1.0553, + "mean_token_accuracy": 0.6839043498039246, + "num_tokens": 191716482.0, + "step": 7407 + }, + { + "epoch": 0.813529540962003, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2883076667785645, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7027019262313843, + "num_tokens": 191741044.0, + "step": 7408 + }, + { + "epoch": 0.8136393586646168, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.071823835372925, + "learning_rate": 1e-06, + "loss": 1.1094, + "mean_token_accuracy": 0.6732197403907776, + "num_tokens": 191772667.0, + "step": 7409 + }, + { + "epoch": 0.8137491763672304, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0016684532165527, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7126480340957642, + "num_tokens": 191801028.0, + "step": 7410 + }, + { + "epoch": 0.8138589940698441, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.213163375854492, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7070274353027344, + "num_tokens": 191825693.0, + "step": 7411 + }, + { + "epoch": 0.8139688117724577, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1641831398010254, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7173313498497009, + "num_tokens": 191850044.0, + "step": 7412 + }, + { + "epoch": 0.8140786294750714, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.319836378097534, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7088085412979126, + "num_tokens": 191873183.0, + "step": 7413 + }, + { + "epoch": 0.814188447177685, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0647149085998535, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7092570066452026, + "num_tokens": 191899944.0, + "step": 7414 + }, + { + "epoch": 0.8142982648802987, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1149966716766357, + "learning_rate": 1e-06, + "loss": 1.0586, + "mean_token_accuracy": 0.6785051226615906, + "num_tokens": 191927847.0, + "step": 7415 + }, + { + "epoch": 0.8144080825829124, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1883625984191895, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7034707069396973, + "num_tokens": 191953831.0, + "step": 7416 + }, + { + "epoch": 0.8145179002855261, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1895511150360107, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.6996424198150635, + "num_tokens": 191979703.0, + "step": 7417 + }, + { + "epoch": 0.8146277179881397, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1142585277557373, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7231201529502869, + "num_tokens": 192007014.0, + "step": 7418 + }, + { + "epoch": 0.8147375356907534, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.118638277053833, + "learning_rate": 1e-06, + "loss": 1.0146, + "mean_token_accuracy": 0.686517596244812, + "num_tokens": 192034396.0, + "step": 7419 + }, + { + "epoch": 0.814847353393367, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.985040307044983, + "learning_rate": 1e-06, + "loss": 1.0134, + "mean_token_accuracy": 0.6897386312484741, + "num_tokens": 192065478.0, + "step": 7420 + }, + { + "epoch": 0.8149571710959806, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1710236072540283, + "learning_rate": 1e-06, + "loss": 1.0482, + "mean_token_accuracy": 0.6929036974906921, + "num_tokens": 192091541.0, + "step": 7421 + }, + { + "epoch": 0.8150669887985943, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3968660831451416, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7189860343933105, + "num_tokens": 192113111.0, + "step": 7422 + }, + { + "epoch": 0.8151768065012079, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.18692684173584, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6944766044616699, + "num_tokens": 192139874.0, + "step": 7423 + }, + { + "epoch": 0.8152866242038217, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.945540428161621, + "learning_rate": 1e-06, + "loss": 1.0565, + "mean_token_accuracy": 0.6862974166870117, + "num_tokens": 192173021.0, + "step": 7424 + }, + { + "epoch": 0.8153964419064353, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0880885124206543, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7271169424057007, + "num_tokens": 192200042.0, + "step": 7425 + }, + { + "epoch": 0.815506259609049, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9245306253433228, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.7041730880737305, + "num_tokens": 192233652.0, + "step": 7426 + }, + { + "epoch": 0.8156160773116626, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2112338542938232, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6936161518096924, + "num_tokens": 192259177.0, + "step": 7427 + }, + { + "epoch": 0.8157258950142763, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.017922878265381, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6895541548728943, + "num_tokens": 192289215.0, + "step": 7428 + }, + { + "epoch": 0.8158357127168899, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.182108163833618, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7284868955612183, + "num_tokens": 192311361.0, + "step": 7429 + }, + { + "epoch": 0.8159455304195036, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2102580070495605, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7189277410507202, + "num_tokens": 192336211.0, + "step": 7430 + }, + { + "epoch": 0.8160553481221173, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2708897590637207, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.715324878692627, + "num_tokens": 192359588.0, + "step": 7431 + }, + { + "epoch": 0.816165165824731, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1602611541748047, + "learning_rate": 1e-06, + "loss": 1.009, + "mean_token_accuracy": 0.6938039660453796, + "num_tokens": 192387109.0, + "step": 7432 + }, + { + "epoch": 0.8162749835273446, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1692895889282227, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.7046140432357788, + "num_tokens": 192414021.0, + "step": 7433 + }, + { + "epoch": 0.8163848012299583, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9591648578643799, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6952269077301025, + "num_tokens": 192444406.0, + "step": 7434 + }, + { + "epoch": 0.8164946189325719, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0569887161254883, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.6887215971946716, + "num_tokens": 192472462.0, + "step": 7435 + }, + { + "epoch": 0.8166044366351856, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.069913387298584, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.6985906362533569, + "num_tokens": 192501003.0, + "step": 7436 + }, + { + "epoch": 0.8167142543377992, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3196518421173096, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7213330268859863, + "num_tokens": 192521744.0, + "step": 7437 + }, + { + "epoch": 0.816824072040413, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9154143333435059, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7202320694923401, + "num_tokens": 192551757.0, + "step": 7438 + }, + { + "epoch": 0.8169338897430266, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.932080626487732, + "learning_rate": 1e-06, + "loss": 1.0232, + "mean_token_accuracy": 0.6921056509017944, + "num_tokens": 192585787.0, + "step": 7439 + }, + { + "epoch": 0.8170437074456403, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5054547786712646, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7192583680152893, + "num_tokens": 192605935.0, + "step": 7440 + }, + { + "epoch": 0.8171535251482539, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.267418384552002, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7129645943641663, + "num_tokens": 192629921.0, + "step": 7441 + }, + { + "epoch": 0.8172633428508675, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.970275640487671, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7176809906959534, + "num_tokens": 192660071.0, + "step": 7442 + }, + { + "epoch": 0.8173731605534812, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.384791851043701, + "learning_rate": 1e-06, + "loss": 1.0549, + "mean_token_accuracy": 0.6787979006767273, + "num_tokens": 192686540.0, + "step": 7443 + }, + { + "epoch": 0.8174829782560948, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.995743751525879, + "learning_rate": 1e-06, + "loss": 1.0479, + "mean_token_accuracy": 0.6890829801559448, + "num_tokens": 192718722.0, + "step": 7444 + }, + { + "epoch": 0.8175927959587086, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3899049758911133, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.6994011998176575, + "num_tokens": 192743632.0, + "step": 7445 + }, + { + "epoch": 0.8177026136613222, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.123609781265259, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7312307953834534, + "num_tokens": 192768444.0, + "step": 7446 + }, + { + "epoch": 0.8178124313639359, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1769142150878906, + "learning_rate": 1e-06, + "loss": 1.0647, + "mean_token_accuracy": 0.687691867351532, + "num_tokens": 192795814.0, + "step": 7447 + }, + { + "epoch": 0.8179222490665495, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0717694759368896, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7078423500061035, + "num_tokens": 192821027.0, + "step": 7448 + }, + { + "epoch": 0.8180320667691632, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.21258282661438, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6914283037185669, + "num_tokens": 192846075.0, + "step": 7449 + }, + { + "epoch": 0.8181418844717768, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8322267532348633, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7182695269584656, + "num_tokens": 192875195.0, + "step": 7450 + }, + { + "epoch": 0.8182517021743905, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.18965482711792, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.703160285949707, + "num_tokens": 192899409.0, + "step": 7451 + }, + { + "epoch": 0.8183615198770041, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.105221748352051, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7097594141960144, + "num_tokens": 192924224.0, + "step": 7452 + }, + { + "epoch": 0.8184713375796179, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4915826320648193, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7060929536819458, + "num_tokens": 192944988.0, + "step": 7453 + }, + { + "epoch": 0.8185811552822315, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.301140069961548, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7308828830718994, + "num_tokens": 192967089.0, + "step": 7454 + }, + { + "epoch": 0.8186909729848452, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.120483636856079, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7054835557937622, + "num_tokens": 192993389.0, + "step": 7455 + }, + { + "epoch": 0.8188007906874588, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.321154832839966, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7111291885375977, + "num_tokens": 193016217.0, + "step": 7456 + }, + { + "epoch": 0.8189106083900725, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1178643703460693, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6973039507865906, + "num_tokens": 193043656.0, + "step": 7457 + }, + { + "epoch": 0.8190204260926861, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3164258003234863, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7110081911087036, + "num_tokens": 193066359.0, + "step": 7458 + }, + { + "epoch": 0.8191302437952998, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.065211296081543, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.6902080774307251, + "num_tokens": 193096540.0, + "step": 7459 + }, + { + "epoch": 0.8192400614979135, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.292900323867798, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7029646039009094, + "num_tokens": 193119171.0, + "step": 7460 + }, + { + "epoch": 0.8193498792005272, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9722181558609009, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7022667527198792, + "num_tokens": 193149032.0, + "step": 7461 + }, + { + "epoch": 0.8194596969031408, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.131469488143921, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7097743153572083, + "num_tokens": 193174665.0, + "step": 7462 + }, + { + "epoch": 0.8195695146057544, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.314483642578125, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7107378244400024, + "num_tokens": 193198821.0, + "step": 7463 + }, + { + "epoch": 0.8196793323083681, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3219354152679443, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7523930072784424, + "num_tokens": 193219550.0, + "step": 7464 + }, + { + "epoch": 0.8197891500109817, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.137369394302368, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7196555137634277, + "num_tokens": 193245055.0, + "step": 7465 + }, + { + "epoch": 0.8198989677135954, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.925459623336792, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7172909379005432, + "num_tokens": 193274167.0, + "step": 7466 + }, + { + "epoch": 0.8200087854162091, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.222761869430542, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7231594324111938, + "num_tokens": 193297555.0, + "step": 7467 + }, + { + "epoch": 0.8201186031188228, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1483137607574463, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7101729512214661, + "num_tokens": 193324070.0, + "step": 7468 + }, + { + "epoch": 0.8202284208214364, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.398179054260254, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.6986292004585266, + "num_tokens": 193347653.0, + "step": 7469 + }, + { + "epoch": 0.8203382385240501, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1714248657226562, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.710660457611084, + "num_tokens": 193372170.0, + "step": 7470 + }, + { + "epoch": 0.8204480562266637, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2105228900909424, + "learning_rate": 1e-06, + "loss": 1.0935, + "mean_token_accuracy": 0.6693036556243896, + "num_tokens": 193404706.0, + "step": 7471 + }, + { + "epoch": 0.8205578739292774, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1389946937561035, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.7176805138587952, + "num_tokens": 193431115.0, + "step": 7472 + }, + { + "epoch": 0.820667691631891, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.186845302581787, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.6950569152832031, + "num_tokens": 193458507.0, + "step": 7473 + }, + { + "epoch": 0.8207775093345048, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0637435913085938, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.7026634812355042, + "num_tokens": 193488381.0, + "step": 7474 + }, + { + "epoch": 0.8208873270371184, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.048663854598999, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6969080567359924, + "num_tokens": 193519142.0, + "step": 7475 + }, + { + "epoch": 0.8209971447397321, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2591161727905273, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7162344455718994, + "num_tokens": 193543492.0, + "step": 7476 + }, + { + "epoch": 0.8211069624423457, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2281908988952637, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6990703344345093, + "num_tokens": 193567255.0, + "step": 7477 + }, + { + "epoch": 0.8212167801449594, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.428476572036743, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7102559208869934, + "num_tokens": 193588674.0, + "step": 7478 + }, + { + "epoch": 0.821326597847573, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.065176486968994, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.6957957744598389, + "num_tokens": 193617385.0, + "step": 7479 + }, + { + "epoch": 0.8214364155501866, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.137070894241333, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7333279848098755, + "num_tokens": 193641889.0, + "step": 7480 + }, + { + "epoch": 0.8215462332528004, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4375431537628174, + "learning_rate": 1e-06, + "loss": 1.0188, + "mean_token_accuracy": 0.6908925771713257, + "num_tokens": 193663586.0, + "step": 7481 + }, + { + "epoch": 0.821656050955414, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.410830020904541, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7090371251106262, + "num_tokens": 193685218.0, + "step": 7482 + }, + { + "epoch": 0.8217658686580277, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.217939853668213, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7195206880569458, + "num_tokens": 193708146.0, + "step": 7483 + }, + { + "epoch": 0.8218756863606413, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.493380308151245, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7140045166015625, + "num_tokens": 193727761.0, + "step": 7484 + }, + { + "epoch": 0.821985504063255, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2425053119659424, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7009949684143066, + "num_tokens": 193751345.0, + "step": 7485 + }, + { + "epoch": 0.8220953217658686, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0567638874053955, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7303916215896606, + "num_tokens": 193781026.0, + "step": 7486 + }, + { + "epoch": 0.8222051394684823, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1680402755737305, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7140200138092041, + "num_tokens": 193807025.0, + "step": 7487 + }, + { + "epoch": 0.8223149571710959, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.108381986618042, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7343685030937195, + "num_tokens": 193831146.0, + "step": 7488 + }, + { + "epoch": 0.8224247748737097, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.218312978744507, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7318785190582275, + "num_tokens": 193856325.0, + "step": 7489 + }, + { + "epoch": 0.8225345925763233, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1472249031066895, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.6989636421203613, + "num_tokens": 193881662.0, + "step": 7490 + }, + { + "epoch": 0.822644410278937, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0978963375091553, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7022124528884888, + "num_tokens": 193909299.0, + "step": 7491 + }, + { + "epoch": 0.8227542279815506, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8504043817520142, + "learning_rate": 1e-06, + "loss": 1.1057, + "mean_token_accuracy": 0.669145941734314, + "num_tokens": 193945491.0, + "step": 7492 + }, + { + "epoch": 0.8228640456841643, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9929590225219727, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.6993092894554138, + "num_tokens": 193972892.0, + "step": 7493 + }, + { + "epoch": 0.8229738633867779, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4139769077301025, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.699372410774231, + "num_tokens": 193994630.0, + "step": 7494 + }, + { + "epoch": 0.8230836810893916, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.347625970840454, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7267568707466125, + "num_tokens": 194016765.0, + "step": 7495 + }, + { + "epoch": 0.8231934987920053, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1165223121643066, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7136763334274292, + "num_tokens": 194041630.0, + "step": 7496 + }, + { + "epoch": 0.823303316494619, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.66898775100708, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7045630812644958, + "num_tokens": 194059924.0, + "step": 7497 + }, + { + "epoch": 0.8234131341972326, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.110339403152466, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.706702709197998, + "num_tokens": 194087610.0, + "step": 7498 + }, + { + "epoch": 0.8235229518998463, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.138033866882324, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7214757800102234, + "num_tokens": 194112974.0, + "step": 7499 + }, + { + "epoch": 0.8236327696024599, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8742239475250244, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7254905700683594, + "num_tokens": 194145084.0, + "step": 7500 + }, + { + "epoch": 0.8237425873050735, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0245590209960938, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7091554403305054, + "num_tokens": 194170949.0, + "step": 7501 + }, + { + "epoch": 0.8238524050076872, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3348441123962402, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7106049060821533, + "num_tokens": 194194584.0, + "step": 7502 + }, + { + "epoch": 0.823962222710301, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1943092346191406, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7021067142486572, + "num_tokens": 194219238.0, + "step": 7503 + }, + { + "epoch": 0.8240720404129146, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1727347373962402, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7173561453819275, + "num_tokens": 194245490.0, + "step": 7504 + }, + { + "epoch": 0.8241818581155282, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9986423254013062, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.6984083652496338, + "num_tokens": 194277489.0, + "step": 7505 + }, + { + "epoch": 0.8242916758181419, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0009994506835938, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7273277640342712, + "num_tokens": 194304550.0, + "step": 7506 + }, + { + "epoch": 0.8244014935207555, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2677361965179443, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7250016927719116, + "num_tokens": 194328241.0, + "step": 7507 + }, + { + "epoch": 0.8245113112233692, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2640631198883057, + "learning_rate": 1e-06, + "loss": 1.0857, + "mean_token_accuracy": 0.6888341903686523, + "num_tokens": 194353809.0, + "step": 7508 + }, + { + "epoch": 0.8246211289259828, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.45310115814209, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7410919666290283, + "num_tokens": 194374441.0, + "step": 7509 + }, + { + "epoch": 0.8247309466285966, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1020078659057617, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7169626355171204, + "num_tokens": 194401876.0, + "step": 7510 + }, + { + "epoch": 0.8248407643312102, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3518574237823486, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7287329435348511, + "num_tokens": 194423064.0, + "step": 7511 + }, + { + "epoch": 0.8249505820338239, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.112762689590454, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7212948203086853, + "num_tokens": 194449636.0, + "step": 7512 + }, + { + "epoch": 0.8250603997364375, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.117985963821411, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7179192304611206, + "num_tokens": 194476248.0, + "step": 7513 + }, + { + "epoch": 0.8251702174390512, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0873210430145264, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7005013227462769, + "num_tokens": 194504609.0, + "step": 7514 + }, + { + "epoch": 0.8252800351416648, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2443645000457764, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7108950018882751, + "num_tokens": 194526147.0, + "step": 7515 + }, + { + "epoch": 0.8253898528442785, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1498098373413086, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.7027788162231445, + "num_tokens": 194552878.0, + "step": 7516 + }, + { + "epoch": 0.8254996705468921, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9940091371536255, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6993657350540161, + "num_tokens": 194581847.0, + "step": 7517 + }, + { + "epoch": 0.8256094882495059, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9803794622421265, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7078977823257446, + "num_tokens": 194610977.0, + "step": 7518 + }, + { + "epoch": 0.8257193059521195, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.217582941055298, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7555208206176758, + "num_tokens": 194633139.0, + "step": 7519 + }, + { + "epoch": 0.8258291236547332, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2080187797546387, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6973618268966675, + "num_tokens": 194658417.0, + "step": 7520 + }, + { + "epoch": 0.8259389413573468, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3084235191345215, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7077394723892212, + "num_tokens": 194682723.0, + "step": 7521 + }, + { + "epoch": 0.8260487590599604, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2083702087402344, + "learning_rate": 1e-06, + "loss": 1.0829, + "mean_token_accuracy": 0.6748422980308533, + "num_tokens": 194709719.0, + "step": 7522 + }, + { + "epoch": 0.8261585767625741, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2427573204040527, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7234807014465332, + "num_tokens": 194735545.0, + "step": 7523 + }, + { + "epoch": 0.8262683944651877, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0618178844451904, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7105317711830139, + "num_tokens": 194764463.0, + "step": 7524 + }, + { + "epoch": 0.8263782121678015, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2314844131469727, + "learning_rate": 1e-06, + "loss": 0.9948, + "mean_token_accuracy": 0.7056843042373657, + "num_tokens": 194788477.0, + "step": 7525 + }, + { + "epoch": 0.8264880298704151, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.016448736190796, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7141919136047363, + "num_tokens": 194818075.0, + "step": 7526 + }, + { + "epoch": 0.8265978475730288, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1027700901031494, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7119043469429016, + "num_tokens": 194844525.0, + "step": 7527 + }, + { + "epoch": 0.8267076652756424, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9379057884216309, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6924242973327637, + "num_tokens": 194875526.0, + "step": 7528 + }, + { + "epoch": 0.8268174829782561, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0435659885406494, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.706674337387085, + "num_tokens": 194902455.0, + "step": 7529 + }, + { + "epoch": 0.8269273006808697, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2760417461395264, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7292296886444092, + "num_tokens": 194924955.0, + "step": 7530 + }, + { + "epoch": 0.8270371183834834, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8945149183273315, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7161394357681274, + "num_tokens": 194956923.0, + "step": 7531 + }, + { + "epoch": 0.8271469360860971, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.090771436691284, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7131508588790894, + "num_tokens": 194982918.0, + "step": 7532 + }, + { + "epoch": 0.8272567537887108, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.230083703994751, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6964802145957947, + "num_tokens": 195006390.0, + "step": 7533 + }, + { + "epoch": 0.8273665714913244, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2035350799560547, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.722008466720581, + "num_tokens": 195030450.0, + "step": 7534 + }, + { + "epoch": 0.8274763891939381, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.11846661567688, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.6856240630149841, + "num_tokens": 195057157.0, + "step": 7535 + }, + { + "epoch": 0.8275862068965517, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.131990671157837, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.6892077922821045, + "num_tokens": 195086748.0, + "step": 7536 + }, + { + "epoch": 0.8276960245991654, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9690630435943604, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7059261798858643, + "num_tokens": 195116699.0, + "step": 7537 + }, + { + "epoch": 0.827805842301779, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0492706298828125, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7107150554656982, + "num_tokens": 195144108.0, + "step": 7538 + }, + { + "epoch": 0.8279156600043928, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2170660495758057, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7133423089981079, + "num_tokens": 195167984.0, + "step": 7539 + }, + { + "epoch": 0.8280254777070064, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.655799388885498, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7419170141220093, + "num_tokens": 195186092.0, + "step": 7540 + }, + { + "epoch": 0.82813529540962, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.068002700805664, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7058559656143188, + "num_tokens": 195213490.0, + "step": 7541 + }, + { + "epoch": 0.8282451131122337, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.481788396835327, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7054677605628967, + "num_tokens": 195235215.0, + "step": 7542 + }, + { + "epoch": 0.8283549308148473, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9044795036315918, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6889282464981079, + "num_tokens": 195268719.0, + "step": 7543 + }, + { + "epoch": 0.828464748517461, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.012416124343872, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6863186359405518, + "num_tokens": 195302496.0, + "step": 7544 + }, + { + "epoch": 0.8285745662200746, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2086739540100098, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.7048342227935791, + "num_tokens": 195328694.0, + "step": 7545 + }, + { + "epoch": 0.8286843839226883, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2419514656066895, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.71720951795578, + "num_tokens": 195352880.0, + "step": 7546 + }, + { + "epoch": 0.828794201625302, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.039602518081665, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7350096702575684, + "num_tokens": 195381758.0, + "step": 7547 + }, + { + "epoch": 0.8289040193279157, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2355730533599854, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6936917304992676, + "num_tokens": 195406281.0, + "step": 7548 + }, + { + "epoch": 0.8290138370305293, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.231926441192627, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6944628953933716, + "num_tokens": 195431651.0, + "step": 7549 + }, + { + "epoch": 0.829123654733143, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0322165489196777, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7131506204605103, + "num_tokens": 195460148.0, + "step": 7550 + }, + { + "epoch": 0.8292334724357566, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1836354732513428, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7452746629714966, + "num_tokens": 195484865.0, + "step": 7551 + }, + { + "epoch": 0.8293432901383703, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2066357135772705, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7114918231964111, + "num_tokens": 195510871.0, + "step": 7552 + }, + { + "epoch": 0.8294531078409839, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.050611734390259, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6974468231201172, + "num_tokens": 195540329.0, + "step": 7553 + }, + { + "epoch": 0.8295629255435977, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2735254764556885, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7307533621788025, + "num_tokens": 195563569.0, + "step": 7554 + }, + { + "epoch": 0.8296727432462113, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.510444402694702, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7102957963943481, + "num_tokens": 195584615.0, + "step": 7555 + }, + { + "epoch": 0.829782560948825, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.065382957458496, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6870524883270264, + "num_tokens": 195612512.0, + "step": 7556 + }, + { + "epoch": 0.8298923786514386, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.480093002319336, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7175149917602539, + "num_tokens": 195632874.0, + "step": 7557 + }, + { + "epoch": 0.8300021963540523, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.409128189086914, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6970312595367432, + "num_tokens": 195655467.0, + "step": 7558 + }, + { + "epoch": 0.8301120140566659, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9341853857040405, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.7084330320358276, + "num_tokens": 195687886.0, + "step": 7559 + }, + { + "epoch": 0.8302218317592795, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2928202152252197, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.7059048414230347, + "num_tokens": 195711794.0, + "step": 7560 + }, + { + "epoch": 0.8303316494618933, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.039769411087036, + "learning_rate": 1e-06, + "loss": 1.0673, + "mean_token_accuracy": 0.6803684234619141, + "num_tokens": 195742931.0, + "step": 7561 + }, + { + "epoch": 0.830441467164507, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.560100793838501, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7165406346321106, + "num_tokens": 195762619.0, + "step": 7562 + }, + { + "epoch": 0.8305512848671206, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.279642105102539, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7289226055145264, + "num_tokens": 195784610.0, + "step": 7563 + }, + { + "epoch": 0.8306611025697342, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1285173892974854, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7060564160346985, + "num_tokens": 195810897.0, + "step": 7564 + }, + { + "epoch": 0.8307709202723479, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8404003381729126, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6867293119430542, + "num_tokens": 195846421.0, + "step": 7565 + }, + { + "epoch": 0.8308807379749615, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.316305160522461, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7015314102172852, + "num_tokens": 195868398.0, + "step": 7566 + }, + { + "epoch": 0.8309905556775752, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.38063383102417, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.69847172498703, + "num_tokens": 195889234.0, + "step": 7567 + }, + { + "epoch": 0.8311003733801889, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.52921986579895, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7060871124267578, + "num_tokens": 195908887.0, + "step": 7568 + }, + { + "epoch": 0.8312101910828026, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.430492401123047, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6927011013031006, + "num_tokens": 195932480.0, + "step": 7569 + }, + { + "epoch": 0.8313200087854162, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.138667345046997, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.7041001915931702, + "num_tokens": 195959682.0, + "step": 7570 + }, + { + "epoch": 0.8314298264880299, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.034010171890259, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6879922151565552, + "num_tokens": 195992913.0, + "step": 7571 + }, + { + "epoch": 0.8315396441906435, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.390996217727661, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7105107307434082, + "num_tokens": 196013906.0, + "step": 7572 + }, + { + "epoch": 0.8316494618932572, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3653383255004883, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6869863271713257, + "num_tokens": 196039076.0, + "step": 7573 + }, + { + "epoch": 0.8317592795958708, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9568865299224854, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7022184133529663, + "num_tokens": 196070521.0, + "step": 7574 + }, + { + "epoch": 0.8318690972984845, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.141035556793213, + "learning_rate": 1e-06, + "loss": 1.0632, + "mean_token_accuracy": 0.6780660152435303, + "num_tokens": 196097730.0, + "step": 7575 + }, + { + "epoch": 0.8319789150010982, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.304905652999878, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.739999532699585, + "num_tokens": 196117974.0, + "step": 7576 + }, + { + "epoch": 0.8320887327037119, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1349823474884033, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7057958841323853, + "num_tokens": 196146404.0, + "step": 7577 + }, + { + "epoch": 0.8321985504063255, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1777689456939697, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7149354815483093, + "num_tokens": 196171219.0, + "step": 7578 + }, + { + "epoch": 0.8323083681089392, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.370234251022339, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6853974461555481, + "num_tokens": 196194458.0, + "step": 7579 + }, + { + "epoch": 0.8324181858115528, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1442983150482178, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7262755036354065, + "num_tokens": 196219471.0, + "step": 7580 + }, + { + "epoch": 0.8325280035141664, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1859469413757324, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6945735216140747, + "num_tokens": 196243947.0, + "step": 7581 + }, + { + "epoch": 0.8326378212167801, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.101303815841675, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7019204497337341, + "num_tokens": 196270203.0, + "step": 7582 + }, + { + "epoch": 0.8327476389193939, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.194873094558716, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7119297981262207, + "num_tokens": 196297066.0, + "step": 7583 + }, + { + "epoch": 0.8328574566220075, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3818039894104004, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7152878046035767, + "num_tokens": 196318379.0, + "step": 7584 + }, + { + "epoch": 0.8329672743246211, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.448207139968872, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7195355296134949, + "num_tokens": 196338137.0, + "step": 7585 + }, + { + "epoch": 0.8330770920272348, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.046163320541382, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6971005201339722, + "num_tokens": 196367424.0, + "step": 7586 + }, + { + "epoch": 0.8331869097298484, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1341662406921387, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7181912660598755, + "num_tokens": 196391314.0, + "step": 7587 + }, + { + "epoch": 0.8332967274324621, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.155334949493408, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7010750770568848, + "num_tokens": 196416374.0, + "step": 7588 + }, + { + "epoch": 0.8334065451350757, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1123249530792236, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7311184406280518, + "num_tokens": 196441639.0, + "step": 7589 + }, + { + "epoch": 0.8335163628376895, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8946433067321777, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7171118259429932, + "num_tokens": 196471450.0, + "step": 7590 + }, + { + "epoch": 0.8336261805403031, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9694534540176392, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6907435655593872, + "num_tokens": 196500994.0, + "step": 7591 + }, + { + "epoch": 0.8337359982429168, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0878186225891113, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7136660814285278, + "num_tokens": 196526947.0, + "step": 7592 + }, + { + "epoch": 0.8338458159455304, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.309502124786377, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7082357406616211, + "num_tokens": 196550249.0, + "step": 7593 + }, + { + "epoch": 0.8339556336481441, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2541840076446533, + "learning_rate": 1e-06, + "loss": 1.0691, + "mean_token_accuracy": 0.686996340751648, + "num_tokens": 196577281.0, + "step": 7594 + }, + { + "epoch": 0.8340654513507577, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.193507194519043, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7093009948730469, + "num_tokens": 196602095.0, + "step": 7595 + }, + { + "epoch": 0.8341752690533714, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9940587282180786, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6965546607971191, + "num_tokens": 196629629.0, + "step": 7596 + }, + { + "epoch": 0.8342850867559851, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4475576877593994, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7078877687454224, + "num_tokens": 196652009.0, + "step": 7597 + }, + { + "epoch": 0.8343949044585988, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1653096675872803, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7258830666542053, + "num_tokens": 196677268.0, + "step": 7598 + }, + { + "epoch": 0.8345047221612124, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1975362300872803, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6973909139633179, + "num_tokens": 196703110.0, + "step": 7599 + }, + { + "epoch": 0.834614539863826, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8540552854537964, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.70298832654953, + "num_tokens": 196735845.0, + "step": 7600 + }, + { + "epoch": 0.8347243575664397, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0782525539398193, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7030202150344849, + "num_tokens": 196764382.0, + "step": 7601 + }, + { + "epoch": 0.8348341752690533, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3062210083007812, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.722923755645752, + "num_tokens": 196785955.0, + "step": 7602 + }, + { + "epoch": 0.834943992971667, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.044264078140259, + "learning_rate": 1e-06, + "loss": 1.0856, + "mean_token_accuracy": 0.6754595637321472, + "num_tokens": 196815121.0, + "step": 7603 + }, + { + "epoch": 0.8350538106742806, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.917340636253357, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7130571603775024, + "num_tokens": 196844968.0, + "step": 7604 + }, + { + "epoch": 0.8351636283768944, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.150510311126709, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7040996551513672, + "num_tokens": 196872347.0, + "step": 7605 + }, + { + "epoch": 0.835273446079508, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3591055870056152, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6967606544494629, + "num_tokens": 196896899.0, + "step": 7606 + }, + { + "epoch": 0.8353832637821217, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0682830810546875, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7038466334342957, + "num_tokens": 196925898.0, + "step": 7607 + }, + { + "epoch": 0.8354930814847353, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2604377269744873, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7004416584968567, + "num_tokens": 196949512.0, + "step": 7608 + }, + { + "epoch": 0.835602899187349, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.511587381362915, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7230507135391235, + "num_tokens": 196969131.0, + "step": 7609 + }, + { + "epoch": 0.8357127168899626, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.096353769302368, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.693927526473999, + "num_tokens": 196996637.0, + "step": 7610 + }, + { + "epoch": 0.8358225345925763, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4117298126220703, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7092618942260742, + "num_tokens": 197017944.0, + "step": 7611 + }, + { + "epoch": 0.83593235229519, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3899757862091064, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7182186245918274, + "num_tokens": 197039417.0, + "step": 7612 + }, + { + "epoch": 0.8360421699978037, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.145768642425537, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7109030485153198, + "num_tokens": 197063264.0, + "step": 7613 + }, + { + "epoch": 0.8361519877004173, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3066818714141846, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7307068109512329, + "num_tokens": 197085752.0, + "step": 7614 + }, + { + "epoch": 0.836261805403031, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.23933744430542, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6934995651245117, + "num_tokens": 197114138.0, + "step": 7615 + }, + { + "epoch": 0.8363716231056446, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3745810985565186, + "learning_rate": 1e-06, + "loss": 1.0405, + "mean_token_accuracy": 0.6827911138534546, + "num_tokens": 197136693.0, + "step": 7616 + }, + { + "epoch": 0.8364814408082583, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1230688095092773, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7234004735946655, + "num_tokens": 197161869.0, + "step": 7617 + }, + { + "epoch": 0.8365912585108719, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.007312536239624, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7143672704696655, + "num_tokens": 197192424.0, + "step": 7618 + }, + { + "epoch": 0.8367010762134857, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.941021203994751, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.711593508720398, + "num_tokens": 197221800.0, + "step": 7619 + }, + { + "epoch": 0.8368108939160993, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.08667254447937, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7091832160949707, + "num_tokens": 197249437.0, + "step": 7620 + }, + { + "epoch": 0.836920711618713, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.338901996612549, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7169805765151978, + "num_tokens": 197271744.0, + "step": 7621 + }, + { + "epoch": 0.8370305293213266, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.235504388809204, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7107409238815308, + "num_tokens": 197295207.0, + "step": 7622 + }, + { + "epoch": 0.8371403470239402, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3415050506591797, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7095078825950623, + "num_tokens": 197319632.0, + "step": 7623 + }, + { + "epoch": 0.8372501647265539, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1449809074401855, + "learning_rate": 1e-06, + "loss": 1.0391, + "mean_token_accuracy": 0.6852474808692932, + "num_tokens": 197350363.0, + "step": 7624 + }, + { + "epoch": 0.8373599824291675, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.184748649597168, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7243468761444092, + "num_tokens": 197375405.0, + "step": 7625 + }, + { + "epoch": 0.8374698001317813, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1498892307281494, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6950393915176392, + "num_tokens": 197401800.0, + "step": 7626 + }, + { + "epoch": 0.8375796178343949, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2552196979522705, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6939346790313721, + "num_tokens": 197425361.0, + "step": 7627 + }, + { + "epoch": 0.8376894355370086, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0757198333740234, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7117539644241333, + "num_tokens": 197451855.0, + "step": 7628 + }, + { + "epoch": 0.8377992532396222, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.33738112449646, + "learning_rate": 1e-06, + "loss": 1.0255, + "mean_token_accuracy": 0.6924761533737183, + "num_tokens": 197475433.0, + "step": 7629 + }, + { + "epoch": 0.8379090709422359, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5858235359191895, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.717525839805603, + "num_tokens": 197495708.0, + "step": 7630 + }, + { + "epoch": 0.8380188886448495, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9007805585861206, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6972933411598206, + "num_tokens": 197529731.0, + "step": 7631 + }, + { + "epoch": 0.8381287063474632, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.25788950920105, + "learning_rate": 1e-06, + "loss": 1.0072, + "mean_token_accuracy": 0.6891767382621765, + "num_tokens": 197555343.0, + "step": 7632 + }, + { + "epoch": 0.8382385240500769, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0068089962005615, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7130642533302307, + "num_tokens": 197585097.0, + "step": 7633 + }, + { + "epoch": 0.8383483417526906, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0706090927124023, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7420021295547485, + "num_tokens": 197610696.0, + "step": 7634 + }, + { + "epoch": 0.8384581594553042, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.022084951400757, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7024240493774414, + "num_tokens": 197638076.0, + "step": 7635 + }, + { + "epoch": 0.8385679771579179, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.492678165435791, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7092263102531433, + "num_tokens": 197659044.0, + "step": 7636 + }, + { + "epoch": 0.8386777948605315, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1389358043670654, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7060238718986511, + "num_tokens": 197684359.0, + "step": 7637 + }, + { + "epoch": 0.8387876125631452, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.395692825317383, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7110586166381836, + "num_tokens": 197706907.0, + "step": 7638 + }, + { + "epoch": 0.8388974302657588, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.105590343475342, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.708898663520813, + "num_tokens": 197734448.0, + "step": 7639 + }, + { + "epoch": 0.8390072479683724, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0978198051452637, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7062755823135376, + "num_tokens": 197760172.0, + "step": 7640 + }, + { + "epoch": 0.8391170656709862, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2104923725128174, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7061643600463867, + "num_tokens": 197786884.0, + "step": 7641 + }, + { + "epoch": 0.8392268833735999, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.289152145385742, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7112284302711487, + "num_tokens": 197809095.0, + "step": 7642 + }, + { + "epoch": 0.8393367010762135, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1325063705444336, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7126351594924927, + "num_tokens": 197835389.0, + "step": 7643 + }, + { + "epoch": 0.8394465187788271, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.455756187438965, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7329651117324829, + "num_tokens": 197857238.0, + "step": 7644 + }, + { + "epoch": 0.8395563364814408, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.842395305633545, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6950012445449829, + "num_tokens": 197891332.0, + "step": 7645 + }, + { + "epoch": 0.8396661541840544, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.177462339401245, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6924537420272827, + "num_tokens": 197917516.0, + "step": 7646 + }, + { + "epoch": 0.8397759718866681, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3454222679138184, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7228044867515564, + "num_tokens": 197937808.0, + "step": 7647 + }, + { + "epoch": 0.8398857895892818, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9854915142059326, + "learning_rate": 1e-06, + "loss": 1.0818, + "mean_token_accuracy": 0.672858476638794, + "num_tokens": 197970058.0, + "step": 7648 + }, + { + "epoch": 0.8399956072918955, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.04294753074646, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6932241916656494, + "num_tokens": 198000588.0, + "step": 7649 + }, + { + "epoch": 0.8401054249945091, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.943115234375, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6956862807273865, + "num_tokens": 198030599.0, + "step": 7650 + }, + { + "epoch": 0.8402152426971228, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.168180227279663, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7106573581695557, + "num_tokens": 198055412.0, + "step": 7651 + }, + { + "epoch": 0.8403250603997364, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.177574634552002, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7228572368621826, + "num_tokens": 198079265.0, + "step": 7652 + }, + { + "epoch": 0.8404348781023501, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.533461093902588, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.712674617767334, + "num_tokens": 198099381.0, + "step": 7653 + }, + { + "epoch": 0.8405446958049637, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2636983394622803, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7014387845993042, + "num_tokens": 198122592.0, + "step": 7654 + }, + { + "epoch": 0.8406545135075775, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3856773376464844, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7092846035957336, + "num_tokens": 198146233.0, + "step": 7655 + }, + { + "epoch": 0.8407643312101911, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.314851999282837, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7100332379341125, + "num_tokens": 198170128.0, + "step": 7656 + }, + { + "epoch": 0.8408741489128048, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2113029956817627, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.727255642414093, + "num_tokens": 198195310.0, + "step": 7657 + }, + { + "epoch": 0.8409839666154184, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1363625526428223, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7002569437026978, + "num_tokens": 198221303.0, + "step": 7658 + }, + { + "epoch": 0.8410937843180321, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1582820415496826, + "learning_rate": 1e-06, + "loss": 1.0626, + "mean_token_accuracy": 0.690416157245636, + "num_tokens": 198247562.0, + "step": 7659 + }, + { + "epoch": 0.8412036020206457, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0585901737213135, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6883931159973145, + "num_tokens": 198276906.0, + "step": 7660 + }, + { + "epoch": 0.8413134197232593, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6316800117492676, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7141443490982056, + "num_tokens": 198295104.0, + "step": 7661 + }, + { + "epoch": 0.8414232374258731, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0582058429718018, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7028241157531738, + "num_tokens": 198323331.0, + "step": 7662 + }, + { + "epoch": 0.8415330551284868, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.172024965286255, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7049764394760132, + "num_tokens": 198350668.0, + "step": 7663 + }, + { + "epoch": 0.8416428728311004, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9085402488708496, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7155388593673706, + "num_tokens": 198381292.0, + "step": 7664 + }, + { + "epoch": 0.841752690533714, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2450969219207764, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7120639085769653, + "num_tokens": 198406274.0, + "step": 7665 + }, + { + "epoch": 0.8418625082363277, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2564003467559814, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.6995221972465515, + "num_tokens": 198430130.0, + "step": 7666 + }, + { + "epoch": 0.8419723259389413, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.949936866760254, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.726349949836731, + "num_tokens": 198459777.0, + "step": 7667 + }, + { + "epoch": 0.842082143641555, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1894004344940186, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.7010509967803955, + "num_tokens": 198485192.0, + "step": 7668 + }, + { + "epoch": 0.8421919613441686, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5013513565063477, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7308366298675537, + "num_tokens": 198503231.0, + "step": 7669 + }, + { + "epoch": 0.8423017790467824, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.266982316970825, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7203196883201599, + "num_tokens": 198528863.0, + "step": 7670 + }, + { + "epoch": 0.842411596749396, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1648762226104736, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7170904278755188, + "num_tokens": 198553857.0, + "step": 7671 + }, + { + "epoch": 0.8425214144520097, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.061554431915283, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.7002687454223633, + "num_tokens": 198583558.0, + "step": 7672 + }, + { + "epoch": 0.8426312321546233, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1803431510925293, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7283203601837158, + "num_tokens": 198609291.0, + "step": 7673 + }, + { + "epoch": 0.842741049857237, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0929033756256104, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7121209502220154, + "num_tokens": 198636900.0, + "step": 7674 + }, + { + "epoch": 0.8428508675598506, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9301435947418213, + "learning_rate": 1e-06, + "loss": 1.0081, + "mean_token_accuracy": 0.6924324035644531, + "num_tokens": 198668433.0, + "step": 7675 + }, + { + "epoch": 0.8429606852624643, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.193147659301758, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7158063054084778, + "num_tokens": 198694562.0, + "step": 7676 + }, + { + "epoch": 0.843070502965078, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.394772529602051, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7074589729309082, + "num_tokens": 198717369.0, + "step": 7677 + }, + { + "epoch": 0.8431803206676917, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0667006969451904, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7125078439712524, + "num_tokens": 198744856.0, + "step": 7678 + }, + { + "epoch": 0.8432901383703053, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1640570163726807, + "learning_rate": 1e-06, + "loss": 1.0344, + "mean_token_accuracy": 0.6899532079696655, + "num_tokens": 198771538.0, + "step": 7679 + }, + { + "epoch": 0.843399956072919, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.239003896713257, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.693016529083252, + "num_tokens": 198796885.0, + "step": 7680 + }, + { + "epoch": 0.8435097737755326, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2956552505493164, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7160248756408691, + "num_tokens": 198818940.0, + "step": 7681 + }, + { + "epoch": 0.8436195914781462, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.192417860031128, + "learning_rate": 1e-06, + "loss": 1.0475, + "mean_token_accuracy": 0.677958071231842, + "num_tokens": 198845612.0, + "step": 7682 + }, + { + "epoch": 0.8437294091807599, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0401625633239746, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7096184492111206, + "num_tokens": 198874429.0, + "step": 7683 + }, + { + "epoch": 0.8438392268833736, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.338700294494629, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7114530801773071, + "num_tokens": 198899257.0, + "step": 7684 + }, + { + "epoch": 0.8439490445859873, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3974833488464355, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.715219259262085, + "num_tokens": 198923303.0, + "step": 7685 + }, + { + "epoch": 0.8440588622886009, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1906771659851074, + "learning_rate": 1e-06, + "loss": 1.1111, + "mean_token_accuracy": 0.6616817712783813, + "num_tokens": 198953795.0, + "step": 7686 + }, + { + "epoch": 0.8441686799912146, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.131709337234497, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7034297585487366, + "num_tokens": 198981725.0, + "step": 7687 + }, + { + "epoch": 0.8442784976938282, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2063350677490234, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7124873399734497, + "num_tokens": 199005153.0, + "step": 7688 + }, + { + "epoch": 0.8443883153964419, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.363748073577881, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7277930974960327, + "num_tokens": 199027042.0, + "step": 7689 + }, + { + "epoch": 0.8444981330990555, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.378553628921509, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7115252614021301, + "num_tokens": 199050146.0, + "step": 7690 + }, + { + "epoch": 0.8446079508016693, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4099254608154297, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.69910728931427, + "num_tokens": 199073404.0, + "step": 7691 + }, + { + "epoch": 0.8447177685042829, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1397786140441895, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6993722915649414, + "num_tokens": 199101008.0, + "step": 7692 + }, + { + "epoch": 0.8448275862068966, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2603049278259277, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6946427822113037, + "num_tokens": 199126165.0, + "step": 7693 + }, + { + "epoch": 0.8449374039095102, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0480563640594482, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7059807777404785, + "num_tokens": 199153533.0, + "step": 7694 + }, + { + "epoch": 0.8450472216121239, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0326428413391113, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7388228178024292, + "num_tokens": 199179424.0, + "step": 7695 + }, + { + "epoch": 0.8451570393147375, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.6982526779174805, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7012907266616821, + "num_tokens": 199202764.0, + "step": 7696 + }, + { + "epoch": 0.8452668570173512, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2578136920928955, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7181961536407471, + "num_tokens": 199227940.0, + "step": 7697 + }, + { + "epoch": 0.8453766747199648, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.184558868408203, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7113065719604492, + "num_tokens": 199252255.0, + "step": 7698 + }, + { + "epoch": 0.8454864924225786, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0633533000946045, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7128128409385681, + "num_tokens": 199278817.0, + "step": 7699 + }, + { + "epoch": 0.8455963101251922, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1848621368408203, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7104352712631226, + "num_tokens": 199304937.0, + "step": 7700 + }, + { + "epoch": 0.8457061278278059, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8381954431533813, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7310041189193726, + "num_tokens": 199337317.0, + "step": 7701 + }, + { + "epoch": 0.8458159455304195, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1765174865722656, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7082116603851318, + "num_tokens": 199363824.0, + "step": 7702 + }, + { + "epoch": 0.8459257632330331, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.215301990509033, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7076207399368286, + "num_tokens": 199388320.0, + "step": 7703 + }, + { + "epoch": 0.8460355809356468, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1604738235473633, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6978687644004822, + "num_tokens": 199415523.0, + "step": 7704 + }, + { + "epoch": 0.8461453986382604, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.268137216567993, + "learning_rate": 1e-06, + "loss": 1.0811, + "mean_token_accuracy": 0.6898601651191711, + "num_tokens": 199441645.0, + "step": 7705 + }, + { + "epoch": 0.8462552163408742, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5279972553253174, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7192825078964233, + "num_tokens": 199462863.0, + "step": 7706 + }, + { + "epoch": 0.8463650340434878, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3870487213134766, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7300875782966614, + "num_tokens": 199484387.0, + "step": 7707 + }, + { + "epoch": 0.8464748517461015, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1189494132995605, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7047015428543091, + "num_tokens": 199513679.0, + "step": 7708 + }, + { + "epoch": 0.8465846694487151, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1438722610473633, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6881759762763977, + "num_tokens": 199542808.0, + "step": 7709 + }, + { + "epoch": 0.8466944871513288, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3243870735168457, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6931962966918945, + "num_tokens": 199566968.0, + "step": 7710 + }, + { + "epoch": 0.8468043048539424, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1836066246032715, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7189449071884155, + "num_tokens": 199593692.0, + "step": 7711 + }, + { + "epoch": 0.8469141225565561, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9842748641967773, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7046927213668823, + "num_tokens": 199622402.0, + "step": 7712 + }, + { + "epoch": 0.8470239402591698, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.717902660369873, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7206064462661743, + "num_tokens": 199641522.0, + "step": 7713 + }, + { + "epoch": 0.8471337579617835, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.119375705718994, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7115138173103333, + "num_tokens": 199667927.0, + "step": 7714 + }, + { + "epoch": 0.8472435756643971, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3242008686065674, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6895650625228882, + "num_tokens": 199693552.0, + "step": 7715 + }, + { + "epoch": 0.8473533933670108, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2781646251678467, + "learning_rate": 1e-06, + "loss": 1.0537, + "mean_token_accuracy": 0.6850064992904663, + "num_tokens": 199716886.0, + "step": 7716 + }, + { + "epoch": 0.8474632110696244, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4143729209899902, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.712013840675354, + "num_tokens": 199737928.0, + "step": 7717 + }, + { + "epoch": 0.8475730287722381, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0432329177856445, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7124912738800049, + "num_tokens": 199766914.0, + "step": 7718 + }, + { + "epoch": 0.8476828464748517, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.934927225112915, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7081406116485596, + "num_tokens": 199797223.0, + "step": 7719 + }, + { + "epoch": 0.8477926641774655, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2655375003814697, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7363818883895874, + "num_tokens": 199819060.0, + "step": 7720 + }, + { + "epoch": 0.8479024818800791, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2292656898498535, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.712775468826294, + "num_tokens": 199845241.0, + "step": 7721 + }, + { + "epoch": 0.8480122995826928, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.253859758377075, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7250693440437317, + "num_tokens": 199868686.0, + "step": 7722 + }, + { + "epoch": 0.8481221172853064, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.538544178009033, + "learning_rate": 1e-06, + "loss": 1.0667, + "mean_token_accuracy": 0.6826298236846924, + "num_tokens": 199892312.0, + "step": 7723 + }, + { + "epoch": 0.84823193498792, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.275080442428589, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7097203731536865, + "num_tokens": 199917331.0, + "step": 7724 + }, + { + "epoch": 0.8483417526905337, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2368836402893066, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7098520994186401, + "num_tokens": 199942069.0, + "step": 7725 + }, + { + "epoch": 0.8484515703931473, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2059099674224854, + "learning_rate": 1e-06, + "loss": 1.0209, + "mean_token_accuracy": 0.6934087872505188, + "num_tokens": 199967921.0, + "step": 7726 + }, + { + "epoch": 0.848561388095761, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1363682746887207, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7203530073165894, + "num_tokens": 199993258.0, + "step": 7727 + }, + { + "epoch": 0.8486712057983747, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9451876878738403, + "learning_rate": 1e-06, + "loss": 1.06, + "mean_token_accuracy": 0.6761903762817383, + "num_tokens": 200025395.0, + "step": 7728 + }, + { + "epoch": 0.8487810235009884, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.567547082901001, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.694161057472229, + "num_tokens": 200047088.0, + "step": 7729 + }, + { + "epoch": 0.848890841203602, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.116208791732788, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7180653810501099, + "num_tokens": 200073044.0, + "step": 7730 + }, + { + "epoch": 0.8490006589062157, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9801331758499146, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7043387293815613, + "num_tokens": 200103693.0, + "step": 7731 + }, + { + "epoch": 0.8491104766088293, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1263911724090576, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7189196348190308, + "num_tokens": 200129605.0, + "step": 7732 + }, + { + "epoch": 0.849220294311443, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0691349506378174, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.7017605304718018, + "num_tokens": 200159419.0, + "step": 7733 + }, + { + "epoch": 0.8493301120140566, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.274674654006958, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.685710072517395, + "num_tokens": 200183858.0, + "step": 7734 + }, + { + "epoch": 0.8494399297166704, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.006556272506714, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6859561204910278, + "num_tokens": 200215080.0, + "step": 7735 + }, + { + "epoch": 0.849549747419284, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.416698694229126, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.692026674747467, + "num_tokens": 200236944.0, + "step": 7736 + }, + { + "epoch": 0.8496595651218977, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1291143894195557, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.7038564682006836, + "num_tokens": 200265185.0, + "step": 7737 + }, + { + "epoch": 0.8497693828245113, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.993796944618225, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7061216235160828, + "num_tokens": 200293716.0, + "step": 7738 + }, + { + "epoch": 0.849879200527125, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.194981336593628, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7033724188804626, + "num_tokens": 200318403.0, + "step": 7739 + }, + { + "epoch": 0.8499890182297386, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0526254177093506, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7341823577880859, + "num_tokens": 200343584.0, + "step": 7740 + }, + { + "epoch": 0.8500988359323522, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8830069303512573, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7073062062263489, + "num_tokens": 200375251.0, + "step": 7741 + }, + { + "epoch": 0.850208653634966, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.040900945663452, + "learning_rate": 1e-06, + "loss": 1.0367, + "mean_token_accuracy": 0.6935927271842957, + "num_tokens": 200406054.0, + "step": 7742 + }, + { + "epoch": 0.8503184713375797, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4176502227783203, + "learning_rate": 1e-06, + "loss": 1.0263, + "mean_token_accuracy": 0.7077349424362183, + "num_tokens": 200427875.0, + "step": 7743 + }, + { + "epoch": 0.8504282890401933, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.015907049179077, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7273603081703186, + "num_tokens": 200454987.0, + "step": 7744 + }, + { + "epoch": 0.8505381067428069, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1970386505126953, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7307140827178955, + "num_tokens": 200478848.0, + "step": 7745 + }, + { + "epoch": 0.8506479244454206, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.193401336669922, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7165805101394653, + "num_tokens": 200502938.0, + "step": 7746 + }, + { + "epoch": 0.8507577421480342, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2059237957000732, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7309523224830627, + "num_tokens": 200527962.0, + "step": 7747 + }, + { + "epoch": 0.8508675598506479, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.185987949371338, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7067238092422485, + "num_tokens": 200554384.0, + "step": 7748 + }, + { + "epoch": 0.8509773775532616, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.178467273712158, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.6970983743667603, + "num_tokens": 200580256.0, + "step": 7749 + }, + { + "epoch": 0.8510871952558753, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4126532077789307, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.702423095703125, + "num_tokens": 200602913.0, + "step": 7750 + }, + { + "epoch": 0.8511970129584889, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0117244720458984, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.6974467635154724, + "num_tokens": 200637997.0, + "step": 7751 + }, + { + "epoch": 0.8513068306611026, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.248403310775757, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7200307250022888, + "num_tokens": 200663680.0, + "step": 7752 + }, + { + "epoch": 0.8514166483637162, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.118443727493286, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.723644495010376, + "num_tokens": 200687698.0, + "step": 7753 + }, + { + "epoch": 0.8515264660663299, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0518200397491455, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6949758529663086, + "num_tokens": 200718240.0, + "step": 7754 + }, + { + "epoch": 0.8516362837689435, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2294018268585205, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6860466599464417, + "num_tokens": 200743003.0, + "step": 7755 + }, + { + "epoch": 0.8517461014715572, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1293938159942627, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6845170855522156, + "num_tokens": 200772187.0, + "step": 7756 + }, + { + "epoch": 0.8518559191741709, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.219571828842163, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7240046262741089, + "num_tokens": 200800494.0, + "step": 7757 + }, + { + "epoch": 0.8519657368767846, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.413635492324829, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7217264175415039, + "num_tokens": 200820127.0, + "step": 7758 + }, + { + "epoch": 0.8520755545793982, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5595476627349854, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.735519528388977, + "num_tokens": 200838496.0, + "step": 7759 + }, + { + "epoch": 0.8521853722820119, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2331390380859375, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.7056162357330322, + "num_tokens": 200863720.0, + "step": 7760 + }, + { + "epoch": 0.8522951899846255, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.236829996109009, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.7100242972373962, + "num_tokens": 200890496.0, + "step": 7761 + }, + { + "epoch": 0.8524050076872391, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.018266201019287, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7134828567504883, + "num_tokens": 200920864.0, + "step": 7762 + }, + { + "epoch": 0.8525148253898528, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.287414073944092, + "learning_rate": 1e-06, + "loss": 1.0457, + "mean_token_accuracy": 0.6821466684341431, + "num_tokens": 200945141.0, + "step": 7763 + }, + { + "epoch": 0.8526246430924665, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1176815032958984, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7089230418205261, + "num_tokens": 200973178.0, + "step": 7764 + }, + { + "epoch": 0.8527344607950802, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.213844060897827, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7209633588790894, + "num_tokens": 200996047.0, + "step": 7765 + }, + { + "epoch": 0.8528442784976938, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2040252685546875, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7021070718765259, + "num_tokens": 201021146.0, + "step": 7766 + }, + { + "epoch": 0.8529540962003075, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.220954179763794, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7218352556228638, + "num_tokens": 201046614.0, + "step": 7767 + }, + { + "epoch": 0.8530639139029211, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.238992691040039, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7361417412757874, + "num_tokens": 201068484.0, + "step": 7768 + }, + { + "epoch": 0.8531737316055348, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3461737632751465, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7151543498039246, + "num_tokens": 201089170.0, + "step": 7769 + }, + { + "epoch": 0.8532835493081484, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3372561931610107, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6934850215911865, + "num_tokens": 201112542.0, + "step": 7770 + }, + { + "epoch": 0.8533933670107622, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2146201133728027, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7355446219444275, + "num_tokens": 201136180.0, + "step": 7771 + }, + { + "epoch": 0.8535031847133758, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.372363805770874, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7058626413345337, + "num_tokens": 201159136.0, + "step": 7772 + }, + { + "epoch": 0.8536130024159895, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.033021926879883, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.6959216594696045, + "num_tokens": 201188466.0, + "step": 7773 + }, + { + "epoch": 0.8537228201186031, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1165502071380615, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6897100210189819, + "num_tokens": 201214870.0, + "step": 7774 + }, + { + "epoch": 0.8538326378212168, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.323709487915039, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7161770462989807, + "num_tokens": 201238516.0, + "step": 7775 + }, + { + "epoch": 0.8539424555238304, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1612792015075684, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7089434862136841, + "num_tokens": 201264492.0, + "step": 7776 + }, + { + "epoch": 0.8540522732264441, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8065742254257202, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6880875825881958, + "num_tokens": 201300458.0, + "step": 7777 + }, + { + "epoch": 0.8541620909290578, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.7303239107131958, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7126667499542236, + "num_tokens": 201337032.0, + "step": 7778 + }, + { + "epoch": 0.8542719086316715, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3480024337768555, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7112733721733093, + "num_tokens": 201360124.0, + "step": 7779 + }, + { + "epoch": 0.8543817263342851, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.569573402404785, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7067123651504517, + "num_tokens": 201379881.0, + "step": 7780 + }, + { + "epoch": 0.8544915440368988, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.125924587249756, + "learning_rate": 1e-06, + "loss": 1.0802, + "mean_token_accuracy": 0.67143714427948, + "num_tokens": 201407020.0, + "step": 7781 + }, + { + "epoch": 0.8546013617395124, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.411407709121704, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7175378799438477, + "num_tokens": 201430308.0, + "step": 7782 + }, + { + "epoch": 0.854711179442126, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2328929901123047, + "learning_rate": 1e-06, + "loss": 1.0101, + "mean_token_accuracy": 0.6916794776916504, + "num_tokens": 201456423.0, + "step": 7783 + }, + { + "epoch": 0.8548209971447397, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1855618953704834, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7186951041221619, + "num_tokens": 201479965.0, + "step": 7784 + }, + { + "epoch": 0.8549308148473534, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1553151607513428, + "learning_rate": 1e-06, + "loss": 1.0545, + "mean_token_accuracy": 0.6867251992225647, + "num_tokens": 201506398.0, + "step": 7785 + }, + { + "epoch": 0.8550406325499671, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0727133750915527, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7069455981254578, + "num_tokens": 201533902.0, + "step": 7786 + }, + { + "epoch": 0.8551504502525807, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.328397035598755, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7206832766532898, + "num_tokens": 201555508.0, + "step": 7787 + }, + { + "epoch": 0.8552602679551944, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2343149185180664, + "learning_rate": 1e-06, + "loss": 1.0751, + "mean_token_accuracy": 0.6777474880218506, + "num_tokens": 201583580.0, + "step": 7788 + }, + { + "epoch": 0.855370085657808, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0879054069519043, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6995720863342285, + "num_tokens": 201614497.0, + "step": 7789 + }, + { + "epoch": 0.8554799033604217, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2062692642211914, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.714735746383667, + "num_tokens": 201637724.0, + "step": 7790 + }, + { + "epoch": 0.8555897210630353, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.074190139770508, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7215765714645386, + "num_tokens": 201663407.0, + "step": 7791 + }, + { + "epoch": 0.855699538765649, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.512737989425659, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7016770243644714, + "num_tokens": 201683632.0, + "step": 7792 + }, + { + "epoch": 0.8558093564682627, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4138827323913574, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7204093933105469, + "num_tokens": 201704739.0, + "step": 7793 + }, + { + "epoch": 0.8559191741708764, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.413034677505493, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7103050947189331, + "num_tokens": 201727432.0, + "step": 7794 + }, + { + "epoch": 0.85602899187349, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3499536514282227, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7065112590789795, + "num_tokens": 201751819.0, + "step": 7795 + }, + { + "epoch": 0.8561388095761037, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8280566930770874, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7133855819702148, + "num_tokens": 201783478.0, + "step": 7796 + }, + { + "epoch": 0.8562486272787173, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1295502185821533, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7114428281784058, + "num_tokens": 201808626.0, + "step": 7797 + }, + { + "epoch": 0.856358444981331, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.343836784362793, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7152566909790039, + "num_tokens": 201830947.0, + "step": 7798 + }, + { + "epoch": 0.8564682626839446, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2784969806671143, + "learning_rate": 1e-06, + "loss": 1.0094, + "mean_token_accuracy": 0.6966378688812256, + "num_tokens": 201855299.0, + "step": 7799 + }, + { + "epoch": 0.8565780803865584, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1849794387817383, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7148619294166565, + "num_tokens": 201881901.0, + "step": 7800 + }, + { + "epoch": 0.856687898089172, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9664607048034668, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7017465233802795, + "num_tokens": 201914176.0, + "step": 7801 + }, + { + "epoch": 0.8567977157917857, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0141642093658447, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7028393745422363, + "num_tokens": 201941746.0, + "step": 7802 + }, + { + "epoch": 0.8569075334943993, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.066359758377075, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7237153649330139, + "num_tokens": 201968023.0, + "step": 7803 + }, + { + "epoch": 0.8570173511970129, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.100377321243286, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7125370502471924, + "num_tokens": 201992326.0, + "step": 7804 + }, + { + "epoch": 0.8571271688996266, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.7728527784347534, + "learning_rate": 1e-06, + "loss": 1.0794, + "mean_token_accuracy": 0.6701264381408691, + "num_tokens": 202030782.0, + "step": 7805 + }, + { + "epoch": 0.8572369866022402, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0550005435943604, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7059365510940552, + "num_tokens": 202058757.0, + "step": 7806 + }, + { + "epoch": 0.857346804304854, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2784347534179688, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7038160562515259, + "num_tokens": 202083103.0, + "step": 7807 + }, + { + "epoch": 0.8574566220074676, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.070500373840332, + "learning_rate": 1e-06, + "loss": 1.0076, + "mean_token_accuracy": 0.6949107646942139, + "num_tokens": 202112823.0, + "step": 7808 + }, + { + "epoch": 0.8575664397100813, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9084786176681519, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7257612943649292, + "num_tokens": 202143862.0, + "step": 7809 + }, + { + "epoch": 0.8576762574126949, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8937687873840332, + "learning_rate": 1e-06, + "loss": 1.031, + "mean_token_accuracy": 0.6848818063735962, + "num_tokens": 202176427.0, + "step": 7810 + }, + { + "epoch": 0.8577860751153086, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.039580821990967, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7238961458206177, + "num_tokens": 202205462.0, + "step": 7811 + }, + { + "epoch": 0.8578958928179222, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.322244644165039, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7114923000335693, + "num_tokens": 202228124.0, + "step": 7812 + }, + { + "epoch": 0.8580057105205359, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2367725372314453, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7117664813995361, + "num_tokens": 202253879.0, + "step": 7813 + }, + { + "epoch": 0.8581155282231496, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9829000234603882, + "learning_rate": 1e-06, + "loss": 0.9938, + "mean_token_accuracy": 0.7015843391418457, + "num_tokens": 202283863.0, + "step": 7814 + }, + { + "epoch": 0.8582253459257633, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2012124061584473, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7108275294303894, + "num_tokens": 202308816.0, + "step": 7815 + }, + { + "epoch": 0.8583351636283769, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5828680992126465, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7239421606063843, + "num_tokens": 202328376.0, + "step": 7816 + }, + { + "epoch": 0.8584449813309906, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0826468467712402, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7112472653388977, + "num_tokens": 202355544.0, + "step": 7817 + }, + { + "epoch": 0.8585547990336042, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0414276123046875, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7298755645751953, + "num_tokens": 202381441.0, + "step": 7818 + }, + { + "epoch": 0.8586646167362179, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.173218250274658, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7308778166770935, + "num_tokens": 202408014.0, + "step": 7819 + }, + { + "epoch": 0.8587744344388315, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1040143966674805, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7455751299858093, + "num_tokens": 202435919.0, + "step": 7820 + }, + { + "epoch": 0.8588842521414451, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.240051031112671, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7208924293518066, + "num_tokens": 202458627.0, + "step": 7821 + }, + { + "epoch": 0.8589940698440589, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.17378568649292, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7107799053192139, + "num_tokens": 202484299.0, + "step": 7822 + }, + { + "epoch": 0.8591038875466726, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1004176139831543, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7114958167076111, + "num_tokens": 202510414.0, + "step": 7823 + }, + { + "epoch": 0.8592137052492862, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.283567428588867, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7150472402572632, + "num_tokens": 202533513.0, + "step": 7824 + }, + { + "epoch": 0.8593235229518998, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.24690318107605, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.6998781561851501, + "num_tokens": 202559780.0, + "step": 7825 + }, + { + "epoch": 0.8594333406545135, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.95883047580719, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7031781673431396, + "num_tokens": 202589175.0, + "step": 7826 + }, + { + "epoch": 0.8595431583571271, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.429328203201294, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7173458337783813, + "num_tokens": 202611568.0, + "step": 7827 + }, + { + "epoch": 0.8596529760597408, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9962729215621948, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6793678402900696, + "num_tokens": 202640644.0, + "step": 7828 + }, + { + "epoch": 0.8597627937623545, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8432552814483643, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7020542621612549, + "num_tokens": 202674639.0, + "step": 7829 + }, + { + "epoch": 0.8598726114649682, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1970770359039307, + "learning_rate": 1e-06, + "loss": 1.0363, + "mean_token_accuracy": 0.6927543878555298, + "num_tokens": 202700287.0, + "step": 7830 + }, + { + "epoch": 0.8599824291675818, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8209218978881836, + "learning_rate": 1e-06, + "loss": 1.0448, + "mean_token_accuracy": 0.6824658513069153, + "num_tokens": 202736629.0, + "step": 7831 + }, + { + "epoch": 0.8600922468701955, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2922370433807373, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7183676362037659, + "num_tokens": 202757926.0, + "step": 7832 + }, + { + "epoch": 0.8602020645728091, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.174459457397461, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.6984781622886658, + "num_tokens": 202784249.0, + "step": 7833 + }, + { + "epoch": 0.8603118822754228, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2024617195129395, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.6980122327804565, + "num_tokens": 202809701.0, + "step": 7834 + }, + { + "epoch": 0.8604216999780364, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.045909881591797, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7015172243118286, + "num_tokens": 202837782.0, + "step": 7835 + }, + { + "epoch": 0.8605315176806502, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1416261196136475, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.725783109664917, + "num_tokens": 202864615.0, + "step": 7836 + }, + { + "epoch": 0.8606413353832638, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0715296268463135, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7082456350326538, + "num_tokens": 202891672.0, + "step": 7837 + }, + { + "epoch": 0.8607511530858775, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0370678901672363, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6983887553215027, + "num_tokens": 202922336.0, + "step": 7838 + }, + { + "epoch": 0.8608609707884911, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.094487190246582, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.6974714398384094, + "num_tokens": 202949143.0, + "step": 7839 + }, + { + "epoch": 0.8609707884911048, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4123623371124268, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7420540452003479, + "num_tokens": 202967791.0, + "step": 7840 + }, + { + "epoch": 0.8610806061937184, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2403156757354736, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6935335993766785, + "num_tokens": 202990516.0, + "step": 7841 + }, + { + "epoch": 0.861190423896332, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2137486934661865, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7010518312454224, + "num_tokens": 203014639.0, + "step": 7842 + }, + { + "epoch": 0.8613002415989458, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.056645631790161, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7121422290802002, + "num_tokens": 203042275.0, + "step": 7843 + }, + { + "epoch": 0.8614100593015594, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2902050018310547, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.697705090045929, + "num_tokens": 203066848.0, + "step": 7844 + }, + { + "epoch": 0.8615198770041731, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0188379287719727, + "learning_rate": 1e-06, + "loss": 1.0339, + "mean_token_accuracy": 0.6872189044952393, + "num_tokens": 203097751.0, + "step": 7845 + }, + { + "epoch": 0.8616296947067867, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.169008493423462, + "learning_rate": 1e-06, + "loss": 1.0237, + "mean_token_accuracy": 0.6926559209823608, + "num_tokens": 203124755.0, + "step": 7846 + }, + { + "epoch": 0.8617395124094004, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4042673110961914, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.7055923938751221, + "num_tokens": 203146497.0, + "step": 7847 + }, + { + "epoch": 0.861849330112014, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3826684951782227, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7143135070800781, + "num_tokens": 203167468.0, + "step": 7848 + }, + { + "epoch": 0.8619591478146277, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9247344732284546, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7197335362434387, + "num_tokens": 203198814.0, + "step": 7849 + }, + { + "epoch": 0.8620689655172413, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1369287967681885, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7056900262832642, + "num_tokens": 203224152.0, + "step": 7850 + }, + { + "epoch": 0.8621787832198551, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.157503843307495, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7334215641021729, + "num_tokens": 203248679.0, + "step": 7851 + }, + { + "epoch": 0.8622886009224687, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9709056615829468, + "learning_rate": 1e-06, + "loss": 1.0249, + "mean_token_accuracy": 0.6842519044876099, + "num_tokens": 203279320.0, + "step": 7852 + }, + { + "epoch": 0.8623984186250824, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2110838890075684, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7172988653182983, + "num_tokens": 203303420.0, + "step": 7853 + }, + { + "epoch": 0.862508236327696, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2921571731567383, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7151278257369995, + "num_tokens": 203327137.0, + "step": 7854 + }, + { + "epoch": 0.8626180540303097, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.139153242111206, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7175341844558716, + "num_tokens": 203353839.0, + "step": 7855 + }, + { + "epoch": 0.8627278717329233, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0145864486694336, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.689974308013916, + "num_tokens": 203383795.0, + "step": 7856 + }, + { + "epoch": 0.862837689435537, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.327068328857422, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7159575819969177, + "num_tokens": 203407162.0, + "step": 7857 + }, + { + "epoch": 0.8629475071381507, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.186760187149048, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7008030414581299, + "num_tokens": 203434099.0, + "step": 7858 + }, + { + "epoch": 0.8630573248407644, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3260889053344727, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7172058820724487, + "num_tokens": 203456548.0, + "step": 7859 + }, + { + "epoch": 0.863167142543378, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1089084148406982, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7176350355148315, + "num_tokens": 203484169.0, + "step": 7860 + }, + { + "epoch": 0.8632769602459917, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.263009786605835, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6963744759559631, + "num_tokens": 203509125.0, + "step": 7861 + }, + { + "epoch": 0.8633867779486053, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0899505615234375, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7154210805892944, + "num_tokens": 203534755.0, + "step": 7862 + }, + { + "epoch": 0.863496595651219, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2466444969177246, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6969937086105347, + "num_tokens": 203558568.0, + "step": 7863 + }, + { + "epoch": 0.8636064133538326, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2598977088928223, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7092782258987427, + "num_tokens": 203582837.0, + "step": 7864 + }, + { + "epoch": 0.8637162310564463, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0639498233795166, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7169748544692993, + "num_tokens": 203611467.0, + "step": 7865 + }, + { + "epoch": 0.86382604875906, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2635152339935303, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7122453451156616, + "num_tokens": 203636023.0, + "step": 7866 + }, + { + "epoch": 0.8639358664616736, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2310609817504883, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.702229917049408, + "num_tokens": 203660568.0, + "step": 7867 + }, + { + "epoch": 0.8640456841642873, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1765377521514893, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6933122873306274, + "num_tokens": 203688505.0, + "step": 7868 + }, + { + "epoch": 0.8641555018669009, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.051028251647949, + "learning_rate": 1e-06, + "loss": 1.0624, + "mean_token_accuracy": 0.6827550530433655, + "num_tokens": 203718239.0, + "step": 7869 + }, + { + "epoch": 0.8642653195695146, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1424858570098877, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7174134254455566, + "num_tokens": 203743844.0, + "step": 7870 + }, + { + "epoch": 0.8643751372721282, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.297548770904541, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7097563147544861, + "num_tokens": 203767153.0, + "step": 7871 + }, + { + "epoch": 0.864484954974742, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0188982486724854, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7072798609733582, + "num_tokens": 203794917.0, + "step": 7872 + }, + { + "epoch": 0.8645947726773556, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.351860761642456, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7058461308479309, + "num_tokens": 203817801.0, + "step": 7873 + }, + { + "epoch": 0.8647045903799693, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.230497360229492, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7017874717712402, + "num_tokens": 203841656.0, + "step": 7874 + }, + { + "epoch": 0.8648144080825829, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3850245475769043, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7205315232276917, + "num_tokens": 203864128.0, + "step": 7875 + }, + { + "epoch": 0.8649242257851966, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.272946357727051, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7261027097702026, + "num_tokens": 203886915.0, + "step": 7876 + }, + { + "epoch": 0.8650340434878102, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1105785369873047, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7165640592575073, + "num_tokens": 203912550.0, + "step": 7877 + }, + { + "epoch": 0.8651438611904239, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3437979221343994, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7107398509979248, + "num_tokens": 203935089.0, + "step": 7878 + }, + { + "epoch": 0.8652536788930375, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0125601291656494, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7375377416610718, + "num_tokens": 203963748.0, + "step": 7879 + }, + { + "epoch": 0.8653634965956513, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4645509719848633, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7248561382293701, + "num_tokens": 203984878.0, + "step": 7880 + }, + { + "epoch": 0.8654733142982649, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0247793197631836, + "learning_rate": 1e-06, + "loss": 1.0759, + "mean_token_accuracy": 0.6820754408836365, + "num_tokens": 204014695.0, + "step": 7881 + }, + { + "epoch": 0.8655831320008786, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.7606210708618164, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7216368317604065, + "num_tokens": 204035281.0, + "step": 7882 + }, + { + "epoch": 0.8656929497034922, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.202011823654175, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7181682586669922, + "num_tokens": 204059679.0, + "step": 7883 + }, + { + "epoch": 0.8658027674061058, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.404085397720337, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7168641090393066, + "num_tokens": 204081454.0, + "step": 7884 + }, + { + "epoch": 0.8659125851087195, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3324592113494873, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7246665954589844, + "num_tokens": 204103325.0, + "step": 7885 + }, + { + "epoch": 0.8660224028113331, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.543876886367798, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7056099772453308, + "num_tokens": 204124511.0, + "step": 7886 + }, + { + "epoch": 0.8661322205139469, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.083082914352417, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7085533142089844, + "num_tokens": 204151632.0, + "step": 7887 + }, + { + "epoch": 0.8662420382165605, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.115168333053589, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.6876699328422546, + "num_tokens": 204177986.0, + "step": 7888 + }, + { + "epoch": 0.8663518559191742, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1784749031066895, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6999009251594543, + "num_tokens": 204205111.0, + "step": 7889 + }, + { + "epoch": 0.8664616736217878, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.267697811126709, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7057349681854248, + "num_tokens": 204227601.0, + "step": 7890 + }, + { + "epoch": 0.8665714913244015, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2448770999908447, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7234021425247192, + "num_tokens": 204250245.0, + "step": 7891 + }, + { + "epoch": 0.8666813090270151, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2967278957366943, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7101422548294067, + "num_tokens": 204274328.0, + "step": 7892 + }, + { + "epoch": 0.8667911267296288, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3527071475982666, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7211173176765442, + "num_tokens": 204297123.0, + "step": 7893 + }, + { + "epoch": 0.8669009444322425, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9313857555389404, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7097797393798828, + "num_tokens": 204327408.0, + "step": 7894 + }, + { + "epoch": 0.8670107621348562, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4408671855926514, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7190848588943481, + "num_tokens": 204349246.0, + "step": 7895 + }, + { + "epoch": 0.8671205798374698, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.568704843521118, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7294836640357971, + "num_tokens": 204367333.0, + "step": 7896 + }, + { + "epoch": 0.8672303975400835, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.8691858053207397, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7270347476005554, + "num_tokens": 204400348.0, + "step": 7897 + }, + { + "epoch": 0.8673402152426971, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.232196807861328, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6888049244880676, + "num_tokens": 204425388.0, + "step": 7898 + }, + { + "epoch": 0.8674500329453108, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.414820909500122, + "learning_rate": 1e-06, + "loss": 1.021, + "mean_token_accuracy": 0.6943252086639404, + "num_tokens": 204449858.0, + "step": 7899 + }, + { + "epoch": 0.8675598506479244, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2244107723236084, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6869187355041504, + "num_tokens": 204476824.0, + "step": 7900 + }, + { + "epoch": 0.8676696683505382, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9637486934661865, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.686692476272583, + "num_tokens": 204505730.0, + "step": 7901 + }, + { + "epoch": 0.8677794860531518, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.5972964763641357, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7066474556922913, + "num_tokens": 204526003.0, + "step": 7902 + }, + { + "epoch": 0.8678893037557655, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4301462173461914, + "learning_rate": 1e-06, + "loss": 1.0382, + "mean_token_accuracy": 0.6980282068252563, + "num_tokens": 204548218.0, + "step": 7903 + }, + { + "epoch": 0.8679991214583791, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0839755535125732, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6962183117866516, + "num_tokens": 204576206.0, + "step": 7904 + }, + { + "epoch": 0.8681089391609927, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.964591145515442, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6985807418823242, + "num_tokens": 204608378.0, + "step": 7905 + }, + { + "epoch": 0.8682187568636064, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.535048007965088, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6889199018478394, + "num_tokens": 204628158.0, + "step": 7906 + }, + { + "epoch": 0.86832857456622, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3658926486968994, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.716792106628418, + "num_tokens": 204650319.0, + "step": 7907 + }, + { + "epoch": 0.8684383922688337, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.7366349697113037, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7223634123802185, + "num_tokens": 204668863.0, + "step": 7908 + }, + { + "epoch": 0.8685482099714474, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2472431659698486, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.720940113067627, + "num_tokens": 204692670.0, + "step": 7909 + }, + { + "epoch": 0.8686580276740611, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3767101764678955, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6902594566345215, + "num_tokens": 204715673.0, + "step": 7910 + }, + { + "epoch": 0.8687678453766747, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3429994583129883, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7341157793998718, + "num_tokens": 204738523.0, + "step": 7911 + }, + { + "epoch": 0.8688776630792884, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.055870532989502, + "learning_rate": 1e-06, + "loss": 1.0653, + "mean_token_accuracy": 0.6741505265235901, + "num_tokens": 204767872.0, + "step": 7912 + }, + { + "epoch": 0.868987480781902, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1443488597869873, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6970750689506531, + "num_tokens": 204795170.0, + "step": 7913 + }, + { + "epoch": 0.8690972984845157, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.061502456665039, + "learning_rate": 1e-06, + "loss": 1.0319, + "mean_token_accuracy": 0.6851383447647095, + "num_tokens": 204825372.0, + "step": 7914 + }, + { + "epoch": 0.8692071161871293, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0513923168182373, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6998456120491028, + "num_tokens": 204853010.0, + "step": 7915 + }, + { + "epoch": 0.8693169338897431, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.025707721710205, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7046703100204468, + "num_tokens": 204883722.0, + "step": 7916 + }, + { + "epoch": 0.8694267515923567, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.213712215423584, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7098620533943176, + "num_tokens": 204908704.0, + "step": 7917 + }, + { + "epoch": 0.8695365692949704, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2656056880950928, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7135887742042542, + "num_tokens": 204931118.0, + "step": 7918 + }, + { + "epoch": 0.869646386997584, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0755910873413086, + "learning_rate": 1e-06, + "loss": 1.0411, + "mean_token_accuracy": 0.6793230772018433, + "num_tokens": 204959245.0, + "step": 7919 + }, + { + "epoch": 0.8697562047001977, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.499734401702881, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7039079666137695, + "num_tokens": 204980960.0, + "step": 7920 + }, + { + "epoch": 0.8698660224028113, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9446465969085693, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7143046855926514, + "num_tokens": 205012385.0, + "step": 7921 + }, + { + "epoch": 0.869975840105425, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9911062717437744, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7084367275238037, + "num_tokens": 205042316.0, + "step": 7922 + }, + { + "epoch": 0.8700856578080387, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.107897996902466, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7057301998138428, + "num_tokens": 205069593.0, + "step": 7923 + }, + { + "epoch": 0.8701954755106523, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1556997299194336, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7117930054664612, + "num_tokens": 205095869.0, + "step": 7924 + }, + { + "epoch": 0.870305293213266, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2884738445281982, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6909988522529602, + "num_tokens": 205120650.0, + "step": 7925 + }, + { + "epoch": 0.8704151109158796, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3411073684692383, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.6992762088775635, + "num_tokens": 205145191.0, + "step": 7926 + }, + { + "epoch": 0.8705249286184933, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.231653928756714, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7166060209274292, + "num_tokens": 205171048.0, + "step": 7927 + }, + { + "epoch": 0.8706347463211069, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0681698322296143, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7172932028770447, + "num_tokens": 205197124.0, + "step": 7928 + }, + { + "epoch": 0.8707445640237206, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9217621088027954, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7081633806228638, + "num_tokens": 205228334.0, + "step": 7929 + }, + { + "epoch": 0.8708543817263343, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2094674110412598, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7243820428848267, + "num_tokens": 205252781.0, + "step": 7930 + }, + { + "epoch": 0.870964199428948, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1556737422943115, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.7034479379653931, + "num_tokens": 205279568.0, + "step": 7931 + }, + { + "epoch": 0.8710740171315616, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4655723571777344, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7204375267028809, + "num_tokens": 205298739.0, + "step": 7932 + }, + { + "epoch": 0.8711838348341753, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2224247455596924, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7264040112495422, + "num_tokens": 205323281.0, + "step": 7933 + }, + { + "epoch": 0.8712936525367889, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3990745544433594, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7065706253051758, + "num_tokens": 205346345.0, + "step": 7934 + }, + { + "epoch": 0.8714034702394026, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3608405590057373, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7163478136062622, + "num_tokens": 205367389.0, + "step": 7935 + }, + { + "epoch": 0.8715132879420162, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.230485200881958, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7020705938339233, + "num_tokens": 205390695.0, + "step": 7936 + }, + { + "epoch": 0.87162310564463, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.212506055831909, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.6938874125480652, + "num_tokens": 205415769.0, + "step": 7937 + }, + { + "epoch": 0.8717329233472436, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2345011234283447, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7194546461105347, + "num_tokens": 205438970.0, + "step": 7938 + }, + { + "epoch": 0.8718427410498573, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1831462383270264, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7072032690048218, + "num_tokens": 205463899.0, + "step": 7939 + }, + { + "epoch": 0.8719525587524709, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0308732986450195, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6991268396377563, + "num_tokens": 205493986.0, + "step": 7940 + }, + { + "epoch": 0.8720623764550846, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.462324857711792, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7075656652450562, + "num_tokens": 205516520.0, + "step": 7941 + }, + { + "epoch": 0.8721721941576982, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1727263927459717, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6922445297241211, + "num_tokens": 205544956.0, + "step": 7942 + }, + { + "epoch": 0.8722820118603118, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.4785263538360596, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7042804956436157, + "num_tokens": 205565308.0, + "step": 7943 + }, + { + "epoch": 0.8723918295629255, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0448110103607178, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.699530303478241, + "num_tokens": 205592370.0, + "step": 7944 + }, + { + "epoch": 0.8725016472655392, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.974244236946106, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6976257562637329, + "num_tokens": 205622412.0, + "step": 7945 + }, + { + "epoch": 0.8726114649681529, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3171207904815674, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.706436038017273, + "num_tokens": 205646197.0, + "step": 7946 + }, + { + "epoch": 0.8727212826707665, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.243091344833374, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7117839455604553, + "num_tokens": 205669733.0, + "step": 7947 + }, + { + "epoch": 0.8728311003733802, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9834754467010498, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7067502737045288, + "num_tokens": 205697984.0, + "step": 7948 + }, + { + "epoch": 0.8729409180759938, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3408453464508057, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6881677508354187, + "num_tokens": 205722254.0, + "step": 7949 + }, + { + "epoch": 0.8730507357786075, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.187234878540039, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7077890634536743, + "num_tokens": 205748538.0, + "step": 7950 + }, + { + "epoch": 0.8731605534812211, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2319931983947754, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.736165463924408, + "num_tokens": 205772601.0, + "step": 7951 + }, + { + "epoch": 0.8732703711838349, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2163138389587402, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.6948899030685425, + "num_tokens": 205796423.0, + "step": 7952 + }, + { + "epoch": 0.8733801888864485, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.007901906967163, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7267673015594482, + "num_tokens": 205826257.0, + "step": 7953 + }, + { + "epoch": 0.8734900065890622, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.190051555633545, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.6970040798187256, + "num_tokens": 205850383.0, + "step": 7954 + }, + { + "epoch": 0.8735998242916758, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.573244333267212, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.706139326095581, + "num_tokens": 205869212.0, + "step": 7955 + }, + { + "epoch": 0.8737096419942895, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.384545087814331, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7111667990684509, + "num_tokens": 205891962.0, + "step": 7956 + }, + { + "epoch": 0.8738194596969031, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3195230960845947, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7174720764160156, + "num_tokens": 205916887.0, + "step": 7957 + }, + { + "epoch": 0.8739292773995168, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0478515625, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7113739252090454, + "num_tokens": 205943337.0, + "step": 7958 + }, + { + "epoch": 0.8740390951021305, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.122189998626709, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.7021770477294922, + "num_tokens": 205969059.0, + "step": 7959 + }, + { + "epoch": 0.8741489128047442, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 1.9598448276519775, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7218624353408813, + "num_tokens": 205998041.0, + "step": 7960 + }, + { + "epoch": 0.8742587305073578, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2244458198547363, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7180546522140503, + "num_tokens": 206022857.0, + "step": 7961 + }, + { + "epoch": 0.8743685482099715, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0440614223480225, + "learning_rate": 1e-06, + "loss": 1.0679, + "mean_token_accuracy": 0.6853786110877991, + "num_tokens": 206051159.0, + "step": 7962 + }, + { + "epoch": 0.8744783659125851, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.0116119384765625, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7194859385490417, + "num_tokens": 206079039.0, + "step": 7963 + }, + { + "epoch": 0.8745881836151987, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.211127758026123, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.705918550491333, + "num_tokens": 206103595.0, + "step": 7964 + }, + { + "epoch": 0.8746980013178124, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.260019540786743, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7010056376457214, + "num_tokens": 206128831.0, + "step": 7965 + }, + { + "epoch": 0.8748078190204261, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2110517024993896, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7029892802238464, + "num_tokens": 206153310.0, + "step": 7966 + }, + { + "epoch": 0.8749176367230398, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.027365207672119, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7264119982719421, + "num_tokens": 206179791.0, + "step": 7967 + }, + { + "epoch": 0.8750274544256534, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.5583858489990234, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7178821563720703, + "num_tokens": 206198392.0, + "step": 7968 + }, + { + "epoch": 0.8751372721282671, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.1041128635406494, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.7092239856719971, + "num_tokens": 206226572.0, + "step": 7969 + }, + { + "epoch": 0.8752470898308807, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3503873348236084, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7351565957069397, + "num_tokens": 206247127.0, + "step": 7970 + }, + { + "epoch": 0.8753569075334944, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.2027804851531982, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7147665023803711, + "num_tokens": 206271724.0, + "step": 7971 + }, + { + "epoch": 0.875466725236108, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9391993284225464, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.6996564269065857, + "num_tokens": 206302578.0, + "step": 7972 + }, + { + "epoch": 0.8755765429387217, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.272494077682495, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7140307426452637, + "num_tokens": 206328020.0, + "step": 7973 + }, + { + "epoch": 0.8756863606413354, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.096761465072632, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7202090620994568, + "num_tokens": 206353497.0, + "step": 7974 + }, + { + "epoch": 0.8757961783439491, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.1688404083251953, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7095951437950134, + "num_tokens": 206377663.0, + "step": 7975 + }, + { + "epoch": 0.8759059960465627, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.105015277862549, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7043107151985168, + "num_tokens": 206404111.0, + "step": 7976 + }, + { + "epoch": 0.8760158137491764, + "ewc_loss": 4.0531158447265625e-06, + "grad_norm": 2.3726866245269775, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7074031233787537, + "num_tokens": 206425714.0, + "step": 7977 + }, + { + "epoch": 0.87612563145179, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.076814889907837, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7002643346786499, + "num_tokens": 206454801.0, + "step": 7978 + }, + { + "epoch": 0.8762354491544037, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2550618648529053, + "learning_rate": 1e-06, + "loss": 1.0598, + "mean_token_accuracy": 0.687731146812439, + "num_tokens": 206480724.0, + "step": 7979 + }, + { + "epoch": 0.8763452668570173, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0700645446777344, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7156389951705933, + "num_tokens": 206508803.0, + "step": 7980 + }, + { + "epoch": 0.8764550845596311, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2396316528320312, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7041835784912109, + "num_tokens": 206532791.0, + "step": 7981 + }, + { + "epoch": 0.8765649022622447, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.7103750705718994, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7098010778427124, + "num_tokens": 206550452.0, + "step": 7982 + }, + { + "epoch": 0.8766747199648584, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.144972562789917, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.703851580619812, + "num_tokens": 206575311.0, + "step": 7983 + }, + { + "epoch": 0.876784537667472, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.93013596534729, + "learning_rate": 1e-06, + "loss": 1.0281, + "mean_token_accuracy": 0.6938619613647461, + "num_tokens": 206606387.0, + "step": 7984 + }, + { + "epoch": 0.8768943553700856, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.108832597732544, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7247845530509949, + "num_tokens": 206631078.0, + "step": 7985 + }, + { + "epoch": 0.8770041730726993, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2545268535614014, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7040836811065674, + "num_tokens": 206656326.0, + "step": 7986 + }, + { + "epoch": 0.8771139907753129, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9904013872146606, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6991655826568604, + "num_tokens": 206687052.0, + "step": 7987 + }, + { + "epoch": 0.8772238084779267, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.4670798778533936, + "learning_rate": 1e-06, + "loss": 1.0604, + "mean_token_accuracy": 0.6765010952949524, + "num_tokens": 206708479.0, + "step": 7988 + }, + { + "epoch": 0.8773336261805403, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.101250410079956, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.692722499370575, + "num_tokens": 206734578.0, + "step": 7989 + }, + { + "epoch": 0.877443443883154, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.035989999771118, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7220231294631958, + "num_tokens": 206761012.0, + "step": 7990 + }, + { + "epoch": 0.8775532615857676, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.269381284713745, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7287825345993042, + "num_tokens": 206782739.0, + "step": 7991 + }, + { + "epoch": 0.8776630792883813, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.914317011833191, + "learning_rate": 1e-06, + "loss": 1.0494, + "mean_token_accuracy": 0.6796834468841553, + "num_tokens": 206816363.0, + "step": 7992 + }, + { + "epoch": 0.8777728969909949, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.236628293991089, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7210861444473267, + "num_tokens": 206840364.0, + "step": 7993 + }, + { + "epoch": 0.8778827146936086, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2943971157073975, + "learning_rate": 1e-06, + "loss": 1.0511, + "mean_token_accuracy": 0.6889796853065491, + "num_tokens": 206868088.0, + "step": 7994 + }, + { + "epoch": 0.8779925323962223, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.475707769393921, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7076963186264038, + "num_tokens": 206889482.0, + "step": 7995 + }, + { + "epoch": 0.878102350098836, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3047759532928467, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7079827785491943, + "num_tokens": 206912349.0, + "step": 7996 + }, + { + "epoch": 0.8782121678014496, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0907700061798096, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7211211919784546, + "num_tokens": 206937852.0, + "step": 7997 + }, + { + "epoch": 0.8783219855040633, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.226717233657837, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7311928272247314, + "num_tokens": 206961008.0, + "step": 7998 + }, + { + "epoch": 0.8784318032066769, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3096516132354736, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7000830173492432, + "num_tokens": 206983322.0, + "step": 7999 + }, + { + "epoch": 0.8785416209092906, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.248464822769165, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7060901522636414, + "num_tokens": 207007351.0, + "step": 8000 + }, + { + "epoch": 0.8786514386119042, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9453686475753784, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7109144330024719, + "num_tokens": 207038395.0, + "step": 8001 + }, + { + "epoch": 0.8787612563145178, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.300271511077881, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7110298871994019, + "num_tokens": 207062632.0, + "step": 8002 + }, + { + "epoch": 0.8788710740171316, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9473799467086792, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.688050389289856, + "num_tokens": 207092243.0, + "step": 8003 + }, + { + "epoch": 0.8789808917197452, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9531543254852295, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.7004590630531311, + "num_tokens": 207122819.0, + "step": 8004 + }, + { + "epoch": 0.8790907094223589, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.271726369857788, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7155214548110962, + "num_tokens": 207147033.0, + "step": 8005 + }, + { + "epoch": 0.8792005271249725, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.176318883895874, + "learning_rate": 1e-06, + "loss": 1.0417, + "mean_token_accuracy": 0.689871072769165, + "num_tokens": 207172566.0, + "step": 8006 + }, + { + "epoch": 0.8793103448275862, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9905470609664917, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7040976881980896, + "num_tokens": 207201224.0, + "step": 8007 + }, + { + "epoch": 0.8794201625301998, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0640134811401367, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.722713053226471, + "num_tokens": 207228077.0, + "step": 8008 + }, + { + "epoch": 0.8795299802328135, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0623817443847656, + "learning_rate": 1e-06, + "loss": 1.0335, + "mean_token_accuracy": 0.6812858581542969, + "num_tokens": 207256588.0, + "step": 8009 + }, + { + "epoch": 0.8796397979354272, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.246199607849121, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.719851553440094, + "num_tokens": 207278042.0, + "step": 8010 + }, + { + "epoch": 0.8797496156380409, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.1683201789855957, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.6887451410293579, + "num_tokens": 207305518.0, + "step": 8011 + }, + { + "epoch": 0.8798594333406545, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9392383098602295, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7104828357696533, + "num_tokens": 207335128.0, + "step": 8012 + }, + { + "epoch": 0.8799692510432682, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.4356300830841064, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7174724340438843, + "num_tokens": 207357774.0, + "step": 8013 + }, + { + "epoch": 0.8800790687458818, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.186671733856201, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7123351693153381, + "num_tokens": 207383683.0, + "step": 8014 + }, + { + "epoch": 0.8801888864484955, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.9439325332641602, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7132796049118042, + "num_tokens": 207413913.0, + "step": 8015 + }, + { + "epoch": 0.8802987041511091, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.8526374101638794, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7081338167190552, + "num_tokens": 207449426.0, + "step": 8016 + }, + { + "epoch": 0.8804085218537229, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.1400306224823, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7077029943466187, + "num_tokens": 207476016.0, + "step": 8017 + }, + { + "epoch": 0.8805183395563365, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3491716384887695, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.716272234916687, + "num_tokens": 207497955.0, + "step": 8018 + }, + { + "epoch": 0.8806281572589502, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.460273265838623, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7412053942680359, + "num_tokens": 207516932.0, + "step": 8019 + }, + { + "epoch": 0.8807379749615638, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.103945255279541, + "learning_rate": 1e-06, + "loss": 1.0387, + "mean_token_accuracy": 0.6888641119003296, + "num_tokens": 207545331.0, + "step": 8020 + }, + { + "epoch": 0.8808477926641775, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3284037113189697, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.713251531124115, + "num_tokens": 207568324.0, + "step": 8021 + }, + { + "epoch": 0.8809576103667911, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.242039680480957, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7229958772659302, + "num_tokens": 207590390.0, + "step": 8022 + }, + { + "epoch": 0.8810674280694047, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.10679030418396, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7274411916732788, + "num_tokens": 207614861.0, + "step": 8023 + }, + { + "epoch": 0.8811772457720185, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0634572505950928, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7027336359024048, + "num_tokens": 207643294.0, + "step": 8024 + }, + { + "epoch": 0.8812870634746321, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.489389181137085, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7033964395523071, + "num_tokens": 207663473.0, + "step": 8025 + }, + { + "epoch": 0.8813968811772458, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.073064088821411, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7125709056854248, + "num_tokens": 207693923.0, + "step": 8026 + }, + { + "epoch": 0.8815066988798594, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.0876076221466064, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7310200929641724, + "num_tokens": 207722358.0, + "step": 8027 + }, + { + "epoch": 0.8816165165824731, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.1953554153442383, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7092812657356262, + "num_tokens": 207748579.0, + "step": 8028 + }, + { + "epoch": 0.8817263342850867, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.3410940170288086, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7224386930465698, + "num_tokens": 207771070.0, + "step": 8029 + }, + { + "epoch": 0.8818361519877004, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.5658681392669678, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7199931144714355, + "num_tokens": 207789767.0, + "step": 8030 + }, + { + "epoch": 0.881945969690314, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.305537700653076, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7218904495239258, + "num_tokens": 207810612.0, + "step": 8031 + }, + { + "epoch": 0.8820557873929278, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2282259464263916, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6965253353118896, + "num_tokens": 207836096.0, + "step": 8032 + }, + { + "epoch": 0.8821656050955414, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 1.84824800491333, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7033056020736694, + "num_tokens": 207868083.0, + "step": 8033 + }, + { + "epoch": 0.8822754227981551, + "ewc_loss": 4.082918167114258e-06, + "grad_norm": 2.2102255821228027, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.721766471862793, + "num_tokens": 207892881.0, + "step": 8034 + }, + { + "epoch": 0.8823852405007687, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.938103199005127, + "learning_rate": 1e-06, + "loss": 0.9924, + "mean_token_accuracy": 0.6967340707778931, + "num_tokens": 207924786.0, + "step": 8035 + }, + { + "epoch": 0.8824950582033824, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.536964178085327, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7244707345962524, + "num_tokens": 207946705.0, + "step": 8036 + }, + { + "epoch": 0.882604875905996, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.2768805027008057, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7337180376052856, + "num_tokens": 207968344.0, + "step": 8037 + }, + { + "epoch": 0.8827146936086097, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.308101177215576, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7119063138961792, + "num_tokens": 207991588.0, + "step": 8038 + }, + { + "epoch": 0.8828245113112234, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.0699658393859863, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6964848041534424, + "num_tokens": 208020156.0, + "step": 8039 + }, + { + "epoch": 0.8829343290138371, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.5760691165924072, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7115936279296875, + "num_tokens": 208039776.0, + "step": 8040 + }, + { + "epoch": 0.8830441467164507, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.8082224130630493, + "learning_rate": 1e-06, + "loss": 1.0675, + "mean_token_accuracy": 0.6833020448684692, + "num_tokens": 208078399.0, + "step": 8041 + }, + { + "epoch": 0.8831539644190644, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.412550926208496, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7149344682693481, + "num_tokens": 208098993.0, + "step": 8042 + }, + { + "epoch": 0.883263782121678, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.9517637491226196, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7088656425476074, + "num_tokens": 208132064.0, + "step": 8043 + }, + { + "epoch": 0.8833735998242916, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.117587089538574, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.730710506439209, + "num_tokens": 208158784.0, + "step": 8044 + }, + { + "epoch": 0.8834834175269053, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1845433712005615, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.703606128692627, + "num_tokens": 208184449.0, + "step": 8045 + }, + { + "epoch": 0.883593235229519, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.441204309463501, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7215894460678101, + "num_tokens": 208204624.0, + "step": 8046 + }, + { + "epoch": 0.8837030529321327, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.42435884475708, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7008876204490662, + "num_tokens": 208227598.0, + "step": 8047 + }, + { + "epoch": 0.8838128706347463, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.3300607204437256, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7283060550689697, + "num_tokens": 208251223.0, + "step": 8048 + }, + { + "epoch": 0.88392268833736, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.5247085094451904, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7024260759353638, + "num_tokens": 208271542.0, + "step": 8049 + }, + { + "epoch": 0.8840325060399736, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1197314262390137, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7021006941795349, + "num_tokens": 208300464.0, + "step": 8050 + }, + { + "epoch": 0.8841423237425873, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.484734058380127, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7176731824874878, + "num_tokens": 208320189.0, + "step": 8051 + }, + { + "epoch": 0.8842521414452009, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.880591630935669, + "learning_rate": 1e-06, + "loss": 1.0651, + "mean_token_accuracy": 0.691744327545166, + "num_tokens": 208355152.0, + "step": 8052 + }, + { + "epoch": 0.8843619591478147, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.0264687538146973, + "learning_rate": 1e-06, + "loss": 1.0192, + "mean_token_accuracy": 0.6881954073905945, + "num_tokens": 208383984.0, + "step": 8053 + }, + { + "epoch": 0.8844717768504283, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.071702480316162, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.717176079750061, + "num_tokens": 208410915.0, + "step": 8054 + }, + { + "epoch": 0.884581594553042, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1398963928222656, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7463311553001404, + "num_tokens": 208436148.0, + "step": 8055 + }, + { + "epoch": 0.8846914122556556, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.2134740352630615, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7110095620155334, + "num_tokens": 208465224.0, + "step": 8056 + }, + { + "epoch": 0.8848012299582693, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.9054893255233765, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7343120574951172, + "num_tokens": 208494111.0, + "step": 8057 + }, + { + "epoch": 0.8849110476608829, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.3585164546966553, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7262773513793945, + "num_tokens": 208514390.0, + "step": 8058 + }, + { + "epoch": 0.8850208653634966, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.9878251552581787, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.7000947594642639, + "num_tokens": 208543421.0, + "step": 8059 + }, + { + "epoch": 0.8851306830661102, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1720027923583984, + "learning_rate": 1e-06, + "loss": 1.0524, + "mean_token_accuracy": 0.6872354745864868, + "num_tokens": 208570744.0, + "step": 8060 + }, + { + "epoch": 0.885240500768724, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1486620903015137, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7186566591262817, + "num_tokens": 208596170.0, + "step": 8061 + }, + { + "epoch": 0.8853503184713376, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.0885708332061768, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6972784399986267, + "num_tokens": 208625206.0, + "step": 8062 + }, + { + "epoch": 0.8854601361739513, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.394972085952759, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7256695032119751, + "num_tokens": 208645910.0, + "step": 8063 + }, + { + "epoch": 0.8855699538765649, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.0478005409240723, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7070233821868896, + "num_tokens": 208672805.0, + "step": 8064 + }, + { + "epoch": 0.8856797715791785, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.9911426305770874, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.7061824798583984, + "num_tokens": 208703429.0, + "step": 8065 + }, + { + "epoch": 0.8857895892817922, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1453776359558105, + "learning_rate": 1e-06, + "loss": 1.0617, + "mean_token_accuracy": 0.6842265725135803, + "num_tokens": 208732360.0, + "step": 8066 + }, + { + "epoch": 0.8858994069844058, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1413445472717285, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.7010775804519653, + "num_tokens": 208759178.0, + "step": 8067 + }, + { + "epoch": 0.8860092246870196, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.458806037902832, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7156450748443604, + "num_tokens": 208778917.0, + "step": 8068 + }, + { + "epoch": 0.8861190423896332, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.0280275344848633, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.711789608001709, + "num_tokens": 208808384.0, + "step": 8069 + }, + { + "epoch": 0.8862288600922469, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.4360921382904053, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7142566442489624, + "num_tokens": 208829418.0, + "step": 8070 + }, + { + "epoch": 0.8863386777948605, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.076631784439087, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.7037259340286255, + "num_tokens": 208857521.0, + "step": 8071 + }, + { + "epoch": 0.8864484954974742, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.2290024757385254, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.6978199481964111, + "num_tokens": 208881906.0, + "step": 8072 + }, + { + "epoch": 0.8865583132000878, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 1.8565641641616821, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7344335317611694, + "num_tokens": 208914194.0, + "step": 8073 + }, + { + "epoch": 0.8866681309027015, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.2505617141723633, + "learning_rate": 1e-06, + "loss": 1.0483, + "mean_token_accuracy": 0.6849249005317688, + "num_tokens": 208939704.0, + "step": 8074 + }, + { + "epoch": 0.8867779486053152, + "ewc_loss": 4.112720489501953e-06, + "grad_norm": 2.1864421367645264, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7321468591690063, + "num_tokens": 208963714.0, + "step": 8075 + }, + { + "epoch": 0.8868877663079289, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.103736639022827, + "learning_rate": 1e-06, + "loss": 1.0119, + "mean_token_accuracy": 0.6910998821258545, + "num_tokens": 208992718.0, + "step": 8076 + }, + { + "epoch": 0.8869975840105425, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0139482021331787, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7169392108917236, + "num_tokens": 209021300.0, + "step": 8077 + }, + { + "epoch": 0.8871074017131562, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.2306418418884277, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.699426531791687, + "num_tokens": 209048287.0, + "step": 8078 + }, + { + "epoch": 0.8872172194157698, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.451897144317627, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7060344815254211, + "num_tokens": 209069545.0, + "step": 8079 + }, + { + "epoch": 0.8873270371183835, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.090695858001709, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.702828049659729, + "num_tokens": 209095028.0, + "step": 8080 + }, + { + "epoch": 0.8874368548209971, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1961183547973633, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7246318459510803, + "num_tokens": 209118193.0, + "step": 8081 + }, + { + "epoch": 0.8875466725236109, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0076730251312256, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7080531716346741, + "num_tokens": 209146980.0, + "step": 8082 + }, + { + "epoch": 0.8876564902262245, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.095013380050659, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7015267014503479, + "num_tokens": 209174812.0, + "step": 8083 + }, + { + "epoch": 0.8877663079288381, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1949830055236816, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7204763889312744, + "num_tokens": 209200809.0, + "step": 8084 + }, + { + "epoch": 0.8878761256314518, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1340649127960205, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7319329380989075, + "num_tokens": 209225031.0, + "step": 8085 + }, + { + "epoch": 0.8879859433340654, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.164776086807251, + "learning_rate": 1e-06, + "loss": 1.0563, + "mean_token_accuracy": 0.6795860528945923, + "num_tokens": 209253581.0, + "step": 8086 + }, + { + "epoch": 0.8880957610366791, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.103104591369629, + "learning_rate": 1e-06, + "loss": 1.0273, + "mean_token_accuracy": 0.6854984760284424, + "num_tokens": 209283539.0, + "step": 8087 + }, + { + "epoch": 0.8882055787392927, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.981743574142456, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.6903308629989624, + "num_tokens": 209313896.0, + "step": 8088 + }, + { + "epoch": 0.8883153964419065, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0504679679870605, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.722212016582489, + "num_tokens": 209338742.0, + "step": 8089 + }, + { + "epoch": 0.8884252141445201, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.448974370956421, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7124724388122559, + "num_tokens": 209358816.0, + "step": 8090 + }, + { + "epoch": 0.8885350318471338, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1976280212402344, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7121928334236145, + "num_tokens": 209383834.0, + "step": 8091 + }, + { + "epoch": 0.8886448495497474, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.116521120071411, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.6968975067138672, + "num_tokens": 209410554.0, + "step": 8092 + }, + { + "epoch": 0.8887546672523611, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.48085355758667, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7167885899543762, + "num_tokens": 209430014.0, + "step": 8093 + }, + { + "epoch": 0.8888644849549747, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3501033782958984, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7090392708778381, + "num_tokens": 209450901.0, + "step": 8094 + }, + { + "epoch": 0.8889743026575884, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0523269176483154, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7000526785850525, + "num_tokens": 209480423.0, + "step": 8095 + }, + { + "epoch": 0.889084120360202, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.012413263320923, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7195756435394287, + "num_tokens": 209509147.0, + "step": 8096 + }, + { + "epoch": 0.8891939380628158, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9370049238204956, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7138850092887878, + "num_tokens": 209539620.0, + "step": 8097 + }, + { + "epoch": 0.8893037557654294, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3446669578552246, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7330114841461182, + "num_tokens": 209560177.0, + "step": 8098 + }, + { + "epoch": 0.8894135734680431, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1331334114074707, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7142454981803894, + "num_tokens": 209585869.0, + "step": 8099 + }, + { + "epoch": 0.8895233911706567, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.4507439136505127, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6987653970718384, + "num_tokens": 209607506.0, + "step": 8100 + }, + { + "epoch": 0.8896332088732704, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1930723190307617, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7033594846725464, + "num_tokens": 209634975.0, + "step": 8101 + }, + { + "epoch": 0.889743026575884, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.142580986022949, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7104908227920532, + "num_tokens": 209661090.0, + "step": 8102 + }, + { + "epoch": 0.8898528442784976, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.132395029067993, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7044203281402588, + "num_tokens": 209687591.0, + "step": 8103 + }, + { + "epoch": 0.8899626619811114, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3658883571624756, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7181355953216553, + "num_tokens": 209709367.0, + "step": 8104 + }, + { + "epoch": 0.890072479683725, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9517000913619995, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7090332508087158, + "num_tokens": 209740280.0, + "step": 8105 + }, + { + "epoch": 0.8901822973863387, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.069634437561035, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7268154621124268, + "num_tokens": 209766365.0, + "step": 8106 + }, + { + "epoch": 0.8902921150889523, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1886940002441406, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7096390724182129, + "num_tokens": 209792007.0, + "step": 8107 + }, + { + "epoch": 0.890401932791566, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9164115190505981, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7093839645385742, + "num_tokens": 209820382.0, + "step": 8108 + }, + { + "epoch": 0.8905117504941796, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.294896364212036, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7007262706756592, + "num_tokens": 209844582.0, + "step": 8109 + }, + { + "epoch": 0.8906215681967933, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9948495626449585, + "learning_rate": 1e-06, + "loss": 1.0894, + "mean_token_accuracy": 0.6727586984634399, + "num_tokens": 209876586.0, + "step": 8110 + }, + { + "epoch": 0.890731385899407, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.234034776687622, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.721189558506012, + "num_tokens": 209901208.0, + "step": 8111 + }, + { + "epoch": 0.8908412036020207, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.2726569175720215, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7427259683609009, + "num_tokens": 209922889.0, + "step": 8112 + }, + { + "epoch": 0.8909510213046343, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.124392509460449, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7081899046897888, + "num_tokens": 209948901.0, + "step": 8113 + }, + { + "epoch": 0.891060839007248, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.2742104530334473, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.708085298538208, + "num_tokens": 209972783.0, + "step": 8114 + }, + { + "epoch": 0.8911706567098616, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9757049083709717, + "learning_rate": 1e-06, + "loss": 1.0519, + "mean_token_accuracy": 0.6810542345046997, + "num_tokens": 210003463.0, + "step": 8115 + }, + { + "epoch": 0.8912804744124753, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.285447597503662, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7110021114349365, + "num_tokens": 210026671.0, + "step": 8116 + }, + { + "epoch": 0.8913902921150889, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1793060302734375, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7270194292068481, + "num_tokens": 210049798.0, + "step": 8117 + }, + { + "epoch": 0.8915001098177027, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.128455877304077, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6996911764144897, + "num_tokens": 210076020.0, + "step": 8118 + }, + { + "epoch": 0.8916099275203163, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9930329322814941, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6917434930801392, + "num_tokens": 210106761.0, + "step": 8119 + }, + { + "epoch": 0.89171974522293, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1841275691986084, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7241601347923279, + "num_tokens": 210132344.0, + "step": 8120 + }, + { + "epoch": 0.8918295629255436, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.174791097640991, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7054445743560791, + "num_tokens": 210159175.0, + "step": 8121 + }, + { + "epoch": 0.8919393806281573, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0811595916748047, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6912283301353455, + "num_tokens": 210189422.0, + "step": 8122 + }, + { + "epoch": 0.8920491983307709, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9924719333648682, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.6885584592819214, + "num_tokens": 210220355.0, + "step": 8123 + }, + { + "epoch": 0.8921590160333845, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.232163190841675, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.715125322341919, + "num_tokens": 210243366.0, + "step": 8124 + }, + { + "epoch": 0.8922688337359982, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 1.9822425842285156, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7012509107589722, + "num_tokens": 210273857.0, + "step": 8125 + }, + { + "epoch": 0.892378651438612, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.341878652572632, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.695624053478241, + "num_tokens": 210298244.0, + "step": 8126 + }, + { + "epoch": 0.8924884691412256, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.0658812522888184, + "learning_rate": 1e-06, + "loss": 1.0444, + "mean_token_accuracy": 0.686153769493103, + "num_tokens": 210329923.0, + "step": 8127 + }, + { + "epoch": 0.8925982868438392, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.7941665649414062, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.7025220990180969, + "num_tokens": 210358538.0, + "step": 8128 + }, + { + "epoch": 0.8927081045464529, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.016737699508667, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7086747884750366, + "num_tokens": 210386966.0, + "step": 8129 + }, + { + "epoch": 0.8928179222490665, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.127610445022583, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.6955524682998657, + "num_tokens": 210414246.0, + "step": 8130 + }, + { + "epoch": 0.8929277399516802, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.3203518390655518, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6953425407409668, + "num_tokens": 210437784.0, + "step": 8131 + }, + { + "epoch": 0.8930375576542938, + "ewc_loss": 4.172325134277344e-06, + "grad_norm": 2.1436350345611572, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7027114629745483, + "num_tokens": 210463775.0, + "step": 8132 + }, + { + "epoch": 0.8931473753569076, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.087531566619873, + "learning_rate": 1e-06, + "loss": 1.029, + "mean_token_accuracy": 0.6956480741500854, + "num_tokens": 210492247.0, + "step": 8133 + }, + { + "epoch": 0.8932571930595212, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.267159938812256, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.7011343836784363, + "num_tokens": 210516236.0, + "step": 8134 + }, + { + "epoch": 0.8933670107621349, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.5206685066223145, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7246502637863159, + "num_tokens": 210539167.0, + "step": 8135 + }, + { + "epoch": 0.8934768284647485, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.331389904022217, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7181808948516846, + "num_tokens": 210562684.0, + "step": 8136 + }, + { + "epoch": 0.8935866461673622, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9068620204925537, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7229863405227661, + "num_tokens": 210596516.0, + "step": 8137 + }, + { + "epoch": 0.8936964638699758, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1106650829315186, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7239515781402588, + "num_tokens": 210623295.0, + "step": 8138 + }, + { + "epoch": 0.8938062815725895, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.4058632850646973, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7252830266952515, + "num_tokens": 210643221.0, + "step": 8139 + }, + { + "epoch": 0.8939160992752032, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1388814449310303, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.6976180076599121, + "num_tokens": 210670898.0, + "step": 8140 + }, + { + "epoch": 0.8940259169778169, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.367469072341919, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7073588371276855, + "num_tokens": 210694162.0, + "step": 8141 + }, + { + "epoch": 0.8941357346804305, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1666994094848633, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6968319416046143, + "num_tokens": 210719930.0, + "step": 8142 + }, + { + "epoch": 0.8942455523830442, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.234126329421997, + "learning_rate": 1e-06, + "loss": 1.0342, + "mean_token_accuracy": 0.6873564720153809, + "num_tokens": 210747367.0, + "step": 8143 + }, + { + "epoch": 0.8943553700856578, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.153449535369873, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.7093514204025269, + "num_tokens": 210773069.0, + "step": 8144 + }, + { + "epoch": 0.8944651877882714, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1037182807922363, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.713559627532959, + "num_tokens": 210801344.0, + "step": 8145 + }, + { + "epoch": 0.8945750054908851, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.8425065279006958, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7173571586608887, + "num_tokens": 210835548.0, + "step": 8146 + }, + { + "epoch": 0.8946848231934988, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2882745265960693, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.726394534111023, + "num_tokens": 210857281.0, + "step": 8147 + }, + { + "epoch": 0.8947946408961125, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1320245265960693, + "learning_rate": 1e-06, + "loss": 0.9874, + "mean_token_accuracy": 0.7083855867385864, + "num_tokens": 210884292.0, + "step": 8148 + }, + { + "epoch": 0.8949044585987261, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2134037017822266, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7096225023269653, + "num_tokens": 210908448.0, + "step": 8149 + }, + { + "epoch": 0.8950142763013398, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3357269763946533, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.6931266784667969, + "num_tokens": 210932876.0, + "step": 8150 + }, + { + "epoch": 0.8951240940039534, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9123249053955078, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7150384783744812, + "num_tokens": 210963703.0, + "step": 8151 + }, + { + "epoch": 0.8952339117065671, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1636695861816406, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7110065221786499, + "num_tokens": 210990458.0, + "step": 8152 + }, + { + "epoch": 0.8953437294091807, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9389822483062744, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6961911916732788, + "num_tokens": 211021700.0, + "step": 8153 + }, + { + "epoch": 0.8954535471117944, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3471717834472656, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7065134048461914, + "num_tokens": 211044523.0, + "step": 8154 + }, + { + "epoch": 0.8955633648144081, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.177367687225342, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7120673060417175, + "num_tokens": 211067809.0, + "step": 8155 + }, + { + "epoch": 0.8956731825170218, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.21391224861145, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7044638991355896, + "num_tokens": 211093559.0, + "step": 8156 + }, + { + "epoch": 0.8957830002196354, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.195941209793091, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7203465700149536, + "num_tokens": 211120767.0, + "step": 8157 + }, + { + "epoch": 0.8958928179222491, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.199192523956299, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7163587808609009, + "num_tokens": 211144996.0, + "step": 8158 + }, + { + "epoch": 0.8960026356248627, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9934296607971191, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7185889482498169, + "num_tokens": 211175424.0, + "step": 8159 + }, + { + "epoch": 0.8961124533274764, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9807226657867432, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7062365412712097, + "num_tokens": 211204392.0, + "step": 8160 + }, + { + "epoch": 0.89622227103009, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.262834310531616, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7038155794143677, + "num_tokens": 211229550.0, + "step": 8161 + }, + { + "epoch": 0.8963320887327038, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.153712749481201, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7214682698249817, + "num_tokens": 211255545.0, + "step": 8162 + }, + { + "epoch": 0.8964419064353174, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.018228054046631, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7003556489944458, + "num_tokens": 211283602.0, + "step": 8163 + }, + { + "epoch": 0.896551724137931, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.53869366645813, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.707673966884613, + "num_tokens": 211302935.0, + "step": 8164 + }, + { + "epoch": 0.8966615418405447, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.628161668777466, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7074725031852722, + "num_tokens": 211321835.0, + "step": 8165 + }, + { + "epoch": 0.8967713595431583, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.010664701461792, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7002276182174683, + "num_tokens": 211351891.0, + "step": 8166 + }, + { + "epoch": 0.896881177245772, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.367410898208618, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6994967460632324, + "num_tokens": 211372892.0, + "step": 8167 + }, + { + "epoch": 0.8969909949483856, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2351083755493164, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7179701328277588, + "num_tokens": 211394671.0, + "step": 8168 + }, + { + "epoch": 0.8971008126509994, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.010338068008423, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7156823873519897, + "num_tokens": 211422879.0, + "step": 8169 + }, + { + "epoch": 0.897210630353613, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0446043014526367, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7055871486663818, + "num_tokens": 211450565.0, + "step": 8170 + }, + { + "epoch": 0.8973204480562267, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.38549542427063, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7038129568099976, + "num_tokens": 211471829.0, + "step": 8171 + }, + { + "epoch": 0.8974302657588403, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3144419193267822, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7259476184844971, + "num_tokens": 211495522.0, + "step": 8172 + }, + { + "epoch": 0.897540083461454, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.132004976272583, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7134215831756592, + "num_tokens": 211522860.0, + "step": 8173 + }, + { + "epoch": 0.8976499011640676, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.164146661758423, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.6972413063049316, + "num_tokens": 211548583.0, + "step": 8174 + }, + { + "epoch": 0.8977597188666813, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1817235946655273, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.7085709571838379, + "num_tokens": 211572444.0, + "step": 8175 + }, + { + "epoch": 0.897869536569295, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3804092407226562, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7265214323997498, + "num_tokens": 211594694.0, + "step": 8176 + }, + { + "epoch": 0.8979793542719087, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.5460920333862305, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7057961821556091, + "num_tokens": 211614529.0, + "step": 8177 + }, + { + "epoch": 0.8980891719745223, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.197293519973755, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7151302099227905, + "num_tokens": 211640705.0, + "step": 8178 + }, + { + "epoch": 0.898198989677136, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2870523929595947, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.71396404504776, + "num_tokens": 211664393.0, + "step": 8179 + }, + { + "epoch": 0.8983088073797496, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1372523307800293, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7169634699821472, + "num_tokens": 211689982.0, + "step": 8180 + }, + { + "epoch": 0.8984186250823633, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.567352294921875, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7129766941070557, + "num_tokens": 211708693.0, + "step": 8181 + }, + { + "epoch": 0.8985284427849769, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.314418077468872, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6970337629318237, + "num_tokens": 211732231.0, + "step": 8182 + }, + { + "epoch": 0.8986382604875905, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9498835802078247, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7184341549873352, + "num_tokens": 211760191.0, + "step": 8183 + }, + { + "epoch": 0.8987480781902043, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.138338565826416, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6980597972869873, + "num_tokens": 211786463.0, + "step": 8184 + }, + { + "epoch": 0.898857895892818, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1814680099487305, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7126656770706177, + "num_tokens": 211812926.0, + "step": 8185 + }, + { + "epoch": 0.8989677135954316, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2962605953216553, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7316619157791138, + "num_tokens": 211834368.0, + "step": 8186 + }, + { + "epoch": 0.8990775312980452, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.5666756629943848, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7167408466339111, + "num_tokens": 211853863.0, + "step": 8187 + }, + { + "epoch": 0.8991873490006589, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2889318466186523, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6998811364173889, + "num_tokens": 211877620.0, + "step": 8188 + }, + { + "epoch": 0.8992971667032725, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0438802242279053, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6938002705574036, + "num_tokens": 211908189.0, + "step": 8189 + }, + { + "epoch": 0.8994069844058862, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2050065994262695, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.6997222900390625, + "num_tokens": 211935489.0, + "step": 8190 + }, + { + "epoch": 0.8995168021084999, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.381934404373169, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.713138222694397, + "num_tokens": 211958611.0, + "step": 8191 + }, + { + "epoch": 0.8996266198111136, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1137847900390625, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7089501023292542, + "num_tokens": 211983462.0, + "step": 8192 + }, + { + "epoch": 0.8997364375137272, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.00970721244812, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.726976752281189, + "num_tokens": 212011557.0, + "step": 8193 + }, + { + "epoch": 0.8998462552163409, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.8925987482070923, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7265720963478088, + "num_tokens": 212042302.0, + "step": 8194 + }, + { + "epoch": 0.8999560729189545, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.311249256134033, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7057676315307617, + "num_tokens": 212065357.0, + "step": 8195 + }, + { + "epoch": 0.9000658906215682, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1714718341827393, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7073612213134766, + "num_tokens": 212092573.0, + "step": 8196 + }, + { + "epoch": 0.9001757083241818, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0614171028137207, + "learning_rate": 1e-06, + "loss": 1.0549, + "mean_token_accuracy": 0.6815057396888733, + "num_tokens": 212120452.0, + "step": 8197 + }, + { + "epoch": 0.9002855260267956, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.112839698791504, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7198082804679871, + "num_tokens": 212146205.0, + "step": 8198 + }, + { + "epoch": 0.9003953437294092, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2688405513763428, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.719480574131012, + "num_tokens": 212170022.0, + "step": 8199 + }, + { + "epoch": 0.9005051614320229, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0973150730133057, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.706106960773468, + "num_tokens": 212197763.0, + "step": 8200 + }, + { + "epoch": 0.9006149791346365, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.2536838054656982, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6990957260131836, + "num_tokens": 212223065.0, + "step": 8201 + }, + { + "epoch": 0.9007247968372502, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.316605567932129, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7073673009872437, + "num_tokens": 212245488.0, + "step": 8202 + }, + { + "epoch": 0.9008346145398638, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1071972846984863, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6969894170761108, + "num_tokens": 212274724.0, + "step": 8203 + }, + { + "epoch": 0.9009444322424774, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.235556125640869, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7360809445381165, + "num_tokens": 212296850.0, + "step": 8204 + }, + { + "epoch": 0.9010542499450912, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1006133556365967, + "learning_rate": 1e-06, + "loss": 1.012, + "mean_token_accuracy": 0.6864315867424011, + "num_tokens": 212324861.0, + "step": 8205 + }, + { + "epoch": 0.9011640676477048, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.347430944442749, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.6963485479354858, + "num_tokens": 212350599.0, + "step": 8206 + }, + { + "epoch": 0.9012738853503185, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3069047927856445, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7145352363586426, + "num_tokens": 212374498.0, + "step": 8207 + }, + { + "epoch": 0.9013837030529321, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.254849910736084, + "learning_rate": 1e-06, + "loss": 1.0184, + "mean_token_accuracy": 0.6922487020492554, + "num_tokens": 212400734.0, + "step": 8208 + }, + { + "epoch": 0.9014935207555458, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.082036018371582, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.728724479675293, + "num_tokens": 212428540.0, + "step": 8209 + }, + { + "epoch": 0.9016033384581594, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3683996200561523, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7229505181312561, + "num_tokens": 212450911.0, + "step": 8210 + }, + { + "epoch": 0.9017131561607731, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3207218647003174, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.735040545463562, + "num_tokens": 212473053.0, + "step": 8211 + }, + { + "epoch": 0.9018229738633867, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.246401071548462, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6997087001800537, + "num_tokens": 212498830.0, + "step": 8212 + }, + { + "epoch": 0.9019327915660005, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9226385354995728, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6981979012489319, + "num_tokens": 212530401.0, + "step": 8213 + }, + { + "epoch": 0.9020426092686141, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.100020170211792, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.6898049712181091, + "num_tokens": 212558520.0, + "step": 8214 + }, + { + "epoch": 0.9021524269712278, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.102585554122925, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6935842037200928, + "num_tokens": 212586735.0, + "step": 8215 + }, + { + "epoch": 0.9022622446738414, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.090020179748535, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6992127895355225, + "num_tokens": 212613134.0, + "step": 8216 + }, + { + "epoch": 0.9023720623764551, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1959638595581055, + "learning_rate": 1e-06, + "loss": 1.0182, + "mean_token_accuracy": 0.7012549042701721, + "num_tokens": 212638391.0, + "step": 8217 + }, + { + "epoch": 0.9024818800790687, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.241309881210327, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7189295291900635, + "num_tokens": 212663725.0, + "step": 8218 + }, + { + "epoch": 0.9025916977816824, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.056685447692871, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.711796224117279, + "num_tokens": 212693559.0, + "step": 8219 + }, + { + "epoch": 0.9027015154842961, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.255429267883301, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7097494006156921, + "num_tokens": 212716262.0, + "step": 8220 + }, + { + "epoch": 0.9028113331869098, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9810246229171753, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7167832851409912, + "num_tokens": 212747179.0, + "step": 8221 + }, + { + "epoch": 0.9029211508895234, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0332939624786377, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7111687064170837, + "num_tokens": 212774828.0, + "step": 8222 + }, + { + "epoch": 0.903030968592137, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.944541096687317, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.6945133209228516, + "num_tokens": 212808125.0, + "step": 8223 + }, + { + "epoch": 0.9031407862947507, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1931145191192627, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.6879065036773682, + "num_tokens": 212834200.0, + "step": 8224 + }, + { + "epoch": 0.9032506039973643, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1294641494750977, + "learning_rate": 1e-06, + "loss": 1.0536, + "mean_token_accuracy": 0.6841535568237305, + "num_tokens": 212861783.0, + "step": 8225 + }, + { + "epoch": 0.903360421699978, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.950492262840271, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6938084959983826, + "num_tokens": 212893688.0, + "step": 8226 + }, + { + "epoch": 0.9034702394025917, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9983406066894531, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7095329761505127, + "num_tokens": 212923304.0, + "step": 8227 + }, + { + "epoch": 0.9035800571052054, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.030395746231079, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6884755492210388, + "num_tokens": 212955307.0, + "step": 8228 + }, + { + "epoch": 0.903689874807819, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.273451805114746, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7098774909973145, + "num_tokens": 212976112.0, + "step": 8229 + }, + { + "epoch": 0.9037996925104327, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.5715243816375732, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7169492244720459, + "num_tokens": 212995197.0, + "step": 8230 + }, + { + "epoch": 0.9039095102130463, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.1049647331237793, + "learning_rate": 1e-06, + "loss": 1.0921, + "mean_token_accuracy": 0.6774070858955383, + "num_tokens": 213024273.0, + "step": 8231 + }, + { + "epoch": 0.90401932791566, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 1.9344727993011475, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7173976302146912, + "num_tokens": 213053233.0, + "step": 8232 + }, + { + "epoch": 0.9041291456182736, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.4242091178894043, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7284002900123596, + "num_tokens": 213074531.0, + "step": 8233 + }, + { + "epoch": 0.9042389633208874, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.0369606018066406, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7050662636756897, + "num_tokens": 213103128.0, + "step": 8234 + }, + { + "epoch": 0.904348781023501, + "ewc_loss": 4.202127456665039e-06, + "grad_norm": 2.3518645763397217, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7157366275787354, + "num_tokens": 213124284.0, + "step": 8235 + }, + { + "epoch": 0.9044585987261147, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.001659631729126, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7403954267501831, + "num_tokens": 213153427.0, + "step": 8236 + }, + { + "epoch": 0.9045684164287283, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.099492073059082, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6978400349617004, + "num_tokens": 213180649.0, + "step": 8237 + }, + { + "epoch": 0.904678234131342, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.936409592628479, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.694328248500824, + "num_tokens": 213212215.0, + "step": 8238 + }, + { + "epoch": 0.9047880518339556, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9974074363708496, + "learning_rate": 1e-06, + "loss": 1.0204, + "mean_token_accuracy": 0.6918030977249146, + "num_tokens": 213242563.0, + "step": 8239 + }, + { + "epoch": 0.9048978695365693, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1348047256469727, + "learning_rate": 1e-06, + "loss": 1.0275, + "mean_token_accuracy": 0.6886378526687622, + "num_tokens": 213271046.0, + "step": 8240 + }, + { + "epoch": 0.9050076872391829, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9728999137878418, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.721296489238739, + "num_tokens": 213301255.0, + "step": 8241 + }, + { + "epoch": 0.9051175049417967, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2848827838897705, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.729300320148468, + "num_tokens": 213324833.0, + "step": 8242 + }, + { + "epoch": 0.9052273226444103, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0733768939971924, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.694546103477478, + "num_tokens": 213351024.0, + "step": 8243 + }, + { + "epoch": 0.905337140347024, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0762956142425537, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6909202933311462, + "num_tokens": 213379797.0, + "step": 8244 + }, + { + "epoch": 0.9054469580496376, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0669240951538086, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6893841624259949, + "num_tokens": 213409556.0, + "step": 8245 + }, + { + "epoch": 0.9055567757522512, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1755900382995605, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7092458009719849, + "num_tokens": 213433519.0, + "step": 8246 + }, + { + "epoch": 0.9056665934548649, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0185704231262207, + "learning_rate": 1e-06, + "loss": 1.0215, + "mean_token_accuracy": 0.6932563185691833, + "num_tokens": 213462154.0, + "step": 8247 + }, + { + "epoch": 0.9057764111574785, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.216527223587036, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7122483253479004, + "num_tokens": 213485521.0, + "step": 8248 + }, + { + "epoch": 0.9058862288600923, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.052917242050171, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7029261589050293, + "num_tokens": 213514699.0, + "step": 8249 + }, + { + "epoch": 0.9059960465627059, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2346811294555664, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.6969078779220581, + "num_tokens": 213539877.0, + "step": 8250 + }, + { + "epoch": 0.9061058642653196, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1099698543548584, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.699201226234436, + "num_tokens": 213567267.0, + "step": 8251 + }, + { + "epoch": 0.9062156819679332, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.161388874053955, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7189925909042358, + "num_tokens": 213593174.0, + "step": 8252 + }, + { + "epoch": 0.9063254996705469, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.163360834121704, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7043517231941223, + "num_tokens": 213619039.0, + "step": 8253 + }, + { + "epoch": 0.9064353173731605, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2082996368408203, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7175685167312622, + "num_tokens": 213644412.0, + "step": 8254 + }, + { + "epoch": 0.9065451350757742, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.25276255607605, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7143597602844238, + "num_tokens": 213668556.0, + "step": 8255 + }, + { + "epoch": 0.9066549527783879, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.966713547706604, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7461766004562378, + "num_tokens": 213694420.0, + "step": 8256 + }, + { + "epoch": 0.9067647704810016, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2282357215881348, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6910247206687927, + "num_tokens": 213721246.0, + "step": 8257 + }, + { + "epoch": 0.9068745881836152, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1036124229431152, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7176494002342224, + "num_tokens": 213747046.0, + "step": 8258 + }, + { + "epoch": 0.9069844058862289, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2815511226654053, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.7003535628318787, + "num_tokens": 213771105.0, + "step": 8259 + }, + { + "epoch": 0.9070942235888425, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1507012844085693, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7256949543952942, + "num_tokens": 213795878.0, + "step": 8260 + }, + { + "epoch": 0.9072040412914562, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.051668643951416, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6944118738174438, + "num_tokens": 213824083.0, + "step": 8261 + }, + { + "epoch": 0.9073138589940698, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0599098205566406, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7182949781417847, + "num_tokens": 213851743.0, + "step": 8262 + }, + { + "epoch": 0.9074236766966836, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.4512338638305664, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7091161012649536, + "num_tokens": 213872726.0, + "step": 8263 + }, + { + "epoch": 0.9075334943992972, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9107805490493774, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7006810903549194, + "num_tokens": 213904422.0, + "step": 8264 + }, + { + "epoch": 0.9076433121019108, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2385904788970947, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7089704275131226, + "num_tokens": 213929673.0, + "step": 8265 + }, + { + "epoch": 0.9077531298045245, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.241878032684326, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7155704498291016, + "num_tokens": 213952516.0, + "step": 8266 + }, + { + "epoch": 0.9078629475071381, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9132791757583618, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.693732738494873, + "num_tokens": 213982810.0, + "step": 8267 + }, + { + "epoch": 0.9079727652097518, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.084784984588623, + "learning_rate": 1e-06, + "loss": 1.0497, + "mean_token_accuracy": 0.6868748664855957, + "num_tokens": 214010916.0, + "step": 8268 + }, + { + "epoch": 0.9080825829123654, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2594964504241943, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7038343548774719, + "num_tokens": 214035488.0, + "step": 8269 + }, + { + "epoch": 0.9081924006149792, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.967426061630249, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7235790491104126, + "num_tokens": 214064658.0, + "step": 8270 + }, + { + "epoch": 0.9083022183175928, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1898016929626465, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7046886086463928, + "num_tokens": 214090493.0, + "step": 8271 + }, + { + "epoch": 0.9084120360202065, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.4500889778137207, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.723471999168396, + "num_tokens": 214109319.0, + "step": 8272 + }, + { + "epoch": 0.9085218537228201, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1307544708251953, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7067141532897949, + "num_tokens": 214135566.0, + "step": 8273 + }, + { + "epoch": 0.9086316714254338, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2828257083892822, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7174570560455322, + "num_tokens": 214158638.0, + "step": 8274 + }, + { + "epoch": 0.9087414891280474, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.044536828994751, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.6964550614356995, + "num_tokens": 214186129.0, + "step": 8275 + }, + { + "epoch": 0.9088513068306611, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.0058798789978027, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.699510395526886, + "num_tokens": 214216102.0, + "step": 8276 + }, + { + "epoch": 0.9089611245332747, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.037116765975952, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7045283913612366, + "num_tokens": 214244267.0, + "step": 8277 + }, + { + "epoch": 0.9090709422358885, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9671990871429443, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6956784725189209, + "num_tokens": 214274438.0, + "step": 8278 + }, + { + "epoch": 0.9091807599385021, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 1.9173694849014282, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6945712566375732, + "num_tokens": 214306660.0, + "step": 8279 + }, + { + "epoch": 0.9092905776411158, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.2222554683685303, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7445660829544067, + "num_tokens": 214331101.0, + "step": 8280 + }, + { + "epoch": 0.9094003953437294, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.10233211517334, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7220873236656189, + "num_tokens": 214358743.0, + "step": 8281 + }, + { + "epoch": 0.909510213046343, + "ewc_loss": 4.231929779052734e-06, + "grad_norm": 2.1471054553985596, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7016156911849976, + "num_tokens": 214386075.0, + "step": 8282 + }, + { + "epoch": 0.9096200307489567, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.669837474822998, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7103068828582764, + "num_tokens": 214405293.0, + "step": 8283 + }, + { + "epoch": 0.9097298484515703, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0437052249908447, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7068158984184265, + "num_tokens": 214432685.0, + "step": 8284 + }, + { + "epoch": 0.9098396661541841, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.193939208984375, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7032350301742554, + "num_tokens": 214458723.0, + "step": 8285 + }, + { + "epoch": 0.9099494838567977, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.063640594482422, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7161121368408203, + "num_tokens": 214486020.0, + "step": 8286 + }, + { + "epoch": 0.9100593015594114, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3709967136383057, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7049842476844788, + "num_tokens": 214509254.0, + "step": 8287 + }, + { + "epoch": 0.910169119262025, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0477449893951416, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6881958842277527, + "num_tokens": 214540063.0, + "step": 8288 + }, + { + "epoch": 0.9102789369646387, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2332212924957275, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.698638916015625, + "num_tokens": 214565092.0, + "step": 8289 + }, + { + "epoch": 0.9103887546672523, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8681740760803223, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7048116326332092, + "num_tokens": 214598113.0, + "step": 8290 + }, + { + "epoch": 0.910498572369866, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2706542015075684, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7098737955093384, + "num_tokens": 214621759.0, + "step": 8291 + }, + { + "epoch": 0.9106083900724797, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.23884916305542, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7216300368309021, + "num_tokens": 214646977.0, + "step": 8292 + }, + { + "epoch": 0.9107182077750934, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3279402256011963, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7361313104629517, + "num_tokens": 214667009.0, + "step": 8293 + }, + { + "epoch": 0.910828025477707, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.286379337310791, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.695044755935669, + "num_tokens": 214690203.0, + "step": 8294 + }, + { + "epoch": 0.9109378431803207, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.4483110904693604, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7280113697052002, + "num_tokens": 214711257.0, + "step": 8295 + }, + { + "epoch": 0.9110476608829343, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9860666990280151, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7062329053878784, + "num_tokens": 214740590.0, + "step": 8296 + }, + { + "epoch": 0.911157478585548, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.940129280090332, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7152167558670044, + "num_tokens": 214772673.0, + "step": 8297 + }, + { + "epoch": 0.9112672962881616, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.253577947616577, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6977884769439697, + "num_tokens": 214796315.0, + "step": 8298 + }, + { + "epoch": 0.9113771139907754, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2813899517059326, + "learning_rate": 1e-06, + "loss": 1.0405, + "mean_token_accuracy": 0.6867373585700989, + "num_tokens": 214822372.0, + "step": 8299 + }, + { + "epoch": 0.911486931693389, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9236282110214233, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6931649446487427, + "num_tokens": 214855213.0, + "step": 8300 + }, + { + "epoch": 0.9115967493960027, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8338598012924194, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7052587270736694, + "num_tokens": 214890209.0, + "step": 8301 + }, + { + "epoch": 0.9117065670986163, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9341683387756348, + "learning_rate": 1e-06, + "loss": 1.0563, + "mean_token_accuracy": 0.6868735551834106, + "num_tokens": 214924274.0, + "step": 8302 + }, + { + "epoch": 0.91181638480123, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.33864164352417, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.700541615486145, + "num_tokens": 214946732.0, + "step": 8303 + }, + { + "epoch": 0.9119262025038436, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.5548582077026367, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7339630126953125, + "num_tokens": 214965979.0, + "step": 8304 + }, + { + "epoch": 0.9120360202064572, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.356914520263672, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7042285203933716, + "num_tokens": 214989147.0, + "step": 8305 + }, + { + "epoch": 0.9121458379090709, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.323883295059204, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7150625586509705, + "num_tokens": 215012035.0, + "step": 8306 + }, + { + "epoch": 0.9122556556116846, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.267963409423828, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7253544330596924, + "num_tokens": 215036677.0, + "step": 8307 + }, + { + "epoch": 0.9123654733142983, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3373026847839355, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7269071936607361, + "num_tokens": 215057133.0, + "step": 8308 + }, + { + "epoch": 0.9124752910169119, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.114349126815796, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.6972750425338745, + "num_tokens": 215082831.0, + "step": 8309 + }, + { + "epoch": 0.9125851087195256, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1918890476226807, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7069475650787354, + "num_tokens": 215106983.0, + "step": 8310 + }, + { + "epoch": 0.9126949264221392, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1813371181488037, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.722906231880188, + "num_tokens": 215131950.0, + "step": 8311 + }, + { + "epoch": 0.9128047441247529, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.372422933578491, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7097213268280029, + "num_tokens": 215153394.0, + "step": 8312 + }, + { + "epoch": 0.9129145618273665, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0421080589294434, + "learning_rate": 1e-06, + "loss": 1.0466, + "mean_token_accuracy": 0.678451657295227, + "num_tokens": 215184587.0, + "step": 8313 + }, + { + "epoch": 0.9130243795299803, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1277103424072266, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.700209379196167, + "num_tokens": 215212894.0, + "step": 8314 + }, + { + "epoch": 0.9131341972325939, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1645805835723877, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6964355707168579, + "num_tokens": 215240789.0, + "step": 8315 + }, + { + "epoch": 0.9132440149352076, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3006842136383057, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6867897510528564, + "num_tokens": 215265364.0, + "step": 8316 + }, + { + "epoch": 0.9133538326378212, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.521198034286499, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7101690769195557, + "num_tokens": 215285525.0, + "step": 8317 + }, + { + "epoch": 0.9134636503404349, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2544615268707275, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7207568883895874, + "num_tokens": 215308355.0, + "step": 8318 + }, + { + "epoch": 0.9135734680430485, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3494632244110107, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.6973713636398315, + "num_tokens": 215336239.0, + "step": 8319 + }, + { + "epoch": 0.9136832857456622, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.991189956665039, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7082492709159851, + "num_tokens": 215360536.0, + "step": 8320 + }, + { + "epoch": 0.9137931034482759, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2017197608947754, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7115212678909302, + "num_tokens": 215386149.0, + "step": 8321 + }, + { + "epoch": 0.9139029211508896, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.367236614227295, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7220977544784546, + "num_tokens": 215407934.0, + "step": 8322 + }, + { + "epoch": 0.9140127388535032, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.236931800842285, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7295268774032593, + "num_tokens": 215435377.0, + "step": 8323 + }, + { + "epoch": 0.9141225565561168, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3395869731903076, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.699108362197876, + "num_tokens": 215461261.0, + "step": 8324 + }, + { + "epoch": 0.9142323742587305, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.333570957183838, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.723524808883667, + "num_tokens": 215483834.0, + "step": 8325 + }, + { + "epoch": 0.9143421919613441, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1600191593170166, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6979562640190125, + "num_tokens": 215508844.0, + "step": 8326 + }, + { + "epoch": 0.9144520096639578, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9958776235580444, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7310320734977722, + "num_tokens": 215536103.0, + "step": 8327 + }, + { + "epoch": 0.9145618273665715, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.4975578784942627, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7299957275390625, + "num_tokens": 215555919.0, + "step": 8328 + }, + { + "epoch": 0.9146716450691852, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2917046546936035, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.7074368000030518, + "num_tokens": 215580581.0, + "step": 8329 + }, + { + "epoch": 0.9147814627717988, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0890350341796875, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7240920066833496, + "num_tokens": 215605680.0, + "step": 8330 + }, + { + "epoch": 0.9148912804744125, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0701653957366943, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.6936919689178467, + "num_tokens": 215634982.0, + "step": 8331 + }, + { + "epoch": 0.9150010981770261, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9321157932281494, + "learning_rate": 1e-06, + "loss": 1.0172, + "mean_token_accuracy": 0.6928188800811768, + "num_tokens": 215667249.0, + "step": 8332 + }, + { + "epoch": 0.9151109158796398, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2724480628967285, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7235912680625916, + "num_tokens": 215690494.0, + "step": 8333 + }, + { + "epoch": 0.9152207335822534, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0842461585998535, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6979417204856873, + "num_tokens": 215716541.0, + "step": 8334 + }, + { + "epoch": 0.9153305512848671, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.190967559814453, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7076742649078369, + "num_tokens": 215741018.0, + "step": 8335 + }, + { + "epoch": 0.9154403689874808, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.27262806892395, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7153339982032776, + "num_tokens": 215765566.0, + "step": 8336 + }, + { + "epoch": 0.9155501866900945, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.297839879989624, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.719977080821991, + "num_tokens": 215789434.0, + "step": 8337 + }, + { + "epoch": 0.9156600043927081, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0930354595184326, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7170485854148865, + "num_tokens": 215816457.0, + "step": 8338 + }, + { + "epoch": 0.9157698220953218, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.324101686477661, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7256104946136475, + "num_tokens": 215839046.0, + "step": 8339 + }, + { + "epoch": 0.9158796397979354, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1896002292633057, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7266685962677002, + "num_tokens": 215861389.0, + "step": 8340 + }, + { + "epoch": 0.915989457500549, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.439605951309204, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7175760865211487, + "num_tokens": 215881662.0, + "step": 8341 + }, + { + "epoch": 0.9160992752031627, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.4344022274017334, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.723901093006134, + "num_tokens": 215901617.0, + "step": 8342 + }, + { + "epoch": 0.9162090929057765, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.18221378326416, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7150405049324036, + "num_tokens": 215927452.0, + "step": 8343 + }, + { + "epoch": 0.9163189106083901, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.207292318344116, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7152639627456665, + "num_tokens": 215953475.0, + "step": 8344 + }, + { + "epoch": 0.9164287283110037, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3321011066436768, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7021100521087646, + "num_tokens": 215978069.0, + "step": 8345 + }, + { + "epoch": 0.9165385460136174, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.09045672416687, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.715916633605957, + "num_tokens": 216005963.0, + "step": 8346 + }, + { + "epoch": 0.916648363716231, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1464242935180664, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7165100574493408, + "num_tokens": 216032217.0, + "step": 8347 + }, + { + "epoch": 0.9167581814188447, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.077432632446289, + "learning_rate": 1e-06, + "loss": 1.0295, + "mean_token_accuracy": 0.681623101234436, + "num_tokens": 216061479.0, + "step": 8348 + }, + { + "epoch": 0.9168679991214583, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.031020402908325, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7069128751754761, + "num_tokens": 216090079.0, + "step": 8349 + }, + { + "epoch": 0.9169778168240721, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1627843379974365, + "learning_rate": 1e-06, + "loss": 1.0307, + "mean_token_accuracy": 0.6874131560325623, + "num_tokens": 216117217.0, + "step": 8350 + }, + { + "epoch": 0.9170876345266857, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.069851875305176, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7247480750083923, + "num_tokens": 216144407.0, + "step": 8351 + }, + { + "epoch": 0.9171974522292994, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2716987133026123, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7259083390235901, + "num_tokens": 216168373.0, + "step": 8352 + }, + { + "epoch": 0.917307269931913, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3281443119049072, + "learning_rate": 1e-06, + "loss": 1.073, + "mean_token_accuracy": 0.6714795827865601, + "num_tokens": 216193251.0, + "step": 8353 + }, + { + "epoch": 0.9174170876345267, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.48046875, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7265380620956421, + "num_tokens": 216213242.0, + "step": 8354 + }, + { + "epoch": 0.9175269053371403, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0162696838378906, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7171800136566162, + "num_tokens": 216240426.0, + "step": 8355 + }, + { + "epoch": 0.917636723039754, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.011384963989258, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6971088647842407, + "num_tokens": 216270015.0, + "step": 8356 + }, + { + "epoch": 0.9177465407423677, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9739327430725098, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.704150915145874, + "num_tokens": 216300090.0, + "step": 8357 + }, + { + "epoch": 0.9178563584449814, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.95807945728302, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7116999626159668, + "num_tokens": 216330103.0, + "step": 8358 + }, + { + "epoch": 0.917966176147595, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0844125747680664, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7208163142204285, + "num_tokens": 216356619.0, + "step": 8359 + }, + { + "epoch": 0.9180759938502087, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.417171001434326, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7085497379302979, + "num_tokens": 216379171.0, + "step": 8360 + }, + { + "epoch": 0.9181858115528223, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0009727478027344, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6947870254516602, + "num_tokens": 216409387.0, + "step": 8361 + }, + { + "epoch": 0.918295629255436, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2167110443115234, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7167525887489319, + "num_tokens": 216433635.0, + "step": 8362 + }, + { + "epoch": 0.9184054469580496, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9233965873718262, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.6980046033859253, + "num_tokens": 216466485.0, + "step": 8363 + }, + { + "epoch": 0.9185152646606632, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.242196798324585, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7160383462905884, + "num_tokens": 216488907.0, + "step": 8364 + }, + { + "epoch": 0.918625082363277, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8932278156280518, + "learning_rate": 1e-06, + "loss": 0.9954, + "mean_token_accuracy": 0.6993750333786011, + "num_tokens": 216521259.0, + "step": 8365 + }, + { + "epoch": 0.9187349000658906, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.62994384765625, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7293024659156799, + "num_tokens": 216538864.0, + "step": 8366 + }, + { + "epoch": 0.9188447177685043, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.039426803588867, + "learning_rate": 1e-06, + "loss": 0.9654, + "mean_token_accuracy": 0.7083258628845215, + "num_tokens": 216568362.0, + "step": 8367 + }, + { + "epoch": 0.9189545354711179, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9442862272262573, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.705592691898346, + "num_tokens": 216598722.0, + "step": 8368 + }, + { + "epoch": 0.9190643531737316, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.213289260864258, + "learning_rate": 1e-06, + "loss": 0.9981, + "mean_token_accuracy": 0.697341799736023, + "num_tokens": 216623241.0, + "step": 8369 + }, + { + "epoch": 0.9191741708763452, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.057485580444336, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7239284515380859, + "num_tokens": 216650065.0, + "step": 8370 + }, + { + "epoch": 0.9192839885789589, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1916654109954834, + "learning_rate": 1e-06, + "loss": 1.0402, + "mean_token_accuracy": 0.6896023154258728, + "num_tokens": 216675984.0, + "step": 8371 + }, + { + "epoch": 0.9193938062815726, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9718157052993774, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7047579288482666, + "num_tokens": 216706335.0, + "step": 8372 + }, + { + "epoch": 0.9195036239841863, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8995822668075562, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6865930557250977, + "num_tokens": 216738678.0, + "step": 8373 + }, + { + "epoch": 0.9196134416867999, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9930297136306763, + "learning_rate": 1e-06, + "loss": 0.7558, + "mean_token_accuracy": 0.7645672559738159, + "num_tokens": 216764265.0, + "step": 8374 + }, + { + "epoch": 0.9197232593894136, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.301705837249756, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7308343052864075, + "num_tokens": 216787019.0, + "step": 8375 + }, + { + "epoch": 0.9198330770920272, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1842522621154785, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7069120407104492, + "num_tokens": 216810780.0, + "step": 8376 + }, + { + "epoch": 0.9199428947946409, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3911101818084717, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6931120157241821, + "num_tokens": 216833727.0, + "step": 8377 + }, + { + "epoch": 0.9200527124972545, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.180436372756958, + "learning_rate": 1e-06, + "loss": 1.0151, + "mean_token_accuracy": 0.6880462765693665, + "num_tokens": 216860498.0, + "step": 8378 + }, + { + "epoch": 0.9201625301998683, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.614650249481201, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7232751846313477, + "num_tokens": 216880677.0, + "step": 8379 + }, + { + "epoch": 0.9202723479024819, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.277451515197754, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7146366834640503, + "num_tokens": 216903748.0, + "step": 8380 + }, + { + "epoch": 0.9203821656050956, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1624228954315186, + "learning_rate": 1e-06, + "loss": 1.0013, + "mean_token_accuracy": 0.690201461315155, + "num_tokens": 216935680.0, + "step": 8381 + }, + { + "epoch": 0.9204919833077092, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1674845218658447, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7079509496688843, + "num_tokens": 216960966.0, + "step": 8382 + }, + { + "epoch": 0.9206018010103229, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2083263397216797, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.715074896812439, + "num_tokens": 216984655.0, + "step": 8383 + }, + { + "epoch": 0.9207116187129365, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9656533002853394, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7261165380477905, + "num_tokens": 217013859.0, + "step": 8384 + }, + { + "epoch": 0.9208214364155501, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.329648971557617, + "learning_rate": 1e-06, + "loss": 1.0135, + "mean_token_accuracy": 0.691218912601471, + "num_tokens": 217039191.0, + "step": 8385 + }, + { + "epoch": 0.9209312541181639, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1823089122772217, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.705660343170166, + "num_tokens": 217065249.0, + "step": 8386 + }, + { + "epoch": 0.9210410718207775, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1753673553466797, + "learning_rate": 1e-06, + "loss": 0.9857, + "mean_token_accuracy": 0.6988683342933655, + "num_tokens": 217092553.0, + "step": 8387 + }, + { + "epoch": 0.9211508895233912, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.201586961746216, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7133137583732605, + "num_tokens": 217117661.0, + "step": 8388 + }, + { + "epoch": 0.9212607072260048, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.119791269302368, + "learning_rate": 1e-06, + "loss": 1.0381, + "mean_token_accuracy": 0.6833515167236328, + "num_tokens": 217146021.0, + "step": 8389 + }, + { + "epoch": 0.9213705249286185, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1983301639556885, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6984367370605469, + "num_tokens": 217171003.0, + "step": 8390 + }, + { + "epoch": 0.9214803426312321, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.23622989654541, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7005635499954224, + "num_tokens": 217194716.0, + "step": 8391 + }, + { + "epoch": 0.9215901603338458, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.179116725921631, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.6944817304611206, + "num_tokens": 217220352.0, + "step": 8392 + }, + { + "epoch": 0.9216999780364594, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2538979053497314, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.729181706905365, + "num_tokens": 217241491.0, + "step": 8393 + }, + { + "epoch": 0.9218097957390732, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3317737579345703, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.725160539150238, + "num_tokens": 217261562.0, + "step": 8394 + }, + { + "epoch": 0.9219196134416868, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8480567932128906, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7099592685699463, + "num_tokens": 217294792.0, + "step": 8395 + }, + { + "epoch": 0.9220294311443005, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.019256114959717, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7124032974243164, + "num_tokens": 217322645.0, + "step": 8396 + }, + { + "epoch": 0.9221392488469141, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.864970326423645, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7019855976104736, + "num_tokens": 217357087.0, + "step": 8397 + }, + { + "epoch": 0.9222490665495278, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.256174325942993, + "learning_rate": 1e-06, + "loss": 1.0968, + "mean_token_accuracy": 0.6737149953842163, + "num_tokens": 217383181.0, + "step": 8398 + }, + { + "epoch": 0.9223588842521414, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.6700265407562256, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7154602408409119, + "num_tokens": 217401323.0, + "step": 8399 + }, + { + "epoch": 0.922468701954755, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9865162372589111, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7022914886474609, + "num_tokens": 217430911.0, + "step": 8400 + }, + { + "epoch": 0.9225785196573688, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9726753234863281, + "learning_rate": 1e-06, + "loss": 1.0083, + "mean_token_accuracy": 0.6932841539382935, + "num_tokens": 217460849.0, + "step": 8401 + }, + { + "epoch": 0.9226883373599825, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.09354829788208, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7170906662940979, + "num_tokens": 217487254.0, + "step": 8402 + }, + { + "epoch": 0.9227981550625961, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1090002059936523, + "learning_rate": 1e-06, + "loss": 1.0685, + "mean_token_accuracy": 0.6781949400901794, + "num_tokens": 217515029.0, + "step": 8403 + }, + { + "epoch": 0.9229079727652097, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2054224014282227, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7000782489776611, + "num_tokens": 217539472.0, + "step": 8404 + }, + { + "epoch": 0.9230177904678234, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1159703731536865, + "learning_rate": 1e-06, + "loss": 1.0797, + "mean_token_accuracy": 0.6980315446853638, + "num_tokens": 217565277.0, + "step": 8405 + }, + { + "epoch": 0.923127608170437, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.135408401489258, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7017835378646851, + "num_tokens": 217591244.0, + "step": 8406 + }, + { + "epoch": 0.9232374258730507, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0200369358062744, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7071882486343384, + "num_tokens": 217619176.0, + "step": 8407 + }, + { + "epoch": 0.9233472435756644, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.989696979522705, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.721767008304596, + "num_tokens": 217646909.0, + "step": 8408 + }, + { + "epoch": 0.9234570612782781, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.445178747177124, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7125226259231567, + "num_tokens": 217667989.0, + "step": 8409 + }, + { + "epoch": 0.9235668789808917, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0538432598114014, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7243268489837646, + "num_tokens": 217694095.0, + "step": 8410 + }, + { + "epoch": 0.9236766966835054, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0301637649536133, + "learning_rate": 1e-06, + "loss": 1.0333, + "mean_token_accuracy": 0.6847810745239258, + "num_tokens": 217723548.0, + "step": 8411 + }, + { + "epoch": 0.923786514386119, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2989041805267334, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7089765071868896, + "num_tokens": 217747359.0, + "step": 8412 + }, + { + "epoch": 0.9238963320887327, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2517521381378174, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.700248122215271, + "num_tokens": 217772721.0, + "step": 8413 + }, + { + "epoch": 0.9240061497913463, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2456464767456055, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.691512942314148, + "num_tokens": 217797517.0, + "step": 8414 + }, + { + "epoch": 0.9241159674939601, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.014836549758911, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6946192979812622, + "num_tokens": 217826190.0, + "step": 8415 + }, + { + "epoch": 0.9242257851965737, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2698354721069336, + "learning_rate": 1e-06, + "loss": 1.0255, + "mean_token_accuracy": 0.6866455078125, + "num_tokens": 217852227.0, + "step": 8416 + }, + { + "epoch": 0.9243356028991874, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0841805934906006, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7035728693008423, + "num_tokens": 217880720.0, + "step": 8417 + }, + { + "epoch": 0.924445420601801, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.4147257804870605, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7082991600036621, + "num_tokens": 217901861.0, + "step": 8418 + }, + { + "epoch": 0.9245552383044147, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.0398457050323486, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.695382833480835, + "num_tokens": 217931798.0, + "step": 8419 + }, + { + "epoch": 0.9246650560070283, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1003291606903076, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.6920932531356812, + "num_tokens": 217959279.0, + "step": 8420 + }, + { + "epoch": 0.924774873709642, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2218499183654785, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7066113948822021, + "num_tokens": 217984120.0, + "step": 8421 + }, + { + "epoch": 0.9248846914122557, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1120798587799072, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7298022508621216, + "num_tokens": 218010085.0, + "step": 8422 + }, + { + "epoch": 0.9249945091148694, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.9272364377975464, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7053508162498474, + "num_tokens": 218042423.0, + "step": 8423 + }, + { + "epoch": 0.925104326817483, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.2765793800354004, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7001726627349854, + "num_tokens": 218067588.0, + "step": 8424 + }, + { + "epoch": 0.9252141445200966, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 1.8964601755142212, + "learning_rate": 1e-06, + "loss": 1.1299, + "mean_token_accuracy": 0.6645755767822266, + "num_tokens": 218101759.0, + "step": 8425 + }, + { + "epoch": 0.9253239622227103, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.1245522499084473, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7020686864852905, + "num_tokens": 218129303.0, + "step": 8426 + }, + { + "epoch": 0.9254337799253239, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.27123761177063, + "learning_rate": 1e-06, + "loss": 0.9766, + "mean_token_accuracy": 0.7022902965545654, + "num_tokens": 218154835.0, + "step": 8427 + }, + { + "epoch": 0.9255435976279376, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.053060531616211, + "learning_rate": 1e-06, + "loss": 1.0363, + "mean_token_accuracy": 0.6840142607688904, + "num_tokens": 218183636.0, + "step": 8428 + }, + { + "epoch": 0.9256534153305512, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.3924946784973145, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7133803963661194, + "num_tokens": 218207019.0, + "step": 8429 + }, + { + "epoch": 0.925763233033165, + "ewc_loss": 4.26173210144043e-06, + "grad_norm": 2.261141538619995, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7194865345954895, + "num_tokens": 218229583.0, + "step": 8430 + }, + { + "epoch": 0.9258730507357786, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.139528512954712, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7039311528205872, + "num_tokens": 218256194.0, + "step": 8431 + }, + { + "epoch": 0.9259828684383923, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.338571310043335, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.685576319694519, + "num_tokens": 218280046.0, + "step": 8432 + }, + { + "epoch": 0.9260926861410059, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.4910547733306885, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7181298732757568, + "num_tokens": 218299251.0, + "step": 8433 + }, + { + "epoch": 0.9262025038436196, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2728066444396973, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7132188677787781, + "num_tokens": 218323170.0, + "step": 8434 + }, + { + "epoch": 0.9263123215462332, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.304311990737915, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7129695415496826, + "num_tokens": 218346012.0, + "step": 8435 + }, + { + "epoch": 0.9264221392488469, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.433621883392334, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7208276987075806, + "num_tokens": 218368155.0, + "step": 8436 + }, + { + "epoch": 0.9265319569514606, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0345380306243896, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7001078724861145, + "num_tokens": 218394595.0, + "step": 8437 + }, + { + "epoch": 0.9266417746540743, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.242877244949341, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7071866989135742, + "num_tokens": 218419929.0, + "step": 8438 + }, + { + "epoch": 0.9267515923566879, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3190557956695557, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7348734140396118, + "num_tokens": 218441566.0, + "step": 8439 + }, + { + "epoch": 0.9268614100593016, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.343660593032837, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7189345955848694, + "num_tokens": 218464451.0, + "step": 8440 + }, + { + "epoch": 0.9269712277619152, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.049752950668335, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7021909952163696, + "num_tokens": 218492548.0, + "step": 8441 + }, + { + "epoch": 0.9270810454645289, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2188477516174316, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7104437351226807, + "num_tokens": 218519220.0, + "step": 8442 + }, + { + "epoch": 0.9271908631671425, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3790385723114014, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.6990765333175659, + "num_tokens": 218541685.0, + "step": 8443 + }, + { + "epoch": 0.9273006808697563, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.435624122619629, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7256343364715576, + "num_tokens": 218565855.0, + "step": 8444 + }, + { + "epoch": 0.9274104985723699, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0877838134765625, + "learning_rate": 1e-06, + "loss": 1.0407, + "mean_token_accuracy": 0.6937633156776428, + "num_tokens": 218592633.0, + "step": 8445 + }, + { + "epoch": 0.9275203162749835, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.080777645111084, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7082091569900513, + "num_tokens": 218618417.0, + "step": 8446 + }, + { + "epoch": 0.9276301339775972, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.221205711364746, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7208861112594604, + "num_tokens": 218641999.0, + "step": 8447 + }, + { + "epoch": 0.9277399516802108, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.077451467514038, + "learning_rate": 1e-06, + "loss": 1.0501, + "mean_token_accuracy": 0.6863372325897217, + "num_tokens": 218670529.0, + "step": 8448 + }, + { + "epoch": 0.9278497693828245, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.383720874786377, + "learning_rate": 1e-06, + "loss": 1.0728, + "mean_token_accuracy": 0.6846863031387329, + "num_tokens": 218695484.0, + "step": 8449 + }, + { + "epoch": 0.9279595870854381, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0189273357391357, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.7070093154907227, + "num_tokens": 218727725.0, + "step": 8450 + }, + { + "epoch": 0.9280694047880519, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.011586904525757, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7218927145004272, + "num_tokens": 218754505.0, + "step": 8451 + }, + { + "epoch": 0.9281792224906655, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1826868057250977, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7087845802307129, + "num_tokens": 218778201.0, + "step": 8452 + }, + { + "epoch": 0.9282890401932792, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.261913776397705, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7195032835006714, + "num_tokens": 218801071.0, + "step": 8453 + }, + { + "epoch": 0.9283988578958928, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3603310585021973, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7033560276031494, + "num_tokens": 218821290.0, + "step": 8454 + }, + { + "epoch": 0.9285086755985065, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0475358963012695, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7227667570114136, + "num_tokens": 218847790.0, + "step": 8455 + }, + { + "epoch": 0.9286184933011201, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.170949697494507, + "learning_rate": 1e-06, + "loss": 1.0346, + "mean_token_accuracy": 0.6819744110107422, + "num_tokens": 218872798.0, + "step": 8456 + }, + { + "epoch": 0.9287283110037338, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2127184867858887, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.6966947317123413, + "num_tokens": 218896411.0, + "step": 8457 + }, + { + "epoch": 0.9288381287063474, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2645297050476074, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7029789686203003, + "num_tokens": 218919199.0, + "step": 8458 + }, + { + "epoch": 0.9289479464089612, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1180145740509033, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7075884938240051, + "num_tokens": 218947282.0, + "step": 8459 + }, + { + "epoch": 0.9290577641115748, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.207064151763916, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7067369818687439, + "num_tokens": 218970788.0, + "step": 8460 + }, + { + "epoch": 0.9291675818141885, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.986413598060608, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7181985378265381, + "num_tokens": 219001070.0, + "step": 8461 + }, + { + "epoch": 0.9292773995168021, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.21235728263855, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.6785416603088379, + "num_tokens": 219029358.0, + "step": 8462 + }, + { + "epoch": 0.9293872172194158, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1147234439849854, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7299808263778687, + "num_tokens": 219055155.0, + "step": 8463 + }, + { + "epoch": 0.9294970349220294, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.8540204763412476, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6885156035423279, + "num_tokens": 219088784.0, + "step": 8464 + }, + { + "epoch": 0.929606852624643, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.8916661739349365, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.722314178943634, + "num_tokens": 219119790.0, + "step": 8465 + }, + { + "epoch": 0.9297166703272568, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.203695774078369, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7129141092300415, + "num_tokens": 219142801.0, + "step": 8466 + }, + { + "epoch": 0.9298264880298704, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0911219120025635, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7329087853431702, + "num_tokens": 219168134.0, + "step": 8467 + }, + { + "epoch": 0.9299363057324841, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.263089179992676, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7327375411987305, + "num_tokens": 219191748.0, + "step": 8468 + }, + { + "epoch": 0.9300461234350977, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.4736685752868652, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6993777751922607, + "num_tokens": 219214834.0, + "step": 8469 + }, + { + "epoch": 0.9301559411377114, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.20393443107605, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7149944305419922, + "num_tokens": 219243265.0, + "step": 8470 + }, + { + "epoch": 0.930265758840325, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.8588528633117676, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7176494598388672, + "num_tokens": 219275706.0, + "step": 8471 + }, + { + "epoch": 0.9303755765429387, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.032663106918335, + "learning_rate": 1e-06, + "loss": 1.0599, + "mean_token_accuracy": 0.6744258403778076, + "num_tokens": 219305391.0, + "step": 8472 + }, + { + "epoch": 0.9304853942455524, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.207951068878174, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6984869241714478, + "num_tokens": 219331399.0, + "step": 8473 + }, + { + "epoch": 0.9305952119481661, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0294291973114014, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7288936376571655, + "num_tokens": 219358967.0, + "step": 8474 + }, + { + "epoch": 0.9307050296507797, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.8152461051940918, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.702549934387207, + "num_tokens": 219394839.0, + "step": 8475 + }, + { + "epoch": 0.9308148473533934, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.534883975982666, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7191317081451416, + "num_tokens": 219413954.0, + "step": 8476 + }, + { + "epoch": 0.930924665056007, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1208407878875732, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7042653560638428, + "num_tokens": 219440917.0, + "step": 8477 + }, + { + "epoch": 0.9310344827586207, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2295126914978027, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6885310411453247, + "num_tokens": 219469897.0, + "step": 8478 + }, + { + "epoch": 0.9311443004612343, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2837531566619873, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7126941680908203, + "num_tokens": 219491251.0, + "step": 8479 + }, + { + "epoch": 0.9312541181638481, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.393176794052124, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7086788415908813, + "num_tokens": 219512466.0, + "step": 8480 + }, + { + "epoch": 0.9313639358664617, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1476495265960693, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7075023651123047, + "num_tokens": 219539230.0, + "step": 8481 + }, + { + "epoch": 0.9314737535690754, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0703842639923096, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7134414911270142, + "num_tokens": 219567470.0, + "step": 8482 + }, + { + "epoch": 0.931583571271689, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.0428197383880615, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7130968570709229, + "num_tokens": 219595741.0, + "step": 8483 + }, + { + "epoch": 0.9316933889743026, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.208935260772705, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7250508069992065, + "num_tokens": 219620741.0, + "step": 8484 + }, + { + "epoch": 0.9318032066769163, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.153780221939087, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7023749351501465, + "num_tokens": 219648414.0, + "step": 8485 + }, + { + "epoch": 0.9319130243795299, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.05253005027771, + "learning_rate": 1e-06, + "loss": 1.0085, + "mean_token_accuracy": 0.698072075843811, + "num_tokens": 219675933.0, + "step": 8486 + }, + { + "epoch": 0.9320228420821436, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.166616916656494, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.7000134587287903, + "num_tokens": 219702289.0, + "step": 8487 + }, + { + "epoch": 0.9321326597847573, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2783310413360596, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7002036571502686, + "num_tokens": 219724969.0, + "step": 8488 + }, + { + "epoch": 0.932242477487371, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.4492924213409424, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.726118803024292, + "num_tokens": 219745247.0, + "step": 8489 + }, + { + "epoch": 0.9323522951899846, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3707995414733887, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7213082909584045, + "num_tokens": 219767121.0, + "step": 8490 + }, + { + "epoch": 0.9324621128925983, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.086301565170288, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7074935436248779, + "num_tokens": 219793110.0, + "step": 8491 + }, + { + "epoch": 0.9325719305952119, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.2039730548858643, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7196952104568481, + "num_tokens": 219817431.0, + "step": 8492 + }, + { + "epoch": 0.9326817482978256, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.280266046524048, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7131629586219788, + "num_tokens": 219840958.0, + "step": 8493 + }, + { + "epoch": 0.9327915660004392, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.346278429031372, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7043874263763428, + "num_tokens": 219864220.0, + "step": 8494 + }, + { + "epoch": 0.932901383703053, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.299339532852173, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7283819913864136, + "num_tokens": 219885350.0, + "step": 8495 + }, + { + "epoch": 0.9330112014056666, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.157266139984131, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.7055494785308838, + "num_tokens": 219910183.0, + "step": 8496 + }, + { + "epoch": 0.9331210191082803, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3241965770721436, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7144507169723511, + "num_tokens": 219935143.0, + "step": 8497 + }, + { + "epoch": 0.9332308368108939, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.1159894466400146, + "learning_rate": 1e-06, + "loss": 1.0185, + "mean_token_accuracy": 0.688376247882843, + "num_tokens": 219962271.0, + "step": 8498 + }, + { + "epoch": 0.9333406545135076, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 1.944740653038025, + "learning_rate": 1e-06, + "loss": 1.0521, + "mean_token_accuracy": 0.6859737634658813, + "num_tokens": 219996977.0, + "step": 8499 + }, + { + "epoch": 0.9334504722161212, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.3148128986358643, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7098882794380188, + "num_tokens": 220019017.0, + "step": 8500 + }, + { + "epoch": 0.9335602899187349, + "ewc_loss": 4.291534423828125e-06, + "grad_norm": 2.098281145095825, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7208766937255859, + "num_tokens": 220044241.0, + "step": 8501 + }, + { + "epoch": 0.9336701076213486, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.433171510696411, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6983641982078552, + "num_tokens": 220066675.0, + "step": 8502 + }, + { + "epoch": 0.9337799253239623, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.4079902172088623, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7236499190330505, + "num_tokens": 220087579.0, + "step": 8503 + }, + { + "epoch": 0.9338897430265759, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.657201051712036, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7142671346664429, + "num_tokens": 220105762.0, + "step": 8504 + }, + { + "epoch": 0.9339995607291895, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3995816707611084, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7116602659225464, + "num_tokens": 220127377.0, + "step": 8505 + }, + { + "epoch": 0.9341093784318032, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.990504264831543, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.718108057975769, + "num_tokens": 220156035.0, + "step": 8506 + }, + { + "epoch": 0.9342191961344168, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.9721993207931519, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.710114598274231, + "num_tokens": 220186364.0, + "step": 8507 + }, + { + "epoch": 0.9343290138370305, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0502333641052246, + "learning_rate": 1e-06, + "loss": 1.0288, + "mean_token_accuracy": 0.6910821795463562, + "num_tokens": 220216923.0, + "step": 8508 + }, + { + "epoch": 0.9344388315396442, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.099926471710205, + "learning_rate": 1e-06, + "loss": 1.0334, + "mean_token_accuracy": 0.6846843957901001, + "num_tokens": 220243427.0, + "step": 8509 + }, + { + "epoch": 0.9345486492422579, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.100512742996216, + "learning_rate": 1e-06, + "loss": 1.0546, + "mean_token_accuracy": 0.683985710144043, + "num_tokens": 220270601.0, + "step": 8510 + }, + { + "epoch": 0.9346584669448715, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1147491931915283, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.7019587755203247, + "num_tokens": 220299537.0, + "step": 8511 + }, + { + "epoch": 0.9347682846474852, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3767013549804688, + "learning_rate": 1e-06, + "loss": 1.0624, + "mean_token_accuracy": 0.6790447235107422, + "num_tokens": 220325252.0, + "step": 8512 + }, + { + "epoch": 0.9348781023500988, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.007216453552246, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7136411666870117, + "num_tokens": 220351982.0, + "step": 8513 + }, + { + "epoch": 0.9349879200527125, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.949266791343689, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7051125764846802, + "num_tokens": 220381152.0, + "step": 8514 + }, + { + "epoch": 0.9350977377553261, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.5460665225982666, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7335811257362366, + "num_tokens": 220399949.0, + "step": 8515 + }, + { + "epoch": 0.9352075554579398, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.89519464969635, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7119100093841553, + "num_tokens": 220432231.0, + "step": 8516 + }, + { + "epoch": 0.9353173731605535, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2520487308502197, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7176230549812317, + "num_tokens": 220455363.0, + "step": 8517 + }, + { + "epoch": 0.9354271908631672, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0008063316345215, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.6968574523925781, + "num_tokens": 220485388.0, + "step": 8518 + }, + { + "epoch": 0.9355370085657808, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.137314796447754, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7005770206451416, + "num_tokens": 220510686.0, + "step": 8519 + }, + { + "epoch": 0.9356468262683945, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.110992193222046, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7009557485580444, + "num_tokens": 220538470.0, + "step": 8520 + }, + { + "epoch": 0.9357566439710081, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3893909454345703, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7285361886024475, + "num_tokens": 220559701.0, + "step": 8521 + }, + { + "epoch": 0.9358664616736218, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0486342906951904, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7259455323219299, + "num_tokens": 220587287.0, + "step": 8522 + }, + { + "epoch": 0.9359762793762354, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.197572946548462, + "learning_rate": 1e-06, + "loss": 0.9824, + "mean_token_accuracy": 0.7051737904548645, + "num_tokens": 220612885.0, + "step": 8523 + }, + { + "epoch": 0.9360860970788492, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.004634380340576, + "learning_rate": 1e-06, + "loss": 1.0362, + "mean_token_accuracy": 0.6857221126556396, + "num_tokens": 220642379.0, + "step": 8524 + }, + { + "epoch": 0.9361959147814628, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.030881881713867, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.7001841068267822, + "num_tokens": 220670345.0, + "step": 8525 + }, + { + "epoch": 0.9363057324840764, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2410597801208496, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6882556676864624, + "num_tokens": 220695748.0, + "step": 8526 + }, + { + "epoch": 0.9364155501866901, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0456948280334473, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.6979688405990601, + "num_tokens": 220725066.0, + "step": 8527 + }, + { + "epoch": 0.9365253678893037, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2861170768737793, + "learning_rate": 1e-06, + "loss": 1.0302, + "mean_token_accuracy": 0.6916763782501221, + "num_tokens": 220750281.0, + "step": 8528 + }, + { + "epoch": 0.9366351855919174, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.165400505065918, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7075381875038147, + "num_tokens": 220778300.0, + "step": 8529 + }, + { + "epoch": 0.936745003294531, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2032530307769775, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7079725861549377, + "num_tokens": 220803943.0, + "step": 8530 + }, + { + "epoch": 0.9368548209971448, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.149238109588623, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7236692905426025, + "num_tokens": 220827576.0, + "step": 8531 + }, + { + "epoch": 0.9369646386997584, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.496286392211914, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.6974901556968689, + "num_tokens": 220849305.0, + "step": 8532 + }, + { + "epoch": 0.9370744564023721, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2349965572357178, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6927181482315063, + "num_tokens": 220875563.0, + "step": 8533 + }, + { + "epoch": 0.9371842741049857, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.254026412963867, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7108098864555359, + "num_tokens": 220898919.0, + "step": 8534 + }, + { + "epoch": 0.9372940918075994, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1369409561157227, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.6951276659965515, + "num_tokens": 220925388.0, + "step": 8535 + }, + { + "epoch": 0.937403909510213, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3972136974334717, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7081843614578247, + "num_tokens": 220946587.0, + "step": 8536 + }, + { + "epoch": 0.9375137272128267, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2928450107574463, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.6992953419685364, + "num_tokens": 220970097.0, + "step": 8537 + }, + { + "epoch": 0.9376235449154404, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.029426336288452, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.6960406303405762, + "num_tokens": 220999343.0, + "step": 8538 + }, + { + "epoch": 0.9377333626180541, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.058480978012085, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7061708569526672, + "num_tokens": 221027252.0, + "step": 8539 + }, + { + "epoch": 0.9378431803206677, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.333448886871338, + "learning_rate": 1e-06, + "loss": 1.0458, + "mean_token_accuracy": 0.7023499011993408, + "num_tokens": 221052625.0, + "step": 8540 + }, + { + "epoch": 0.9379529980232814, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.27530574798584, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7080450654029846, + "num_tokens": 221076412.0, + "step": 8541 + }, + { + "epoch": 0.938062815725895, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.160956621170044, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7242076992988586, + "num_tokens": 221104149.0, + "step": 8542 + }, + { + "epoch": 0.9381726334285087, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.032472610473633, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7007791996002197, + "num_tokens": 221132434.0, + "step": 8543 + }, + { + "epoch": 0.9382824511311223, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.161416530609131, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7067475318908691, + "num_tokens": 221157135.0, + "step": 8544 + }, + { + "epoch": 0.9383922688337359, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.360564947128296, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7243858575820923, + "num_tokens": 221177932.0, + "step": 8545 + }, + { + "epoch": 0.9385020865363497, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3604254722595215, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7145916223526001, + "num_tokens": 221200677.0, + "step": 8546 + }, + { + "epoch": 0.9386119042389633, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0587971210479736, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7119070887565613, + "num_tokens": 221226510.0, + "step": 8547 + }, + { + "epoch": 0.938721721941577, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0627522468566895, + "learning_rate": 1e-06, + "loss": 1.0345, + "mean_token_accuracy": 0.6833782196044922, + "num_tokens": 221254321.0, + "step": 8548 + }, + { + "epoch": 0.9388315396441906, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3652875423431396, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7068454027175903, + "num_tokens": 221277301.0, + "step": 8549 + }, + { + "epoch": 0.9389413573468043, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.403306007385254, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.72696453332901, + "num_tokens": 221298457.0, + "step": 8550 + }, + { + "epoch": 0.9390511750494179, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.161304473876953, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7110697031021118, + "num_tokens": 221323999.0, + "step": 8551 + }, + { + "epoch": 0.9391609927520316, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.035492420196533, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7201522588729858, + "num_tokens": 221350861.0, + "step": 8552 + }, + { + "epoch": 0.9392708104546453, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0597124099731445, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.6993141174316406, + "num_tokens": 221379066.0, + "step": 8553 + }, + { + "epoch": 0.939380628157259, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1546080112457275, + "learning_rate": 1e-06, + "loss": 1.0858, + "mean_token_accuracy": 0.6764211058616638, + "num_tokens": 221405645.0, + "step": 8554 + }, + { + "epoch": 0.9394904458598726, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.9250673055648804, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7085950970649719, + "num_tokens": 221434456.0, + "step": 8555 + }, + { + "epoch": 0.9396002635624863, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.306502103805542, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7264152765274048, + "num_tokens": 221455517.0, + "step": 8556 + }, + { + "epoch": 0.9397100812650999, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.4675660133361816, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7113891839981079, + "num_tokens": 221475373.0, + "step": 8557 + }, + { + "epoch": 0.9398198989677136, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1164748668670654, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7121149897575378, + "num_tokens": 221503028.0, + "step": 8558 + }, + { + "epoch": 0.9399297166703272, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.357013463973999, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7159042954444885, + "num_tokens": 221526144.0, + "step": 8559 + }, + { + "epoch": 0.940039534372941, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.503643274307251, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6953514814376831, + "num_tokens": 221547183.0, + "step": 8560 + }, + { + "epoch": 0.9401493520755546, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2952053546905518, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7490240335464478, + "num_tokens": 221569164.0, + "step": 8561 + }, + { + "epoch": 0.9402591697781683, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0657620429992676, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7060295343399048, + "num_tokens": 221597868.0, + "step": 8562 + }, + { + "epoch": 0.9403689874807819, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1882452964782715, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7052832245826721, + "num_tokens": 221624517.0, + "step": 8563 + }, + { + "epoch": 0.9404788051833955, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.9920463562011719, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.7038052082061768, + "num_tokens": 221654198.0, + "step": 8564 + }, + { + "epoch": 0.9405886228860092, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.424420118331909, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.712613582611084, + "num_tokens": 221676354.0, + "step": 8565 + }, + { + "epoch": 0.9406984405886228, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.9070152044296265, + "learning_rate": 1e-06, + "loss": 1.0541, + "mean_token_accuracy": 0.6796526908874512, + "num_tokens": 221711475.0, + "step": 8566 + }, + { + "epoch": 0.9408082582912366, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.179959774017334, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7197276949882507, + "num_tokens": 221738072.0, + "step": 8567 + }, + { + "epoch": 0.9409180759938502, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1253087520599365, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.6943663358688354, + "num_tokens": 221765201.0, + "step": 8568 + }, + { + "epoch": 0.9410278936964639, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.115675210952759, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7053680419921875, + "num_tokens": 221789983.0, + "step": 8569 + }, + { + "epoch": 0.9411377113990775, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0186848640441895, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.711720883846283, + "num_tokens": 221818675.0, + "step": 8570 + }, + { + "epoch": 0.9412475291016912, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 1.9333778619766235, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6978005170822144, + "num_tokens": 221849966.0, + "step": 8571 + }, + { + "epoch": 0.9413573468043048, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.1180307865142822, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7088218927383423, + "num_tokens": 221877967.0, + "step": 8572 + }, + { + "epoch": 0.9414671645069185, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2133116722106934, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.73084557056427, + "num_tokens": 221901832.0, + "step": 8573 + }, + { + "epoch": 0.9415769822095322, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.3027663230895996, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7101269364356995, + "num_tokens": 221926364.0, + "step": 8574 + }, + { + "epoch": 0.9416867999121459, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.2272307872772217, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7242496609687805, + "num_tokens": 221949997.0, + "step": 8575 + }, + { + "epoch": 0.9417966176147595, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.6172709465026855, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7154266834259033, + "num_tokens": 221972166.0, + "step": 8576 + }, + { + "epoch": 0.9419064353173732, + "ewc_loss": 4.32133674621582e-06, + "grad_norm": 2.0353505611419678, + "learning_rate": 1e-06, + "loss": 1.0069, + "mean_token_accuracy": 0.7028701305389404, + "num_tokens": 222000577.0, + "step": 8577 + }, + { + "epoch": 0.9420162530199868, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1048567295074463, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7126956582069397, + "num_tokens": 222028566.0, + "step": 8578 + }, + { + "epoch": 0.9421260707226005, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2611753940582275, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7285480499267578, + "num_tokens": 222052475.0, + "step": 8579 + }, + { + "epoch": 0.9422358884252141, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 1.9986025094985962, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.6965542435646057, + "num_tokens": 222080988.0, + "step": 8580 + }, + { + "epoch": 0.9423457061278278, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0995612144470215, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.7038835287094116, + "num_tokens": 222108697.0, + "step": 8581 + }, + { + "epoch": 0.9424555238304415, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.741089344024658, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7242955565452576, + "num_tokens": 222126981.0, + "step": 8582 + }, + { + "epoch": 0.9425653415330552, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0747907161712646, + "learning_rate": 1e-06, + "loss": 1.1078, + "mean_token_accuracy": 0.6720761656761169, + "num_tokens": 222155767.0, + "step": 8583 + }, + { + "epoch": 0.9426751592356688, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.486370325088501, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7006615400314331, + "num_tokens": 222177533.0, + "step": 8584 + }, + { + "epoch": 0.9427849769382824, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.108645439147949, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7152423858642578, + "num_tokens": 222201844.0, + "step": 8585 + }, + { + "epoch": 0.9428947946408961, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3324990272521973, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7154244780540466, + "num_tokens": 222224495.0, + "step": 8586 + }, + { + "epoch": 0.9430046123435097, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.612240791320801, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.705429196357727, + "num_tokens": 222243560.0, + "step": 8587 + }, + { + "epoch": 0.9431144300461234, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 1.9842854738235474, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7213529348373413, + "num_tokens": 222274129.0, + "step": 8588 + }, + { + "epoch": 0.9432242477487371, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.042623519897461, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7192125916481018, + "num_tokens": 222305237.0, + "step": 8589 + }, + { + "epoch": 0.9433340654513508, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3918776512145996, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7031246423721313, + "num_tokens": 222328774.0, + "step": 8590 + }, + { + "epoch": 0.9434438831539644, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1211390495300293, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.725365400314331, + "num_tokens": 222352974.0, + "step": 8591 + }, + { + "epoch": 0.9435537008565781, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.299901247024536, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7046736478805542, + "num_tokens": 222375481.0, + "step": 8592 + }, + { + "epoch": 0.9436635185591917, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.223409652709961, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7056412100791931, + "num_tokens": 222399146.0, + "step": 8593 + }, + { + "epoch": 0.9437733362618054, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 1.8331336975097656, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6950275897979736, + "num_tokens": 222432584.0, + "step": 8594 + }, + { + "epoch": 0.943883153964419, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.424806833267212, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7059530019760132, + "num_tokens": 222454366.0, + "step": 8595 + }, + { + "epoch": 0.9439929716670328, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.274578332901001, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7098041772842407, + "num_tokens": 222479174.0, + "step": 8596 + }, + { + "epoch": 0.9441027893696464, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2395379543304443, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7254509925842285, + "num_tokens": 222503908.0, + "step": 8597 + }, + { + "epoch": 0.9442126070722601, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.6477551460266113, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7428230047225952, + "num_tokens": 222520848.0, + "step": 8598 + }, + { + "epoch": 0.9443224247748737, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.524275779724121, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7309803366661072, + "num_tokens": 222538364.0, + "step": 8599 + }, + { + "epoch": 0.9444322424774874, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.093966245651245, + "learning_rate": 1e-06, + "loss": 0.9985, + "mean_token_accuracy": 0.6944948434829712, + "num_tokens": 222565875.0, + "step": 8600 + }, + { + "epoch": 0.944542060180101, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2598462104797363, + "learning_rate": 1e-06, + "loss": 1.0171, + "mean_token_accuracy": 0.6850298047065735, + "num_tokens": 222591050.0, + "step": 8601 + }, + { + "epoch": 0.9446518778827147, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1522436141967773, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7162132263183594, + "num_tokens": 222616289.0, + "step": 8602 + }, + { + "epoch": 0.9447616955853284, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0380451679229736, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7053658366203308, + "num_tokens": 222642934.0, + "step": 8603 + }, + { + "epoch": 0.944871513287942, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0843498706817627, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7335207462310791, + "num_tokens": 222669971.0, + "step": 8604 + }, + { + "epoch": 0.9449813309905557, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3291516304016113, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7169076204299927, + "num_tokens": 222693886.0, + "step": 8605 + }, + { + "epoch": 0.9450911486931693, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2225043773651123, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7032737731933594, + "num_tokens": 222718751.0, + "step": 8606 + }, + { + "epoch": 0.945200966395783, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.4190807342529297, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.728851318359375, + "num_tokens": 222739723.0, + "step": 8607 + }, + { + "epoch": 0.9453107840983966, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.27534556388855, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7161214351654053, + "num_tokens": 222764431.0, + "step": 8608 + }, + { + "epoch": 0.9454206018010103, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.363166093826294, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7214174270629883, + "num_tokens": 222786067.0, + "step": 8609 + }, + { + "epoch": 0.9455304195036239, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.187649965286255, + "learning_rate": 1e-06, + "loss": 1.0323, + "mean_token_accuracy": 0.6889638900756836, + "num_tokens": 222811849.0, + "step": 8610 + }, + { + "epoch": 0.9456402372062377, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2393503189086914, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6910188794136047, + "num_tokens": 222837098.0, + "step": 8611 + }, + { + "epoch": 0.9457500549088513, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1139869689941406, + "learning_rate": 1e-06, + "loss": 1.0568, + "mean_token_accuracy": 0.6925070285797119, + "num_tokens": 222867209.0, + "step": 8612 + }, + { + "epoch": 0.945859872611465, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.218524694442749, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6851449012756348, + "num_tokens": 222891024.0, + "step": 8613 + }, + { + "epoch": 0.9459696903140786, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1935529708862305, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.716394305229187, + "num_tokens": 222914495.0, + "step": 8614 + }, + { + "epoch": 0.9460795080166923, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.4983272552490234, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7217526435852051, + "num_tokens": 222933415.0, + "step": 8615 + }, + { + "epoch": 0.9461893257193059, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.309333086013794, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7081787586212158, + "num_tokens": 222956258.0, + "step": 8616 + }, + { + "epoch": 0.9462991434219196, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.5899789333343506, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7182204723358154, + "num_tokens": 222976109.0, + "step": 8617 + }, + { + "epoch": 0.9464089611245333, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1884896755218506, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.702194094657898, + "num_tokens": 223002419.0, + "step": 8618 + }, + { + "epoch": 0.946518778827147, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0914130210876465, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.7031136751174927, + "num_tokens": 223029426.0, + "step": 8619 + }, + { + "epoch": 0.9466285965297606, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1259541511535645, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7246352434158325, + "num_tokens": 223055601.0, + "step": 8620 + }, + { + "epoch": 0.9467384142323743, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0115609169006348, + "learning_rate": 1e-06, + "loss": 1.0343, + "mean_token_accuracy": 0.6906263828277588, + "num_tokens": 223086401.0, + "step": 8621 + }, + { + "epoch": 0.9468482319349879, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.1743969917297363, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7075490951538086, + "num_tokens": 223112709.0, + "step": 8622 + }, + { + "epoch": 0.9469580496376016, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3704397678375244, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7139843702316284, + "num_tokens": 223132726.0, + "step": 8623 + }, + { + "epoch": 0.9470678673402152, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.6099374294281006, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7270694971084595, + "num_tokens": 223151442.0, + "step": 8624 + }, + { + "epoch": 0.947177685042829, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.2357239723205566, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7068161964416504, + "num_tokens": 223174241.0, + "step": 8625 + }, + { + "epoch": 0.9472875027454426, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.3338780403137207, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7212101221084595, + "num_tokens": 223196993.0, + "step": 8626 + }, + { + "epoch": 0.9473973204480562, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 1.9795458316802979, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.701518177986145, + "num_tokens": 223226771.0, + "step": 8627 + }, + { + "epoch": 0.9475071381506699, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.265735387802124, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.715468168258667, + "num_tokens": 223249487.0, + "step": 8628 + }, + { + "epoch": 0.9476169558532835, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.133237600326538, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6984591484069824, + "num_tokens": 223275629.0, + "step": 8629 + }, + { + "epoch": 0.9477267735558972, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.0870280265808105, + "learning_rate": 1e-06, + "loss": 1.0392, + "mean_token_accuracy": 0.6905021667480469, + "num_tokens": 223304025.0, + "step": 8630 + }, + { + "epoch": 0.9478365912585108, + "ewc_loss": 4.351139068603516e-06, + "grad_norm": 2.073493719100952, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.70557701587677, + "num_tokens": 223333077.0, + "step": 8631 + }, + { + "epoch": 0.9479464089611246, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.367920398712158, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7013748288154602, + "num_tokens": 223356647.0, + "step": 8632 + }, + { + "epoch": 0.9480562266637382, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.791776418685913, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.728223443031311, + "num_tokens": 223374548.0, + "step": 8633 + }, + { + "epoch": 0.9481660443663519, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.260788917541504, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7202166318893433, + "num_tokens": 223400800.0, + "step": 8634 + }, + { + "epoch": 0.9482758620689655, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.6845247745513916, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7288923263549805, + "num_tokens": 223419768.0, + "step": 8635 + }, + { + "epoch": 0.9483856797715792, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0347042083740234, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7155270576477051, + "num_tokens": 223446443.0, + "step": 8636 + }, + { + "epoch": 0.9484954974741928, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0828795433044434, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7184132933616638, + "num_tokens": 223473510.0, + "step": 8637 + }, + { + "epoch": 0.9486053151768065, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.1720733642578125, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7373207807540894, + "num_tokens": 223503095.0, + "step": 8638 + }, + { + "epoch": 0.9487151328794201, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.091357946395874, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.6990616321563721, + "num_tokens": 223532652.0, + "step": 8639 + }, + { + "epoch": 0.9488249505820339, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.08756422996521, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7166241407394409, + "num_tokens": 223560980.0, + "step": 8640 + }, + { + "epoch": 0.9489347682846475, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.012375593185425, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7293699979782104, + "num_tokens": 223586348.0, + "step": 8641 + }, + { + "epoch": 0.9490445859872612, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.471114158630371, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7125245332717896, + "num_tokens": 223607708.0, + "step": 8642 + }, + { + "epoch": 0.9491544036898748, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.41568660736084, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7331841588020325, + "num_tokens": 223628312.0, + "step": 8643 + }, + { + "epoch": 0.9492642213924884, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.112779378890991, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7082109451293945, + "num_tokens": 223656073.0, + "step": 8644 + }, + { + "epoch": 0.9493740390951021, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.1677355766296387, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7131112217903137, + "num_tokens": 223681909.0, + "step": 8645 + }, + { + "epoch": 0.9494838567977157, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.221501350402832, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6947214603424072, + "num_tokens": 223706150.0, + "step": 8646 + }, + { + "epoch": 0.9495936745003295, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.296889066696167, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7041932344436646, + "num_tokens": 223730524.0, + "step": 8647 + }, + { + "epoch": 0.9497034922029431, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.641083002090454, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.737489640712738, + "num_tokens": 223748705.0, + "step": 8648 + }, + { + "epoch": 0.9498133099055568, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 1.8811092376708984, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7024494409561157, + "num_tokens": 223784072.0, + "step": 8649 + }, + { + "epoch": 0.9499231276081704, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.104870557785034, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7381267547607422, + "num_tokens": 223811086.0, + "step": 8650 + }, + { + "epoch": 0.9500329453107841, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.283005475997925, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7156215310096741, + "num_tokens": 223835135.0, + "step": 8651 + }, + { + "epoch": 0.9501427630133977, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.2275054454803467, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6911667585372925, + "num_tokens": 223861787.0, + "step": 8652 + }, + { + "epoch": 0.9502525807160114, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.051983594894409, + "learning_rate": 1e-06, + "loss": 1.0248, + "mean_token_accuracy": 0.688037633895874, + "num_tokens": 223891162.0, + "step": 8653 + }, + { + "epoch": 0.9503623984186251, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.373060941696167, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7201945185661316, + "num_tokens": 223912433.0, + "step": 8654 + }, + { + "epoch": 0.9504722161212388, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.312911033630371, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6875758171081543, + "num_tokens": 223938108.0, + "step": 8655 + }, + { + "epoch": 0.9505820338238524, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0569286346435547, + "learning_rate": 1e-06, + "loss": 1.0455, + "mean_token_accuracy": 0.6883678436279297, + "num_tokens": 223965637.0, + "step": 8656 + }, + { + "epoch": 0.9506918515264661, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.552215576171875, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7131941914558411, + "num_tokens": 223985296.0, + "step": 8657 + }, + { + "epoch": 0.9508016692290797, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.07169246673584, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.69880211353302, + "num_tokens": 224013684.0, + "step": 8658 + }, + { + "epoch": 0.9509114869316934, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.030222177505493, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.701994001865387, + "num_tokens": 224041925.0, + "step": 8659 + }, + { + "epoch": 0.951021304634307, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.431234121322632, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7050670385360718, + "num_tokens": 224062807.0, + "step": 8660 + }, + { + "epoch": 0.9511311223369208, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0952107906341553, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.7066996097564697, + "num_tokens": 224090444.0, + "step": 8661 + }, + { + "epoch": 0.9512409400395344, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.342193126678467, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6945282220840454, + "num_tokens": 224115116.0, + "step": 8662 + }, + { + "epoch": 0.9513507577421481, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.008680582046509, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.6918958425521851, + "num_tokens": 224144631.0, + "step": 8663 + }, + { + "epoch": 0.9514605754447617, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0389404296875, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7088698148727417, + "num_tokens": 224173130.0, + "step": 8664 + }, + { + "epoch": 0.9515703931473753, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.292344808578491, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7300293445587158, + "num_tokens": 224196040.0, + "step": 8665 + }, + { + "epoch": 0.951680210849989, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.0601046085357666, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7357380390167236, + "num_tokens": 224221596.0, + "step": 8666 + }, + { + "epoch": 0.9517900285526026, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.203404664993286, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7148368954658508, + "num_tokens": 224247333.0, + "step": 8667 + }, + { + "epoch": 0.9518998462552163, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.136523485183716, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7189925909042358, + "num_tokens": 224274178.0, + "step": 8668 + }, + { + "epoch": 0.95200966395783, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 1.9797297716140747, + "learning_rate": 1e-06, + "loss": 1.0015, + "mean_token_accuracy": 0.6960723400115967, + "num_tokens": 224304474.0, + "step": 8669 + }, + { + "epoch": 0.9521194816604437, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.297311544418335, + "learning_rate": 1e-06, + "loss": 1.0463, + "mean_token_accuracy": 0.6879752278327942, + "num_tokens": 224329214.0, + "step": 8670 + }, + { + "epoch": 0.9522292993630573, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.1859116554260254, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7087647318840027, + "num_tokens": 224353752.0, + "step": 8671 + }, + { + "epoch": 0.952339117065671, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.711674928665161, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7370563745498657, + "num_tokens": 224369261.0, + "step": 8672 + }, + { + "epoch": 0.9524489347682846, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.094658613204956, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7086837291717529, + "num_tokens": 224397273.0, + "step": 8673 + }, + { + "epoch": 0.9525587524708983, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.348975896835327, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7107097506523132, + "num_tokens": 224419111.0, + "step": 8674 + }, + { + "epoch": 0.9526685701735119, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.1255693435668945, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.713346004486084, + "num_tokens": 224445100.0, + "step": 8675 + }, + { + "epoch": 0.9527783878761257, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.2018401622772217, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7295942306518555, + "num_tokens": 224469047.0, + "step": 8676 + }, + { + "epoch": 0.9528882055787393, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.300191640853882, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7204011678695679, + "num_tokens": 224491089.0, + "step": 8677 + }, + { + "epoch": 0.952998023281353, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1788620948791504, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7055249810218811, + "num_tokens": 224515134.0, + "step": 8678 + }, + { + "epoch": 0.9531078409839666, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9885051250457764, + "learning_rate": 1e-06, + "loss": 1.0691, + "mean_token_accuracy": 0.6796554327011108, + "num_tokens": 224548089.0, + "step": 8679 + }, + { + "epoch": 0.9532176586865803, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1871867179870605, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7108802199363708, + "num_tokens": 224571705.0, + "step": 8680 + }, + { + "epoch": 0.9533274763891939, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.185495376586914, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6906806230545044, + "num_tokens": 224596889.0, + "step": 8681 + }, + { + "epoch": 0.9534372940918076, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.7995648384094238, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7160499691963196, + "num_tokens": 224632269.0, + "step": 8682 + }, + { + "epoch": 0.9535471117944213, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.050229787826538, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6989647150039673, + "num_tokens": 224659968.0, + "step": 8683 + }, + { + "epoch": 0.953656929497035, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9538960456848145, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7172275185585022, + "num_tokens": 224689653.0, + "step": 8684 + }, + { + "epoch": 0.9537667471996486, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.992708683013916, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.6977144479751587, + "num_tokens": 224717652.0, + "step": 8685 + }, + { + "epoch": 0.9538765649022622, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9030128717422485, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.692237377166748, + "num_tokens": 224751602.0, + "step": 8686 + }, + { + "epoch": 0.9539863826048759, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.23997163772583, + "learning_rate": 1e-06, + "loss": 1.0201, + "mean_token_accuracy": 0.6960256099700928, + "num_tokens": 224775812.0, + "step": 8687 + }, + { + "epoch": 0.9540962003074895, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2512645721435547, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7247211933135986, + "num_tokens": 224798433.0, + "step": 8688 + }, + { + "epoch": 0.9542060180101032, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0900070667266846, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6959712505340576, + "num_tokens": 224826605.0, + "step": 8689 + }, + { + "epoch": 0.9543158357127169, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.145881175994873, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7273595929145813, + "num_tokens": 224850608.0, + "step": 8690 + }, + { + "epoch": 0.9544256534153306, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.4731059074401855, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7084493637084961, + "num_tokens": 224872229.0, + "step": 8691 + }, + { + "epoch": 0.9545354711179442, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.115104913711548, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7225306630134583, + "num_tokens": 224899062.0, + "step": 8692 + }, + { + "epoch": 0.9546452888205579, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.430546760559082, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7149839401245117, + "num_tokens": 224920953.0, + "step": 8693 + }, + { + "epoch": 0.9547551065231715, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.2956156730651855, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7014147043228149, + "num_tokens": 224945827.0, + "step": 8694 + }, + { + "epoch": 0.9548649242257852, + "ewc_loss": 4.380941390991211e-06, + "grad_norm": 2.112682342529297, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7031587362289429, + "num_tokens": 224975422.0, + "step": 8695 + }, + { + "epoch": 0.9549747419283988, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.205803632736206, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6976941823959351, + "num_tokens": 225001592.0, + "step": 8696 + }, + { + "epoch": 0.9550845596310125, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.5719945430755615, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7206999063491821, + "num_tokens": 225022802.0, + "step": 8697 + }, + { + "epoch": 0.9551943773336262, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.250086784362793, + "learning_rate": 1e-06, + "loss": 1.0379, + "mean_token_accuracy": 0.6868878602981567, + "num_tokens": 225048842.0, + "step": 8698 + }, + { + "epoch": 0.9553041950362399, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9492321014404297, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7090954184532166, + "num_tokens": 225080619.0, + "step": 8699 + }, + { + "epoch": 0.9554140127388535, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.936765432357788, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7113333344459534, + "num_tokens": 225110529.0, + "step": 8700 + }, + { + "epoch": 0.9555238304414672, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1191930770874023, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7370826601982117, + "num_tokens": 225136463.0, + "step": 8701 + }, + { + "epoch": 0.9556336481440808, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2038896083831787, + "learning_rate": 1e-06, + "loss": 1.0191, + "mean_token_accuracy": 0.6895617842674255, + "num_tokens": 225161023.0, + "step": 8702 + }, + { + "epoch": 0.9557434658466945, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.129891872406006, + "learning_rate": 1e-06, + "loss": 1.0976, + "mean_token_accuracy": 0.6734311580657959, + "num_tokens": 225188822.0, + "step": 8703 + }, + { + "epoch": 0.9558532835493081, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9681276082992554, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7060544490814209, + "num_tokens": 225219426.0, + "step": 8704 + }, + { + "epoch": 0.9559631012519219, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.337092399597168, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7132916450500488, + "num_tokens": 225241096.0, + "step": 8705 + }, + { + "epoch": 0.9560729189545355, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2350029945373535, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7068734169006348, + "num_tokens": 225264747.0, + "step": 8706 + }, + { + "epoch": 0.9561827366571491, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.251910448074341, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7067509889602661, + "num_tokens": 225287683.0, + "step": 8707 + }, + { + "epoch": 0.9562925543597628, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0648114681243896, + "learning_rate": 1e-06, + "loss": 1.0233, + "mean_token_accuracy": 0.6918754577636719, + "num_tokens": 225316574.0, + "step": 8708 + }, + { + "epoch": 0.9564023720623764, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.123891592025757, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7098425626754761, + "num_tokens": 225343257.0, + "step": 8709 + }, + { + "epoch": 0.9565121897649901, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2751548290252686, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7399391531944275, + "num_tokens": 225366372.0, + "step": 8710 + }, + { + "epoch": 0.9566220074676037, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0142972469329834, + "learning_rate": 1e-06, + "loss": 1.0583, + "mean_token_accuracy": 0.6897279024124146, + "num_tokens": 225397796.0, + "step": 8711 + }, + { + "epoch": 0.9567318251702175, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0119004249572754, + "learning_rate": 1e-06, + "loss": 1.114, + "mean_token_accuracy": 0.6790544986724854, + "num_tokens": 225431624.0, + "step": 8712 + }, + { + "epoch": 0.9568416428728311, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0748088359832764, + "learning_rate": 1e-06, + "loss": 1.0781, + "mean_token_accuracy": 0.6790421009063721, + "num_tokens": 225459819.0, + "step": 8713 + }, + { + "epoch": 0.9569514605754448, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1051244735717773, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7082386016845703, + "num_tokens": 225487993.0, + "step": 8714 + }, + { + "epoch": 0.9570612782780584, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.133073329925537, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7193430662155151, + "num_tokens": 225513334.0, + "step": 8715 + }, + { + "epoch": 0.9571710959806721, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.21439266204834, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.712211012840271, + "num_tokens": 225537623.0, + "step": 8716 + }, + { + "epoch": 0.9572809136832857, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.561401128768921, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7077765464782715, + "num_tokens": 225556885.0, + "step": 8717 + }, + { + "epoch": 0.9573907313858994, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2990779876708984, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.6995046138763428, + "num_tokens": 225582847.0, + "step": 8718 + }, + { + "epoch": 0.9575005490885131, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.66949462890625, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7365037202835083, + "num_tokens": 225598872.0, + "step": 8719 + }, + { + "epoch": 0.9576103667911268, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2329347133636475, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6986449360847473, + "num_tokens": 225624823.0, + "step": 8720 + }, + { + "epoch": 0.9577201844937404, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2078206539154053, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.7014048099517822, + "num_tokens": 225649755.0, + "step": 8721 + }, + { + "epoch": 0.9578300021963541, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.829444169998169, + "learning_rate": 1e-06, + "loss": 1.0214, + "mean_token_accuracy": 0.6883132457733154, + "num_tokens": 225683143.0, + "step": 8722 + }, + { + "epoch": 0.9579398198989677, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.8769493103027344, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.722987174987793, + "num_tokens": 225714652.0, + "step": 8723 + }, + { + "epoch": 0.9580496376015813, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.2444419860839844, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7201939821243286, + "num_tokens": 225737649.0, + "step": 8724 + }, + { + "epoch": 0.958159455304195, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.166787624359131, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7174668312072754, + "num_tokens": 225762807.0, + "step": 8725 + }, + { + "epoch": 0.9582692730068088, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.098104476928711, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.6891675591468811, + "num_tokens": 225790806.0, + "step": 8726 + }, + { + "epoch": 0.9583790907094224, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0568103790283203, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7155380249023438, + "num_tokens": 225818475.0, + "step": 8727 + }, + { + "epoch": 0.958488908412036, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.4017317295074463, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7121937274932861, + "num_tokens": 225838758.0, + "step": 8728 + }, + { + "epoch": 0.9585987261146497, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.001615524291992, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7091220617294312, + "num_tokens": 225869813.0, + "step": 8729 + }, + { + "epoch": 0.9587085438172633, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.203094959259033, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7204154133796692, + "num_tokens": 225893794.0, + "step": 8730 + }, + { + "epoch": 0.958818361519877, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0808284282684326, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.6992090940475464, + "num_tokens": 225920531.0, + "step": 8731 + }, + { + "epoch": 0.9589281792224906, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 1.9622077941894531, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7121252417564392, + "num_tokens": 225951936.0, + "step": 8732 + }, + { + "epoch": 0.9590379969251043, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.351288318634033, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7155588865280151, + "num_tokens": 225974222.0, + "step": 8733 + }, + { + "epoch": 0.959147814627718, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.2778971195220947, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7268746495246887, + "num_tokens": 225997213.0, + "step": 8734 + }, + { + "epoch": 0.9592576323303317, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.4324214458465576, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7056128978729248, + "num_tokens": 226017402.0, + "step": 8735 + }, + { + "epoch": 0.9593674500329453, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1257638931274414, + "learning_rate": 1e-06, + "loss": 1.038, + "mean_token_accuracy": 0.68662428855896, + "num_tokens": 226046124.0, + "step": 8736 + }, + { + "epoch": 0.959477267735559, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.193645477294922, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.6992675065994263, + "num_tokens": 226072832.0, + "step": 8737 + }, + { + "epoch": 0.9595870854381726, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.0160396099090576, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7138271331787109, + "num_tokens": 226099350.0, + "step": 8738 + }, + { + "epoch": 0.9596969031407863, + "ewc_loss": 4.410743713378906e-06, + "grad_norm": 2.1515371799468994, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7108035683631897, + "num_tokens": 226123899.0, + "step": 8739 + }, + { + "epoch": 0.9598067208433999, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.0304226875305176, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7126123905181885, + "num_tokens": 226151450.0, + "step": 8740 + }, + { + "epoch": 0.9599165385460137, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.003174304962158, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7190229892730713, + "num_tokens": 226180002.0, + "step": 8741 + }, + { + "epoch": 0.9600263562486273, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 1.9984545707702637, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6955653429031372, + "num_tokens": 226207642.0, + "step": 8742 + }, + { + "epoch": 0.960136173951241, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.297438859939575, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7069293260574341, + "num_tokens": 226233048.0, + "step": 8743 + }, + { + "epoch": 0.9602459916538546, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.2735605239868164, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7256259918212891, + "num_tokens": 226254190.0, + "step": 8744 + }, + { + "epoch": 0.9603558093564682, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.1158010959625244, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7034692764282227, + "num_tokens": 226280390.0, + "step": 8745 + }, + { + "epoch": 0.9604656270590819, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.033670425415039, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.7049213647842407, + "num_tokens": 226308489.0, + "step": 8746 + }, + { + "epoch": 0.9605754447616955, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.137056589126587, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7161451578140259, + "num_tokens": 226334100.0, + "step": 8747 + }, + { + "epoch": 0.9606852624643093, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.3978569507598877, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7221633195877075, + "num_tokens": 226355150.0, + "step": 8748 + }, + { + "epoch": 0.9607950801669229, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.0293619632720947, + "learning_rate": 1e-06, + "loss": 1.0588, + "mean_token_accuracy": 0.6848611235618591, + "num_tokens": 226385231.0, + "step": 8749 + }, + { + "epoch": 0.9609048978695366, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 1.9134124517440796, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7116004228591919, + "num_tokens": 226416209.0, + "step": 8750 + }, + { + "epoch": 0.9610147155721502, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.519726276397705, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7137070894241333, + "num_tokens": 226435635.0, + "step": 8751 + }, + { + "epoch": 0.9611245332747639, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 1.8957399129867554, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7030922770500183, + "num_tokens": 226468132.0, + "step": 8752 + }, + { + "epoch": 0.9612343509773775, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.1584606170654297, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.726720929145813, + "num_tokens": 226492763.0, + "step": 8753 + }, + { + "epoch": 0.9613441686799912, + "ewc_loss": 4.470348358154297e-06, + "grad_norm": 2.2137815952301025, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7088963389396667, + "num_tokens": 226518240.0, + "step": 8754 + }, + { + "epoch": 0.9614539863826049, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.451235771179199, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7138863801956177, + "num_tokens": 226538071.0, + "step": 8755 + }, + { + "epoch": 0.9615638040852186, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.832208275794983, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7086374759674072, + "num_tokens": 226567951.0, + "step": 8756 + }, + { + "epoch": 0.9616736217878322, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.6146247386932373, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7132410407066345, + "num_tokens": 226585283.0, + "step": 8757 + }, + { + "epoch": 0.9617834394904459, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9904048442840576, + "learning_rate": 1e-06, + "loss": 0.9796, + "mean_token_accuracy": 0.7032381296157837, + "num_tokens": 226614720.0, + "step": 8758 + }, + { + "epoch": 0.9618932571930595, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.460797071456909, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7099392414093018, + "num_tokens": 226636638.0, + "step": 8759 + }, + { + "epoch": 0.9620030748956732, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.057004451751709, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6969727873802185, + "num_tokens": 226664693.0, + "step": 8760 + }, + { + "epoch": 0.9621128925982868, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.193054437637329, + "learning_rate": 1e-06, + "loss": 1.0502, + "mean_token_accuracy": 0.6788802742958069, + "num_tokens": 226691569.0, + "step": 8761 + }, + { + "epoch": 0.9622227103009005, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.596275568008423, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7261636853218079, + "num_tokens": 226708968.0, + "step": 8762 + }, + { + "epoch": 0.9623325280035142, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9742884635925293, + "learning_rate": 1e-06, + "loss": 1.0228, + "mean_token_accuracy": 0.6932896971702576, + "num_tokens": 226739371.0, + "step": 8763 + }, + { + "epoch": 0.9624423457061279, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.139194965362549, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.699393630027771, + "num_tokens": 226765112.0, + "step": 8764 + }, + { + "epoch": 0.9625521634087415, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1712381839752197, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7056159973144531, + "num_tokens": 226790665.0, + "step": 8765 + }, + { + "epoch": 0.9626619811113551, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0214436054229736, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7073973417282104, + "num_tokens": 226820430.0, + "step": 8766 + }, + { + "epoch": 0.9627717988139688, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9605438709259033, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6957809925079346, + "num_tokens": 226852716.0, + "step": 8767 + }, + { + "epoch": 0.9628816165165824, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0345072746276855, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7177355885505676, + "num_tokens": 226880918.0, + "step": 8768 + }, + { + "epoch": 0.9629914342191961, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.230642318725586, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7043590545654297, + "num_tokens": 226904758.0, + "step": 8769 + }, + { + "epoch": 0.9631012519218098, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.131920337677002, + "learning_rate": 1e-06, + "loss": 1.0156, + "mean_token_accuracy": 0.6930941343307495, + "num_tokens": 226933464.0, + "step": 8770 + }, + { + "epoch": 0.9632110696244235, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0517172813415527, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7166937589645386, + "num_tokens": 226961442.0, + "step": 8771 + }, + { + "epoch": 0.9633208873270371, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.190962553024292, + "learning_rate": 1e-06, + "loss": 0.9986, + "mean_token_accuracy": 0.7018142938613892, + "num_tokens": 226987718.0, + "step": 8772 + }, + { + "epoch": 0.9634307050296508, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1972763538360596, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7018032670021057, + "num_tokens": 227014707.0, + "step": 8773 + }, + { + "epoch": 0.9635405227322644, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0830318927764893, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6973758339881897, + "num_tokens": 227042658.0, + "step": 8774 + }, + { + "epoch": 0.9636503404348781, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.219517230987549, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7199227809906006, + "num_tokens": 227065980.0, + "step": 8775 + }, + { + "epoch": 0.9637601581374917, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1589279174804688, + "learning_rate": 1e-06, + "loss": 1.007, + "mean_token_accuracy": 0.6884372234344482, + "num_tokens": 227092798.0, + "step": 8776 + }, + { + "epoch": 0.9638699758401055, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1006717681884766, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.7000652551651001, + "num_tokens": 227119037.0, + "step": 8777 + }, + { + "epoch": 0.9639797935427191, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9958502054214478, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7127230763435364, + "num_tokens": 227147385.0, + "step": 8778 + }, + { + "epoch": 0.9640896112453328, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0537095069885254, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7009755373001099, + "num_tokens": 227175846.0, + "step": 8779 + }, + { + "epoch": 0.9641994289479464, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0868797302246094, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6936428546905518, + "num_tokens": 227203360.0, + "step": 8780 + }, + { + "epoch": 0.9643092466505601, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.177732467651367, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7002668380737305, + "num_tokens": 227228358.0, + "step": 8781 + }, + { + "epoch": 0.9644190643531737, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.076233386993408, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7248820066452026, + "num_tokens": 227255492.0, + "step": 8782 + }, + { + "epoch": 0.9645288820557874, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.418100357055664, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.714625358581543, + "num_tokens": 227276939.0, + "step": 8783 + }, + { + "epoch": 0.9646386997584011, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9985944032669067, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.6983532905578613, + "num_tokens": 227305598.0, + "step": 8784 + }, + { + "epoch": 0.9647485174610148, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.253561496734619, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.7010592818260193, + "num_tokens": 227329581.0, + "step": 8785 + }, + { + "epoch": 0.9648583351636284, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1797502040863037, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7107986807823181, + "num_tokens": 227354844.0, + "step": 8786 + }, + { + "epoch": 0.964968152866242, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.035304546356201, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7014089822769165, + "num_tokens": 227383559.0, + "step": 8787 + }, + { + "epoch": 0.9650779705688557, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.2071926593780518, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7176588773727417, + "num_tokens": 227407240.0, + "step": 8788 + }, + { + "epoch": 0.9651877882714693, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9457398653030396, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7080584764480591, + "num_tokens": 227438972.0, + "step": 8789 + }, + { + "epoch": 0.965297605974083, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0466816425323486, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.7052460312843323, + "num_tokens": 227466755.0, + "step": 8790 + }, + { + "epoch": 0.9654074236766966, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.4820001125335693, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7369505167007446, + "num_tokens": 227486042.0, + "step": 8791 + }, + { + "epoch": 0.9655172413793104, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0095224380493164, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7322912216186523, + "num_tokens": 227513569.0, + "step": 8792 + }, + { + "epoch": 0.965627059081924, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.4392611980438232, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7034721374511719, + "num_tokens": 227536632.0, + "step": 8793 + }, + { + "epoch": 0.9657368767845377, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.0389606952667236, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.691494345664978, + "num_tokens": 227565842.0, + "step": 8794 + }, + { + "epoch": 0.9658466944871513, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9750949144363403, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7046371698379517, + "num_tokens": 227595085.0, + "step": 8795 + }, + { + "epoch": 0.965956512189765, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.185419797897339, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7123918533325195, + "num_tokens": 227619531.0, + "step": 8796 + }, + { + "epoch": 0.9660663298923786, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.2075889110565186, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6942389011383057, + "num_tokens": 227643118.0, + "step": 8797 + }, + { + "epoch": 0.9661761475949923, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1238598823547363, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7140551805496216, + "num_tokens": 227668000.0, + "step": 8798 + }, + { + "epoch": 0.966285965297606, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.245374917984009, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7287014722824097, + "num_tokens": 227690524.0, + "step": 8799 + }, + { + "epoch": 0.9663957830002197, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.187119483947754, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7171444296836853, + "num_tokens": 227714620.0, + "step": 8800 + }, + { + "epoch": 0.9665056007028333, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.1455793380737305, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7249160408973694, + "num_tokens": 227739464.0, + "step": 8801 + }, + { + "epoch": 0.966615418405447, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.182879686355591, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.7012884616851807, + "num_tokens": 227764732.0, + "step": 8802 + }, + { + "epoch": 0.9667252361080606, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.9331132173538208, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6839901208877563, + "num_tokens": 227795769.0, + "step": 8803 + }, + { + "epoch": 0.9668350538106742, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 2.2647881507873535, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7159013748168945, + "num_tokens": 227820704.0, + "step": 8804 + }, + { + "epoch": 0.9669448715132879, + "ewc_loss": 4.500150680541992e-06, + "grad_norm": 1.8823710680007935, + "learning_rate": 1e-06, + "loss": 1.0583, + "mean_token_accuracy": 0.691265344619751, + "num_tokens": 227857518.0, + "step": 8805 + }, + { + "epoch": 0.9670546892159017, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.1316895484924316, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7202985882759094, + "num_tokens": 227881946.0, + "step": 8806 + }, + { + "epoch": 0.9671645069185153, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.0533018112182617, + "learning_rate": 1e-06, + "loss": 1.0408, + "mean_token_accuracy": 0.6882627010345459, + "num_tokens": 227912468.0, + "step": 8807 + }, + { + "epoch": 0.967274324621129, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.302053213119507, + "learning_rate": 1e-06, + "loss": 1.0625, + "mean_token_accuracy": 0.6820027828216553, + "num_tokens": 227938208.0, + "step": 8808 + }, + { + "epoch": 0.9673841423237426, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.1842358112335205, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6948928833007812, + "num_tokens": 227964581.0, + "step": 8809 + }, + { + "epoch": 0.9674939600263562, + "ewc_loss": 4.5299530029296875e-06, + "grad_norm": 2.029803514480591, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7065178155899048, + "num_tokens": 227995051.0, + "step": 8810 + }, + { + "epoch": 0.9676037777289699, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.7921528816223145, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7040097713470459, + "num_tokens": 228011822.0, + "step": 8811 + }, + { + "epoch": 0.9677135954315835, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.115081787109375, + "learning_rate": 1e-06, + "loss": 1.0741, + "mean_token_accuracy": 0.6733044385910034, + "num_tokens": 228041097.0, + "step": 8812 + }, + { + "epoch": 0.9678234131341973, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.8854691982269287, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7071935534477234, + "num_tokens": 228073745.0, + "step": 8813 + }, + { + "epoch": 0.9679332308368109, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3375954627990723, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7156647443771362, + "num_tokens": 228096488.0, + "step": 8814 + }, + { + "epoch": 0.9680430485394246, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.011611223220825, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.6967108845710754, + "num_tokens": 228126843.0, + "step": 8815 + }, + { + "epoch": 0.9681528662420382, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.304152488708496, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7252771258354187, + "num_tokens": 228148818.0, + "step": 8816 + }, + { + "epoch": 0.9682626839446519, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2481205463409424, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7290949821472168, + "num_tokens": 228171522.0, + "step": 8817 + }, + { + "epoch": 0.9683725016472655, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.301973819732666, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.6909207105636597, + "num_tokens": 228197254.0, + "step": 8818 + }, + { + "epoch": 0.9684823193498792, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2970032691955566, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7025936245918274, + "num_tokens": 228219818.0, + "step": 8819 + }, + { + "epoch": 0.9685921370524928, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.377500295639038, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7178837060928345, + "num_tokens": 228239719.0, + "step": 8820 + }, + { + "epoch": 0.9687019547551066, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2332215309143066, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7256520390510559, + "num_tokens": 228262713.0, + "step": 8821 + }, + { + "epoch": 0.9688117724577202, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1333351135253906, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.7074543833732605, + "num_tokens": 228289415.0, + "step": 8822 + }, + { + "epoch": 0.9689215901603339, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.081361770629883, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7070122957229614, + "num_tokens": 228315429.0, + "step": 8823 + }, + { + "epoch": 0.9690314078629475, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1686816215515137, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7412782907485962, + "num_tokens": 228339247.0, + "step": 8824 + }, + { + "epoch": 0.9691412255655611, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1451287269592285, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.724068284034729, + "num_tokens": 228364715.0, + "step": 8825 + }, + { + "epoch": 0.9692510432681748, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4414479732513428, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7269972562789917, + "num_tokens": 228386350.0, + "step": 8826 + }, + { + "epoch": 0.9693608609707884, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9525258541107178, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7070250511169434, + "num_tokens": 228418504.0, + "step": 8827 + }, + { + "epoch": 0.9694706786734022, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2428510189056396, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7170002460479736, + "num_tokens": 228443214.0, + "step": 8828 + }, + { + "epoch": 0.9695804963760158, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.342047929763794, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7181294560432434, + "num_tokens": 228465758.0, + "step": 8829 + }, + { + "epoch": 0.9696903140786295, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0106639862060547, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7178993821144104, + "num_tokens": 228492918.0, + "step": 8830 + }, + { + "epoch": 0.9698001317812431, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2303049564361572, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6961631178855896, + "num_tokens": 228517881.0, + "step": 8831 + }, + { + "epoch": 0.9699099494838568, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.925036907196045, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7125575542449951, + "num_tokens": 228548073.0, + "step": 8832 + }, + { + "epoch": 0.9700197671864704, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4435911178588867, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7174581289291382, + "num_tokens": 228568379.0, + "step": 8833 + }, + { + "epoch": 0.9701295848890841, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1125760078430176, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7085431218147278, + "num_tokens": 228595878.0, + "step": 8834 + }, + { + "epoch": 0.9702394025916978, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1409173011779785, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.7002029418945312, + "num_tokens": 228626734.0, + "step": 8835 + }, + { + "epoch": 0.9703492202943115, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1507246494293213, + "learning_rate": 1e-06, + "loss": 1.0337, + "mean_token_accuracy": 0.6913087368011475, + "num_tokens": 228657659.0, + "step": 8836 + }, + { + "epoch": 0.9704590379969251, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3546717166900635, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7110097408294678, + "num_tokens": 228679738.0, + "step": 8837 + }, + { + "epoch": 0.9705688556995388, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0914015769958496, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7104769349098206, + "num_tokens": 228706159.0, + "step": 8838 + }, + { + "epoch": 0.9706786734021524, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.439218044281006, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7105722427368164, + "num_tokens": 228727542.0, + "step": 8839 + }, + { + "epoch": 0.9707884911047661, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2266552448272705, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.695927083492279, + "num_tokens": 228751295.0, + "step": 8840 + }, + { + "epoch": 0.9708983088073797, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.266152858734131, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.7012953758239746, + "num_tokens": 228778417.0, + "step": 8841 + }, + { + "epoch": 0.9710081265099935, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.8465709686279297, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.697490930557251, + "num_tokens": 228812038.0, + "step": 8842 + }, + { + "epoch": 0.9711179442126071, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0863847732543945, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7089060544967651, + "num_tokens": 228839443.0, + "step": 8843 + }, + { + "epoch": 0.9712277619152208, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1741628646850586, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7206676602363586, + "num_tokens": 228863670.0, + "step": 8844 + }, + { + "epoch": 0.9713375796178344, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.8064005374908447, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.7056893110275269, + "num_tokens": 228899048.0, + "step": 8845 + }, + { + "epoch": 0.971447397320448, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.161475658416748, + "learning_rate": 1e-06, + "loss": 1.0219, + "mean_token_accuracy": 0.6841564178466797, + "num_tokens": 228925711.0, + "step": 8846 + }, + { + "epoch": 0.9715572150230617, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.338848352432251, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7113810777664185, + "num_tokens": 228948934.0, + "step": 8847 + }, + { + "epoch": 0.9716670327256753, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.069852113723755, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7120741009712219, + "num_tokens": 228975185.0, + "step": 8848 + }, + { + "epoch": 0.971776850428289, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1251845359802246, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7298943400382996, + "num_tokens": 229001724.0, + "step": 8849 + }, + { + "epoch": 0.9718866681309027, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.583958148956299, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7273726463317871, + "num_tokens": 229023802.0, + "step": 8850 + }, + { + "epoch": 0.9719964858335164, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9995429515838623, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.723397970199585, + "num_tokens": 229053214.0, + "step": 8851 + }, + { + "epoch": 0.97210630353613, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9747583866119385, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7124647498130798, + "num_tokens": 229082556.0, + "step": 8852 + }, + { + "epoch": 0.9722161212387437, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.412017345428467, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.7012903690338135, + "num_tokens": 229102678.0, + "step": 8853 + }, + { + "epoch": 0.9723259389413573, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.310361623764038, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7092827558517456, + "num_tokens": 229127059.0, + "step": 8854 + }, + { + "epoch": 0.972435756643971, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2512052059173584, + "learning_rate": 1e-06, + "loss": 1.0181, + "mean_token_accuracy": 0.7035470604896545, + "num_tokens": 229151902.0, + "step": 8855 + }, + { + "epoch": 0.9725455743465846, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.270226240158081, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.720863401889801, + "num_tokens": 229174950.0, + "step": 8856 + }, + { + "epoch": 0.9726553920491984, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3336164951324463, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7192972302436829, + "num_tokens": 229196120.0, + "step": 8857 + }, + { + "epoch": 0.972765209751812, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1442606449127197, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7000987529754639, + "num_tokens": 229223840.0, + "step": 8858 + }, + { + "epoch": 0.9728750274544257, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4730021953582764, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.715589165687561, + "num_tokens": 229243153.0, + "step": 8859 + }, + { + "epoch": 0.9729848451570393, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1497418880462646, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7010291814804077, + "num_tokens": 229270777.0, + "step": 8860 + }, + { + "epoch": 0.973094662859653, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3641228675842285, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7141678333282471, + "num_tokens": 229292593.0, + "step": 8861 + }, + { + "epoch": 0.9732044805622666, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.271869659423828, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7135002613067627, + "num_tokens": 229317275.0, + "step": 8862 + }, + { + "epoch": 0.9733142982648803, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1650075912475586, + "learning_rate": 1e-06, + "loss": 1.0327, + "mean_token_accuracy": 0.6934443116188049, + "num_tokens": 229343841.0, + "step": 8863 + }, + { + "epoch": 0.973424115967494, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.29836368560791, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6970854997634888, + "num_tokens": 229368220.0, + "step": 8864 + }, + { + "epoch": 0.9735339336701077, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4837963581085205, + "learning_rate": 1e-06, + "loss": 0.9968, + "mean_token_accuracy": 0.6938279271125793, + "num_tokens": 229389388.0, + "step": 8865 + }, + { + "epoch": 0.9736437513727213, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0286648273468018, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6980650424957275, + "num_tokens": 229419230.0, + "step": 8866 + }, + { + "epoch": 0.973753569075335, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9953752756118774, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.6930029988288879, + "num_tokens": 229448049.0, + "step": 8867 + }, + { + "epoch": 0.9738633867779486, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.265029191970825, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.6967878341674805, + "num_tokens": 229472943.0, + "step": 8868 + }, + { + "epoch": 0.9739732044805622, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4452295303344727, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.725033164024353, + "num_tokens": 229492958.0, + "step": 8869 + }, + { + "epoch": 0.9740830221831759, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.999906301498413, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.7541438937187195, + "num_tokens": 229518904.0, + "step": 8870 + }, + { + "epoch": 0.9741928398857896, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0740389823913574, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.705594539642334, + "num_tokens": 229546570.0, + "step": 8871 + }, + { + "epoch": 0.9743026575884033, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.340264320373535, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7164108753204346, + "num_tokens": 229567965.0, + "step": 8872 + }, + { + "epoch": 0.9744124752910169, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.019357681274414, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7124074101448059, + "num_tokens": 229596198.0, + "step": 8873 + }, + { + "epoch": 0.9745222929936306, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.043076515197754, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7064830660820007, + "num_tokens": 229626224.0, + "step": 8874 + }, + { + "epoch": 0.9746321106962442, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.244398355484009, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6906038522720337, + "num_tokens": 229652706.0, + "step": 8875 + }, + { + "epoch": 0.9747419283988579, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1020355224609375, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7069947719573975, + "num_tokens": 229679203.0, + "step": 8876 + }, + { + "epoch": 0.9748517461014715, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3407130241394043, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7003852128982544, + "num_tokens": 229702096.0, + "step": 8877 + }, + { + "epoch": 0.9749615638040853, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.478440523147583, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.6990253329277039, + "num_tokens": 229721871.0, + "step": 8878 + }, + { + "epoch": 0.9750713815066989, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.318081855773926, + "learning_rate": 1e-06, + "loss": 1.0347, + "mean_token_accuracy": 0.6872762441635132, + "num_tokens": 229745519.0, + "step": 8879 + }, + { + "epoch": 0.9751811992093126, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.157310962677002, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7187598943710327, + "num_tokens": 229771220.0, + "step": 8880 + }, + { + "epoch": 0.9752910169119262, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2976818084716797, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6927276253700256, + "num_tokens": 229795634.0, + "step": 8881 + }, + { + "epoch": 0.9754008346145399, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4152326583862305, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7369434833526611, + "num_tokens": 229815447.0, + "step": 8882 + }, + { + "epoch": 0.9755106523171535, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9079937934875488, + "learning_rate": 1e-06, + "loss": 1.1157, + "mean_token_accuracy": 0.6648929119110107, + "num_tokens": 229846624.0, + "step": 8883 + }, + { + "epoch": 0.9756204700197671, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.237691879272461, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.710970938205719, + "num_tokens": 229870063.0, + "step": 8884 + }, + { + "epoch": 0.9757302877223808, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.9715057611465454, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7071533203125, + "num_tokens": 229898350.0, + "step": 8885 + }, + { + "epoch": 0.9758401054249946, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.351430892944336, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6934658288955688, + "num_tokens": 229921538.0, + "step": 8886 + }, + { + "epoch": 0.9759499231276082, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2092905044555664, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.6987844705581665, + "num_tokens": 229945119.0, + "step": 8887 + }, + { + "epoch": 0.9760597408302218, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3210792541503906, + "learning_rate": 1e-06, + "loss": 1.0273, + "mean_token_accuracy": 0.6907117366790771, + "num_tokens": 229968920.0, + "step": 8888 + }, + { + "epoch": 0.9761695585328355, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.8874081373214722, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7095901966094971, + "num_tokens": 230003802.0, + "step": 8889 + }, + { + "epoch": 0.9762793762354491, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.194187641143799, + "learning_rate": 1e-06, + "loss": 1.0436, + "mean_token_accuracy": 0.7004655599594116, + "num_tokens": 230030227.0, + "step": 8890 + }, + { + "epoch": 0.9763891939380628, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4622080326080322, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7162582278251648, + "num_tokens": 230053413.0, + "step": 8891 + }, + { + "epoch": 0.9764990116406764, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.366452932357788, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7042773365974426, + "num_tokens": 230076540.0, + "step": 8892 + }, + { + "epoch": 0.9766088293432902, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1389424800872803, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.709277868270874, + "num_tokens": 230102079.0, + "step": 8893 + }, + { + "epoch": 0.9767186470459038, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1112220287323, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6967196464538574, + "num_tokens": 230129033.0, + "step": 8894 + }, + { + "epoch": 0.9768284647485175, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.040295124053955, + "learning_rate": 1e-06, + "loss": 1.0097, + "mean_token_accuracy": 0.6915429830551147, + "num_tokens": 230158669.0, + "step": 8895 + }, + { + "epoch": 0.9769382824511311, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.4891536235809326, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7239859104156494, + "num_tokens": 230178979.0, + "step": 8896 + }, + { + "epoch": 0.9770481001537448, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1694250106811523, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7136472463607788, + "num_tokens": 230203677.0, + "step": 8897 + }, + { + "epoch": 0.9771579178563584, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.434823513031006, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7409405708312988, + "num_tokens": 230223685.0, + "step": 8898 + }, + { + "epoch": 0.9772677355589721, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.372467517852783, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7270922660827637, + "num_tokens": 230243836.0, + "step": 8899 + }, + { + "epoch": 0.9773775532615858, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2230520248413086, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.6948877573013306, + "num_tokens": 230267068.0, + "step": 8900 + }, + { + "epoch": 0.9774873709641995, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1122000217437744, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6973164081573486, + "num_tokens": 230293163.0, + "step": 8901 + }, + { + "epoch": 0.9775971886668131, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 1.965602159500122, + "learning_rate": 1e-06, + "loss": 1.0485, + "mean_token_accuracy": 0.6833475828170776, + "num_tokens": 230325348.0, + "step": 8902 + }, + { + "epoch": 0.9777070063694268, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.5070478916168213, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7329401969909668, + "num_tokens": 230345616.0, + "step": 8903 + }, + { + "epoch": 0.9778168240720404, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.393390417098999, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7153983116149902, + "num_tokens": 230366751.0, + "step": 8904 + }, + { + "epoch": 0.977926641774654, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.110199451446533, + "learning_rate": 1e-06, + "loss": 1.0475, + "mean_token_accuracy": 0.6850535869598389, + "num_tokens": 230395863.0, + "step": 8905 + }, + { + "epoch": 0.9780364594772677, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.6849536895751953, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7267075777053833, + "num_tokens": 230414203.0, + "step": 8906 + }, + { + "epoch": 0.9781462771798815, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1586530208587646, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7129819393157959, + "num_tokens": 230440005.0, + "step": 8907 + }, + { + "epoch": 0.9782560948824951, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.2925753593444824, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7029811143875122, + "num_tokens": 230464148.0, + "step": 8908 + }, + { + "epoch": 0.9783659125851087, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.3631858825683594, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7119094729423523, + "num_tokens": 230484594.0, + "step": 8909 + }, + { + "epoch": 0.9784757302877224, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0904948711395264, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7095423340797424, + "num_tokens": 230510260.0, + "step": 8910 + }, + { + "epoch": 0.978585547990336, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.0174129009246826, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.6969566941261292, + "num_tokens": 230539461.0, + "step": 8911 + }, + { + "epoch": 0.9786953656929497, + "ewc_loss": 4.559755325317383e-06, + "grad_norm": 2.1551389694213867, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7077014446258545, + "num_tokens": 230566495.0, + "step": 8912 + }, + { + "epoch": 0.9788051833955633, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 1.8102004528045654, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6939508318901062, + "num_tokens": 230601093.0, + "step": 8913 + }, + { + "epoch": 0.978915001098177, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 1.9181519746780396, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7094346284866333, + "num_tokens": 230635869.0, + "step": 8914 + }, + { + "epoch": 0.9790248188007907, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0289814472198486, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6943525075912476, + "num_tokens": 230663962.0, + "step": 8915 + }, + { + "epoch": 0.9791346365034044, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 1.9358279705047607, + "learning_rate": 1e-06, + "loss": 1.0332, + "mean_token_accuracy": 0.6904362440109253, + "num_tokens": 230696912.0, + "step": 8916 + }, + { + "epoch": 0.979244454206018, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0774967670440674, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7023434638977051, + "num_tokens": 230725370.0, + "step": 8917 + }, + { + "epoch": 0.9793542719086317, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.193321466445923, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7072837352752686, + "num_tokens": 230749997.0, + "step": 8918 + }, + { + "epoch": 0.9794640896112453, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.136199951171875, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7376347184181213, + "num_tokens": 230774837.0, + "step": 8919 + }, + { + "epoch": 0.979573907313859, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.274400234222412, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7235276699066162, + "num_tokens": 230799299.0, + "step": 8920 + }, + { + "epoch": 0.9796837250164726, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3219878673553467, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7046217918395996, + "num_tokens": 230822987.0, + "step": 8921 + }, + { + "epoch": 0.9797935427190864, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.505566358566284, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7067805528640747, + "num_tokens": 230843790.0, + "step": 8922 + }, + { + "epoch": 0.9799033604217, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.2456576824188232, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7213023900985718, + "num_tokens": 230866674.0, + "step": 8923 + }, + { + "epoch": 0.9800131781243137, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.1085386276245117, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6956882476806641, + "num_tokens": 230896171.0, + "step": 8924 + }, + { + "epoch": 0.9801229958269273, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 1.9429655075073242, + "learning_rate": 1e-06, + "loss": 1.013, + "mean_token_accuracy": 0.6924688816070557, + "num_tokens": 230927449.0, + "step": 8925 + }, + { + "epoch": 0.980232813529541, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3757753372192383, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7140340209007263, + "num_tokens": 230949584.0, + "step": 8926 + }, + { + "epoch": 0.9803426312321546, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3660178184509277, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7224600315093994, + "num_tokens": 230971160.0, + "step": 8927 + }, + { + "epoch": 0.9804524489347682, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.051652431488037, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.713202714920044, + "num_tokens": 230998463.0, + "step": 8928 + }, + { + "epoch": 0.980562266637382, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0157487392425537, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7302429676055908, + "num_tokens": 231027874.0, + "step": 8929 + }, + { + "epoch": 0.9806720843399956, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0941290855407715, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7048115730285645, + "num_tokens": 231054479.0, + "step": 8930 + }, + { + "epoch": 0.9807819020426093, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0094573497772217, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7011101245880127, + "num_tokens": 231084304.0, + "step": 8931 + }, + { + "epoch": 0.9808917197452229, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.2058660984039307, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7100926041603088, + "num_tokens": 231109727.0, + "step": 8932 + }, + { + "epoch": 0.9810015374478366, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.171332836151123, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.6977396011352539, + "num_tokens": 231138001.0, + "step": 8933 + }, + { + "epoch": 0.9811113551504502, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.4427318572998047, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.706139326095581, + "num_tokens": 231161019.0, + "step": 8934 + }, + { + "epoch": 0.9812211728530639, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3991928100585938, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7209808826446533, + "num_tokens": 231182904.0, + "step": 8935 + }, + { + "epoch": 0.9813309905556776, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3705551624298096, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7131320834159851, + "num_tokens": 231205834.0, + "step": 8936 + }, + { + "epoch": 0.9814408082582913, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.3375542163848877, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7176506519317627, + "num_tokens": 231229297.0, + "step": 8937 + }, + { + "epoch": 0.9815506259609049, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0869140625, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7164939045906067, + "num_tokens": 231253732.0, + "step": 8938 + }, + { + "epoch": 0.9816604436635186, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.1400108337402344, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.7008681893348694, + "num_tokens": 231281803.0, + "step": 8939 + }, + { + "epoch": 0.9817702613661322, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 1.9493153095245361, + "learning_rate": 1e-06, + "loss": 1.024, + "mean_token_accuracy": 0.6889932155609131, + "num_tokens": 231312159.0, + "step": 8940 + }, + { + "epoch": 0.9818800790687459, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.0936262607574463, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7186849117279053, + "num_tokens": 231339780.0, + "step": 8941 + }, + { + "epoch": 0.9819898967713595, + "ewc_loss": 4.589557647705078e-06, + "grad_norm": 2.1779439449310303, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7272286415100098, + "num_tokens": 231365522.0, + "step": 8942 + }, + { + "epoch": 0.9820997144739732, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.142397880554199, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6971019506454468, + "num_tokens": 231392658.0, + "step": 8943 + }, + { + "epoch": 0.9822095321765869, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9868004322052002, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6995633840560913, + "num_tokens": 231421138.0, + "step": 8944 + }, + { + "epoch": 0.9823193498792006, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.001223564147949, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.7016778588294983, + "num_tokens": 231449858.0, + "step": 8945 + }, + { + "epoch": 0.9824291675818142, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.406934976577759, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7018848657608032, + "num_tokens": 231471215.0, + "step": 8946 + }, + { + "epoch": 0.9825389852844278, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2574429512023926, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7531247138977051, + "num_tokens": 231493172.0, + "step": 8947 + }, + { + "epoch": 0.9826488029870415, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2987277507781982, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.712285578250885, + "num_tokens": 231517220.0, + "step": 8948 + }, + { + "epoch": 0.9827586206896551, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0143966674804688, + "learning_rate": 1e-06, + "loss": 1.0303, + "mean_token_accuracy": 0.6949337124824524, + "num_tokens": 231546201.0, + "step": 8949 + }, + { + "epoch": 0.9828684383922688, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2469258308410645, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7199600338935852, + "num_tokens": 231569750.0, + "step": 8950 + }, + { + "epoch": 0.9829782560948825, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.429396867752075, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7032485008239746, + "num_tokens": 231590161.0, + "step": 8951 + }, + { + "epoch": 0.9830880737974962, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9864870309829712, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7042223215103149, + "num_tokens": 231621120.0, + "step": 8952 + }, + { + "epoch": 0.9831978915001098, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9777936935424805, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7062821388244629, + "num_tokens": 231648905.0, + "step": 8953 + }, + { + "epoch": 0.9833077092027235, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2425975799560547, + "learning_rate": 1e-06, + "loss": 1.0318, + "mean_token_accuracy": 0.6885920763015747, + "num_tokens": 231676121.0, + "step": 8954 + }, + { + "epoch": 0.9834175269053371, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0727288722991943, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7129355669021606, + "num_tokens": 231703666.0, + "step": 8955 + }, + { + "epoch": 0.9835273446079508, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.248823404312134, + "learning_rate": 1e-06, + "loss": 1.0431, + "mean_token_accuracy": 0.678744912147522, + "num_tokens": 231729719.0, + "step": 8956 + }, + { + "epoch": 0.9836371623105644, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1207518577575684, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6971381306648254, + "num_tokens": 231756062.0, + "step": 8957 + }, + { + "epoch": 0.9837469800131782, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9906185865402222, + "learning_rate": 1e-06, + "loss": 1.0942, + "mean_token_accuracy": 0.6677820682525635, + "num_tokens": 231787721.0, + "step": 8958 + }, + { + "epoch": 0.9838567977157918, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.14772629737854, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6983282566070557, + "num_tokens": 231813870.0, + "step": 8959 + }, + { + "epoch": 0.9839666154184055, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0945701599121094, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7120468020439148, + "num_tokens": 231840917.0, + "step": 8960 + }, + { + "epoch": 0.9840764331210191, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.332947015762329, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7169511318206787, + "num_tokens": 231862843.0, + "step": 8961 + }, + { + "epoch": 0.9841862508236328, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0647757053375244, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7146703004837036, + "num_tokens": 231889043.0, + "step": 8962 + }, + { + "epoch": 0.9842960685262464, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0986897945404053, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.706788957118988, + "num_tokens": 231916725.0, + "step": 8963 + }, + { + "epoch": 0.98440588622886, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.011107921600342, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7202784419059753, + "num_tokens": 231946104.0, + "step": 8964 + }, + { + "epoch": 0.9845157039314738, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.425105094909668, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7051699161529541, + "num_tokens": 231967785.0, + "step": 8965 + }, + { + "epoch": 0.9846255216340875, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0372416973114014, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7042359113693237, + "num_tokens": 231994741.0, + "step": 8966 + }, + { + "epoch": 0.9847353393367011, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.433283567428589, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.6970047950744629, + "num_tokens": 232016422.0, + "step": 8967 + }, + { + "epoch": 0.9848451570393147, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2471237182617188, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7044488787651062, + "num_tokens": 232040771.0, + "step": 8968 + }, + { + "epoch": 0.9849549747419284, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0795605182647705, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7183845639228821, + "num_tokens": 232066750.0, + "step": 8969 + }, + { + "epoch": 0.985064792444542, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2261242866516113, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7094454765319824, + "num_tokens": 232090449.0, + "step": 8970 + }, + { + "epoch": 0.9851746101471557, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.893144130706787, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6909297108650208, + "num_tokens": 232122063.0, + "step": 8971 + }, + { + "epoch": 0.9852844278497693, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.064629077911377, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7151483297348022, + "num_tokens": 232149959.0, + "step": 8972 + }, + { + "epoch": 0.9853942455523831, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.4955413341522217, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7285562753677368, + "num_tokens": 232169056.0, + "step": 8973 + }, + { + "epoch": 0.9855040632549967, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2024712562561035, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7099690437316895, + "num_tokens": 232194796.0, + "step": 8974 + }, + { + "epoch": 0.9856138809576104, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.050644636154175, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7096030712127686, + "num_tokens": 232224152.0, + "step": 8975 + }, + { + "epoch": 0.985723698660224, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2412047386169434, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.6910607814788818, + "num_tokens": 232248823.0, + "step": 8976 + }, + { + "epoch": 0.9858335163628377, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9410737752914429, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.69681316614151, + "num_tokens": 232278960.0, + "step": 8977 + }, + { + "epoch": 0.9859433340654513, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.457308769226074, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7291841506958008, + "num_tokens": 232299002.0, + "step": 8978 + }, + { + "epoch": 0.986053151768065, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.058070182800293, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7289828658103943, + "num_tokens": 232323783.0, + "step": 8979 + }, + { + "epoch": 0.9861629694706787, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.25085711479187, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7297947406768799, + "num_tokens": 232345943.0, + "step": 8980 + }, + { + "epoch": 0.9862727871732924, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2197351455688477, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7008300423622131, + "num_tokens": 232370383.0, + "step": 8981 + }, + { + "epoch": 0.986382604875906, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0467162132263184, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6908237934112549, + "num_tokens": 232399776.0, + "step": 8982 + }, + { + "epoch": 0.9864924225785197, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.972644329071045, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.704252302646637, + "num_tokens": 232429891.0, + "step": 8983 + }, + { + "epoch": 0.9866022402811333, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2805943489074707, + "learning_rate": 1e-06, + "loss": 1.07, + "mean_token_accuracy": 0.678130567073822, + "num_tokens": 232456867.0, + "step": 8984 + }, + { + "epoch": 0.986712057983747, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2084290981292725, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7320483326911926, + "num_tokens": 232481328.0, + "step": 8985 + }, + { + "epoch": 0.9868218756863606, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.126342296600342, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7157787084579468, + "num_tokens": 232504646.0, + "step": 8986 + }, + { + "epoch": 0.9869316933889744, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.994119644165039, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.6997978091239929, + "num_tokens": 232535282.0, + "step": 8987 + }, + { + "epoch": 0.987041511091588, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.175809621810913, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7096988558769226, + "num_tokens": 232561180.0, + "step": 8988 + }, + { + "epoch": 0.9871513287942016, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.118072509765625, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.7081794738769531, + "num_tokens": 232588832.0, + "step": 8989 + }, + { + "epoch": 0.9872611464968153, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.064183235168457, + "learning_rate": 1e-06, + "loss": 0.9988, + "mean_token_accuracy": 0.697868287563324, + "num_tokens": 232616528.0, + "step": 8990 + }, + { + "epoch": 0.9873709641994289, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1955223083496094, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.7015761137008667, + "num_tokens": 232643620.0, + "step": 8991 + }, + { + "epoch": 0.9874807819020426, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0201079845428467, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7087830305099487, + "num_tokens": 232672758.0, + "step": 8992 + }, + { + "epoch": 0.9875905996046562, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1960678100585938, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7036780118942261, + "num_tokens": 232698818.0, + "step": 8993 + }, + { + "epoch": 0.98770041730727, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.162229537963867, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.702903687953949, + "num_tokens": 232723684.0, + "step": 8994 + }, + { + "epoch": 0.9878102350098836, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1435182094573975, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7190548181533813, + "num_tokens": 232748578.0, + "step": 8995 + }, + { + "epoch": 0.9879200527124973, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.033494472503662, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7119448184967041, + "num_tokens": 232777194.0, + "step": 8996 + }, + { + "epoch": 0.9880298704151109, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.963666319847107, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6971925497055054, + "num_tokens": 232806509.0, + "step": 8997 + }, + { + "epoch": 0.9881396881177246, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.362478494644165, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6965317130088806, + "num_tokens": 232829602.0, + "step": 8998 + }, + { + "epoch": 0.9882495058203382, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.34391450881958, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7103532552719116, + "num_tokens": 232853101.0, + "step": 8999 + }, + { + "epoch": 0.9883593235229519, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2586112022399902, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7158259153366089, + "num_tokens": 232876847.0, + "step": 9000 + }, + { + "epoch": 0.9884691412255655, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3321025371551514, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7293065786361694, + "num_tokens": 232897886.0, + "step": 9001 + }, + { + "epoch": 0.9885789589281793, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1748242378234863, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6953451037406921, + "num_tokens": 232923874.0, + "step": 9002 + }, + { + "epoch": 0.9886887766307929, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.233447551727295, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7111616134643555, + "num_tokens": 232948175.0, + "step": 9003 + }, + { + "epoch": 0.9887985943334066, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3650875091552734, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.6966220140457153, + "num_tokens": 232970566.0, + "step": 9004 + }, + { + "epoch": 0.9889084120360202, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.336362361907959, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6938530206680298, + "num_tokens": 232993237.0, + "step": 9005 + }, + { + "epoch": 0.9890182297386338, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.8836188316345215, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6927051544189453, + "num_tokens": 233027525.0, + "step": 9006 + }, + { + "epoch": 0.9891280474412475, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.923028588294983, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7034157514572144, + "num_tokens": 233061174.0, + "step": 9007 + }, + { + "epoch": 0.9892378651438611, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2209179401397705, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6971827149391174, + "num_tokens": 233087169.0, + "step": 9008 + }, + { + "epoch": 0.9893476828464749, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1624233722686768, + "learning_rate": 1e-06, + "loss": 0.979, + "mean_token_accuracy": 0.7001063823699951, + "num_tokens": 233113169.0, + "step": 9009 + }, + { + "epoch": 0.9894575005490885, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9141789674758911, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7045465707778931, + "num_tokens": 233145830.0, + "step": 9010 + }, + { + "epoch": 0.9895673182517022, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.105562686920166, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6963756084442139, + "num_tokens": 233172642.0, + "step": 9011 + }, + { + "epoch": 0.9896771359543158, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2754108905792236, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7244282960891724, + "num_tokens": 233194751.0, + "step": 9012 + }, + { + "epoch": 0.9897869536569295, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.900275468826294, + "learning_rate": 1e-06, + "loss": 1.0019, + "mean_token_accuracy": 0.7052096128463745, + "num_tokens": 233225997.0, + "step": 9013 + }, + { + "epoch": 0.9898967713595431, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0909388065338135, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7062168121337891, + "num_tokens": 233252009.0, + "step": 9014 + }, + { + "epoch": 0.9900065890621568, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0994908809661865, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7227071523666382, + "num_tokens": 233278741.0, + "step": 9015 + }, + { + "epoch": 0.9901164067647705, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3382415771484375, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7194435596466064, + "num_tokens": 233298654.0, + "step": 9016 + }, + { + "epoch": 0.9902262244673842, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9855402708053589, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7103700041770935, + "num_tokens": 233327394.0, + "step": 9017 + }, + { + "epoch": 0.9903360421699978, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1440622806549072, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7112056612968445, + "num_tokens": 233352462.0, + "step": 9018 + }, + { + "epoch": 0.9904458598726115, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3544270992279053, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7179287672042847, + "num_tokens": 233374329.0, + "step": 9019 + }, + { + "epoch": 0.9905556775752251, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.053196668624878, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7185734510421753, + "num_tokens": 233399704.0, + "step": 9020 + }, + { + "epoch": 0.9906654952778388, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3341102600097656, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7042695879936218, + "num_tokens": 233422103.0, + "step": 9021 + }, + { + "epoch": 0.9907753129804524, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3683648109436035, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7105807065963745, + "num_tokens": 233444041.0, + "step": 9022 + }, + { + "epoch": 0.9908851306830662, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0484201908111572, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7152705192565918, + "num_tokens": 233472497.0, + "step": 9023 + }, + { + "epoch": 0.9909949483856798, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.285869836807251, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7091823220252991, + "num_tokens": 233495761.0, + "step": 9024 + }, + { + "epoch": 0.9911047660882935, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.080695152282715, + "learning_rate": 1e-06, + "loss": 1.0563, + "mean_token_accuracy": 0.6796649098396301, + "num_tokens": 233526528.0, + "step": 9025 + }, + { + "epoch": 0.9912145837909071, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.225271463394165, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7129718065261841, + "num_tokens": 233549552.0, + "step": 9026 + }, + { + "epoch": 0.9913244014935207, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.15809965133667, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7079752087593079, + "num_tokens": 233575314.0, + "step": 9027 + }, + { + "epoch": 0.9914342191961344, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1282529830932617, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7034541368484497, + "num_tokens": 233600910.0, + "step": 9028 + }, + { + "epoch": 0.991544036898748, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.211667776107788, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.7020484209060669, + "num_tokens": 233624970.0, + "step": 9029 + }, + { + "epoch": 0.9916538546013618, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9501413106918335, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6994123458862305, + "num_tokens": 233656091.0, + "step": 9030 + }, + { + "epoch": 0.9917636723039754, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3028533458709717, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7139831781387329, + "num_tokens": 233678512.0, + "step": 9031 + }, + { + "epoch": 0.9918734900065891, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.120603561401367, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7102035880088806, + "num_tokens": 233704187.0, + "step": 9032 + }, + { + "epoch": 0.9919833077092027, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1093854904174805, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.703873872756958, + "num_tokens": 233731448.0, + "step": 9033 + }, + { + "epoch": 0.9920931254118164, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.257460355758667, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7135131359100342, + "num_tokens": 233755151.0, + "step": 9034 + }, + { + "epoch": 0.99220294311443, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.80265474319458, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7125855684280396, + "num_tokens": 233789917.0, + "step": 9035 + }, + { + "epoch": 0.9923127608170437, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.324831008911133, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7136960029602051, + "num_tokens": 233812965.0, + "step": 9036 + }, + { + "epoch": 0.9924225785196573, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.0586767196655273, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7069191932678223, + "num_tokens": 233842025.0, + "step": 9037 + }, + { + "epoch": 0.9925323962222711, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.024444818496704, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.719272792339325, + "num_tokens": 233870113.0, + "step": 9038 + }, + { + "epoch": 0.9926422139248847, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.548959732055664, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7096456289291382, + "num_tokens": 233897948.0, + "step": 9039 + }, + { + "epoch": 0.9927520316274984, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1937038898468018, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.695559561252594, + "num_tokens": 233922903.0, + "step": 9040 + }, + { + "epoch": 0.992861849330112, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.100804567337036, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7332857847213745, + "num_tokens": 233949903.0, + "step": 9041 + }, + { + "epoch": 0.9929716670327257, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.031773805618286, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7008127570152283, + "num_tokens": 233980714.0, + "step": 9042 + }, + { + "epoch": 0.9930814847353393, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.925166130065918, + "learning_rate": 1e-06, + "loss": 1.0548, + "mean_token_accuracy": 0.6855672597885132, + "num_tokens": 234011634.0, + "step": 9043 + }, + { + "epoch": 0.993191302437953, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3111355304718018, + "learning_rate": 1e-06, + "loss": 1.0342, + "mean_token_accuracy": 0.6882712841033936, + "num_tokens": 234035125.0, + "step": 9044 + }, + { + "epoch": 0.9933011201405667, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3756561279296875, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.6996816396713257, + "num_tokens": 234057417.0, + "step": 9045 + }, + { + "epoch": 0.9934109378431804, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.5162627696990967, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6978776454925537, + "num_tokens": 234078743.0, + "step": 9046 + }, + { + "epoch": 0.993520755545794, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1012461185455322, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7305121421813965, + "num_tokens": 234105817.0, + "step": 9047 + }, + { + "epoch": 0.9936305732484076, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.057675838470459, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6950159072875977, + "num_tokens": 234132511.0, + "step": 9048 + }, + { + "epoch": 0.9937403909510213, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1981453895568848, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7224394083023071, + "num_tokens": 234155713.0, + "step": 9049 + }, + { + "epoch": 0.9938502086536349, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9553205966949463, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7120085954666138, + "num_tokens": 234184946.0, + "step": 9050 + }, + { + "epoch": 0.9939600263562486, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.3420097827911377, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7139368057250977, + "num_tokens": 234206926.0, + "step": 9051 + }, + { + "epoch": 0.9940698440588623, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.138892889022827, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.715387225151062, + "num_tokens": 234231917.0, + "step": 9052 + }, + { + "epoch": 0.994179661761476, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1541242599487305, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6986578106880188, + "num_tokens": 234257913.0, + "step": 9053 + }, + { + "epoch": 0.9942894794640896, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2871856689453125, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7242754697799683, + "num_tokens": 234281030.0, + "step": 9054 + }, + { + "epoch": 0.9943992971667033, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.2239086627960205, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7139025330543518, + "num_tokens": 234304409.0, + "step": 9055 + }, + { + "epoch": 0.9945091148693169, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 1.9710042476654053, + "learning_rate": 1e-06, + "loss": 1.069, + "mean_token_accuracy": 0.6736428737640381, + "num_tokens": 234336413.0, + "step": 9056 + }, + { + "epoch": 0.9946189325719306, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.193354845046997, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7177259922027588, + "num_tokens": 234360815.0, + "step": 9057 + }, + { + "epoch": 0.9947287502745442, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1858625411987305, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7159241437911987, + "num_tokens": 234389038.0, + "step": 9058 + }, + { + "epoch": 0.994838567977158, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.386544942855835, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7009069323539734, + "num_tokens": 234410875.0, + "step": 9059 + }, + { + "epoch": 0.9949483856797716, + "ewc_loss": 4.6193599700927734e-06, + "grad_norm": 2.1387100219726562, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7239251136779785, + "num_tokens": 234436468.0, + "step": 9060 + }, + { + "epoch": 0.9950582033823853, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1087327003479004, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7104929685592651, + "num_tokens": 234462935.0, + "step": 9061 + }, + { + "epoch": 0.9951680210849989, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9375847578048706, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7109797596931458, + "num_tokens": 234496078.0, + "step": 9062 + }, + { + "epoch": 0.9952778387876126, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1152520179748535, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6910459399223328, + "num_tokens": 234524273.0, + "step": 9063 + }, + { + "epoch": 0.9953876564902262, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.185606002807617, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7214248776435852, + "num_tokens": 234549620.0, + "step": 9064 + }, + { + "epoch": 0.9954974741928398, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.016860008239746, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7179998159408569, + "num_tokens": 234576957.0, + "step": 9065 + }, + { + "epoch": 0.9956072918954535, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1940035820007324, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7020226716995239, + "num_tokens": 234601466.0, + "step": 9066 + }, + { + "epoch": 0.9957171095980673, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9334933757781982, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.728755533695221, + "num_tokens": 234632198.0, + "step": 9067 + }, + { + "epoch": 0.9958269273006809, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.906657099723816, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7132797241210938, + "num_tokens": 234662027.0, + "step": 9068 + }, + { + "epoch": 0.9959367450032945, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.023859977722168, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7130772471427917, + "num_tokens": 234688572.0, + "step": 9069 + }, + { + "epoch": 0.9960465627059082, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.18117356300354, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6879962682723999, + "num_tokens": 234716877.0, + "step": 9070 + }, + { + "epoch": 0.9961563804085218, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9783387184143066, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7019634246826172, + "num_tokens": 234746332.0, + "step": 9071 + }, + { + "epoch": 0.9962661981111355, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.2577157020568848, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7144275903701782, + "num_tokens": 234769472.0, + "step": 9072 + }, + { + "epoch": 0.9963760158137491, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.162630558013916, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7033096551895142, + "num_tokens": 234794215.0, + "step": 9073 + }, + { + "epoch": 0.9964858335163629, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.968593716621399, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7149755954742432, + "num_tokens": 234821761.0, + "step": 9074 + }, + { + "epoch": 0.9965956512189765, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.13724946975708, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7167947292327881, + "num_tokens": 234848284.0, + "step": 9075 + }, + { + "epoch": 0.9967054689215902, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.606489658355713, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7169510126113892, + "num_tokens": 234866900.0, + "step": 9076 + }, + { + "epoch": 0.9968152866242038, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.255645990371704, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7196451425552368, + "num_tokens": 234890696.0, + "step": 9077 + }, + { + "epoch": 0.9969251043268175, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.586033582687378, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7153980731964111, + "num_tokens": 234909194.0, + "step": 9078 + }, + { + "epoch": 0.9970349220294311, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9732770919799805, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.701555073261261, + "num_tokens": 234938499.0, + "step": 9079 + }, + { + "epoch": 0.9971447397320448, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.635935068130493, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7044057846069336, + "num_tokens": 234956871.0, + "step": 9080 + }, + { + "epoch": 0.9972545574346585, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1374313831329346, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7227028608322144, + "num_tokens": 234983142.0, + "step": 9081 + }, + { + "epoch": 0.9973643751372722, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9193055629730225, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7008556127548218, + "num_tokens": 235015256.0, + "step": 9082 + }, + { + "epoch": 0.9974741928398858, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.121603012084961, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7264635562896729, + "num_tokens": 235039088.0, + "step": 9083 + }, + { + "epoch": 0.9975840105424995, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.48392915725708, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7041766047477722, + "num_tokens": 235059458.0, + "step": 9084 + }, + { + "epoch": 0.9976938282451131, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.2872049808502197, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7216640114784241, + "num_tokens": 235080777.0, + "step": 9085 + }, + { + "epoch": 0.9978036459477267, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.0289251804351807, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7313330769538879, + "num_tokens": 235106971.0, + "step": 9086 + }, + { + "epoch": 0.9979134636503404, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.62251877784729, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.716833770275116, + "num_tokens": 235126459.0, + "step": 9087 + }, + { + "epoch": 0.9980232813529542, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9944586753845215, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7056900858879089, + "num_tokens": 235154795.0, + "step": 9088 + }, + { + "epoch": 0.9981330990555678, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.3823564052581787, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6999579668045044, + "num_tokens": 235177941.0, + "step": 9089 + }, + { + "epoch": 0.9982429167581814, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.0876121520996094, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.6997666358947754, + "num_tokens": 235205466.0, + "step": 9090 + }, + { + "epoch": 0.9983527344607951, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.310163974761963, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7047321200370789, + "num_tokens": 235228890.0, + "step": 9091 + }, + { + "epoch": 0.9984625521634087, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.37910532951355, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7220073938369751, + "num_tokens": 235250592.0, + "step": 9092 + }, + { + "epoch": 0.9985723698660224, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9862678050994873, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7196836471557617, + "num_tokens": 235278580.0, + "step": 9093 + }, + { + "epoch": 0.998682187568636, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.238049030303955, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7226150631904602, + "num_tokens": 235301987.0, + "step": 9094 + }, + { + "epoch": 0.9987920052712497, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.265698194503784, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.6985615491867065, + "num_tokens": 235328062.0, + "step": 9095 + }, + { + "epoch": 0.9989018229738634, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.335035562515259, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7293094396591187, + "num_tokens": 235350562.0, + "step": 9096 + }, + { + "epoch": 0.9990116406764771, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9829350709915161, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7083656787872314, + "num_tokens": 235379626.0, + "step": 9097 + }, + { + "epoch": 0.9991214583790907, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.2226500511169434, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7115477323532104, + "num_tokens": 235405941.0, + "step": 9098 + }, + { + "epoch": 0.9992312760817044, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1098833084106445, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7089388966560364, + "num_tokens": 235432729.0, + "step": 9099 + }, + { + "epoch": 0.999341093784318, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.9612025022506714, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.7031235694885254, + "num_tokens": 235463599.0, + "step": 9100 + }, + { + "epoch": 0.9994509114869317, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.291578769683838, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7075360417366028, + "num_tokens": 235486921.0, + "step": 9101 + }, + { + "epoch": 0.9995607291895453, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.8647937774658203, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6930796504020691, + "num_tokens": 235519812.0, + "step": 9102 + }, + { + "epoch": 0.9996705468921591, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.2457354068756104, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7280271053314209, + "num_tokens": 235543456.0, + "step": 9103 + }, + { + "epoch": 0.9997803645947727, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.0556678771972656, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6947426795959473, + "num_tokens": 235572016.0, + "step": 9104 + }, + { + "epoch": 0.9998901822973864, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.3901166915893555, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7202856540679932, + "num_tokens": 235594912.0, + "step": 9105 + }, + { + "epoch": 1.0, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.3204917907714844, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7091063857078552, + "num_tokens": 235617245.0, + "step": 9106 + }, + { + "epoch": 1.0001098177026138, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.0822677612304688, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7017713189125061, + "num_tokens": 235647042.0, + "step": 9107 + }, + { + "epoch": 1.0002196354052273, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.027217149734497, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.71837317943573, + "num_tokens": 235675620.0, + "step": 9108 + }, + { + "epoch": 1.000329453107841, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.3641397953033447, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7141329050064087, + "num_tokens": 235699190.0, + "step": 9109 + }, + { + "epoch": 1.0004392708104546, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 1.8591965436935425, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7160729169845581, + "num_tokens": 235729568.0, + "step": 9110 + }, + { + "epoch": 1.0005490885130683, + "ewc_loss": 4.649162292480469e-06, + "grad_norm": 2.1419293880462646, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7273591756820679, + "num_tokens": 235753975.0, + "step": 9111 + }, + { + "epoch": 1.0006589062156819, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.284430742263794, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7280110120773315, + "num_tokens": 235774114.0, + "step": 9112 + }, + { + "epoch": 1.0007687239182956, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.130777597427368, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6961269378662109, + "num_tokens": 235799505.0, + "step": 9113 + }, + { + "epoch": 1.0008785416209094, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.0965993404388428, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7190235257148743, + "num_tokens": 235828000.0, + "step": 9114 + }, + { + "epoch": 1.000988359323523, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.408646583557129, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7458623647689819, + "num_tokens": 235851026.0, + "step": 9115 + }, + { + "epoch": 1.0010981770261367, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.242683172225952, + "learning_rate": 1e-06, + "loss": 1.036, + "mean_token_accuracy": 0.6859542727470398, + "num_tokens": 235880806.0, + "step": 9116 + }, + { + "epoch": 1.0012079947287502, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.1425201892852783, + "learning_rate": 1e-06, + "loss": 0.9756, + "mean_token_accuracy": 0.7032886743545532, + "num_tokens": 235913442.0, + "step": 9117 + }, + { + "epoch": 1.001317812431364, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.329618453979492, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7304801940917969, + "num_tokens": 235936145.0, + "step": 9118 + }, + { + "epoch": 1.0014276301339775, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.3910553455352783, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7219155430793762, + "num_tokens": 235958222.0, + "step": 9119 + }, + { + "epoch": 1.0015374478365913, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.3223540782928467, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7292824387550354, + "num_tokens": 235980934.0, + "step": 9120 + }, + { + "epoch": 1.001647265539205, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.2615835666656494, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7091424465179443, + "num_tokens": 236005418.0, + "step": 9121 + }, + { + "epoch": 1.0017570832418186, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.188014507293701, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7243033051490784, + "num_tokens": 236031486.0, + "step": 9122 + }, + { + "epoch": 1.0018669009444323, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.2807974815368652, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.6958310604095459, + "num_tokens": 236058222.0, + "step": 9123 + }, + { + "epoch": 1.0019767186470458, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.2000293731689453, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7341091632843018, + "num_tokens": 236082865.0, + "step": 9124 + }, + { + "epoch": 1.0020865363496596, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 1.9367622137069702, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.6953802108764648, + "num_tokens": 236118019.0, + "step": 9125 + }, + { + "epoch": 1.0021963540522731, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.3178558349609375, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7182235717773438, + "num_tokens": 236140259.0, + "step": 9126 + }, + { + "epoch": 1.002306171754887, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 1.9444040060043335, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7058556079864502, + "num_tokens": 236171193.0, + "step": 9127 + }, + { + "epoch": 1.0024159894575007, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.1498396396636963, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7160129547119141, + "num_tokens": 236198553.0, + "step": 9128 + }, + { + "epoch": 1.0025258071601142, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.265519618988037, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.699080228805542, + "num_tokens": 236225340.0, + "step": 9129 + }, + { + "epoch": 1.002635624862728, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.317798614501953, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7218990325927734, + "num_tokens": 236249433.0, + "step": 9130 + }, + { + "epoch": 1.0027454425653415, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.0376014709472656, + "learning_rate": 1e-06, + "loss": 1.025, + "mean_token_accuracy": 0.6894640326499939, + "num_tokens": 236278592.0, + "step": 9131 + }, + { + "epoch": 1.0028552602679552, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.627163887023926, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7390087842941284, + "num_tokens": 236296550.0, + "step": 9132 + }, + { + "epoch": 1.0029650779705688, + "ewc_loss": 4.678964614868164e-06, + "grad_norm": 2.0384693145751953, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7142919898033142, + "num_tokens": 236325169.0, + "step": 9133 + }, + { + "epoch": 1.0030748956731825, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.005918502807617, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.6968708634376526, + "num_tokens": 236355195.0, + "step": 9134 + }, + { + "epoch": 1.0031847133757963, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 1.8907142877578735, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6971288323402405, + "num_tokens": 236392651.0, + "step": 9135 + }, + { + "epoch": 1.0032945310784098, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 1.8659946918487549, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.709259569644928, + "num_tokens": 236429123.0, + "step": 9136 + }, + { + "epoch": 1.0034043487810236, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2100143432617188, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7239875197410583, + "num_tokens": 236458505.0, + "step": 9137 + }, + { + "epoch": 1.0035141664836371, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2918293476104736, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7236887216567993, + "num_tokens": 236483365.0, + "step": 9138 + }, + { + "epoch": 1.0036239841862509, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1854407787323, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7201501131057739, + "num_tokens": 236507918.0, + "step": 9139 + }, + { + "epoch": 1.0037338018888644, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.7787935733795166, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7070689797401428, + "num_tokens": 236542471.0, + "step": 9140 + }, + { + "epoch": 1.0038436195914782, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.312045097351074, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7287136316299438, + "num_tokens": 236566275.0, + "step": 9141 + }, + { + "epoch": 1.0039534372940917, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.362200975418091, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7240270972251892, + "num_tokens": 236588724.0, + "step": 9142 + }, + { + "epoch": 1.0040632549967055, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.988318681716919, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.694158673286438, + "num_tokens": 236619468.0, + "step": 9143 + }, + { + "epoch": 1.0041730726993192, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4466938972473145, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7168118953704834, + "num_tokens": 236641105.0, + "step": 9144 + }, + { + "epoch": 1.0042828904019327, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4483065605163574, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7220698595046997, + "num_tokens": 236664417.0, + "step": 9145 + }, + { + "epoch": 1.0043927081045465, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4460256099700928, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7232296466827393, + "num_tokens": 236685394.0, + "step": 9146 + }, + { + "epoch": 1.00450252580716, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.395617723464966, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7374719381332397, + "num_tokens": 236709099.0, + "step": 9147 + }, + { + "epoch": 1.0046123435097738, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.069289207458496, + "learning_rate": 1e-06, + "loss": 0.9998, + "mean_token_accuracy": 0.703201174736023, + "num_tokens": 236736328.0, + "step": 9148 + }, + { + "epoch": 1.0047221612123873, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1503686904907227, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7242373824119568, + "num_tokens": 236762359.0, + "step": 9149 + }, + { + "epoch": 1.004831978915001, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.10097336769104, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.7033969163894653, + "num_tokens": 236788664.0, + "step": 9150 + }, + { + "epoch": 1.0049417966176148, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2479588985443115, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7271703481674194, + "num_tokens": 236811440.0, + "step": 9151 + }, + { + "epoch": 1.0050516143202284, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0272231101989746, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.7021927833557129, + "num_tokens": 236844805.0, + "step": 9152 + }, + { + "epoch": 1.0051614320228421, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2326791286468506, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7290047407150269, + "num_tokens": 236869318.0, + "step": 9153 + }, + { + "epoch": 1.0052712497254557, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.855891466140747, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7402231097221375, + "num_tokens": 236901039.0, + "step": 9154 + }, + { + "epoch": 1.0053810674280694, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1650896072387695, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.700180172920227, + "num_tokens": 236928639.0, + "step": 9155 + }, + { + "epoch": 1.005490885130683, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.2609546184539795, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6988561153411865, + "num_tokens": 236955654.0, + "step": 9156 + }, + { + "epoch": 1.0056007028332967, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.2164418697357178, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7288970947265625, + "num_tokens": 236980378.0, + "step": 9157 + }, + { + "epoch": 1.0057105205359105, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.1493077278137207, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.704393744468689, + "num_tokens": 237008385.0, + "step": 9158 + }, + { + "epoch": 1.005820338238524, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.187429428100586, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7104548811912537, + "num_tokens": 237034555.0, + "step": 9159 + }, + { + "epoch": 1.0059301559411378, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.1006290912628174, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.705819845199585, + "num_tokens": 237066324.0, + "step": 9160 + }, + { + "epoch": 1.0060399736437513, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.007289409637451, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7064023017883301, + "num_tokens": 237094989.0, + "step": 9161 + }, + { + "epoch": 1.006149791346365, + "ewc_loss": 4.708766937255859e-06, + "grad_norm": 2.28774356842041, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6960840225219727, + "num_tokens": 237120765.0, + "step": 9162 + }, + { + "epoch": 1.0062596090489786, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9943058490753174, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7089381217956543, + "num_tokens": 237151511.0, + "step": 9163 + }, + { + "epoch": 1.0063694267515924, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3402113914489746, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7116187214851379, + "num_tokens": 237176336.0, + "step": 9164 + }, + { + "epoch": 1.0064792444542061, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.114009380340576, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.692828893661499, + "num_tokens": 237207576.0, + "step": 9165 + }, + { + "epoch": 1.0065890621568196, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1482527256011963, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7121772766113281, + "num_tokens": 237235137.0, + "step": 9166 + }, + { + "epoch": 1.0066988798594334, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.459333896636963, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7244047522544861, + "num_tokens": 237255124.0, + "step": 9167 + }, + { + "epoch": 1.006808697562047, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.453317403793335, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6974594593048096, + "num_tokens": 237278364.0, + "step": 9168 + }, + { + "epoch": 1.0069185152646607, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.296593189239502, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7320707440376282, + "num_tokens": 237302500.0, + "step": 9169 + }, + { + "epoch": 1.0070283329672742, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1830222606658936, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7315881848335266, + "num_tokens": 237327908.0, + "step": 9170 + }, + { + "epoch": 1.007138150669888, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1949448585510254, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.70619797706604, + "num_tokens": 237355523.0, + "step": 9171 + }, + { + "epoch": 1.0072479683725017, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1425957679748535, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7121838331222534, + "num_tokens": 237381925.0, + "step": 9172 + }, + { + "epoch": 1.0073577860751153, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4245150089263916, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7144995331764221, + "num_tokens": 237408436.0, + "step": 9173 + }, + { + "epoch": 1.007467603777729, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.20589542388916, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7193400859832764, + "num_tokens": 237434340.0, + "step": 9174 + }, + { + "epoch": 1.0075774214803426, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.147279739379883, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7410992383956909, + "num_tokens": 237459680.0, + "step": 9175 + }, + { + "epoch": 1.0076872391829563, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9697751998901367, + "learning_rate": 1e-06, + "loss": 1.0068, + "mean_token_accuracy": 0.6928142309188843, + "num_tokens": 237492224.0, + "step": 9176 + }, + { + "epoch": 1.0077970568855699, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1537296772003174, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7458264231681824, + "num_tokens": 237516302.0, + "step": 9177 + }, + { + "epoch": 1.0079068745881836, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2556281089782715, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7139006853103638, + "num_tokens": 237540954.0, + "step": 9178 + }, + { + "epoch": 1.0080166922907974, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2510881423950195, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.714784562587738, + "num_tokens": 237565679.0, + "step": 9179 + }, + { + "epoch": 1.008126509993411, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.009937286376953, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7039319276809692, + "num_tokens": 237592971.0, + "step": 9180 + }, + { + "epoch": 1.0082363276960247, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1672446727752686, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7210676670074463, + "num_tokens": 237618973.0, + "step": 9181 + }, + { + "epoch": 1.0083461453986382, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.076434373855591, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7093320488929749, + "num_tokens": 237649304.0, + "step": 9182 + }, + { + "epoch": 1.008455963101252, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.384036064147949, + "learning_rate": 1e-06, + "loss": 1.0333, + "mean_token_accuracy": 0.6800371408462524, + "num_tokens": 237676775.0, + "step": 9183 + }, + { + "epoch": 1.0085657808038655, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.006013870239258, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7046152949333191, + "num_tokens": 237707669.0, + "step": 9184 + }, + { + "epoch": 1.0086755985064793, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.305325984954834, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7323753833770752, + "num_tokens": 237732335.0, + "step": 9185 + }, + { + "epoch": 1.008785416209093, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.421567678451538, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.716170072555542, + "num_tokens": 237755742.0, + "step": 9186 + }, + { + "epoch": 1.0088952339117065, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3280720710754395, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7184868454933167, + "num_tokens": 237779404.0, + "step": 9187 + }, + { + "epoch": 1.0090050516143203, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4661543369293213, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7274712324142456, + "num_tokens": 237801424.0, + "step": 9188 + }, + { + "epoch": 1.0091148693169338, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4127838611602783, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7145103216171265, + "num_tokens": 237824471.0, + "step": 9189 + }, + { + "epoch": 1.0092246870195476, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9901344776153564, + "learning_rate": 1e-06, + "loss": 1.0033, + "mean_token_accuracy": 0.6913188695907593, + "num_tokens": 237856973.0, + "step": 9190 + }, + { + "epoch": 1.0093345047221611, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.103337049484253, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7089799046516418, + "num_tokens": 237886112.0, + "step": 9191 + }, + { + "epoch": 1.0094443224247749, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.308619976043701, + "learning_rate": 1e-06, + "loss": 1.051, + "mean_token_accuracy": 0.6859188675880432, + "num_tokens": 237912563.0, + "step": 9192 + }, + { + "epoch": 1.0095541401273886, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.5809080600738525, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7285583019256592, + "num_tokens": 237931329.0, + "step": 9193 + }, + { + "epoch": 1.0096639578300022, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.839611530303955, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.725326657295227, + "num_tokens": 237949637.0, + "step": 9194 + }, + { + "epoch": 1.009773775532616, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.6798548698425293, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7410203218460083, + "num_tokens": 237968165.0, + "step": 9195 + }, + { + "epoch": 1.0098835932352295, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2792062759399414, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7091975212097168, + "num_tokens": 237993937.0, + "step": 9196 + }, + { + "epoch": 1.0099934109378432, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.275691270828247, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7205688953399658, + "num_tokens": 238017976.0, + "step": 9197 + }, + { + "epoch": 1.0101032286404568, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9467706680297852, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.6955273151397705, + "num_tokens": 238049863.0, + "step": 9198 + }, + { + "epoch": 1.0102130463430705, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.086958408355713, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.728598952293396, + "num_tokens": 238079151.0, + "step": 9199 + }, + { + "epoch": 1.010322864045684, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.241936206817627, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7287600636482239, + "num_tokens": 238103439.0, + "step": 9200 + }, + { + "epoch": 1.0104326817482978, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.144606590270996, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7052611112594604, + "num_tokens": 238132132.0, + "step": 9201 + }, + { + "epoch": 1.0105424994509116, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.5361812114715576, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7328976988792419, + "num_tokens": 238152749.0, + "step": 9202 + }, + { + "epoch": 1.010652317153525, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0269970893859863, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7289965152740479, + "num_tokens": 238180616.0, + "step": 9203 + }, + { + "epoch": 1.0107621348561389, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3783297538757324, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7110607028007507, + "num_tokens": 238205218.0, + "step": 9204 + }, + { + "epoch": 1.0108719525587524, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.357795000076294, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7191548347473145, + "num_tokens": 238228805.0, + "step": 9205 + }, + { + "epoch": 1.0109817702613662, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.069594144821167, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7042748332023621, + "num_tokens": 238258713.0, + "step": 9206 + }, + { + "epoch": 1.0110915879639797, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.446589231491089, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.755280613899231, + "num_tokens": 238277651.0, + "step": 9207 + }, + { + "epoch": 1.0112014056665934, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.109675407409668, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.6995285153388977, + "num_tokens": 238306098.0, + "step": 9208 + }, + { + "epoch": 1.0113112233692072, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3998043537139893, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7336175441741943, + "num_tokens": 238327071.0, + "step": 9209 + }, + { + "epoch": 1.0114210410718207, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.189897298812866, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7055178284645081, + "num_tokens": 238353322.0, + "step": 9210 + }, + { + "epoch": 1.0115308587744345, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.048382520675659, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7233667969703674, + "num_tokens": 238382018.0, + "step": 9211 + }, + { + "epoch": 1.011640676477048, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.479888916015625, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7341328859329224, + "num_tokens": 238401017.0, + "step": 9212 + }, + { + "epoch": 1.0117504941796618, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2049758434295654, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7194622159004211, + "num_tokens": 238428954.0, + "step": 9213 + }, + { + "epoch": 1.0118603118822753, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.128711462020874, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7358994483947754, + "num_tokens": 238455499.0, + "step": 9214 + }, + { + "epoch": 1.011970129584889, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.15789532661438, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7169332504272461, + "num_tokens": 238481219.0, + "step": 9215 + }, + { + "epoch": 1.0120799472875028, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1495256423950195, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7371984720230103, + "num_tokens": 238506894.0, + "step": 9216 + }, + { + "epoch": 1.0121897649901164, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.243718385696411, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7235623598098755, + "num_tokens": 238534255.0, + "step": 9217 + }, + { + "epoch": 1.0122995826927301, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.5039267539978027, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7234640121459961, + "num_tokens": 238555269.0, + "step": 9218 + }, + { + "epoch": 1.0124094003953437, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9243770837783813, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7071892023086548, + "num_tokens": 238588256.0, + "step": 9219 + }, + { + "epoch": 1.0125192180979574, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.8657118082046509, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7061868906021118, + "num_tokens": 238622800.0, + "step": 9220 + }, + { + "epoch": 1.012629035800571, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.40278697013855, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7267502546310425, + "num_tokens": 238645351.0, + "step": 9221 + }, + { + "epoch": 1.0127388535031847, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3557512760162354, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7056534290313721, + "num_tokens": 238668786.0, + "step": 9222 + }, + { + "epoch": 1.0128486712057985, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1449973583221436, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.727990984916687, + "num_tokens": 238694597.0, + "step": 9223 + }, + { + "epoch": 1.012958488908412, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2773361206054688, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7295703887939453, + "num_tokens": 238719112.0, + "step": 9224 + }, + { + "epoch": 1.0130683066110258, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.8520230054855347, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7224059700965881, + "num_tokens": 238751808.0, + "step": 9225 + }, + { + "epoch": 1.0131781243136393, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.305187463760376, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7248355150222778, + "num_tokens": 238776031.0, + "step": 9226 + }, + { + "epoch": 1.013287942016253, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4973723888397217, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.716206431388855, + "num_tokens": 238796867.0, + "step": 9227 + }, + { + "epoch": 1.0133977597188666, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.29410719871521, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7099908590316772, + "num_tokens": 238822315.0, + "step": 9228 + }, + { + "epoch": 1.0135075774214803, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2397334575653076, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.710110068321228, + "num_tokens": 238847142.0, + "step": 9229 + }, + { + "epoch": 1.013617395124094, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0749001502990723, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7066838145256042, + "num_tokens": 238877845.0, + "step": 9230 + }, + { + "epoch": 1.0137272128267076, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1703827381134033, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.7013064026832581, + "num_tokens": 238904482.0, + "step": 9231 + }, + { + "epoch": 1.0138370305293214, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.329190731048584, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7042151689529419, + "num_tokens": 238928874.0, + "step": 9232 + }, + { + "epoch": 1.013946848231935, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3623790740966797, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7080134153366089, + "num_tokens": 238955467.0, + "step": 9233 + }, + { + "epoch": 1.0140566659345487, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4220733642578125, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7376853823661804, + "num_tokens": 238977613.0, + "step": 9234 + }, + { + "epoch": 1.0141664836371622, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.401428461074829, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7052304744720459, + "num_tokens": 239002498.0, + "step": 9235 + }, + { + "epoch": 1.014276301339776, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.9792675971984863, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.703622579574585, + "num_tokens": 239038191.0, + "step": 9236 + }, + { + "epoch": 1.0143861190423897, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.053532123565674, + "learning_rate": 1e-06, + "loss": 1.0167, + "mean_token_accuracy": 0.6892762184143066, + "num_tokens": 239067476.0, + "step": 9237 + }, + { + "epoch": 1.0144959367450033, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.238431215286255, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7195464968681335, + "num_tokens": 239094249.0, + "step": 9238 + }, + { + "epoch": 1.014605754447617, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2646284103393555, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7163625359535217, + "num_tokens": 239119765.0, + "step": 9239 + }, + { + "epoch": 1.0147155721502306, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.049136161804199, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7155287265777588, + "num_tokens": 239147445.0, + "step": 9240 + }, + { + "epoch": 1.0148253898528443, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3509953022003174, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7282845973968506, + "num_tokens": 239169710.0, + "step": 9241 + }, + { + "epoch": 1.0149352075554579, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4130194187164307, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7314884662628174, + "num_tokens": 239191689.0, + "step": 9242 + }, + { + "epoch": 1.0150450252580716, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1411499977111816, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.6995038390159607, + "num_tokens": 239220203.0, + "step": 9243 + }, + { + "epoch": 1.0151548429606854, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2329938411712646, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7403236031532288, + "num_tokens": 239247003.0, + "step": 9244 + }, + { + "epoch": 1.015264660663299, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4790453910827637, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7355478405952454, + "num_tokens": 239266673.0, + "step": 9245 + }, + { + "epoch": 1.0153744783659127, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2716472148895264, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.709865391254425, + "num_tokens": 239291722.0, + "step": 9246 + }, + { + "epoch": 1.0154842960685262, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2877049446105957, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7229560613632202, + "num_tokens": 239315226.0, + "step": 9247 + }, + { + "epoch": 1.01559411377114, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.261176109313965, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7284979224205017, + "num_tokens": 239341129.0, + "step": 9248 + }, + { + "epoch": 1.0157039314737535, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1300792694091797, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7156804800033569, + "num_tokens": 239366717.0, + "step": 9249 + }, + { + "epoch": 1.0158137491763672, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.221544027328491, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.714150607585907, + "num_tokens": 239392989.0, + "step": 9250 + }, + { + "epoch": 1.015923566878981, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.112185478210449, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.704227864742279, + "num_tokens": 239424703.0, + "step": 9251 + }, + { + "epoch": 1.0160333845815945, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1990275382995605, + "learning_rate": 1e-06, + "loss": 0.9927, + "mean_token_accuracy": 0.6933099031448364, + "num_tokens": 239454364.0, + "step": 9252 + }, + { + "epoch": 1.0161432022842083, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2280988693237305, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6953269243240356, + "num_tokens": 239482975.0, + "step": 9253 + }, + { + "epoch": 1.0162530199868218, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.150796890258789, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6939646005630493, + "num_tokens": 239515619.0, + "step": 9254 + }, + { + "epoch": 1.0163628376894356, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.439966917037964, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7076493501663208, + "num_tokens": 239537698.0, + "step": 9255 + }, + { + "epoch": 1.0164726553920491, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.3220877647399902, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7224098443984985, + "num_tokens": 239562114.0, + "step": 9256 + }, + { + "epoch": 1.0165824730946629, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0438334941864014, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7002032995223999, + "num_tokens": 239590906.0, + "step": 9257 + }, + { + "epoch": 1.0166922907972764, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1896326541900635, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7354897856712341, + "num_tokens": 239615166.0, + "step": 9258 + }, + { + "epoch": 1.0168021084998902, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1162335872650146, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7115275263786316, + "num_tokens": 239641273.0, + "step": 9259 + }, + { + "epoch": 1.016911926202504, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0951383113861084, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7105991840362549, + "num_tokens": 239669932.0, + "step": 9260 + }, + { + "epoch": 1.0170217439051175, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.391089677810669, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7235612273216248, + "num_tokens": 239691431.0, + "step": 9261 + }, + { + "epoch": 1.0171315616077312, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.549609422683716, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7111049890518188, + "num_tokens": 239714438.0, + "step": 9262 + }, + { + "epoch": 1.0172413793103448, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0311405658721924, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7202749848365784, + "num_tokens": 239741908.0, + "step": 9263 + }, + { + "epoch": 1.0173511970129585, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.119534492492676, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7129465937614441, + "num_tokens": 239772040.0, + "step": 9264 + }, + { + "epoch": 1.017461014715572, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.117213726043701, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7345575094223022, + "num_tokens": 239798973.0, + "step": 9265 + }, + { + "epoch": 1.0175708324181858, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.165876626968384, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7282357215881348, + "num_tokens": 239826141.0, + "step": 9266 + }, + { + "epoch": 1.0176806501207996, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.208303928375244, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.7041447162628174, + "num_tokens": 239852736.0, + "step": 9267 + }, + { + "epoch": 1.017790467823413, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.5419836044311523, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7383343577384949, + "num_tokens": 239872701.0, + "step": 9268 + }, + { + "epoch": 1.0179002855260268, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0427985191345215, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7026473879814148, + "num_tokens": 239904145.0, + "step": 9269 + }, + { + "epoch": 1.0180101032286404, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2013375759124756, + "learning_rate": 1e-06, + "loss": 0.777, + "mean_token_accuracy": 0.7485215663909912, + "num_tokens": 239928765.0, + "step": 9270 + }, + { + "epoch": 1.0181199209312541, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2564094066619873, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6969282031059265, + "num_tokens": 239953182.0, + "step": 9271 + }, + { + "epoch": 1.0182297386338677, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4308788776397705, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7205914258956909, + "num_tokens": 239975609.0, + "step": 9272 + }, + { + "epoch": 1.0183395563364814, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.96053946018219, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7236688733100891, + "num_tokens": 240006532.0, + "step": 9273 + }, + { + "epoch": 1.0184493740390952, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.8653900623321533, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7300967574119568, + "num_tokens": 240037463.0, + "step": 9274 + }, + { + "epoch": 1.0185591917417087, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.060434103012085, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6992877721786499, + "num_tokens": 240068775.0, + "step": 9275 + }, + { + "epoch": 1.0186690094443225, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0483877658843994, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7054333686828613, + "num_tokens": 240098457.0, + "step": 9276 + }, + { + "epoch": 1.018778827146936, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.2121431827545166, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7046890258789062, + "num_tokens": 240126114.0, + "step": 9277 + }, + { + "epoch": 1.0188886448495498, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.335533618927002, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7257158756256104, + "num_tokens": 240149000.0, + "step": 9278 + }, + { + "epoch": 1.0189984625521633, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.8515479564666748, + "learning_rate": 1e-06, + "loss": 0.9993, + "mean_token_accuracy": 0.691062331199646, + "num_tokens": 240184417.0, + "step": 9279 + }, + { + "epoch": 1.019108280254777, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0976622104644775, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7110868692398071, + "num_tokens": 240213251.0, + "step": 9280 + }, + { + "epoch": 1.0192180979573908, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.221787214279175, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6912454962730408, + "num_tokens": 240243902.0, + "step": 9281 + }, + { + "epoch": 1.0193279156600044, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.120328187942505, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7179653644561768, + "num_tokens": 240271686.0, + "step": 9282 + }, + { + "epoch": 1.0194377333626181, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.249109983444214, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7006962299346924, + "num_tokens": 240298105.0, + "step": 9283 + }, + { + "epoch": 1.0195475510652316, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1530416011810303, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7216796278953552, + "num_tokens": 240323329.0, + "step": 9284 + }, + { + "epoch": 1.0196573687678454, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.230893611907959, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.71422278881073, + "num_tokens": 240348312.0, + "step": 9285 + }, + { + "epoch": 1.019767186470459, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.119802713394165, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.712939977645874, + "num_tokens": 240373960.0, + "step": 9286 + }, + { + "epoch": 1.0198770041730727, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.464674711227417, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7242766618728638, + "num_tokens": 240395472.0, + "step": 9287 + }, + { + "epoch": 1.0199868218756865, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.980735421180725, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7010858058929443, + "num_tokens": 240427402.0, + "step": 9288 + }, + { + "epoch": 1.0200966395783, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.9082956314086914, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7259413599967957, + "num_tokens": 240461485.0, + "step": 9289 + }, + { + "epoch": 1.0202064572809137, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.257046937942505, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7214512228965759, + "num_tokens": 240488154.0, + "step": 9290 + }, + { + "epoch": 1.0203162749835273, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.310396432876587, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7173365354537964, + "num_tokens": 240515041.0, + "step": 9291 + }, + { + "epoch": 1.020426092686141, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.0421290397644043, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7225572466850281, + "num_tokens": 240543212.0, + "step": 9292 + }, + { + "epoch": 1.0205359103887546, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 1.911548137664795, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7259029150009155, + "num_tokens": 240577429.0, + "step": 9293 + }, + { + "epoch": 1.0206457280913683, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1146934032440186, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7147272229194641, + "num_tokens": 240606022.0, + "step": 9294 + }, + { + "epoch": 1.020755545793982, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.488877058029175, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7119085192680359, + "num_tokens": 240627360.0, + "step": 9295 + }, + { + "epoch": 1.0208653634965956, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.084470748901367, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7081038355827332, + "num_tokens": 240655953.0, + "step": 9296 + }, + { + "epoch": 1.0209751811992094, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1391854286193848, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7186688184738159, + "num_tokens": 240681901.0, + "step": 9297 + }, + { + "epoch": 1.021084998901823, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1039347648620605, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7129982113838196, + "num_tokens": 240708896.0, + "step": 9298 + }, + { + "epoch": 1.0211948166044367, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1757137775421143, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.729888379573822, + "num_tokens": 240734080.0, + "step": 9299 + }, + { + "epoch": 1.0213046343070502, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.4517149925231934, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7074577808380127, + "num_tokens": 240757776.0, + "step": 9300 + }, + { + "epoch": 1.021414452009664, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.488935947418213, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.723786473274231, + "num_tokens": 240779967.0, + "step": 9301 + }, + { + "epoch": 1.0215242697122777, + "ewc_loss": 4.738569259643555e-06, + "grad_norm": 2.1249747276306152, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7087214589118958, + "num_tokens": 240811222.0, + "step": 9302 + }, + { + "epoch": 1.0216340874148913, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.153675079345703, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7173774242401123, + "num_tokens": 240839514.0, + "step": 9303 + }, + { + "epoch": 1.021743905117505, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0684115886688232, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.6995404958724976, + "num_tokens": 240869492.0, + "step": 9304 + }, + { + "epoch": 1.0218537228201185, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.9739062786102295, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6985105276107788, + "num_tokens": 240903330.0, + "step": 9305 + }, + { + "epoch": 1.0219635405227323, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0939223766326904, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7504900693893433, + "num_tokens": 240931670.0, + "step": 9306 + }, + { + "epoch": 1.0220733582253458, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2463784217834473, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.731433629989624, + "num_tokens": 240955765.0, + "step": 9307 + }, + { + "epoch": 1.0221831759279596, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0944364070892334, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7183701992034912, + "num_tokens": 240984698.0, + "step": 9308 + }, + { + "epoch": 1.0222929936305734, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.275792360305786, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7142248749732971, + "num_tokens": 241011486.0, + "step": 9309 + }, + { + "epoch": 1.022402811333187, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2228446006774902, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7063942551612854, + "num_tokens": 241038036.0, + "step": 9310 + }, + { + "epoch": 1.0225126290358006, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3522303104400635, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7199414968490601, + "num_tokens": 241062846.0, + "step": 9311 + }, + { + "epoch": 1.0226224467384142, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2444822788238525, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7267216444015503, + "num_tokens": 241089714.0, + "step": 9312 + }, + { + "epoch": 1.022732264441028, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0238656997680664, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6999306678771973, + "num_tokens": 241122272.0, + "step": 9313 + }, + { + "epoch": 1.0228420821436415, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.532381534576416, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7240533828735352, + "num_tokens": 241143105.0, + "step": 9314 + }, + { + "epoch": 1.0229518998462552, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.161369800567627, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7053333520889282, + "num_tokens": 241170861.0, + "step": 9315 + }, + { + "epoch": 1.023061717548869, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3823254108428955, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7257708311080933, + "num_tokens": 241193185.0, + "step": 9316 + }, + { + "epoch": 1.0231715352514825, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0036966800689697, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.722278356552124, + "num_tokens": 241222262.0, + "step": 9317 + }, + { + "epoch": 1.0232813529540963, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4045910835266113, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7111834287643433, + "num_tokens": 241245067.0, + "step": 9318 + }, + { + "epoch": 1.0233911706567098, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.177537202835083, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7137795090675354, + "num_tokens": 241272467.0, + "step": 9319 + }, + { + "epoch": 1.0235009883593236, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4412879943847656, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7136054039001465, + "num_tokens": 241297534.0, + "step": 9320 + }, + { + "epoch": 1.023610806061937, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0901896953582764, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7126727104187012, + "num_tokens": 241327611.0, + "step": 9321 + }, + { + "epoch": 1.0237206237645509, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.132951021194458, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7121481895446777, + "num_tokens": 241356755.0, + "step": 9322 + }, + { + "epoch": 1.0238304414671644, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4382948875427246, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7183522582054138, + "num_tokens": 241378369.0, + "step": 9323 + }, + { + "epoch": 1.0239402591697782, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2401130199432373, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7171802520751953, + "num_tokens": 241404176.0, + "step": 9324 + }, + { + "epoch": 1.024050076872392, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2848310470581055, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.723299503326416, + "num_tokens": 241429475.0, + "step": 9325 + }, + { + "epoch": 1.0241598945750054, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.506502866744995, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7166715860366821, + "num_tokens": 241450642.0, + "step": 9326 + }, + { + "epoch": 1.0242697122776192, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.993574857711792, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7342345714569092, + "num_tokens": 241480796.0, + "step": 9327 + }, + { + "epoch": 1.0243795299802327, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 3.0304651260375977, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7326290011405945, + "num_tokens": 241498114.0, + "step": 9328 + }, + { + "epoch": 1.0244893476828465, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.5921614170074463, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7307418584823608, + "num_tokens": 241519122.0, + "step": 9329 + }, + { + "epoch": 1.02459916538546, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0208306312561035, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7072430849075317, + "num_tokens": 241550097.0, + "step": 9330 + }, + { + "epoch": 1.0247089830880738, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.144956350326538, + "learning_rate": 1e-06, + "loss": 0.9929, + "mean_token_accuracy": 0.6955543756484985, + "num_tokens": 241578237.0, + "step": 9331 + }, + { + "epoch": 1.0248188007906875, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.362159013748169, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7386282086372375, + "num_tokens": 241599538.0, + "step": 9332 + }, + { + "epoch": 1.024928618493301, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3212027549743652, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7246254682540894, + "num_tokens": 241623433.0, + "step": 9333 + }, + { + "epoch": 1.0250384361959148, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2640421390533447, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7220524549484253, + "num_tokens": 241647005.0, + "step": 9334 + }, + { + "epoch": 1.0251482538985284, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.0841784477233887, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.706131100654602, + "num_tokens": 241676177.0, + "step": 9335 + }, + { + "epoch": 1.0252580716011421, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.134352684020996, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7266557812690735, + "num_tokens": 241705183.0, + "step": 9336 + }, + { + "epoch": 1.0253678893037557, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.528787851333618, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.707403302192688, + "num_tokens": 241728831.0, + "step": 9337 + }, + { + "epoch": 1.0254777070063694, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.9820749759674072, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7315399646759033, + "num_tokens": 241757350.0, + "step": 9338 + }, + { + "epoch": 1.0255875247089832, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.9576400518417358, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7089701294898987, + "num_tokens": 241792208.0, + "step": 9339 + }, + { + "epoch": 1.0256973424115967, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1222541332244873, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7225174307823181, + "num_tokens": 241820190.0, + "step": 9340 + }, + { + "epoch": 1.0258071601142105, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.290964365005493, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.721933126449585, + "num_tokens": 241844640.0, + "step": 9341 + }, + { + "epoch": 1.025916977816824, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4616751670837402, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7336701154708862, + "num_tokens": 241866014.0, + "step": 9342 + }, + { + "epoch": 1.0260267955194378, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.070610284805298, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.6945260167121887, + "num_tokens": 241896370.0, + "step": 9343 + }, + { + "epoch": 1.0261366132220513, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2044014930725098, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7076466083526611, + "num_tokens": 241921445.0, + "step": 9344 + }, + { + "epoch": 1.026246430924665, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.4460206031799316, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7268862724304199, + "num_tokens": 241942105.0, + "step": 9345 + }, + { + "epoch": 1.0263562486272788, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.035262107849121, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7001909613609314, + "num_tokens": 241974155.0, + "step": 9346 + }, + { + "epoch": 1.0264660663298923, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.176114082336426, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7144172191619873, + "num_tokens": 242000788.0, + "step": 9347 + }, + { + "epoch": 1.026575884032506, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.099630832672119, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7263270616531372, + "num_tokens": 242028371.0, + "step": 9348 + }, + { + "epoch": 1.0266857017351196, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1098837852478027, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7271162867546082, + "num_tokens": 242056593.0, + "step": 9349 + }, + { + "epoch": 1.0267955194377334, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.148283004760742, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7396636605262756, + "num_tokens": 242083641.0, + "step": 9350 + }, + { + "epoch": 1.026905337140347, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1301589012145996, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.731705904006958, + "num_tokens": 242109187.0, + "step": 9351 + }, + { + "epoch": 1.0270151548429607, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3669283390045166, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7171087861061096, + "num_tokens": 242131308.0, + "step": 9352 + }, + { + "epoch": 1.0271249725455744, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2797465324401855, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7172614336013794, + "num_tokens": 242158139.0, + "step": 9353 + }, + { + "epoch": 1.027234790248188, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1686713695526123, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7079528570175171, + "num_tokens": 242185517.0, + "step": 9354 + }, + { + "epoch": 1.0273446079508017, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.188347578048706, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7154529094696045, + "num_tokens": 242212944.0, + "step": 9355 + }, + { + "epoch": 1.0274544256534153, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.278045415878296, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.717041552066803, + "num_tokens": 242237921.0, + "step": 9356 + }, + { + "epoch": 1.027564243356029, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.2275259494781494, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7080724239349365, + "num_tokens": 242264084.0, + "step": 9357 + }, + { + "epoch": 1.0276740610586426, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.214487314224243, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6973143815994263, + "num_tokens": 242289515.0, + "step": 9358 + }, + { + "epoch": 1.0277838787612563, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3622777462005615, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7318227291107178, + "num_tokens": 242312676.0, + "step": 9359 + }, + { + "epoch": 1.02789369646387, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1256494522094727, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7098347544670105, + "num_tokens": 242341108.0, + "step": 9360 + }, + { + "epoch": 1.0280035141664836, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3189234733581543, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7167742848396301, + "num_tokens": 242364261.0, + "step": 9361 + }, + { + "epoch": 1.0281133318690974, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.52976655960083, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7206935286521912, + "num_tokens": 242385200.0, + "step": 9362 + }, + { + "epoch": 1.028223149571711, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 1.8977758884429932, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.70035719871521, + "num_tokens": 242421340.0, + "step": 9363 + }, + { + "epoch": 1.0283329672743247, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1170737743377686, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.704802393913269, + "num_tokens": 242449861.0, + "step": 9364 + }, + { + "epoch": 1.0284427849769382, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.174997568130493, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7357931733131409, + "num_tokens": 242476627.0, + "step": 9365 + }, + { + "epoch": 1.028552602679552, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.3349568843841553, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7085407376289368, + "num_tokens": 242502244.0, + "step": 9366 + }, + { + "epoch": 1.0286624203821657, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.282627582550049, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7165333032608032, + "num_tokens": 242528568.0, + "step": 9367 + }, + { + "epoch": 1.0287722380847792, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.1586296558380127, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7319668531417847, + "num_tokens": 242554618.0, + "step": 9368 + }, + { + "epoch": 1.028882055787393, + "ewc_loss": 4.76837158203125e-06, + "grad_norm": 2.258349657058716, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7157182693481445, + "num_tokens": 242576525.0, + "step": 9369 + }, + { + "epoch": 1.0289918734900065, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.083249568939209, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.6979710459709167, + "num_tokens": 242604868.0, + "step": 9370 + }, + { + "epoch": 1.0291016911926203, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.3013546466827393, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7274432182312012, + "num_tokens": 242627934.0, + "step": 9371 + }, + { + "epoch": 1.0292115088952338, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.4843175411224365, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7288000583648682, + "num_tokens": 242648332.0, + "step": 9372 + }, + { + "epoch": 1.0293213265978476, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1773698329925537, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7246091961860657, + "num_tokens": 242675591.0, + "step": 9373 + }, + { + "epoch": 1.0294311443004613, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.307983160018921, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7120248079299927, + "num_tokens": 242699493.0, + "step": 9374 + }, + { + "epoch": 1.0295409620030749, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.120209217071533, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6838412880897522, + "num_tokens": 242726766.0, + "step": 9375 + }, + { + "epoch": 1.0296507797056886, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 1.9076032638549805, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6938731670379639, + "num_tokens": 242761048.0, + "step": 9376 + }, + { + "epoch": 1.0297605974083022, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.4979331493377686, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.7021270394325256, + "num_tokens": 242785961.0, + "step": 9377 + }, + { + "epoch": 1.029870415110916, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.161470651626587, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7103997468948364, + "num_tokens": 242813003.0, + "step": 9378 + }, + { + "epoch": 1.0299802328135295, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.0846352577209473, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6958744525909424, + "num_tokens": 242844550.0, + "step": 9379 + }, + { + "epoch": 1.0300900505161432, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.532672166824341, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7115226984024048, + "num_tokens": 242868020.0, + "step": 9380 + }, + { + "epoch": 1.030199868218757, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.442817211151123, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7073554992675781, + "num_tokens": 242890243.0, + "step": 9381 + }, + { + "epoch": 1.0303096859213705, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.2021260261535645, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7203717231750488, + "num_tokens": 242915586.0, + "step": 9382 + }, + { + "epoch": 1.0304195036239843, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.118433713912964, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7090795040130615, + "num_tokens": 242943042.0, + "step": 9383 + }, + { + "epoch": 1.0305293213265978, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.191016674041748, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7225965261459351, + "num_tokens": 242968983.0, + "step": 9384 + }, + { + "epoch": 1.0306391390292116, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.188369035720825, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7187588810920715, + "num_tokens": 242999537.0, + "step": 9385 + }, + { + "epoch": 1.030748956731825, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.077869176864624, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7086445093154907, + "num_tokens": 243029698.0, + "step": 9386 + }, + { + "epoch": 1.0308587744344389, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.199991464614868, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7222155332565308, + "num_tokens": 243057058.0, + "step": 9387 + }, + { + "epoch": 1.0309685921370524, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1114120483398438, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7241925001144409, + "num_tokens": 243083376.0, + "step": 9388 + }, + { + "epoch": 1.0310784098396661, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.3000240325927734, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7167038917541504, + "num_tokens": 243106844.0, + "step": 9389 + }, + { + "epoch": 1.03118822754228, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 1.9371262788772583, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7120563983917236, + "num_tokens": 243139308.0, + "step": 9390 + }, + { + "epoch": 1.0312980452448934, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1534340381622314, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7083512544631958, + "num_tokens": 243165509.0, + "step": 9391 + }, + { + "epoch": 1.0314078629475072, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.3109617233276367, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7289113998413086, + "num_tokens": 243187638.0, + "step": 9392 + }, + { + "epoch": 1.0315176806501207, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.5533108711242676, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7333875298500061, + "num_tokens": 243207999.0, + "step": 9393 + }, + { + "epoch": 1.0316274983527345, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1837308406829834, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7067099213600159, + "num_tokens": 243236555.0, + "step": 9394 + }, + { + "epoch": 1.031737316055348, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1335365772247314, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7361115217208862, + "num_tokens": 243262821.0, + "step": 9395 + }, + { + "epoch": 1.0318471337579618, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.1073572635650635, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.710686206817627, + "num_tokens": 243292156.0, + "step": 9396 + }, + { + "epoch": 1.0319569514605755, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.2905776500701904, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7144215106964111, + "num_tokens": 243317821.0, + "step": 9397 + }, + { + "epoch": 1.032066769163189, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.212012767791748, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7043567895889282, + "num_tokens": 243345859.0, + "step": 9398 + }, + { + "epoch": 1.0321765868658028, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.2308220863342285, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7055426836013794, + "num_tokens": 243372271.0, + "step": 9399 + }, + { + "epoch": 1.0322864045684164, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.263059616088867, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.696567714214325, + "num_tokens": 243399628.0, + "step": 9400 + }, + { + "epoch": 1.0323962222710301, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.66262149810791, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7380462884902954, + "num_tokens": 243419002.0, + "step": 9401 + }, + { + "epoch": 1.0325060399736437, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 2.0741827487945557, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7095706462860107, + "num_tokens": 243449824.0, + "step": 9402 + }, + { + "epoch": 1.0326158576762574, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.3307571411132812, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7151594161987305, + "num_tokens": 243474777.0, + "step": 9403 + }, + { + "epoch": 1.0327256753788712, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 1.9856258630752563, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.708294689655304, + "num_tokens": 243508762.0, + "step": 9404 + }, + { + "epoch": 1.0328354930814847, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.667314291000366, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7366046905517578, + "num_tokens": 243528461.0, + "step": 9405 + }, + { + "epoch": 1.0329453107840985, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.2485835552215576, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.714794933795929, + "num_tokens": 243556317.0, + "step": 9406 + }, + { + "epoch": 1.033055128486712, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.138705015182495, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7189684510231018, + "num_tokens": 243585250.0, + "step": 9407 + }, + { + "epoch": 1.0331649461893258, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.4875452518463135, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7366028428077698, + "num_tokens": 243606557.0, + "step": 9408 + }, + { + "epoch": 1.0332747638919393, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.140399932861328, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7110961079597473, + "num_tokens": 243632472.0, + "step": 9409 + }, + { + "epoch": 1.033384581594553, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.095534324645996, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.728361964225769, + "num_tokens": 243660567.0, + "step": 9410 + }, + { + "epoch": 1.0334943992971668, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.5307960510253906, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7232784032821655, + "num_tokens": 243681101.0, + "step": 9411 + }, + { + "epoch": 1.0336042169997803, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.129757881164551, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7103404998779297, + "num_tokens": 243710120.0, + "step": 9412 + }, + { + "epoch": 1.033714034702394, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.493429660797119, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7353991866111755, + "num_tokens": 243731564.0, + "step": 9413 + }, + { + "epoch": 1.0338238524050076, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.5371532440185547, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7507427334785461, + "num_tokens": 243752083.0, + "step": 9414 + }, + { + "epoch": 1.0339336701076214, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.3512282371520996, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7534806132316589, + "num_tokens": 243776098.0, + "step": 9415 + }, + { + "epoch": 1.034043487810235, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.414567470550537, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.736667811870575, + "num_tokens": 243798831.0, + "step": 9416 + }, + { + "epoch": 1.0341533055128487, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.0966668128967285, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7106549739837646, + "num_tokens": 243826736.0, + "step": 9417 + }, + { + "epoch": 1.0342631232154624, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.3483407497406006, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7248783707618713, + "num_tokens": 243850406.0, + "step": 9418 + }, + { + "epoch": 1.034372940918076, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.708526372909546, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7473056316375732, + "num_tokens": 243866665.0, + "step": 9419 + }, + { + "epoch": 1.0344827586206897, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.2892165184020996, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7246020436286926, + "num_tokens": 243891260.0, + "step": 9420 + }, + { + "epoch": 1.0345925763233033, + "ewc_loss": 4.798173904418945e-06, + "grad_norm": 2.4128708839416504, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7248079180717468, + "num_tokens": 243912863.0, + "step": 9421 + }, + { + "epoch": 1.034702394025917, + "ewc_loss": 4.827976226806641e-06, + "grad_norm": 1.9735782146453857, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7219071984291077, + "num_tokens": 243941500.0, + "step": 9422 + }, + { + "epoch": 1.0348122117285306, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.3627138137817383, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7024829983711243, + "num_tokens": 243965604.0, + "step": 9423 + }, + { + "epoch": 1.0349220294311443, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2770917415618896, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7274004220962524, + "num_tokens": 243989569.0, + "step": 9424 + }, + { + "epoch": 1.035031847133758, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0883214473724365, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7295162677764893, + "num_tokens": 244018389.0, + "step": 9425 + }, + { + "epoch": 1.0351416648363716, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2917492389678955, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7187902927398682, + "num_tokens": 244042053.0, + "step": 9426 + }, + { + "epoch": 1.0352514825389854, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.045900583267212, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7162622213363647, + "num_tokens": 244072302.0, + "step": 9427 + }, + { + "epoch": 1.035361300241599, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.435269355773926, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7073421478271484, + "num_tokens": 244095758.0, + "step": 9428 + }, + { + "epoch": 1.0354711179442126, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0639708042144775, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7209656238555908, + "num_tokens": 244124923.0, + "step": 9429 + }, + { + "epoch": 1.0355809356468262, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.322850227355957, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7145264148712158, + "num_tokens": 244148357.0, + "step": 9430 + }, + { + "epoch": 1.03569075334944, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.204056978225708, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7182313203811646, + "num_tokens": 244175301.0, + "step": 9431 + }, + { + "epoch": 1.0358005710520537, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2172152996063232, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7215772867202759, + "num_tokens": 244201386.0, + "step": 9432 + }, + { + "epoch": 1.0359103887546672, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2034971714019775, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7099090814590454, + "num_tokens": 244228633.0, + "step": 9433 + }, + { + "epoch": 1.036020206457281, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2906649112701416, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7132958173751831, + "num_tokens": 244256930.0, + "step": 9434 + }, + { + "epoch": 1.0361300241598945, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.3730509281158447, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7493761777877808, + "num_tokens": 244278992.0, + "step": 9435 + }, + { + "epoch": 1.0362398418625083, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0907936096191406, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7187516689300537, + "num_tokens": 244307073.0, + "step": 9436 + }, + { + "epoch": 1.0363496595651218, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2467029094696045, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7159360647201538, + "num_tokens": 244333029.0, + "step": 9437 + }, + { + "epoch": 1.0364594772677356, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.411031723022461, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.7393868565559387, + "num_tokens": 244354581.0, + "step": 9438 + }, + { + "epoch": 1.036569294970349, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.614020586013794, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7160581946372986, + "num_tokens": 244376640.0, + "step": 9439 + }, + { + "epoch": 1.0366791126729629, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.206876754760742, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6978190541267395, + "num_tokens": 244403385.0, + "step": 9440 + }, + { + "epoch": 1.0367889303755766, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2706942558288574, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7063718438148499, + "num_tokens": 244431391.0, + "step": 9441 + }, + { + "epoch": 1.0368987480781902, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1115522384643555, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7076956033706665, + "num_tokens": 244458610.0, + "step": 9442 + }, + { + "epoch": 1.037008565780804, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0209710597991943, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.707314670085907, + "num_tokens": 244490252.0, + "step": 9443 + }, + { + "epoch": 1.0371183834834174, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.192617654800415, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.727014422416687, + "num_tokens": 244516220.0, + "step": 9444 + }, + { + "epoch": 1.0372282011860312, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1691951751708984, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.718583345413208, + "num_tokens": 244544253.0, + "step": 9445 + }, + { + "epoch": 1.037338018888645, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.3941099643707275, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7216917276382446, + "num_tokens": 244567126.0, + "step": 9446 + }, + { + "epoch": 1.0374478365912585, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.082728147506714, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7226371765136719, + "num_tokens": 244595282.0, + "step": 9447 + }, + { + "epoch": 1.0375576542938723, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.573519229888916, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7038705348968506, + "num_tokens": 244617865.0, + "step": 9448 + }, + { + "epoch": 1.0376674719964858, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.085501194000244, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7155965566635132, + "num_tokens": 244645296.0, + "step": 9449 + }, + { + "epoch": 1.0377772896990995, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.6774251461029053, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7384191751480103, + "num_tokens": 244663788.0, + "step": 9450 + }, + { + "epoch": 1.037887107401713, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.483656167984009, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7203153371810913, + "num_tokens": 244685999.0, + "step": 9451 + }, + { + "epoch": 1.0379969251043268, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1830086708068848, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7043800354003906, + "num_tokens": 244713331.0, + "step": 9452 + }, + { + "epoch": 1.0381067428069404, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1469814777374268, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7277684807777405, + "num_tokens": 244740153.0, + "step": 9453 + }, + { + "epoch": 1.0382165605095541, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.374594211578369, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7214996218681335, + "num_tokens": 244763602.0, + "step": 9454 + }, + { + "epoch": 1.0383263782121679, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.004251003265381, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7272365093231201, + "num_tokens": 244793724.0, + "step": 9455 + }, + { + "epoch": 1.0384361959147814, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1959033012390137, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7381193041801453, + "num_tokens": 244818177.0, + "step": 9456 + }, + { + "epoch": 1.0385460136173952, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 1.9741240739822388, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7282819151878357, + "num_tokens": 244848557.0, + "step": 9457 + }, + { + "epoch": 1.0386558313200087, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2161381244659424, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7278847694396973, + "num_tokens": 244874269.0, + "step": 9458 + }, + { + "epoch": 1.0387656490226225, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.4598186016082764, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7282088398933411, + "num_tokens": 244894552.0, + "step": 9459 + }, + { + "epoch": 1.038875466725236, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.273850202560425, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7341431379318237, + "num_tokens": 244919365.0, + "step": 9460 + }, + { + "epoch": 1.0389852844278498, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2417562007904053, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7135112285614014, + "num_tokens": 244943450.0, + "step": 9461 + }, + { + "epoch": 1.0390951021304635, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.345400810241699, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7168367505073547, + "num_tokens": 244966856.0, + "step": 9462 + }, + { + "epoch": 1.039204919833077, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0712499618530273, + "learning_rate": 1e-06, + "loss": 1.0236, + "mean_token_accuracy": 0.6981502771377563, + "num_tokens": 244998059.0, + "step": 9463 + }, + { + "epoch": 1.0393147375356908, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.280254602432251, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7085542678833008, + "num_tokens": 245022171.0, + "step": 9464 + }, + { + "epoch": 1.0394245552383043, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2140300273895264, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7279372215270996, + "num_tokens": 245049263.0, + "step": 9465 + }, + { + "epoch": 1.039534372940918, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.303323268890381, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7389793395996094, + "num_tokens": 245073793.0, + "step": 9466 + }, + { + "epoch": 1.0396441906435316, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 1.9959088563919067, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7184003591537476, + "num_tokens": 245103029.0, + "step": 9467 + }, + { + "epoch": 1.0397540083461454, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2642319202423096, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7171041369438171, + "num_tokens": 245128495.0, + "step": 9468 + }, + { + "epoch": 1.0398638260487592, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.091550827026367, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7242013812065125, + "num_tokens": 245156033.0, + "step": 9469 + }, + { + "epoch": 1.0399736437513727, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.64753794670105, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7381376028060913, + "num_tokens": 245174744.0, + "step": 9470 + }, + { + "epoch": 1.0400834614539864, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.36814022064209, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7102454900741577, + "num_tokens": 245200011.0, + "step": 9471 + }, + { + "epoch": 1.0401932791566, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1420841217041016, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.727859616279602, + "num_tokens": 245227165.0, + "step": 9472 + }, + { + "epoch": 1.0403030968592137, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0368003845214844, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7112916707992554, + "num_tokens": 245257985.0, + "step": 9473 + }, + { + "epoch": 1.0404129145618273, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.4026331901550293, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7505837082862854, + "num_tokens": 245279876.0, + "step": 9474 + }, + { + "epoch": 1.040522732264441, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2937116622924805, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7045391798019409, + "num_tokens": 245305272.0, + "step": 9475 + }, + { + "epoch": 1.0406325499670548, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.3601648807525635, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7156139612197876, + "num_tokens": 245328051.0, + "step": 9476 + }, + { + "epoch": 1.0407423676696683, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0426981449127197, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7146401405334473, + "num_tokens": 245358157.0, + "step": 9477 + }, + { + "epoch": 1.040852185372282, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.10799241065979, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7073361873626709, + "num_tokens": 245384741.0, + "step": 9478 + }, + { + "epoch": 1.0409620030748956, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.2391316890716553, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.706334114074707, + "num_tokens": 245409463.0, + "step": 9479 + }, + { + "epoch": 1.0410718207775094, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.1087210178375244, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7345328330993652, + "num_tokens": 245438723.0, + "step": 9480 + }, + { + "epoch": 1.041181638480123, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.5035455226898193, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7326552867889404, + "num_tokens": 245458902.0, + "step": 9481 + }, + { + "epoch": 1.0412914561827367, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.113917589187622, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7057661414146423, + "num_tokens": 245485163.0, + "step": 9482 + }, + { + "epoch": 1.0414012738853504, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.0883755683898926, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7294262051582336, + "num_tokens": 245511526.0, + "step": 9483 + }, + { + "epoch": 1.041511091587964, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.443228244781494, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7235874533653259, + "num_tokens": 245533940.0, + "step": 9484 + }, + { + "epoch": 1.0416209092905777, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.4744839668273926, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7515314817428589, + "num_tokens": 245553366.0, + "step": 9485 + }, + { + "epoch": 1.0417307269931912, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.534227132797241, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7359758019447327, + "num_tokens": 245573331.0, + "step": 9486 + }, + { + "epoch": 1.041840544695805, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.466773271560669, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7418235540390015, + "num_tokens": 245595441.0, + "step": 9487 + }, + { + "epoch": 1.0419503623984185, + "ewc_loss": 4.857778549194336e-06, + "grad_norm": 2.4877140522003174, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7285910844802856, + "num_tokens": 245616215.0, + "step": 9488 + }, + { + "epoch": 1.0420601801010323, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2085018157958984, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7134711742401123, + "num_tokens": 245641474.0, + "step": 9489 + }, + { + "epoch": 1.042169997803646, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.288656711578369, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7136536836624146, + "num_tokens": 245666626.0, + "step": 9490 + }, + { + "epoch": 1.0422798155062596, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.0889289379119873, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7225725650787354, + "num_tokens": 245693609.0, + "step": 9491 + }, + { + "epoch": 1.0423896332088733, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.4275527000427246, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7185268402099609, + "num_tokens": 245715909.0, + "step": 9492 + }, + { + "epoch": 1.0424994509114869, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2722949981689453, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7263530492782593, + "num_tokens": 245743609.0, + "step": 9493 + }, + { + "epoch": 1.0426092686141006, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.5102975368499756, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7221248745918274, + "num_tokens": 245765807.0, + "step": 9494 + }, + { + "epoch": 1.0427190863167142, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.0357179641723633, + "learning_rate": 1e-06, + "loss": 0.9892, + "mean_token_accuracy": 0.7022235989570618, + "num_tokens": 245798404.0, + "step": 9495 + }, + { + "epoch": 1.042828904019328, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.3955564498901367, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7157331705093384, + "num_tokens": 245820725.0, + "step": 9496 + }, + { + "epoch": 1.0429387217219417, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2808351516723633, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7098451256752014, + "num_tokens": 245844551.0, + "step": 9497 + }, + { + "epoch": 1.0430485394245552, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.1417622566223145, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7321888208389282, + "num_tokens": 245870890.0, + "step": 9498 + }, + { + "epoch": 1.043158357127169, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2040700912475586, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7115399837493896, + "num_tokens": 245896044.0, + "step": 9499 + }, + { + "epoch": 1.0432681748297825, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.248171806335449, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7004808783531189, + "num_tokens": 245923223.0, + "step": 9500 + }, + { + "epoch": 1.0433779925323963, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.203725814819336, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7076756954193115, + "num_tokens": 245950175.0, + "step": 9501 + }, + { + "epoch": 1.0434878102350098, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2615456581115723, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7279422283172607, + "num_tokens": 245974500.0, + "step": 9502 + }, + { + "epoch": 1.0435976279376236, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 1.9833205938339233, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.702491819858551, + "num_tokens": 246005891.0, + "step": 9503 + }, + { + "epoch": 1.043707445640237, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.2986786365509033, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.688589870929718, + "num_tokens": 246030961.0, + "step": 9504 + }, + { + "epoch": 1.0438172633428509, + "ewc_loss": 4.9173831939697266e-06, + "grad_norm": 2.3015434741973877, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7110288143157959, + "num_tokens": 246054915.0, + "step": 9505 + }, + { + "epoch": 1.0439270810454646, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.195204973220825, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7267096042633057, + "num_tokens": 246081890.0, + "step": 9506 + }, + { + "epoch": 1.0440368987480781, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.105376720428467, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7225875854492188, + "num_tokens": 246110456.0, + "step": 9507 + }, + { + "epoch": 1.044146716450692, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.032954692840576, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6976221799850464, + "num_tokens": 246139574.0, + "step": 9508 + }, + { + "epoch": 1.0442565341533054, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.2660648822784424, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7055783867835999, + "num_tokens": 246165445.0, + "step": 9509 + }, + { + "epoch": 1.0443663518559192, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.4518790245056152, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7340030074119568, + "num_tokens": 246185870.0, + "step": 9510 + }, + { + "epoch": 1.0444761695585327, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.030379295349121, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7003285884857178, + "num_tokens": 246216822.0, + "step": 9511 + }, + { + "epoch": 1.0445859872611465, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.465412139892578, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7324179410934448, + "num_tokens": 246239351.0, + "step": 9512 + }, + { + "epoch": 1.0446958049637602, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.397488594055176, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.736403226852417, + "num_tokens": 246262271.0, + "step": 9513 + }, + { + "epoch": 1.0448056226663738, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.3965682983398438, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.715280294418335, + "num_tokens": 246286562.0, + "step": 9514 + }, + { + "epoch": 1.0449154403689875, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.4411821365356445, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7122249007225037, + "num_tokens": 246310265.0, + "step": 9515 + }, + { + "epoch": 1.045025258071601, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.4012277126312256, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7031301856040955, + "num_tokens": 246335195.0, + "step": 9516 + }, + { + "epoch": 1.0451350757742148, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.3340818881988525, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7275601625442505, + "num_tokens": 246360099.0, + "step": 9517 + }, + { + "epoch": 1.0452448934768284, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.2959747314453125, + "learning_rate": 1e-06, + "loss": 1.0226, + "mean_token_accuracy": 0.6864843368530273, + "num_tokens": 246385857.0, + "step": 9518 + }, + { + "epoch": 1.0453547111794421, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.2034173011779785, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7381617426872253, + "num_tokens": 246411918.0, + "step": 9519 + }, + { + "epoch": 1.0454645288820559, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.3969805240631104, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.743992030620575, + "num_tokens": 246433859.0, + "step": 9520 + }, + { + "epoch": 1.0455743465846694, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.4658761024475098, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7336726188659668, + "num_tokens": 246455155.0, + "step": 9521 + }, + { + "epoch": 1.0456841642872832, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.085433006286621, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7174211740493774, + "num_tokens": 246482223.0, + "step": 9522 + }, + { + "epoch": 1.0457939819898967, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.275864839553833, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7054849863052368, + "num_tokens": 246507535.0, + "step": 9523 + }, + { + "epoch": 1.0459037996925105, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.379634141921997, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7167603969573975, + "num_tokens": 246530389.0, + "step": 9524 + }, + { + "epoch": 1.046013617395124, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.1416542530059814, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7102219462394714, + "num_tokens": 246558410.0, + "step": 9525 + }, + { + "epoch": 1.0461234350977378, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.03378963470459, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7326856255531311, + "num_tokens": 246586971.0, + "step": 9526 + }, + { + "epoch": 1.0462332528003515, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.2178001403808594, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.721668541431427, + "num_tokens": 246611921.0, + "step": 9527 + }, + { + "epoch": 1.046343070502965, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.3026745319366455, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7230426073074341, + "num_tokens": 246637183.0, + "step": 9528 + }, + { + "epoch": 1.0464528882055788, + "ewc_loss": 4.976987838745117e-06, + "grad_norm": 2.2155449390411377, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.711112380027771, + "num_tokens": 246661123.0, + "step": 9529 + }, + { + "epoch": 1.0465627059081923, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.254871368408203, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.730086088180542, + "num_tokens": 246684376.0, + "step": 9530 + }, + { + "epoch": 1.046672523610806, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.5943334102630615, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7264964580535889, + "num_tokens": 246703163.0, + "step": 9531 + }, + { + "epoch": 1.0467823413134196, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.2068769931793213, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7319653630256653, + "num_tokens": 246727436.0, + "step": 9532 + }, + { + "epoch": 1.0468921590160334, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.124288320541382, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7233867049217224, + "num_tokens": 246754963.0, + "step": 9533 + }, + { + "epoch": 1.0470019767186471, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.3924362659454346, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7334020137786865, + "num_tokens": 246777203.0, + "step": 9534 + }, + { + "epoch": 1.0471117944212607, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.4185261726379395, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7471468448638916, + "num_tokens": 246800178.0, + "step": 9535 + }, + { + "epoch": 1.0472216121238744, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.225572109222412, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7092400193214417, + "num_tokens": 246827037.0, + "step": 9536 + }, + { + "epoch": 1.047331429826488, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.4711320400238037, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7506287097930908, + "num_tokens": 246848866.0, + "step": 9537 + }, + { + "epoch": 1.0474412475291017, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.4664411544799805, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7240210771560669, + "num_tokens": 246869487.0, + "step": 9538 + }, + { + "epoch": 1.0475510652317153, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.252256155014038, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.700439453125, + "num_tokens": 246897217.0, + "step": 9539 + }, + { + "epoch": 1.047660882934329, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.05063796043396, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7034599184989929, + "num_tokens": 246927339.0, + "step": 9540 + }, + { + "epoch": 1.0477707006369428, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.0175552368164062, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7201708555221558, + "num_tokens": 246957855.0, + "step": 9541 + }, + { + "epoch": 1.0478805183395563, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3694794178009033, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7312675714492798, + "num_tokens": 246979698.0, + "step": 9542 + }, + { + "epoch": 1.04799033604217, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1132447719573975, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7412763833999634, + "num_tokens": 247007272.0, + "step": 9543 + }, + { + "epoch": 1.0481001537447836, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.398824691772461, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7384110689163208, + "num_tokens": 247030032.0, + "step": 9544 + }, + { + "epoch": 1.0482099714473974, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.3229434490203857, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7254880666732788, + "num_tokens": 247053901.0, + "step": 9545 + }, + { + "epoch": 1.048319789150011, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.4298222064971924, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7140759229660034, + "num_tokens": 247076976.0, + "step": 9546 + }, + { + "epoch": 1.0484296068526247, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.5718789100646973, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7375614643096924, + "num_tokens": 247097690.0, + "step": 9547 + }, + { + "epoch": 1.0485394245552384, + "ewc_loss": 5.0067901611328125e-06, + "grad_norm": 2.2869350910186768, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7372702956199646, + "num_tokens": 247121076.0, + "step": 9548 + }, + { + "epoch": 1.048649242257852, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.228581190109253, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.732165515422821, + "num_tokens": 247145963.0, + "step": 9549 + }, + { + "epoch": 1.0487590599604657, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.190237045288086, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7103330492973328, + "num_tokens": 247172368.0, + "step": 9550 + }, + { + "epoch": 1.0488688776630792, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1968023777008057, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.6988728046417236, + "num_tokens": 247201180.0, + "step": 9551 + }, + { + "epoch": 1.048978695365693, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.034729242324829, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7045608758926392, + "num_tokens": 247232563.0, + "step": 9552 + }, + { + "epoch": 1.0490885130683065, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.272611141204834, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6972384452819824, + "num_tokens": 247257790.0, + "step": 9553 + }, + { + "epoch": 1.0491983307709203, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3624160289764404, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7206013202667236, + "num_tokens": 247279700.0, + "step": 9554 + }, + { + "epoch": 1.0493081484735338, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.343381881713867, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7376741170883179, + "num_tokens": 247303222.0, + "step": 9555 + }, + { + "epoch": 1.0494179661761476, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.32572865486145, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7392228841781616, + "num_tokens": 247325782.0, + "step": 9556 + }, + { + "epoch": 1.0495277838787613, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.0699479579925537, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6975825428962708, + "num_tokens": 247356404.0, + "step": 9557 + }, + { + "epoch": 1.0496376015813749, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.9395742416381836, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7020907998085022, + "num_tokens": 247389537.0, + "step": 9558 + }, + { + "epoch": 1.0497474192839886, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.4214930534362793, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.713317334651947, + "num_tokens": 247412214.0, + "step": 9559 + }, + { + "epoch": 1.0498572369866022, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.417128324508667, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7493156790733337, + "num_tokens": 247432745.0, + "step": 9560 + }, + { + "epoch": 1.049967054689216, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.234562635421753, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7195366024971008, + "num_tokens": 247457197.0, + "step": 9561 + }, + { + "epoch": 1.0500768723918297, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.5708885192871094, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.700913667678833, + "num_tokens": 247479184.0, + "step": 9562 + }, + { + "epoch": 1.0501866900944432, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.403045415878296, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7380698919296265, + "num_tokens": 247502682.0, + "step": 9563 + }, + { + "epoch": 1.050296507797057, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.237626075744629, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7068526744842529, + "num_tokens": 247531199.0, + "step": 9564 + }, + { + "epoch": 1.0504063254996705, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1359264850616455, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7005918622016907, + "num_tokens": 247560239.0, + "step": 9565 + }, + { + "epoch": 1.0505161432022843, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3324670791625977, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.70281982421875, + "num_tokens": 247584598.0, + "step": 9566 + }, + { + "epoch": 1.0506259609048978, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.161709785461426, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7263226509094238, + "num_tokens": 247610185.0, + "step": 9567 + }, + { + "epoch": 1.0507357786075116, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.353544235229492, + "learning_rate": 1e-06, + "loss": 0.7669, + "mean_token_accuracy": 0.7528281211853027, + "num_tokens": 247632298.0, + "step": 9568 + }, + { + "epoch": 1.050845596310125, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.03521728515625, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.731072723865509, + "num_tokens": 247659916.0, + "step": 9569 + }, + { + "epoch": 1.0509554140127388, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.089094400405884, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7249947786331177, + "num_tokens": 247688480.0, + "step": 9570 + }, + { + "epoch": 1.0510652317153526, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1734092235565186, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7156660556793213, + "num_tokens": 247716386.0, + "step": 9571 + }, + { + "epoch": 1.0511750494179661, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.285898447036743, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7147535681724548, + "num_tokens": 247741248.0, + "step": 9572 + }, + { + "epoch": 1.05128486712058, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.227376699447632, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7202657461166382, + "num_tokens": 247767135.0, + "step": 9573 + }, + { + "epoch": 1.0513946848231934, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.202505588531494, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7188788652420044, + "num_tokens": 247793866.0, + "step": 9574 + }, + { + "epoch": 1.0515045025258072, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.9865314960479736, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7103625535964966, + "num_tokens": 247826488.0, + "step": 9575 + }, + { + "epoch": 1.0516143202284207, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.4237771034240723, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7048245668411255, + "num_tokens": 247848265.0, + "step": 9576 + }, + { + "epoch": 1.0517241379310345, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.092348337173462, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7132200598716736, + "num_tokens": 247878027.0, + "step": 9577 + }, + { + "epoch": 1.0518339556336482, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.526080369949341, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7316264510154724, + "num_tokens": 247899160.0, + "step": 9578 + }, + { + "epoch": 1.0519437733362618, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3029117584228516, + "learning_rate": 1e-06, + "loss": 1.0274, + "mean_token_accuracy": 0.6857550144195557, + "num_tokens": 247924094.0, + "step": 9579 + }, + { + "epoch": 1.0520535910388755, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.830731749534607, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7024335861206055, + "num_tokens": 247958200.0, + "step": 9580 + }, + { + "epoch": 1.052163408741489, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.274132013320923, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7185595631599426, + "num_tokens": 247982276.0, + "step": 9581 + }, + { + "epoch": 1.0522732264441028, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.147311210632324, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7244643568992615, + "num_tokens": 248010082.0, + "step": 9582 + }, + { + "epoch": 1.0523830441467164, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.2668981552124023, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7361006140708923, + "num_tokens": 248035005.0, + "step": 9583 + }, + { + "epoch": 1.05249286184933, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1868886947631836, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7092860341072083, + "num_tokens": 248061281.0, + "step": 9584 + }, + { + "epoch": 1.0526026795519439, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.2316389083862305, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.706123948097229, + "num_tokens": 248089291.0, + "step": 9585 + }, + { + "epoch": 1.0527124972545574, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.093273401260376, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7190098166465759, + "num_tokens": 248118085.0, + "step": 9586 + }, + { + "epoch": 1.0528223149571712, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3538854122161865, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7156566381454468, + "num_tokens": 248141100.0, + "step": 9587 + }, + { + "epoch": 1.0529321326597847, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.309485673904419, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.72329181432724, + "num_tokens": 248165329.0, + "step": 9588 + }, + { + "epoch": 1.0530419503623984, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.0854620933532715, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.701559841632843, + "num_tokens": 248196463.0, + "step": 9589 + }, + { + "epoch": 1.053151768065012, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.388230562210083, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7380306124687195, + "num_tokens": 248219419.0, + "step": 9590 + }, + { + "epoch": 1.0532615857676257, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.5601632595062256, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.701506495475769, + "num_tokens": 248242837.0, + "step": 9591 + }, + { + "epoch": 1.0533714034702395, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.3545355796813965, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7137593030929565, + "num_tokens": 248267252.0, + "step": 9592 + }, + { + "epoch": 1.053481221172853, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.8500252962112427, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.699458658695221, + "num_tokens": 248303356.0, + "step": 9593 + }, + { + "epoch": 1.0535910388754668, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1795332431793213, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7059851884841919, + "num_tokens": 248331023.0, + "step": 9594 + }, + { + "epoch": 1.0537008565780803, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.9384751319885254, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7220078706741333, + "num_tokens": 248360940.0, + "step": 9595 + }, + { + "epoch": 1.053810674280694, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.133343458175659, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7199360132217407, + "num_tokens": 248390650.0, + "step": 9596 + }, + { + "epoch": 1.0539204919833076, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.1333985328674316, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7198070883750916, + "num_tokens": 248417494.0, + "step": 9597 + }, + { + "epoch": 1.0540303096859214, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.0180039405822754, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7363772988319397, + "num_tokens": 248444723.0, + "step": 9598 + }, + { + "epoch": 1.0541401273885351, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.394479513168335, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7125769257545471, + "num_tokens": 248466685.0, + "step": 9599 + }, + { + "epoch": 1.0542499450911487, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.488161563873291, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.727575421333313, + "num_tokens": 248488630.0, + "step": 9600 + }, + { + "epoch": 1.0543597627937624, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.0683295726776123, + "learning_rate": 1e-06, + "loss": 1.0309, + "mean_token_accuracy": 0.6861883997917175, + "num_tokens": 248518374.0, + "step": 9601 + }, + { + "epoch": 1.054469580496376, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2483067512512207, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7156331539154053, + "num_tokens": 248542665.0, + "step": 9602 + }, + { + "epoch": 1.0545793981989897, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 1.856052041053772, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7255823612213135, + "num_tokens": 248576368.0, + "step": 9603 + }, + { + "epoch": 1.0546892159016032, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 1.9571375846862793, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.6942840814590454, + "num_tokens": 248606520.0, + "step": 9604 + }, + { + "epoch": 1.054799033604217, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.060580253601074, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.6890126466751099, + "num_tokens": 248634950.0, + "step": 9605 + }, + { + "epoch": 1.0549088513068308, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.197470188140869, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7207925319671631, + "num_tokens": 248663549.0, + "step": 9606 + }, + { + "epoch": 1.0550186690094443, + "ewc_loss": 5.036592483520508e-06, + "grad_norm": 2.328636884689331, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7088946104049683, + "num_tokens": 248688326.0, + "step": 9607 + }, + { + "epoch": 1.055128486712058, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.167684555053711, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7111430168151855, + "num_tokens": 248714511.0, + "step": 9608 + }, + { + "epoch": 1.0552383044146716, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.16982364654541, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.72050940990448, + "num_tokens": 248742304.0, + "step": 9609 + }, + { + "epoch": 1.0553481221172853, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.144092082977295, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7389129400253296, + "num_tokens": 248768175.0, + "step": 9610 + }, + { + "epoch": 1.0554579398198989, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.0714967250823975, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7121788263320923, + "num_tokens": 248798423.0, + "step": 9611 + }, + { + "epoch": 1.0555677575225126, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.27170467376709, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7145812511444092, + "num_tokens": 248823545.0, + "step": 9612 + }, + { + "epoch": 1.0556775752251264, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.5821382999420166, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7382122278213501, + "num_tokens": 248842430.0, + "step": 9613 + }, + { + "epoch": 1.05578739292774, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1171834468841553, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.710265040397644, + "num_tokens": 248870096.0, + "step": 9614 + }, + { + "epoch": 1.0558972106303537, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.348083257675171, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7107129096984863, + "num_tokens": 248893581.0, + "step": 9615 + }, + { + "epoch": 1.0560070283329672, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.174442768096924, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7101360559463501, + "num_tokens": 248919001.0, + "step": 9616 + }, + { + "epoch": 1.056116846035581, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1117472648620605, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7092098593711853, + "num_tokens": 248946510.0, + "step": 9617 + }, + { + "epoch": 1.0562266637381945, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.0008323192596436, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7109102010726929, + "num_tokens": 248978612.0, + "step": 9618 + }, + { + "epoch": 1.0563364814408083, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.087648391723633, + "learning_rate": 1e-06, + "loss": 1.0037, + "mean_token_accuracy": 0.6901524066925049, + "num_tokens": 249008011.0, + "step": 9619 + }, + { + "epoch": 1.0564462991434218, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.22963285446167, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7321094274520874, + "num_tokens": 249031806.0, + "step": 9620 + }, + { + "epoch": 1.0565561168460356, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.5021862983703613, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7207198143005371, + "num_tokens": 249052702.0, + "step": 9621 + }, + { + "epoch": 1.0566659345486493, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 1.8983378410339355, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.68821781873703, + "num_tokens": 249087582.0, + "step": 9622 + }, + { + "epoch": 1.0567757522512629, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3705086708068848, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7093913555145264, + "num_tokens": 249111026.0, + "step": 9623 + }, + { + "epoch": 1.0568855699538766, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3496148586273193, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7135071754455566, + "num_tokens": 249133904.0, + "step": 9624 + }, + { + "epoch": 1.0569953876564901, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.712102174758911, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7210295796394348, + "num_tokens": 249153840.0, + "step": 9625 + }, + { + "epoch": 1.057105205359104, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.356837749481201, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7126011252403259, + "num_tokens": 249177898.0, + "step": 9626 + }, + { + "epoch": 1.0572150230617177, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4965975284576416, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7247810363769531, + "num_tokens": 249199392.0, + "step": 9627 + }, + { + "epoch": 1.0573248407643312, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1451268196105957, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7154251337051392, + "num_tokens": 249227251.0, + "step": 9628 + }, + { + "epoch": 1.057434658466945, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.478478193283081, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7396081686019897, + "num_tokens": 249248156.0, + "step": 9629 + }, + { + "epoch": 1.0575444761695585, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.50087571144104, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.6978082656860352, + "num_tokens": 249271245.0, + "step": 9630 + }, + { + "epoch": 1.0576542938721722, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4185824394226074, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6973927617073059, + "num_tokens": 249297210.0, + "step": 9631 + }, + { + "epoch": 1.0577641115747858, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4967856407165527, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7324065566062927, + "num_tokens": 249317807.0, + "step": 9632 + }, + { + "epoch": 1.0578739292773995, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.151458978652954, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.719180703163147, + "num_tokens": 249347204.0, + "step": 9633 + }, + { + "epoch": 1.057983746980013, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.0326390266418457, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7161184549331665, + "num_tokens": 249377018.0, + "step": 9634 + }, + { + "epoch": 1.0580935646826268, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1366336345672607, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7268154621124268, + "num_tokens": 249404706.0, + "step": 9635 + }, + { + "epoch": 1.0582033823852406, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4058234691619873, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7259204983711243, + "num_tokens": 249428576.0, + "step": 9636 + }, + { + "epoch": 1.0583132000878541, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3129334449768066, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7334709167480469, + "num_tokens": 249455495.0, + "step": 9637 + }, + { + "epoch": 1.0584230177904679, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.5987203121185303, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7295932173728943, + "num_tokens": 249474722.0, + "step": 9638 + }, + { + "epoch": 1.0585328354930814, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2814748287200928, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7189661264419556, + "num_tokens": 249501296.0, + "step": 9639 + }, + { + "epoch": 1.0586426531956952, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.5625014305114746, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7345740795135498, + "num_tokens": 249521551.0, + "step": 9640 + }, + { + "epoch": 1.0587524708983087, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3048484325408936, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7252951264381409, + "num_tokens": 249546353.0, + "step": 9641 + }, + { + "epoch": 1.0588622886009225, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.363405466079712, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7096388936042786, + "num_tokens": 249572895.0, + "step": 9642 + }, + { + "epoch": 1.0589721063035362, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.354874610900879, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7402939796447754, + "num_tokens": 249596783.0, + "step": 9643 + }, + { + "epoch": 1.0590819240061498, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2422804832458496, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7170210480690002, + "num_tokens": 249623279.0, + "step": 9644 + }, + { + "epoch": 1.0591917417087635, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.282822847366333, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7152332067489624, + "num_tokens": 249651574.0, + "step": 9645 + }, + { + "epoch": 1.059301559411377, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.26533579826355, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.709179162979126, + "num_tokens": 249677631.0, + "step": 9646 + }, + { + "epoch": 1.0594113771139908, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2946627140045166, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7062017321586609, + "num_tokens": 249703995.0, + "step": 9647 + }, + { + "epoch": 1.0595211948166043, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.274885416030884, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7226996421813965, + "num_tokens": 249728629.0, + "step": 9648 + }, + { + "epoch": 1.059631012519218, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.139193058013916, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6998217701911926, + "num_tokens": 249757537.0, + "step": 9649 + }, + { + "epoch": 1.0597408302218319, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.482569694519043, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7157477140426636, + "num_tokens": 249779905.0, + "step": 9650 + }, + { + "epoch": 1.0598506479244454, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2706642150878906, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7252094149589539, + "num_tokens": 249806155.0, + "step": 9651 + }, + { + "epoch": 1.0599604656270591, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2395336627960205, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.7070927619934082, + "num_tokens": 249836082.0, + "step": 9652 + }, + { + "epoch": 1.0600702833296727, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3549938201904297, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7196770906448364, + "num_tokens": 249860640.0, + "step": 9653 + }, + { + "epoch": 1.0601801010322864, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3334686756134033, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7119894623756409, + "num_tokens": 249884934.0, + "step": 9654 + }, + { + "epoch": 1.0602899187349, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2845451831817627, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7145638465881348, + "num_tokens": 249910064.0, + "step": 9655 + }, + { + "epoch": 1.0603997364375137, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1778156757354736, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.704131007194519, + "num_tokens": 249937239.0, + "step": 9656 + }, + { + "epoch": 1.0605095541401275, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.7539682388305664, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7418756484985352, + "num_tokens": 249955880.0, + "step": 9657 + }, + { + "epoch": 1.060619371842741, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 1.936430811882019, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7018507719039917, + "num_tokens": 249988295.0, + "step": 9658 + }, + { + "epoch": 1.0607291895453548, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.374082088470459, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7262840867042542, + "num_tokens": 250014371.0, + "step": 9659 + }, + { + "epoch": 1.0608390072479683, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4621381759643555, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7322872877120972, + "num_tokens": 250035108.0, + "step": 9660 + }, + { + "epoch": 1.060948824950582, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4192187786102295, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7256542444229126, + "num_tokens": 250057924.0, + "step": 9661 + }, + { + "epoch": 1.0610586426531956, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.014920234680176, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7122076153755188, + "num_tokens": 250088872.0, + "step": 9662 + }, + { + "epoch": 1.0611684603558094, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.6077725887298584, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.6967631578445435, + "num_tokens": 250110129.0, + "step": 9663 + }, + { + "epoch": 1.0612782780584231, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.375816583633423, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7193067073822021, + "num_tokens": 250132915.0, + "step": 9664 + }, + { + "epoch": 1.0613880957610367, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2309319972991943, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.71077561378479, + "num_tokens": 250158548.0, + "step": 9665 + }, + { + "epoch": 1.0614979134636504, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2820653915405273, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7099869251251221, + "num_tokens": 250183701.0, + "step": 9666 + }, + { + "epoch": 1.061607731166264, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.137129783630371, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7132354378700256, + "num_tokens": 250212131.0, + "step": 9667 + }, + { + "epoch": 1.0617175488688777, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.42521595954895, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7383519411087036, + "num_tokens": 250235131.0, + "step": 9668 + }, + { + "epoch": 1.0618273665714912, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.404879331588745, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7195553183555603, + "num_tokens": 250258718.0, + "step": 9669 + }, + { + "epoch": 1.061937184274105, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.267747402191162, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7434245944023132, + "num_tokens": 250283341.0, + "step": 9670 + }, + { + "epoch": 1.0620470019767188, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.280346155166626, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.741246223449707, + "num_tokens": 250304962.0, + "step": 9671 + }, + { + "epoch": 1.0621568196793323, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.267181396484375, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7005765438079834, + "num_tokens": 250332787.0, + "step": 9672 + }, + { + "epoch": 1.062266637381946, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 1.9096014499664307, + "learning_rate": 1e-06, + "loss": 1.0047, + "mean_token_accuracy": 0.6920878887176514, + "num_tokens": 250368623.0, + "step": 9673 + }, + { + "epoch": 1.0623764550845596, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.3676247596740723, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7239795923233032, + "num_tokens": 250392124.0, + "step": 9674 + }, + { + "epoch": 1.0624862727871733, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1805081367492676, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7187902927398682, + "num_tokens": 250417204.0, + "step": 9675 + }, + { + "epoch": 1.0625960904897869, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.216139316558838, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7068791389465332, + "num_tokens": 250444684.0, + "step": 9676 + }, + { + "epoch": 1.0627059081924006, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.310699701309204, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.7041671276092529, + "num_tokens": 250469887.0, + "step": 9677 + }, + { + "epoch": 1.0628157258950144, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1232964992523193, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7235109806060791, + "num_tokens": 250498534.0, + "step": 9678 + }, + { + "epoch": 1.062925543597628, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1737875938415527, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7412528395652771, + "num_tokens": 250525802.0, + "step": 9679 + }, + { + "epoch": 1.0630353613002417, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.13651967048645, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7266192436218262, + "num_tokens": 250553585.0, + "step": 9680 + }, + { + "epoch": 1.0631451790028552, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.468515396118164, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.732326328754425, + "num_tokens": 250576137.0, + "step": 9681 + }, + { + "epoch": 1.063254996705469, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.017021656036377, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6945523023605347, + "num_tokens": 250606474.0, + "step": 9682 + }, + { + "epoch": 1.0633648144080825, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.007960081100464, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7216529846191406, + "num_tokens": 250636724.0, + "step": 9683 + }, + { + "epoch": 1.0634746321106963, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.0972189903259277, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7127229571342468, + "num_tokens": 250665159.0, + "step": 9684 + }, + { + "epoch": 1.0635844498133098, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.331836223602295, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7365195751190186, + "num_tokens": 250687852.0, + "step": 9685 + }, + { + "epoch": 1.0636942675159236, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.1066412925720215, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7328024506568909, + "num_tokens": 250717529.0, + "step": 9686 + }, + { + "epoch": 1.0638040852185373, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4282822608947754, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7145001888275146, + "num_tokens": 250738915.0, + "step": 9687 + }, + { + "epoch": 1.0639139029211508, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4076437950134277, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7311322689056396, + "num_tokens": 250761103.0, + "step": 9688 + }, + { + "epoch": 1.0640237206237646, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.4445366859436035, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.734362781047821, + "num_tokens": 250784246.0, + "step": 9689 + }, + { + "epoch": 1.0641335383263781, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.222208261489868, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7240015864372253, + "num_tokens": 250808922.0, + "step": 9690 + }, + { + "epoch": 1.064243356028992, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.046905755996704, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7019668817520142, + "num_tokens": 250837921.0, + "step": 9691 + }, + { + "epoch": 1.0643531737316057, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.2262446880340576, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7218919992446899, + "num_tokens": 250864218.0, + "step": 9692 + }, + { + "epoch": 1.0644629914342192, + "ewc_loss": 5.066394805908203e-06, + "grad_norm": 2.194373369216919, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7332009077072144, + "num_tokens": 250889667.0, + "step": 9693 + }, + { + "epoch": 1.064572809136833, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.7579503059387207, + "learning_rate": 1e-06, + "loss": 0.7728, + "mean_token_accuracy": 0.75095534324646, + "num_tokens": 250906367.0, + "step": 9694 + }, + { + "epoch": 1.0646826268394465, + "ewc_loss": 5.0961971282958984e-06, + "grad_norm": 2.0967013835906982, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7195439338684082, + "num_tokens": 250935459.0, + "step": 9695 + }, + { + "epoch": 1.0647924445420602, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.3498551845550537, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7247515916824341, + "num_tokens": 250960205.0, + "step": 9696 + }, + { + "epoch": 1.0649022622446738, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.2189769744873047, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7166475057601929, + "num_tokens": 250986395.0, + "step": 9697 + }, + { + "epoch": 1.0650120799472875, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.4349570274353027, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7110428810119629, + "num_tokens": 251010206.0, + "step": 9698 + }, + { + "epoch": 1.065121897649901, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.149062156677246, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7109030485153198, + "num_tokens": 251037764.0, + "step": 9699 + }, + { + "epoch": 1.0652317153525148, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.169107675552368, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7204940319061279, + "num_tokens": 251065632.0, + "step": 9700 + }, + { + "epoch": 1.0653415330551286, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.5259532928466797, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7091987133026123, + "num_tokens": 251087567.0, + "step": 9701 + }, + { + "epoch": 1.065451350757742, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.728585720062256, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7395095229148865, + "num_tokens": 251105169.0, + "step": 9702 + }, + { + "epoch": 1.0655611684603559, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.226950168609619, + "learning_rate": 1e-06, + "loss": 0.743, + "mean_token_accuracy": 0.7614451050758362, + "num_tokens": 251129897.0, + "step": 9703 + }, + { + "epoch": 1.0656709861629694, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.484715223312378, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7284438014030457, + "num_tokens": 251152836.0, + "step": 9704 + }, + { + "epoch": 1.0657808038655832, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.380007028579712, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7169264554977417, + "num_tokens": 251178610.0, + "step": 9705 + }, + { + "epoch": 1.0658906215681967, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.206439733505249, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7147717475891113, + "num_tokens": 251203642.0, + "step": 9706 + }, + { + "epoch": 1.0660004392708105, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.260866641998291, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7193455100059509, + "num_tokens": 251228970.0, + "step": 9707 + }, + { + "epoch": 1.0661102569734242, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1306495666503906, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7208300828933716, + "num_tokens": 251256481.0, + "step": 9708 + }, + { + "epoch": 1.0662200746760377, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.271059513092041, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6986788511276245, + "num_tokens": 251284419.0, + "step": 9709 + }, + { + "epoch": 1.0663298923786515, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.237797260284424, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7191604375839233, + "num_tokens": 251308932.0, + "step": 9710 + }, + { + "epoch": 1.066439710081265, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.073291063308716, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7044092416763306, + "num_tokens": 251339253.0, + "step": 9711 + }, + { + "epoch": 1.0665495277838788, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.5717880725860596, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7270748615264893, + "num_tokens": 251360860.0, + "step": 9712 + }, + { + "epoch": 1.0666593454864923, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.225679397583008, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7263731360435486, + "num_tokens": 251385661.0, + "step": 9713 + }, + { + "epoch": 1.066769163189106, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.257371187210083, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7180793285369873, + "num_tokens": 251411174.0, + "step": 9714 + }, + { + "epoch": 1.0668789808917198, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.5051302909851074, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.72257399559021, + "num_tokens": 251433780.0, + "step": 9715 + }, + { + "epoch": 1.0669887985943334, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.4547066688537598, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7157751321792603, + "num_tokens": 251457958.0, + "step": 9716 + }, + { + "epoch": 1.0670986162969471, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.219770908355713, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7195929884910583, + "num_tokens": 251483937.0, + "step": 9717 + }, + { + "epoch": 1.0672084339995607, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.357003927230835, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7260245084762573, + "num_tokens": 251506303.0, + "step": 9718 + }, + { + "epoch": 1.0673182517021744, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.3549964427948, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7078403830528259, + "num_tokens": 251531605.0, + "step": 9719 + }, + { + "epoch": 1.067428069404788, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2056093215942383, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6927330493927002, + "num_tokens": 251559470.0, + "step": 9720 + }, + { + "epoch": 1.0675378871074017, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.4759163856506348, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7448309659957886, + "num_tokens": 251580254.0, + "step": 9721 + }, + { + "epoch": 1.0676477048100155, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1609678268432617, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7143954038619995, + "num_tokens": 251606712.0, + "step": 9722 + }, + { + "epoch": 1.067757522512629, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.20648193359375, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7072691917419434, + "num_tokens": 251632873.0, + "step": 9723 + }, + { + "epoch": 1.0678673402152428, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1489315032958984, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7050669193267822, + "num_tokens": 251662345.0, + "step": 9724 + }, + { + "epoch": 1.0679771579178563, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 1.9510432481765747, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7224643230438232, + "num_tokens": 251694926.0, + "step": 9725 + }, + { + "epoch": 1.06808697562047, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.409885883331299, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7210435271263123, + "num_tokens": 251717795.0, + "step": 9726 + }, + { + "epoch": 1.0681967933230836, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1044328212738037, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7256981134414673, + "num_tokens": 251744985.0, + "step": 9727 + }, + { + "epoch": 1.0683066110256974, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.25919771194458, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.711298942565918, + "num_tokens": 251769990.0, + "step": 9728 + }, + { + "epoch": 1.068416428728311, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.505094289779663, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7059735059738159, + "num_tokens": 251793256.0, + "step": 9729 + }, + { + "epoch": 1.0685262464309246, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.0682754516601562, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6923609972000122, + "num_tokens": 251824616.0, + "step": 9730 + }, + { + "epoch": 1.0686360641335384, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.383246421813965, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7047219276428223, + "num_tokens": 251849164.0, + "step": 9731 + }, + { + "epoch": 1.068745881836152, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.0124220848083496, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7329801321029663, + "num_tokens": 251879989.0, + "step": 9732 + }, + { + "epoch": 1.0688556995387657, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.161735773086548, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7157279253005981, + "num_tokens": 251906896.0, + "step": 9733 + }, + { + "epoch": 1.0689655172413792, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.117776393890381, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7067492008209229, + "num_tokens": 251935270.0, + "step": 9734 + }, + { + "epoch": 1.069075334943993, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.4866626262664795, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7267196178436279, + "num_tokens": 251958585.0, + "step": 9735 + }, + { + "epoch": 1.0691851526466065, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.0414576530456543, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.685283899307251, + "num_tokens": 251989265.0, + "step": 9736 + }, + { + "epoch": 1.0692949703492203, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.1256775856018066, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7375907897949219, + "num_tokens": 252015010.0, + "step": 9737 + }, + { + "epoch": 1.069404788051834, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.3360533714294434, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7268209457397461, + "num_tokens": 252038183.0, + "step": 9738 + }, + { + "epoch": 1.0695146057544476, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.500223159790039, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7070831060409546, + "num_tokens": 252061193.0, + "step": 9739 + }, + { + "epoch": 1.0696244234570613, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.0590150356292725, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7171802520751953, + "num_tokens": 252091708.0, + "step": 9740 + }, + { + "epoch": 1.0697342411596749, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.5960757732391357, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7371163368225098, + "num_tokens": 252110662.0, + "step": 9741 + }, + { + "epoch": 1.0698440588622886, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.0845675468444824, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7218122482299805, + "num_tokens": 252138537.0, + "step": 9742 + }, + { + "epoch": 1.0699538765649024, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.22346568107605, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7163432836532593, + "num_tokens": 252164974.0, + "step": 9743 + }, + { + "epoch": 1.070063694267516, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.1683197021484375, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7223838567733765, + "num_tokens": 252189644.0, + "step": 9744 + }, + { + "epoch": 1.0701735119701297, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.283590316772461, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6969230771064758, + "num_tokens": 252215816.0, + "step": 9745 + }, + { + "epoch": 1.0702833296727432, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.1381702423095703, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7181817293167114, + "num_tokens": 252242546.0, + "step": 9746 + }, + { + "epoch": 1.070393147375357, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.617915153503418, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.719976544380188, + "num_tokens": 252261287.0, + "step": 9747 + }, + { + "epoch": 1.0705029650779705, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.20507550239563, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7252078056335449, + "num_tokens": 252288336.0, + "step": 9748 + }, + { + "epoch": 1.0706127827805842, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.1177175045013428, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7292418479919434, + "num_tokens": 252315179.0, + "step": 9749 + }, + { + "epoch": 1.0707226004831978, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.221367120742798, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7088517546653748, + "num_tokens": 252341554.0, + "step": 9750 + }, + { + "epoch": 1.0708324181858115, + "ewc_loss": 5.125999450683594e-06, + "grad_norm": 2.1767213344573975, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7084068059921265, + "num_tokens": 252369884.0, + "step": 9751 + }, + { + "epoch": 1.0709422358884253, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.206329345703125, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.723182737827301, + "num_tokens": 252395945.0, + "step": 9752 + }, + { + "epoch": 1.0710520535910388, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2658135890960693, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7204295992851257, + "num_tokens": 252421213.0, + "step": 9753 + }, + { + "epoch": 1.0711618712936526, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.37749981880188, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7170383930206299, + "num_tokens": 252444980.0, + "step": 9754 + }, + { + "epoch": 1.0712716889962661, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2815747261047363, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7115392684936523, + "num_tokens": 252472709.0, + "step": 9755 + }, + { + "epoch": 1.0713815066988799, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 1.9671515226364136, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7175438404083252, + "num_tokens": 252502923.0, + "step": 9756 + }, + { + "epoch": 1.0714913244014934, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1500985622406006, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7105509042739868, + "num_tokens": 252530288.0, + "step": 9757 + }, + { + "epoch": 1.0716011421041072, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.056917667388916, + "learning_rate": 1e-06, + "loss": 1.0198, + "mean_token_accuracy": 0.6962046027183533, + "num_tokens": 252561438.0, + "step": 9758 + }, + { + "epoch": 1.071710959806721, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.0354788303375244, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6945829391479492, + "num_tokens": 252591279.0, + "step": 9759 + }, + { + "epoch": 1.0718207775093345, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1049509048461914, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7218960523605347, + "num_tokens": 252621008.0, + "step": 9760 + }, + { + "epoch": 1.0719305952119482, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.240719795227051, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7134684324264526, + "num_tokens": 252647364.0, + "step": 9761 + }, + { + "epoch": 1.0720404129145618, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.3833212852478027, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7243558168411255, + "num_tokens": 252670834.0, + "step": 9762 + }, + { + "epoch": 1.0721502306171755, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.296126365661621, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7061581611633301, + "num_tokens": 252695647.0, + "step": 9763 + }, + { + "epoch": 1.072260048319789, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.0400843620300293, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7196738123893738, + "num_tokens": 252725425.0, + "step": 9764 + }, + { + "epoch": 1.0723698660224028, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.021758794784546, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7016985416412354, + "num_tokens": 252755765.0, + "step": 9765 + }, + { + "epoch": 1.0724796837250166, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.0821821689605713, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.720168948173523, + "num_tokens": 252784490.0, + "step": 9766 + }, + { + "epoch": 1.07258950142763, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.168180227279663, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7247456312179565, + "num_tokens": 252812843.0, + "step": 9767 + }, + { + "epoch": 1.0726993191302439, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.546037197113037, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.720990777015686, + "num_tokens": 252834022.0, + "step": 9768 + }, + { + "epoch": 1.0728091368328574, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2892866134643555, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7198977470397949, + "num_tokens": 252860473.0, + "step": 9769 + }, + { + "epoch": 1.0729189545354711, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.1117098331451416, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6905565857887268, + "num_tokens": 252892540.0, + "step": 9770 + }, + { + "epoch": 1.0730287722380847, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 1.9405927658081055, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7210625410079956, + "num_tokens": 252925170.0, + "step": 9771 + }, + { + "epoch": 1.0731385899406984, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.483952522277832, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7237608432769775, + "num_tokens": 252946526.0, + "step": 9772 + }, + { + "epoch": 1.0732484076433122, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.410982847213745, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.700173020362854, + "num_tokens": 252974024.0, + "step": 9773 + }, + { + "epoch": 1.0733582253459257, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.4354050159454346, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7203953266143799, + "num_tokens": 252997552.0, + "step": 9774 + }, + { + "epoch": 1.0734680430485395, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.2105278968811035, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7179009914398193, + "num_tokens": 253024018.0, + "step": 9775 + }, + { + "epoch": 1.073577860751153, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 1.8101599216461182, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7460159063339233, + "num_tokens": 253059439.0, + "step": 9776 + }, + { + "epoch": 1.0736876784537668, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 1.8872328996658325, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7084258794784546, + "num_tokens": 253092060.0, + "step": 9777 + }, + { + "epoch": 1.0737974961563803, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.2720065116882324, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7212328910827637, + "num_tokens": 253116732.0, + "step": 9778 + }, + { + "epoch": 1.073907313858994, + "ewc_loss": 5.155801773071289e-06, + "grad_norm": 2.4819164276123047, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7134304046630859, + "num_tokens": 253137444.0, + "step": 9779 + }, + { + "epoch": 1.0740171315616078, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.3462178707122803, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7094736099243164, + "num_tokens": 253163728.0, + "step": 9780 + }, + { + "epoch": 1.0741269492642214, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.552950382232666, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7153230905532837, + "num_tokens": 253185007.0, + "step": 9781 + }, + { + "epoch": 1.0742367669668351, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.230997323989868, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.717721164226532, + "num_tokens": 253210039.0, + "step": 9782 + }, + { + "epoch": 1.0743465846694487, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.30710768699646, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7123663425445557, + "num_tokens": 253233730.0, + "step": 9783 + }, + { + "epoch": 1.0744564023720624, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.5208394527435303, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7464864253997803, + "num_tokens": 253253245.0, + "step": 9784 + }, + { + "epoch": 1.074566220074676, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.311779022216797, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7395046353340149, + "num_tokens": 253276720.0, + "step": 9785 + }, + { + "epoch": 1.0746760377772897, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.0770537853240967, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7123657464981079, + "num_tokens": 253304939.0, + "step": 9786 + }, + { + "epoch": 1.0747858554799035, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.0275044441223145, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.7018149495124817, + "num_tokens": 253337224.0, + "step": 9787 + }, + { + "epoch": 1.074895673182517, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.4020307064056396, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.717754065990448, + "num_tokens": 253361449.0, + "step": 9788 + }, + { + "epoch": 1.0750054908851308, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 2.5240137577056885, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7320132255554199, + "num_tokens": 253382157.0, + "step": 9789 + }, + { + "epoch": 1.0751153085877443, + "ewc_loss": 5.185604095458984e-06, + "grad_norm": 1.9309784173965454, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7079617977142334, + "num_tokens": 253416618.0, + "step": 9790 + }, + { + "epoch": 1.075225126290358, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.4187357425689697, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7311751842498779, + "num_tokens": 253437656.0, + "step": 9791 + }, + { + "epoch": 1.0753349439929716, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.0827314853668213, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7107644081115723, + "num_tokens": 253465676.0, + "step": 9792 + }, + { + "epoch": 1.0754447616955853, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3753018379211426, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7186732292175293, + "num_tokens": 253488247.0, + "step": 9793 + }, + { + "epoch": 1.075554579398199, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.5080337524414062, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7160472273826599, + "num_tokens": 253509582.0, + "step": 9794 + }, + { + "epoch": 1.0756643971008126, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3491365909576416, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7112019062042236, + "num_tokens": 253533499.0, + "step": 9795 + }, + { + "epoch": 1.0757742148034264, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.157710075378418, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.719248354434967, + "num_tokens": 253559942.0, + "step": 9796 + }, + { + "epoch": 1.07588403250604, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.2063183784484863, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7245984673500061, + "num_tokens": 253587048.0, + "step": 9797 + }, + { + "epoch": 1.0759938502086537, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3092308044433594, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7136151194572449, + "num_tokens": 253614210.0, + "step": 9798 + }, + { + "epoch": 1.0761036679112672, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 1.9736419916152954, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7162778377532959, + "num_tokens": 253645388.0, + "step": 9799 + }, + { + "epoch": 1.076213485613881, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.511371374130249, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7290850877761841, + "num_tokens": 253666976.0, + "step": 9800 + }, + { + "epoch": 1.0763233033164945, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.0918586254119873, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6949141025543213, + "num_tokens": 253695884.0, + "step": 9801 + }, + { + "epoch": 1.0764331210191083, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 1.8735398054122925, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.6981271505355835, + "num_tokens": 253730963.0, + "step": 9802 + }, + { + "epoch": 1.076542938721722, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.350780487060547, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7256914377212524, + "num_tokens": 253754933.0, + "step": 9803 + }, + { + "epoch": 1.0766527564243356, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.2467541694641113, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7063997983932495, + "num_tokens": 253782586.0, + "step": 9804 + }, + { + "epoch": 1.0767625741269493, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.000347375869751, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7245026230812073, + "num_tokens": 253816135.0, + "step": 9805 + }, + { + "epoch": 1.0768723918295628, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.261626720428467, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7354022860527039, + "num_tokens": 253840466.0, + "step": 9806 + }, + { + "epoch": 1.0769822095321766, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.174893617630005, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7230604290962219, + "num_tokens": 253866512.0, + "step": 9807 + }, + { + "epoch": 1.0770920272347904, + "ewc_loss": 5.245208740234375e-06, + "grad_norm": 2.842074155807495, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7398175001144409, + "num_tokens": 253890883.0, + "step": 9808 + }, + { + "epoch": 1.077201844937404, + "ewc_loss": 5.245208740234375e-06, + "grad_norm": 2.4512693881988525, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.724470853805542, + "num_tokens": 253913297.0, + "step": 9809 + }, + { + "epoch": 1.0773116626400177, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.12483811378479, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7132999897003174, + "num_tokens": 253941360.0, + "step": 9810 + }, + { + "epoch": 1.0774214803426312, + "ewc_loss": 5.245208740234375e-06, + "grad_norm": 1.9915282726287842, + "learning_rate": 1e-06, + "loss": 0.997, + "mean_token_accuracy": 0.6928174495697021, + "num_tokens": 253971636.0, + "step": 9811 + }, + { + "epoch": 1.077531298045245, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 2.3037075996398926, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7189608812332153, + "num_tokens": 253996545.0, + "step": 9812 + }, + { + "epoch": 1.0776411157478585, + "ewc_loss": 5.21540641784668e-06, + "grad_norm": 1.9852558374404907, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7003233432769775, + "num_tokens": 254030517.0, + "step": 9813 + }, + { + "epoch": 1.0777509334504722, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.6106767654418945, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.712922215461731, + "num_tokens": 254052332.0, + "step": 9814 + }, + { + "epoch": 1.0778607511530858, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3908989429473877, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7186882495880127, + "num_tokens": 254075027.0, + "step": 9815 + }, + { + "epoch": 1.0779705688556995, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.145921468734741, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7058441638946533, + "num_tokens": 254105596.0, + "step": 9816 + }, + { + "epoch": 1.0780803865583133, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.304168224334717, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7175673246383667, + "num_tokens": 254129912.0, + "step": 9817 + }, + { + "epoch": 1.0781902042609268, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.287468194961548, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7396931648254395, + "num_tokens": 254152853.0, + "step": 9818 + }, + { + "epoch": 1.0783000219635406, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.242199659347534, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7162297368049622, + "num_tokens": 254179367.0, + "step": 9819 + }, + { + "epoch": 1.0784098396661541, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2006616592407227, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.702890157699585, + "num_tokens": 254205615.0, + "step": 9820 + }, + { + "epoch": 1.0785196573687679, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1538617610931396, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7294056415557861, + "num_tokens": 254232185.0, + "step": 9821 + }, + { + "epoch": 1.0786294750713814, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1865906715393066, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.7075501680374146, + "num_tokens": 254258918.0, + "step": 9822 + }, + { + "epoch": 1.0787392927739952, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.4743478298187256, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7313756942749023, + "num_tokens": 254281886.0, + "step": 9823 + }, + { + "epoch": 1.078849110476609, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.07582950592041, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.705846905708313, + "num_tokens": 254312053.0, + "step": 9824 + }, + { + "epoch": 1.0789589281792225, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3468215465545654, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7293224334716797, + "num_tokens": 254335024.0, + "step": 9825 + }, + { + "epoch": 1.0790687458818362, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.158787727355957, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7129213213920593, + "num_tokens": 254361962.0, + "step": 9826 + }, + { + "epoch": 1.0791785635844497, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 1.8718022108078003, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7144827842712402, + "num_tokens": 254396842.0, + "step": 9827 + }, + { + "epoch": 1.0792883812870635, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2642059326171875, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7136267423629761, + "num_tokens": 254422731.0, + "step": 9828 + }, + { + "epoch": 1.079398198989677, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.185770034790039, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7346422672271729, + "num_tokens": 254448617.0, + "step": 9829 + }, + { + "epoch": 1.0795080166922908, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.585738182067871, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7070705890655518, + "num_tokens": 254468811.0, + "step": 9830 + }, + { + "epoch": 1.0796178343949046, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3774662017822266, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7137572169303894, + "num_tokens": 254492675.0, + "step": 9831 + }, + { + "epoch": 1.079727652097518, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.6234536170959473, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7223559021949768, + "num_tokens": 254512410.0, + "step": 9832 + }, + { + "epoch": 1.0798374698001318, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.365853786468506, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.717276930809021, + "num_tokens": 254535716.0, + "step": 9833 + }, + { + "epoch": 1.0799472875027454, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.4768381118774414, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7285137176513672, + "num_tokens": 254557037.0, + "step": 9834 + }, + { + "epoch": 1.0800571052053591, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2718257904052734, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.6944675445556641, + "num_tokens": 254584259.0, + "step": 9835 + }, + { + "epoch": 1.0801669229079727, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.235036849975586, + "learning_rate": 1e-06, + "loss": 1.0266, + "mean_token_accuracy": 0.6813175678253174, + "num_tokens": 254614845.0, + "step": 9836 + }, + { + "epoch": 1.0802767406105864, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2042455673217773, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7265220880508423, + "num_tokens": 254640490.0, + "step": 9837 + }, + { + "epoch": 1.0803865583132002, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1460089683532715, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7199666500091553, + "num_tokens": 254667390.0, + "step": 9838 + }, + { + "epoch": 1.0804963760158137, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.4161980152130127, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.70409095287323, + "num_tokens": 254690296.0, + "step": 9839 + }, + { + "epoch": 1.0806061937184275, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.401487112045288, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7051911354064941, + "num_tokens": 254715809.0, + "step": 9840 + }, + { + "epoch": 1.080716011421041, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2573814392089844, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7496844530105591, + "num_tokens": 254741448.0, + "step": 9841 + }, + { + "epoch": 1.0808258291236548, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.203991651535034, + "learning_rate": 1e-06, + "loss": 0.9958, + "mean_token_accuracy": 0.7067631483078003, + "num_tokens": 254767857.0, + "step": 9842 + }, + { + "epoch": 1.0809356468262683, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.342926025390625, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7227027416229248, + "num_tokens": 254792749.0, + "step": 9843 + }, + { + "epoch": 1.081045464528882, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.005230665206909, + "learning_rate": 1e-06, + "loss": 1.0313, + "mean_token_accuracy": 0.6942828297615051, + "num_tokens": 254827509.0, + "step": 9844 + }, + { + "epoch": 1.0811552822314958, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2643918991088867, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7459145784378052, + "num_tokens": 254849998.0, + "step": 9845 + }, + { + "epoch": 1.0812650999341094, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.4377329349517822, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7065650224685669, + "num_tokens": 254873692.0, + "step": 9846 + }, + { + "epoch": 1.081374917636723, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.168018341064453, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7287504076957703, + "num_tokens": 254901458.0, + "step": 9847 + }, + { + "epoch": 1.0814847353393366, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.445369005203247, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7199887037277222, + "num_tokens": 254922737.0, + "step": 9848 + }, + { + "epoch": 1.0815945530419504, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1902122497558594, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7075619697570801, + "num_tokens": 254951216.0, + "step": 9849 + }, + { + "epoch": 1.081704370744564, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3626208305358887, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.715185284614563, + "num_tokens": 254975725.0, + "step": 9850 + }, + { + "epoch": 1.0818141884471777, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 1.8800725936889648, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6993890404701233, + "num_tokens": 255012158.0, + "step": 9851 + }, + { + "epoch": 1.0819240061497912, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.840573787689209, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7395530939102173, + "num_tokens": 255028102.0, + "step": 9852 + }, + { + "epoch": 1.082033823852405, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2027196884155273, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7296895384788513, + "num_tokens": 255053098.0, + "step": 9853 + }, + { + "epoch": 1.0821436415550187, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1212003231048584, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7520988583564758, + "num_tokens": 255079204.0, + "step": 9854 + }, + { + "epoch": 1.0822534592576323, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 1.9557207822799683, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7255876064300537, + "num_tokens": 255109320.0, + "step": 9855 + }, + { + "epoch": 1.082363276960246, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 1.9138379096984863, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6963609457015991, + "num_tokens": 255142634.0, + "step": 9856 + }, + { + "epoch": 1.0824730946628596, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.308069944381714, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.745162844657898, + "num_tokens": 255166313.0, + "step": 9857 + }, + { + "epoch": 1.0825829123654733, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.260923147201538, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7138315439224243, + "num_tokens": 255194625.0, + "step": 9858 + }, + { + "epoch": 1.082692730068087, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.150960683822632, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7318639159202576, + "num_tokens": 255220850.0, + "step": 9859 + }, + { + "epoch": 1.0828025477707006, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.5015954971313477, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7264441251754761, + "num_tokens": 255243028.0, + "step": 9860 + }, + { + "epoch": 1.0829123654733144, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1923537254333496, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.6950681209564209, + "num_tokens": 255272468.0, + "step": 9861 + }, + { + "epoch": 1.083022183175928, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.417513847351074, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7189604640007019, + "num_tokens": 255296927.0, + "step": 9862 + }, + { + "epoch": 1.0831320008785417, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.130984306335449, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7419278621673584, + "num_tokens": 255323354.0, + "step": 9863 + }, + { + "epoch": 1.0832418185811552, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.001826047897339, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.6965550184249878, + "num_tokens": 255356309.0, + "step": 9864 + }, + { + "epoch": 1.083351636283769, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1507632732391357, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7154468894004822, + "num_tokens": 255383162.0, + "step": 9865 + }, + { + "epoch": 1.0834614539863825, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.287721633911133, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7051609754562378, + "num_tokens": 255408460.0, + "step": 9866 + }, + { + "epoch": 1.0835712716889963, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.2821133136749268, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7192255258560181, + "num_tokens": 255434169.0, + "step": 9867 + }, + { + "epoch": 1.08368108939161, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.545074462890625, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7079905271530151, + "num_tokens": 255454970.0, + "step": 9868 + }, + { + "epoch": 1.0837909070942235, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.4605588912963867, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7411348223686218, + "num_tokens": 255475839.0, + "step": 9869 + }, + { + "epoch": 1.0839007247968373, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.0844717025756836, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.6973109245300293, + "num_tokens": 255505850.0, + "step": 9870 + }, + { + "epoch": 1.0840105424994508, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1091506481170654, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.6997983455657959, + "num_tokens": 255536149.0, + "step": 9871 + }, + { + "epoch": 1.0841203602020646, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2525293827056885, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7243610620498657, + "num_tokens": 255561334.0, + "step": 9872 + }, + { + "epoch": 1.0842301779046783, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.146406650543213, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.715259850025177, + "num_tokens": 255589566.0, + "step": 9873 + }, + { + "epoch": 1.0843399956072919, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.3087317943573, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7407506704330444, + "num_tokens": 255613867.0, + "step": 9874 + }, + { + "epoch": 1.0844498133099056, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.5587990283966064, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7280160188674927, + "num_tokens": 255634796.0, + "step": 9875 + }, + { + "epoch": 1.0845596310125192, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 1.977533221244812, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7009406089782715, + "num_tokens": 255666912.0, + "step": 9876 + }, + { + "epoch": 1.084669448715133, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2428252696990967, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7070889472961426, + "num_tokens": 255693497.0, + "step": 9877 + }, + { + "epoch": 1.0847792664177465, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2165467739105225, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7079843282699585, + "num_tokens": 255719485.0, + "step": 9878 + }, + { + "epoch": 1.0848890841203602, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.24411940574646, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7181998491287231, + "num_tokens": 255743656.0, + "step": 9879 + }, + { + "epoch": 1.0849989018229738, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.012906074523926, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7113215327262878, + "num_tokens": 255774053.0, + "step": 9880 + }, + { + "epoch": 1.0851087195255875, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.599836587905884, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7029274702072144, + "num_tokens": 255794035.0, + "step": 9881 + }, + { + "epoch": 1.0852185372282013, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.1578454971313477, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7485276460647583, + "num_tokens": 255819485.0, + "step": 9882 + }, + { + "epoch": 1.0853283549308148, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.352980852127075, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6917281746864319, + "num_tokens": 255844570.0, + "step": 9883 + }, + { + "epoch": 1.0854381726334286, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.1900453567504883, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6940175294876099, + "num_tokens": 255873762.0, + "step": 9884 + }, + { + "epoch": 1.085547990336042, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.482805013656616, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7186762690544128, + "num_tokens": 255896096.0, + "step": 9885 + }, + { + "epoch": 1.0856578080386559, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.0285704135894775, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6879127025604248, + "num_tokens": 255927815.0, + "step": 9886 + }, + { + "epoch": 1.0857676257412694, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3528501987457275, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7308244109153748, + "num_tokens": 255949988.0, + "step": 9887 + }, + { + "epoch": 1.0858774434438832, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.5577797889709473, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7498908042907715, + "num_tokens": 255968301.0, + "step": 9888 + }, + { + "epoch": 1.085987261146497, + "ewc_loss": 5.27501106262207e-06, + "grad_norm": 2.3963754177093506, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7199082374572754, + "num_tokens": 255991431.0, + "step": 9889 + }, + { + "epoch": 1.0860970788491104, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0204901695251465, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6956836581230164, + "num_tokens": 256023317.0, + "step": 9890 + }, + { + "epoch": 1.0862068965517242, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.81284761428833, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7188076972961426, + "num_tokens": 256042239.0, + "step": 9891 + }, + { + "epoch": 1.0863167142543377, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1906847953796387, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7422126531600952, + "num_tokens": 256069235.0, + "step": 9892 + }, + { + "epoch": 1.0864265319569515, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.329096794128418, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7223787307739258, + "num_tokens": 256094727.0, + "step": 9893 + }, + { + "epoch": 1.086536349659565, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.423229932785034, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.733698308467865, + "num_tokens": 256116410.0, + "step": 9894 + }, + { + "epoch": 1.0866461673621788, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2943360805511475, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7334758043289185, + "num_tokens": 256140523.0, + "step": 9895 + }, + { + "epoch": 1.0867559850647925, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.0449512004852295, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7298123240470886, + "num_tokens": 256173084.0, + "step": 9896 + }, + { + "epoch": 1.086865802767406, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.4373996257781982, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7094864845275879, + "num_tokens": 256197211.0, + "step": 9897 + }, + { + "epoch": 1.0869756204700198, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.326151132583618, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7269749045372009, + "num_tokens": 256219446.0, + "step": 9898 + }, + { + "epoch": 1.0870854381726334, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.5056393146514893, + "learning_rate": 1e-06, + "loss": 0.8133, + "mean_token_accuracy": 0.7398461103439331, + "num_tokens": 256240081.0, + "step": 9899 + }, + { + "epoch": 1.0871952558752471, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.079893112182617, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7309489250183105, + "num_tokens": 256268252.0, + "step": 9900 + }, + { + "epoch": 1.0873050735778607, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.568960189819336, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7471145391464233, + "num_tokens": 256288236.0, + "step": 9901 + }, + { + "epoch": 1.0874148912804744, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.6058108806610107, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7352747321128845, + "num_tokens": 256307306.0, + "step": 9902 + }, + { + "epoch": 1.0875247089830882, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.067516565322876, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.6926960945129395, + "num_tokens": 256336305.0, + "step": 9903 + }, + { + "epoch": 1.0876345266857017, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3168423175811768, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7377527952194214, + "num_tokens": 256360289.0, + "step": 9904 + }, + { + "epoch": 1.0877443443883155, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0101873874664307, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7216731309890747, + "num_tokens": 256390078.0, + "step": 9905 + }, + { + "epoch": 1.087854162090929, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.111440420150757, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7116841077804565, + "num_tokens": 256420136.0, + "step": 9906 + }, + { + "epoch": 1.0879639797935428, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1131703853607178, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7219363451004028, + "num_tokens": 256448256.0, + "step": 9907 + }, + { + "epoch": 1.0880737974961563, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.094423770904541, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7057244777679443, + "num_tokens": 256477398.0, + "step": 9908 + }, + { + "epoch": 1.08818361519877, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.413440227508545, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7359742522239685, + "num_tokens": 256499772.0, + "step": 9909 + }, + { + "epoch": 1.0882934329013838, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.140989065170288, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7163289785385132, + "num_tokens": 256528746.0, + "step": 9910 + }, + { + "epoch": 1.0884032506039973, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4559762477874756, + "learning_rate": 1e-06, + "loss": 1.0306, + "mean_token_accuracy": 0.6846559643745422, + "num_tokens": 256553700.0, + "step": 9911 + }, + { + "epoch": 1.088513068306611, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1459827423095703, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7265763282775879, + "num_tokens": 256581267.0, + "step": 9912 + }, + { + "epoch": 1.0886228860092246, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2062432765960693, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7018293142318726, + "num_tokens": 256606458.0, + "step": 9913 + }, + { + "epoch": 1.0887327037118384, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3213412761688232, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7237029671669006, + "num_tokens": 256631026.0, + "step": 9914 + }, + { + "epoch": 1.088842521414452, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0738914012908936, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7091092467308044, + "num_tokens": 256659438.0, + "step": 9915 + }, + { + "epoch": 1.0889523391170657, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.180039167404175, + "learning_rate": 1e-06, + "loss": 1.0106, + "mean_token_accuracy": 0.6881132125854492, + "num_tokens": 256690266.0, + "step": 9916 + }, + { + "epoch": 1.0890621568196792, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1354804039001465, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7302442193031311, + "num_tokens": 256719995.0, + "step": 9917 + }, + { + "epoch": 1.089171974522293, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.5081589221954346, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7142433524131775, + "num_tokens": 256741027.0, + "step": 9918 + }, + { + "epoch": 1.0892817922249067, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.402010679244995, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6887797713279724, + "num_tokens": 256764856.0, + "step": 9919 + }, + { + "epoch": 1.0893916099275203, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1867406368255615, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7044942378997803, + "num_tokens": 256791611.0, + "step": 9920 + }, + { + "epoch": 1.089501427630134, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.353058338165283, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7136013507843018, + "num_tokens": 256816799.0, + "step": 9921 + }, + { + "epoch": 1.0896112453327476, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0912466049194336, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6940334439277649, + "num_tokens": 256844587.0, + "step": 9922 + }, + { + "epoch": 1.0897210630353613, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.246614456176758, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7170305848121643, + "num_tokens": 256869769.0, + "step": 9923 + }, + { + "epoch": 1.089830880737975, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.056107521057129, + "learning_rate": 1e-06, + "loss": 1.0145, + "mean_token_accuracy": 0.694348931312561, + "num_tokens": 256902427.0, + "step": 9924 + }, + { + "epoch": 1.0899406984405886, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3434062004089355, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7136561870574951, + "num_tokens": 256928302.0, + "step": 9925 + }, + { + "epoch": 1.0900505161432024, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.5113024711608887, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7101318836212158, + "num_tokens": 256949129.0, + "step": 9926 + }, + { + "epoch": 1.090160333845816, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.132688522338867, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7161685228347778, + "num_tokens": 256979446.0, + "step": 9927 + }, + { + "epoch": 1.0902701515484297, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.233743906021118, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7200075387954712, + "num_tokens": 257006538.0, + "step": 9928 + }, + { + "epoch": 1.0903799692510432, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.266869306564331, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7361325621604919, + "num_tokens": 257030709.0, + "step": 9929 + }, + { + "epoch": 1.090489786953657, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3011302947998047, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7274636030197144, + "num_tokens": 257054215.0, + "step": 9930 + }, + { + "epoch": 1.0905996046562705, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.159872055053711, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7221153974533081, + "num_tokens": 257080815.0, + "step": 9931 + }, + { + "epoch": 1.0907094223588842, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2020576000213623, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7274038791656494, + "num_tokens": 257105877.0, + "step": 9932 + }, + { + "epoch": 1.090819240061498, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.454214334487915, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7225034832954407, + "num_tokens": 257128601.0, + "step": 9933 + }, + { + "epoch": 1.0909290577641115, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4800920486450195, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7301028370857239, + "num_tokens": 257149597.0, + "step": 9934 + }, + { + "epoch": 1.0910388754667253, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.44754695892334, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7126055359840393, + "num_tokens": 257171312.0, + "step": 9935 + }, + { + "epoch": 1.0911486931693388, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.004356622695923, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7065624594688416, + "num_tokens": 257204631.0, + "step": 9936 + }, + { + "epoch": 1.0912585108719526, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3653531074523926, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.7051130533218384, + "num_tokens": 257228960.0, + "step": 9937 + }, + { + "epoch": 1.0913683285745663, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4084622859954834, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7280123829841614, + "num_tokens": 257251455.0, + "step": 9938 + }, + { + "epoch": 1.0914781462771799, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3283116817474365, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7177504897117615, + "num_tokens": 257277132.0, + "step": 9939 + }, + { + "epoch": 1.0915879639797936, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.284651041030884, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7158463597297668, + "num_tokens": 257301807.0, + "step": 9940 + }, + { + "epoch": 1.0916977816824072, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.963565468788147, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7089341878890991, + "num_tokens": 257333641.0, + "step": 9941 + }, + { + "epoch": 1.091807599385021, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.091446876525879, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7287584543228149, + "num_tokens": 257361424.0, + "step": 9942 + }, + { + "epoch": 1.0919174170876345, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.454946517944336, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7145655155181885, + "num_tokens": 257383212.0, + "step": 9943 + }, + { + "epoch": 1.0920272347902482, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0795133113861084, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7106939554214478, + "num_tokens": 257412798.0, + "step": 9944 + }, + { + "epoch": 1.0921370524928617, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2141706943511963, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7138755321502686, + "num_tokens": 257440412.0, + "step": 9945 + }, + { + "epoch": 1.0922468701954755, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1506989002227783, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7215856313705444, + "num_tokens": 257468699.0, + "step": 9946 + }, + { + "epoch": 1.0923566878980893, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.046564817428589, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7062913179397583, + "num_tokens": 257497606.0, + "step": 9947 + }, + { + "epoch": 1.0924665056007028, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3584868907928467, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7304777503013611, + "num_tokens": 257520184.0, + "step": 9948 + }, + { + "epoch": 1.0925763233033166, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1860077381134033, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6962321996688843, + "num_tokens": 257550058.0, + "step": 9949 + }, + { + "epoch": 1.09268614100593, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1702568531036377, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7133732438087463, + "num_tokens": 257577836.0, + "step": 9950 + }, + { + "epoch": 1.0927959587085438, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.8307245969772339, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7035893201828003, + "num_tokens": 257614681.0, + "step": 9951 + }, + { + "epoch": 1.0929057764111574, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3051040172576904, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7198184132575989, + "num_tokens": 257638113.0, + "step": 9952 + }, + { + "epoch": 1.0930155941137711, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2981598377227783, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7318094968795776, + "num_tokens": 257661459.0, + "step": 9953 + }, + { + "epoch": 1.093125411816385, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.9368207454681396, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6995270252227783, + "num_tokens": 257694517.0, + "step": 9954 + }, + { + "epoch": 1.0932352295189984, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.10774827003479, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7292263507843018, + "num_tokens": 257721198.0, + "step": 9955 + }, + { + "epoch": 1.0933450472216122, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3477303981781006, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7396067380905151, + "num_tokens": 257741074.0, + "step": 9956 + }, + { + "epoch": 1.0934548649242257, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.235962152481079, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7088603377342224, + "num_tokens": 257767076.0, + "step": 9957 + }, + { + "epoch": 1.0935646826268395, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0447330474853516, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6925535798072815, + "num_tokens": 257798867.0, + "step": 9958 + }, + { + "epoch": 1.093674500329453, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.477853298187256, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7280210852622986, + "num_tokens": 257821000.0, + "step": 9959 + }, + { + "epoch": 1.0937843180320668, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2157418727874756, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7378746271133423, + "num_tokens": 257845751.0, + "step": 9960 + }, + { + "epoch": 1.0938941357346805, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.238891839981079, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7231276035308838, + "num_tokens": 257870734.0, + "step": 9961 + }, + { + "epoch": 1.094003953437294, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4871397018432617, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.711555540561676, + "num_tokens": 257892104.0, + "step": 9962 + }, + { + "epoch": 1.0941137711399078, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.269704818725586, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7072175741195679, + "num_tokens": 257918169.0, + "step": 9963 + }, + { + "epoch": 1.0942235888425214, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.6167726516723633, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7265029549598694, + "num_tokens": 257939361.0, + "step": 9964 + }, + { + "epoch": 1.0943334065451351, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4132678508758545, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7208214402198792, + "num_tokens": 257962990.0, + "step": 9965 + }, + { + "epoch": 1.0944432242477486, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.856626510620117, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7521812319755554, + "num_tokens": 257980096.0, + "step": 9966 + }, + { + "epoch": 1.0945530419503624, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.133349657058716, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7394962310791016, + "num_tokens": 258007207.0, + "step": 9967 + }, + { + "epoch": 1.0946628596529762, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.8579216003417969, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.7005698680877686, + "num_tokens": 258043075.0, + "step": 9968 + }, + { + "epoch": 1.0947726773555897, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.997214436531067, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7392683029174805, + "num_tokens": 258073828.0, + "step": 9969 + }, + { + "epoch": 1.0948824950582035, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3291356563568115, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7042489051818848, + "num_tokens": 258100626.0, + "step": 9970 + }, + { + "epoch": 1.094992312760817, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.16511869430542, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7544742822647095, + "num_tokens": 258125951.0, + "step": 9971 + }, + { + "epoch": 1.0951021304634307, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0928966999053955, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7362743616104126, + "num_tokens": 258151984.0, + "step": 9972 + }, + { + "epoch": 1.0952119481660443, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.231502056121826, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7013740539550781, + "num_tokens": 258177525.0, + "step": 9973 + }, + { + "epoch": 1.095321765868658, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.267986297607422, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.714250922203064, + "num_tokens": 258203254.0, + "step": 9974 + }, + { + "epoch": 1.0954315835712718, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.294329881668091, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7107338905334473, + "num_tokens": 258230192.0, + "step": 9975 + }, + { + "epoch": 1.0955414012738853, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.389833927154541, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7010064125061035, + "num_tokens": 258253607.0, + "step": 9976 + }, + { + "epoch": 1.095651218976499, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.460866689682007, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7193782925605774, + "num_tokens": 258276578.0, + "step": 9977 + }, + { + "epoch": 1.0957610366791126, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.287564277648926, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7145506143569946, + "num_tokens": 258302613.0, + "step": 9978 + }, + { + "epoch": 1.0958708543817264, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.8612343072891235, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7068050503730774, + "num_tokens": 258337946.0, + "step": 9979 + }, + { + "epoch": 1.09598067208434, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.0329489707946777, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7027376890182495, + "num_tokens": 258369426.0, + "step": 9980 + }, + { + "epoch": 1.0960904897869537, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4479820728302, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.722956120967865, + "num_tokens": 258391493.0, + "step": 9981 + }, + { + "epoch": 1.0962003074895672, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.182898998260498, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7150280475616455, + "num_tokens": 258417199.0, + "step": 9982 + }, + { + "epoch": 1.096310125192181, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.49534010887146, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7146332263946533, + "num_tokens": 258438100.0, + "step": 9983 + }, + { + "epoch": 1.0964199428947947, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2441630363464355, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7412888407707214, + "num_tokens": 258463474.0, + "step": 9984 + }, + { + "epoch": 1.0965297605974083, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.290764331817627, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7075968980789185, + "num_tokens": 258488155.0, + "step": 9985 + }, + { + "epoch": 1.096639578300022, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3093278408050537, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7241949439048767, + "num_tokens": 258511403.0, + "step": 9986 + }, + { + "epoch": 1.0967493960026355, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4843828678131104, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7043025493621826, + "num_tokens": 258533843.0, + "step": 9987 + }, + { + "epoch": 1.0968592137052493, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2097532749176025, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7221872806549072, + "num_tokens": 258559380.0, + "step": 9988 + }, + { + "epoch": 1.096969031407863, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.231224775314331, + "learning_rate": 1e-06, + "loss": 1.0514, + "mean_token_accuracy": 0.6816874146461487, + "num_tokens": 258589868.0, + "step": 9989 + }, + { + "epoch": 1.0970788491104766, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.365387201309204, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.736821711063385, + "num_tokens": 258611281.0, + "step": 9990 + }, + { + "epoch": 1.0971886668130904, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2037200927734375, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.7009339332580566, + "num_tokens": 258639167.0, + "step": 9991 + }, + { + "epoch": 1.0972984845157039, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2230613231658936, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.719297468662262, + "num_tokens": 258665202.0, + "step": 9992 + }, + { + "epoch": 1.0974083022183176, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.51312518119812, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7319067716598511, + "num_tokens": 258686204.0, + "step": 9993 + }, + { + "epoch": 1.0975181199209312, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.176511526107788, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7300418615341187, + "num_tokens": 258712260.0, + "step": 9994 + }, + { + "epoch": 1.097627937623545, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.6209418773651123, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7266978025436401, + "num_tokens": 258732364.0, + "step": 9995 + }, + { + "epoch": 1.0977377553261585, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2837986946105957, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7493422031402588, + "num_tokens": 258758137.0, + "step": 9996 + }, + { + "epoch": 1.0978475730287722, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.408407688140869, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7181944847106934, + "num_tokens": 258781894.0, + "step": 9997 + }, + { + "epoch": 1.097957390731386, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 1.8791985511779785, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7278255224227905, + "num_tokens": 258816433.0, + "step": 9998 + }, + { + "epoch": 1.0980672084339995, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2886741161346436, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7039777040481567, + "num_tokens": 258842293.0, + "step": 9999 + }, + { + "epoch": 1.0981770261366133, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.326669216156006, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7121554017066956, + "num_tokens": 258868498.0, + "step": 10000 + }, + { + "epoch": 1.0982868438392268, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3017313480377197, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7128972411155701, + "num_tokens": 258894056.0, + "step": 10001 + }, + { + "epoch": 1.0983966615418406, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.6450352668762207, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7437317371368408, + "num_tokens": 258913204.0, + "step": 10002 + }, + { + "epoch": 1.098506479244454, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.510894298553467, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7489563822746277, + "num_tokens": 258934560.0, + "step": 10003 + }, + { + "epoch": 1.0986162969470679, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.5611469745635986, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7380844354629517, + "num_tokens": 258957202.0, + "step": 10004 + }, + { + "epoch": 1.0987261146496816, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.577981472015381, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7246032953262329, + "num_tokens": 258978611.0, + "step": 10005 + }, + { + "epoch": 1.0988359323522952, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.293931484222412, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7139682769775391, + "num_tokens": 259004896.0, + "step": 10006 + }, + { + "epoch": 1.098945750054909, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.389232635498047, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7125886678695679, + "num_tokens": 259029899.0, + "step": 10007 + }, + { + "epoch": 1.0990555677575224, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.0021116733551025, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7261191606521606, + "num_tokens": 259058605.0, + "step": 10008 + }, + { + "epoch": 1.0991653854601362, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2515933513641357, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.717278003692627, + "num_tokens": 259082524.0, + "step": 10009 + }, + { + "epoch": 1.0992752031627497, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2983384132385254, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.723172128200531, + "num_tokens": 259107190.0, + "step": 10010 + }, + { + "epoch": 1.0993850208653635, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 1.957468867301941, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7194251418113708, + "num_tokens": 259140725.0, + "step": 10011 + }, + { + "epoch": 1.0994948385679773, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.303839683532715, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7015309929847717, + "num_tokens": 259166114.0, + "step": 10012 + }, + { + "epoch": 1.0996046562705908, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.3351879119873047, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.721928060054779, + "num_tokens": 259188512.0, + "step": 10013 + }, + { + "epoch": 1.0997144739732045, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.186936855316162, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7088357210159302, + "num_tokens": 259213322.0, + "step": 10014 + }, + { + "epoch": 1.099824291675818, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 1.955473780632019, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7115211486816406, + "num_tokens": 259244247.0, + "step": 10015 + }, + { + "epoch": 1.0999341093784318, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.1161439418792725, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7265887260437012, + "num_tokens": 259272335.0, + "step": 10016 + }, + { + "epoch": 1.1000439270810454, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.123826265335083, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7242552042007446, + "num_tokens": 259299824.0, + "step": 10017 + }, + { + "epoch": 1.1001537447836591, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.241867780685425, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7306061387062073, + "num_tokens": 259322970.0, + "step": 10018 + }, + { + "epoch": 1.1002635624862729, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.158576726913452, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7154118418693542, + "num_tokens": 259352116.0, + "step": 10019 + }, + { + "epoch": 1.1003733801888864, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.059657573699951, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.6976057291030884, + "num_tokens": 259383382.0, + "step": 10020 + }, + { + "epoch": 1.1004831978915002, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 1.9587998390197754, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.718052864074707, + "num_tokens": 259412079.0, + "step": 10021 + }, + { + "epoch": 1.1005930155941137, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 1.9984444379806519, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7128657102584839, + "num_tokens": 259445070.0, + "step": 10022 + }, + { + "epoch": 1.1007028332967275, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.251645803451538, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7190687656402588, + "num_tokens": 259473069.0, + "step": 10023 + }, + { + "epoch": 1.100812650999341, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2470028400421143, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7146742343902588, + "num_tokens": 259498640.0, + "step": 10024 + }, + { + "epoch": 1.1009224687019548, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.1986210346221924, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6940581798553467, + "num_tokens": 259525249.0, + "step": 10025 + }, + { + "epoch": 1.1010322864045685, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.013234853744507, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.695492684841156, + "num_tokens": 259558002.0, + "step": 10026 + }, + { + "epoch": 1.101142104107182, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.064023971557617, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7268061637878418, + "num_tokens": 259587005.0, + "step": 10027 + }, + { + "epoch": 1.1012519218097958, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.476356267929077, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7041783928871155, + "num_tokens": 259608555.0, + "step": 10028 + }, + { + "epoch": 1.1013617395124093, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.2740955352783203, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7159812450408936, + "num_tokens": 259632527.0, + "step": 10029 + }, + { + "epoch": 1.101471557215023, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.193699359893799, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7378457188606262, + "num_tokens": 259657114.0, + "step": 10030 + }, + { + "epoch": 1.1015813749176366, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.085031270980835, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7181020975112915, + "num_tokens": 259687056.0, + "step": 10031 + }, + { + "epoch": 1.1016911926202504, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.296628713607788, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7320001125335693, + "num_tokens": 259711041.0, + "step": 10032 + }, + { + "epoch": 1.1018010103228641, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.221010684967041, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7087509632110596, + "num_tokens": 259735799.0, + "step": 10033 + }, + { + "epoch": 1.1019108280254777, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.3767004013061523, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7212642431259155, + "num_tokens": 259759245.0, + "step": 10034 + }, + { + "epoch": 1.1020206457280914, + "ewc_loss": 5.304813385009766e-06, + "grad_norm": 2.242551565170288, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7100235223770142, + "num_tokens": 259786430.0, + "step": 10035 + }, + { + "epoch": 1.102130463430705, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3148975372314453, + "learning_rate": 1e-06, + "loss": 0.9686, + "mean_token_accuracy": 0.7040069699287415, + "num_tokens": 259812205.0, + "step": 10036 + }, + { + "epoch": 1.1022402811333187, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.4755618572235107, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.720119297504425, + "num_tokens": 259834090.0, + "step": 10037 + }, + { + "epoch": 1.1023500988359323, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.276472568511963, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7186863422393799, + "num_tokens": 259859684.0, + "step": 10038 + }, + { + "epoch": 1.102459916538546, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.137770891189575, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7297264337539673, + "num_tokens": 259887694.0, + "step": 10039 + }, + { + "epoch": 1.1025697342411598, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.199582815170288, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7069071531295776, + "num_tokens": 259915361.0, + "step": 10040 + }, + { + "epoch": 1.1026795519437733, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.376138687133789, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7094541192054749, + "num_tokens": 259939604.0, + "step": 10041 + }, + { + "epoch": 1.102789369646387, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2450695037841797, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7033255100250244, + "num_tokens": 259964570.0, + "step": 10042 + }, + { + "epoch": 1.1028991873490006, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.335326910018921, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7296496629714966, + "num_tokens": 259988233.0, + "step": 10043 + }, + { + "epoch": 1.1030090050516144, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.2050087451934814, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.7026436924934387, + "num_tokens": 260016806.0, + "step": 10044 + }, + { + "epoch": 1.103118822754228, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1980020999908447, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7101074457168579, + "num_tokens": 260044011.0, + "step": 10045 + }, + { + "epoch": 1.1032286404568417, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.3686532974243164, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7240867018699646, + "num_tokens": 260068734.0, + "step": 10046 + }, + { + "epoch": 1.1033384581594552, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.191707134246826, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7273617386817932, + "num_tokens": 260093825.0, + "step": 10047 + }, + { + "epoch": 1.103448275862069, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.097355604171753, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.6924020051956177, + "num_tokens": 260122935.0, + "step": 10048 + }, + { + "epoch": 1.1035580935646827, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.373878240585327, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7324039936065674, + "num_tokens": 260148211.0, + "step": 10049 + }, + { + "epoch": 1.1036679112672962, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.4342403411865234, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7287514209747314, + "num_tokens": 260171463.0, + "step": 10050 + }, + { + "epoch": 1.10377772896991, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.3194735050201416, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7344452738761902, + "num_tokens": 260194896.0, + "step": 10051 + }, + { + "epoch": 1.1038875466725235, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.4798648357391357, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7001185417175293, + "num_tokens": 260218383.0, + "step": 10052 + }, + { + "epoch": 1.1039973643751373, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.285383462905884, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7090137600898743, + "num_tokens": 260244991.0, + "step": 10053 + }, + { + "epoch": 1.104107182077751, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.244183301925659, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7232065200805664, + "num_tokens": 260271284.0, + "step": 10054 + }, + { + "epoch": 1.1042169997803646, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.4862780570983887, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.7016677856445312, + "num_tokens": 260294880.0, + "step": 10055 + }, + { + "epoch": 1.1043268174829783, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.2170889377593994, + "learning_rate": 1e-06, + "loss": 1.0087, + "mean_token_accuracy": 0.6874186992645264, + "num_tokens": 260322375.0, + "step": 10056 + }, + { + "epoch": 1.1044366351855919, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.1196136474609375, + "learning_rate": 1e-06, + "loss": 0.9898, + "mean_token_accuracy": 0.6970109939575195, + "num_tokens": 260352339.0, + "step": 10057 + }, + { + "epoch": 1.1045464528882056, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.4128119945526123, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7144902944564819, + "num_tokens": 260376081.0, + "step": 10058 + }, + { + "epoch": 1.1046562705908192, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.31136417388916, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7105308771133423, + "num_tokens": 260400360.0, + "step": 10059 + }, + { + "epoch": 1.104766088293433, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.214773654937744, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7201482057571411, + "num_tokens": 260427736.0, + "step": 10060 + }, + { + "epoch": 1.1048759059960465, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.287302017211914, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.711488664150238, + "num_tokens": 260451174.0, + "step": 10061 + }, + { + "epoch": 1.1049857236986602, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.378131866455078, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7158523797988892, + "num_tokens": 260474631.0, + "step": 10062 + }, + { + "epoch": 1.105095541401274, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.2431747913360596, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7347925305366516, + "num_tokens": 260499111.0, + "step": 10063 + }, + { + "epoch": 1.1052053591038875, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.417031764984131, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7271474003791809, + "num_tokens": 260520362.0, + "step": 10064 + }, + { + "epoch": 1.1053151768065013, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.1573095321655273, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7293248176574707, + "num_tokens": 260547511.0, + "step": 10065 + }, + { + "epoch": 1.1054249945091148, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.4188289642333984, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7440235614776611, + "num_tokens": 260569000.0, + "step": 10066 + }, + { + "epoch": 1.1055348122117286, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 1.8214386701583862, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7240761518478394, + "num_tokens": 260602601.0, + "step": 10067 + }, + { + "epoch": 1.105644629914342, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.167267084121704, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7035514712333679, + "num_tokens": 260631014.0, + "step": 10068 + }, + { + "epoch": 1.1057544476169558, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.2095131874084473, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7244178056716919, + "num_tokens": 260656564.0, + "step": 10069 + }, + { + "epoch": 1.1058642653195696, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.292863130569458, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7258801460266113, + "num_tokens": 260678529.0, + "step": 10070 + }, + { + "epoch": 1.1059740830221831, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.142534017562866, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7304972410202026, + "num_tokens": 260703650.0, + "step": 10071 + }, + { + "epoch": 1.106083900724797, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.409771203994751, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7202732563018799, + "num_tokens": 260725587.0, + "step": 10072 + }, + { + "epoch": 1.1061937184274104, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.301485300064087, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7396759390830994, + "num_tokens": 260749706.0, + "step": 10073 + }, + { + "epoch": 1.1063035361300242, + "ewc_loss": 5.364418029785156e-06, + "grad_norm": 2.215542793273926, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.737578272819519, + "num_tokens": 260775255.0, + "step": 10074 + }, + { + "epoch": 1.1064133538326377, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.1651406288146973, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7212820053100586, + "num_tokens": 260803395.0, + "step": 10075 + }, + { + "epoch": 1.1065231715352515, + "ewc_loss": 5.334615707397461e-06, + "grad_norm": 2.378330707550049, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7169678211212158, + "num_tokens": 260827511.0, + "step": 10076 + }, + { + "epoch": 1.1066329892378652, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.0475540161132812, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7198714017868042, + "num_tokens": 260857517.0, + "step": 10077 + }, + { + "epoch": 1.1067428069404788, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.398956775665283, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7254489064216614, + "num_tokens": 260881444.0, + "step": 10078 + }, + { + "epoch": 1.1068526246430925, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.2709920406341553, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7245166897773743, + "num_tokens": 260907718.0, + "step": 10079 + }, + { + "epoch": 1.106962442345706, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.2012457847595215, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7144671678543091, + "num_tokens": 260934543.0, + "step": 10080 + }, + { + "epoch": 1.1070722600483198, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.322695732116699, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7073209285736084, + "num_tokens": 260960660.0, + "step": 10081 + }, + { + "epoch": 1.1071820777509334, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.153122663497925, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7206146121025085, + "num_tokens": 260988653.0, + "step": 10082 + }, + { + "epoch": 1.1072918954535471, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.1364452838897705, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7092646360397339, + "num_tokens": 261017363.0, + "step": 10083 + }, + { + "epoch": 1.1074017131561609, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.214412212371826, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7119724750518799, + "num_tokens": 261044320.0, + "step": 10084 + }, + { + "epoch": 1.1075115308587744, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.1602132320404053, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.741108238697052, + "num_tokens": 261069403.0, + "step": 10085 + }, + { + "epoch": 1.1076213485613882, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.699524164199829, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7290472984313965, + "num_tokens": 261088395.0, + "step": 10086 + }, + { + "epoch": 1.1077311662640017, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.218935251235962, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7347443699836731, + "num_tokens": 261113142.0, + "step": 10087 + }, + { + "epoch": 1.1078409839666155, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.4134249687194824, + "learning_rate": 1e-06, + "loss": 1.0065, + "mean_token_accuracy": 0.6927134990692139, + "num_tokens": 261135076.0, + "step": 10088 + }, + { + "epoch": 1.107950801669229, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.214190721511841, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7123093605041504, + "num_tokens": 261159556.0, + "step": 10089 + }, + { + "epoch": 1.1080606193718427, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.398756504058838, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7264019250869751, + "num_tokens": 261185178.0, + "step": 10090 + }, + { + "epoch": 1.1081704370744565, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.181910514831543, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7430029511451721, + "num_tokens": 261210986.0, + "step": 10091 + }, + { + "epoch": 1.10828025477707, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 1.9893155097961426, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7035512328147888, + "num_tokens": 261242510.0, + "step": 10092 + }, + { + "epoch": 1.1083900724796838, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.215238332748413, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7190397381782532, + "num_tokens": 261267792.0, + "step": 10093 + }, + { + "epoch": 1.1084998901822973, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.286665439605713, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7285720705986023, + "num_tokens": 261293488.0, + "step": 10094 + }, + { + "epoch": 1.108609707884911, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.5708401203155518, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7253090739250183, + "num_tokens": 261313720.0, + "step": 10095 + }, + { + "epoch": 1.1087195255875246, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.7130415439605713, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7380131483078003, + "num_tokens": 261331486.0, + "step": 10096 + }, + { + "epoch": 1.1088293432901384, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.4181721210479736, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.718583345413208, + "num_tokens": 261353896.0, + "step": 10097 + }, + { + "epoch": 1.108939160992752, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.5996649265289307, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7132115364074707, + "num_tokens": 261376889.0, + "step": 10098 + }, + { + "epoch": 1.1090489786953657, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.6783905029296875, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7211567163467407, + "num_tokens": 261395160.0, + "step": 10099 + }, + { + "epoch": 1.1091587963979794, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.2414934635162354, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7236804366111755, + "num_tokens": 261421634.0, + "step": 10100 + }, + { + "epoch": 1.109268614100593, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1357057094573975, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7177380323410034, + "num_tokens": 261450939.0, + "step": 10101 + }, + { + "epoch": 1.1093784318032067, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.354632616043091, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7227635383605957, + "num_tokens": 261474246.0, + "step": 10102 + }, + { + "epoch": 1.1094882495058203, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.201258897781372, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7395967245101929, + "num_tokens": 261499381.0, + "step": 10103 + }, + { + "epoch": 1.109598067208434, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.023261308670044, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7251864075660706, + "num_tokens": 261529173.0, + "step": 10104 + }, + { + "epoch": 1.1097078849110478, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 1.9794782400131226, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7137261629104614, + "num_tokens": 261562491.0, + "step": 10105 + }, + { + "epoch": 1.1098177026136613, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.283024549484253, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7236230373382568, + "num_tokens": 261587022.0, + "step": 10106 + }, + { + "epoch": 1.109927520316275, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.2535953521728516, + "learning_rate": 1e-06, + "loss": 0.9903, + "mean_token_accuracy": 0.7089287042617798, + "num_tokens": 261613628.0, + "step": 10107 + }, + { + "epoch": 1.1100373380188886, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.3268368244171143, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7265928387641907, + "num_tokens": 261634952.0, + "step": 10108 + }, + { + "epoch": 1.1101471557215024, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.3352558612823486, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7288013100624084, + "num_tokens": 261657788.0, + "step": 10109 + }, + { + "epoch": 1.110256973424116, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 1.9933829307556152, + "learning_rate": 1e-06, + "loss": 1.0748, + "mean_token_accuracy": 0.6733791828155518, + "num_tokens": 261691464.0, + "step": 10110 + }, + { + "epoch": 1.1103667911267296, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.3673932552337646, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7325634360313416, + "num_tokens": 261713940.0, + "step": 10111 + }, + { + "epoch": 1.1104766088293432, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.272477388381958, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7313138246536255, + "num_tokens": 261739358.0, + "step": 10112 + }, + { + "epoch": 1.110586426531957, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.291355848312378, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7341889142990112, + "num_tokens": 261765350.0, + "step": 10113 + }, + { + "epoch": 1.1106962442345707, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.3485310077667236, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7182233333587646, + "num_tokens": 261790069.0, + "step": 10114 + }, + { + "epoch": 1.1108060619371842, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.2182388305664062, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7235866189002991, + "num_tokens": 261815604.0, + "step": 10115 + }, + { + "epoch": 1.110915879639798, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.189924716949463, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7051617503166199, + "num_tokens": 261841623.0, + "step": 10116 + }, + { + "epoch": 1.1110256973424115, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.5566112995147705, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7003153562545776, + "num_tokens": 261862501.0, + "step": 10117 + }, + { + "epoch": 1.1111355150450253, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1269538402557373, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.7011699080467224, + "num_tokens": 261890576.0, + "step": 10118 + }, + { + "epoch": 1.111245332747639, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.258654832839966, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7342035174369812, + "num_tokens": 261915171.0, + "step": 10119 + }, + { + "epoch": 1.1113551504502526, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.193988561630249, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.734305202960968, + "num_tokens": 261942007.0, + "step": 10120 + }, + { + "epoch": 1.1114649681528663, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.233661651611328, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7085098624229431, + "num_tokens": 261966752.0, + "step": 10121 + }, + { + "epoch": 1.1115747858554799, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.505051851272583, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7150418758392334, + "num_tokens": 261990830.0, + "step": 10122 + }, + { + "epoch": 1.1116846035580936, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.3033604621887207, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7056425213813782, + "num_tokens": 262015820.0, + "step": 10123 + }, + { + "epoch": 1.1117944212607072, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.6698219776153564, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7372090816497803, + "num_tokens": 262034653.0, + "step": 10124 + }, + { + "epoch": 1.111904238963321, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1520845890045166, + "learning_rate": 1e-06, + "loss": 1.0338, + "mean_token_accuracy": 0.6866916418075562, + "num_tokens": 262065968.0, + "step": 10125 + }, + { + "epoch": 1.1120140566659344, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 1.8924267292022705, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7042738199234009, + "num_tokens": 262102326.0, + "step": 10126 + }, + { + "epoch": 1.1121238743685482, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1946167945861816, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7066636085510254, + "num_tokens": 262129269.0, + "step": 10127 + }, + { + "epoch": 1.112233692071162, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1458330154418945, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6811007261276245, + "num_tokens": 262159214.0, + "step": 10128 + }, + { + "epoch": 1.1123435097737755, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.063145637512207, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.7007731199264526, + "num_tokens": 262187605.0, + "step": 10129 + }, + { + "epoch": 1.1124533274763893, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.2512073516845703, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7013134360313416, + "num_tokens": 262215300.0, + "step": 10130 + }, + { + "epoch": 1.1125631451790028, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.3204801082611084, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7253537178039551, + "num_tokens": 262237667.0, + "step": 10131 + }, + { + "epoch": 1.1126729628816165, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.399773597717285, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7210060358047485, + "num_tokens": 262260847.0, + "step": 10132 + }, + { + "epoch": 1.11278278058423, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.1435177326202393, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7349605560302734, + "num_tokens": 262288164.0, + "step": 10133 + }, + { + "epoch": 1.1128925982868438, + "ewc_loss": 5.3942203521728516e-06, + "grad_norm": 2.1235764026641846, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.718184232711792, + "num_tokens": 262315357.0, + "step": 10134 + }, + { + "epoch": 1.1130024159894576, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1704039573669434, + "learning_rate": 1e-06, + "loss": 0.7723, + "mean_token_accuracy": 0.7523994445800781, + "num_tokens": 262340817.0, + "step": 10135 + }, + { + "epoch": 1.1131122336920711, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.0539441108703613, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7225522994995117, + "num_tokens": 262369774.0, + "step": 10136 + }, + { + "epoch": 1.1132220513946849, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.185915231704712, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7067655324935913, + "num_tokens": 262395648.0, + "step": 10137 + }, + { + "epoch": 1.1133318690972984, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1437649726867676, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7175197601318359, + "num_tokens": 262421124.0, + "step": 10138 + }, + { + "epoch": 1.1134416867999122, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1801934242248535, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7383503913879395, + "num_tokens": 262446550.0, + "step": 10139 + }, + { + "epoch": 1.1135515045025257, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1853952407836914, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7343727350234985, + "num_tokens": 262471616.0, + "step": 10140 + }, + { + "epoch": 1.1136613222051395, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.912158727645874, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7226951122283936, + "num_tokens": 262489414.0, + "step": 10141 + }, + { + "epoch": 1.1137711399077532, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 1.818809986114502, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7035733461380005, + "num_tokens": 262526085.0, + "step": 10142 + }, + { + "epoch": 1.1138809576103668, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.277315855026245, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7164633870124817, + "num_tokens": 262550686.0, + "step": 10143 + }, + { + "epoch": 1.1139907753129805, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.246232271194458, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7028151750564575, + "num_tokens": 262575657.0, + "step": 10144 + }, + { + "epoch": 1.114100593015594, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.2154037952423096, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6904544830322266, + "num_tokens": 262603127.0, + "step": 10145 + }, + { + "epoch": 1.1142104107182078, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.2811827659606934, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.6989529132843018, + "num_tokens": 262628529.0, + "step": 10146 + }, + { + "epoch": 1.1143202284208213, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.163311719894409, + "learning_rate": 1e-06, + "loss": 1.0374, + "mean_token_accuracy": 0.6857295036315918, + "num_tokens": 262656454.0, + "step": 10147 + }, + { + "epoch": 1.114430046123435, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.0989112854003906, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7374567985534668, + "num_tokens": 262685153.0, + "step": 10148 + }, + { + "epoch": 1.1145398638260489, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.1145317554473877, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7030249834060669, + "num_tokens": 262713813.0, + "step": 10149 + }, + { + "epoch": 1.1146496815286624, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 1.9899499416351318, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7260420322418213, + "num_tokens": 262743376.0, + "step": 10150 + }, + { + "epoch": 1.1147594992312762, + "ewc_loss": 5.453824996948242e-06, + "grad_norm": 2.144564151763916, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.74010169506073, + "num_tokens": 262770450.0, + "step": 10151 + }, + { + "epoch": 1.1148693169338897, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.2759273052215576, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6920150518417358, + "num_tokens": 262795402.0, + "step": 10152 + }, + { + "epoch": 1.1149791346365034, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.6027233600616455, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.7002090811729431, + "num_tokens": 262818722.0, + "step": 10153 + }, + { + "epoch": 1.115088952339117, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.3214833736419678, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7407525777816772, + "num_tokens": 262841738.0, + "step": 10154 + }, + { + "epoch": 1.1151987700417307, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.2146997451782227, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7441601753234863, + "num_tokens": 262865964.0, + "step": 10155 + }, + { + "epoch": 1.1153085877443445, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.736524820327759, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.703306257724762, + "num_tokens": 262885839.0, + "step": 10156 + }, + { + "epoch": 1.115418405446958, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.3792076110839844, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7235366106033325, + "num_tokens": 262908507.0, + "step": 10157 + }, + { + "epoch": 1.1155282231495718, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.456165313720703, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7212591171264648, + "num_tokens": 262931525.0, + "step": 10158 + }, + { + "epoch": 1.1156380408521853, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.4847288131713867, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7078688740730286, + "num_tokens": 262954414.0, + "step": 10159 + }, + { + "epoch": 1.115747858554799, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.3777174949645996, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7077047228813171, + "num_tokens": 262976802.0, + "step": 10160 + }, + { + "epoch": 1.1158576762574126, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.102440357208252, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7189044952392578, + "num_tokens": 263006436.0, + "step": 10161 + }, + { + "epoch": 1.1159674939600264, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.1983630657196045, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.702855110168457, + "num_tokens": 263033693.0, + "step": 10162 + }, + { + "epoch": 1.11607731166264, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.4976651668548584, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7192806601524353, + "num_tokens": 263055731.0, + "step": 10163 + }, + { + "epoch": 1.1161871293652537, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.2190608978271484, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.710202693939209, + "num_tokens": 263081564.0, + "step": 10164 + }, + { + "epoch": 1.1162969470678674, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.169480800628662, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7217673063278198, + "num_tokens": 263110244.0, + "step": 10165 + }, + { + "epoch": 1.116406764770481, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.30329966545105, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7047719955444336, + "num_tokens": 263135014.0, + "step": 10166 + }, + { + "epoch": 1.1165165824730947, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.521353244781494, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7018420100212097, + "num_tokens": 263156727.0, + "step": 10167 + }, + { + "epoch": 1.1166264001757082, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.1804752349853516, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7229387760162354, + "num_tokens": 263183522.0, + "step": 10168 + }, + { + "epoch": 1.116736217878322, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.285708427429199, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7339811325073242, + "num_tokens": 263208367.0, + "step": 10169 + }, + { + "epoch": 1.1168460355809358, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.5589280128479004, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7535378932952881, + "num_tokens": 263227569.0, + "step": 10170 + }, + { + "epoch": 1.1169558532835493, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.0943524837493896, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7223761081695557, + "num_tokens": 263255747.0, + "step": 10171 + }, + { + "epoch": 1.117065670986163, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.4657232761383057, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7224262356758118, + "num_tokens": 263277067.0, + "step": 10172 + }, + { + "epoch": 1.1171754886887766, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.365640878677368, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7259221076965332, + "num_tokens": 263300135.0, + "step": 10173 + }, + { + "epoch": 1.1172853063913903, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.715463399887085, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7334081530570984, + "num_tokens": 263318790.0, + "step": 10174 + }, + { + "epoch": 1.1173951240940039, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.104374647140503, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.714805543422699, + "num_tokens": 263348357.0, + "step": 10175 + }, + { + "epoch": 1.1175049417966176, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.259324550628662, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7353391051292419, + "num_tokens": 263372010.0, + "step": 10176 + }, + { + "epoch": 1.1176147594992312, + "ewc_loss": 5.4836273193359375e-06, + "grad_norm": 2.2613258361816406, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7133147120475769, + "num_tokens": 263397378.0, + "step": 10177 + }, + { + "epoch": 1.117724577201845, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.2055695056915283, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7221578359603882, + "num_tokens": 263422840.0, + "step": 10178 + }, + { + "epoch": 1.1178343949044587, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.289140462875366, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7249149084091187, + "num_tokens": 263447140.0, + "step": 10179 + }, + { + "epoch": 1.1179442126070722, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.045498847961426, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7040318250656128, + "num_tokens": 263477950.0, + "step": 10180 + }, + { + "epoch": 1.118054030309686, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 1.9846649169921875, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.6979862451553345, + "num_tokens": 263511333.0, + "step": 10181 + }, + { + "epoch": 1.1181638480122995, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.07521915435791, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.712209165096283, + "num_tokens": 263539008.0, + "step": 10182 + }, + { + "epoch": 1.1182736657149133, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.4694631099700928, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7288411259651184, + "num_tokens": 263559854.0, + "step": 10183 + }, + { + "epoch": 1.1183834834175268, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.329413890838623, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7214984893798828, + "num_tokens": 263583069.0, + "step": 10184 + }, + { + "epoch": 1.1184933011201406, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.50673246383667, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7547026872634888, + "num_tokens": 263603230.0, + "step": 10185 + }, + { + "epoch": 1.1186031188227543, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.0805561542510986, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7106037735939026, + "num_tokens": 263632745.0, + "step": 10186 + }, + { + "epoch": 1.1187129365253679, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.3635778427124023, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7091683745384216, + "num_tokens": 263656588.0, + "step": 10187 + }, + { + "epoch": 1.1188227542279816, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.4400675296783447, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7297824025154114, + "num_tokens": 263677098.0, + "step": 10188 + }, + { + "epoch": 1.1189325719305951, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.3107926845550537, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.704249382019043, + "num_tokens": 263700576.0, + "step": 10189 + }, + { + "epoch": 1.119042389633209, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.0674545764923096, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7154509425163269, + "num_tokens": 263728753.0, + "step": 10190 + }, + { + "epoch": 1.1191522073358224, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.079012870788574, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7137840986251831, + "num_tokens": 263754501.0, + "step": 10191 + }, + { + "epoch": 1.1192620250384362, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.1113548278808594, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7081706523895264, + "num_tokens": 263781638.0, + "step": 10192 + }, + { + "epoch": 1.11937184274105, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.3186962604522705, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7030975818634033, + "num_tokens": 263806214.0, + "step": 10193 + }, + { + "epoch": 1.1194816604436635, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.1152522563934326, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7121226787567139, + "num_tokens": 263834287.0, + "step": 10194 + }, + { + "epoch": 1.1195914781462772, + "ewc_loss": 5.543231964111328e-06, + "grad_norm": 2.2243950366973877, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7152813673019409, + "num_tokens": 263861180.0, + "step": 10195 + }, + { + "epoch": 1.1197012958488908, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.4486076831817627, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.731643557548523, + "num_tokens": 263883046.0, + "step": 10196 + }, + { + "epoch": 1.1198111135515045, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.1758859157562256, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7246913909912109, + "num_tokens": 263910091.0, + "step": 10197 + }, + { + "epoch": 1.119920931254118, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.1022491455078125, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7281959652900696, + "num_tokens": 263937454.0, + "step": 10198 + }, + { + "epoch": 1.1200307489567318, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.193864583969116, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7217656970024109, + "num_tokens": 263963185.0, + "step": 10199 + }, + { + "epoch": 1.1201405666593456, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.4857752323150635, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7296555042266846, + "num_tokens": 263985334.0, + "step": 10200 + }, + { + "epoch": 1.1202503843619591, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.1847100257873535, + "learning_rate": 1e-06, + "loss": 0.9832, + "mean_token_accuracy": 0.7023967504501343, + "num_tokens": 264013050.0, + "step": 10201 + }, + { + "epoch": 1.1203602020645729, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.2720131874084473, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7061376571655273, + "num_tokens": 264040065.0, + "step": 10202 + }, + { + "epoch": 1.1204700197671864, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.155090093612671, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.705933153629303, + "num_tokens": 264067824.0, + "step": 10203 + }, + { + "epoch": 1.1205798374698002, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.188645601272583, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7027694582939148, + "num_tokens": 264094078.0, + "step": 10204 + }, + { + "epoch": 1.1206896551724137, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.080883264541626, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.706673264503479, + "num_tokens": 264122206.0, + "step": 10205 + }, + { + "epoch": 1.1207994728750275, + "ewc_loss": 5.5730342864990234e-06, + "grad_norm": 2.324094533920288, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7052677869796753, + "num_tokens": 264146490.0, + "step": 10206 + }, + { + "epoch": 1.1209092905776412, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.919248104095459, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.7444955110549927, + "num_tokens": 264162769.0, + "step": 10207 + }, + { + "epoch": 1.1210191082802548, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.321672201156616, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.724193811416626, + "num_tokens": 264185118.0, + "step": 10208 + }, + { + "epoch": 1.1211289259828685, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.037851572036743, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7123686075210571, + "num_tokens": 264217125.0, + "step": 10209 + }, + { + "epoch": 1.121238743685482, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.146953582763672, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7306467294692993, + "num_tokens": 264243982.0, + "step": 10210 + }, + { + "epoch": 1.1213485613880958, + "ewc_loss": 5.602836608886719e-06, + "grad_norm": 2.3604793548583984, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7027613520622253, + "num_tokens": 264268132.0, + "step": 10211 + }, + { + "epoch": 1.1214583790907093, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.3652865886688232, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7220739126205444, + "num_tokens": 264290913.0, + "step": 10212 + }, + { + "epoch": 1.121568196793323, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.0379509925842285, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7338100671768188, + "num_tokens": 264317797.0, + "step": 10213 + }, + { + "epoch": 1.1216780144959368, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3316454887390137, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7127805948257446, + "num_tokens": 264343062.0, + "step": 10214 + }, + { + "epoch": 1.1217878321985504, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3669090270996094, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7134779691696167, + "num_tokens": 264365747.0, + "step": 10215 + }, + { + "epoch": 1.1218976499011641, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3100783824920654, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.708512544631958, + "num_tokens": 264389155.0, + "step": 10216 + }, + { + "epoch": 1.1220074676037777, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2167296409606934, + "learning_rate": 1e-06, + "loss": 0.76, + "mean_token_accuracy": 0.7635267376899719, + "num_tokens": 264412724.0, + "step": 10217 + }, + { + "epoch": 1.1221172853063914, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.360868215560913, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7043559551239014, + "num_tokens": 264437737.0, + "step": 10218 + }, + { + "epoch": 1.122227103009005, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3689181804656982, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.7008861303329468, + "num_tokens": 264464141.0, + "step": 10219 + }, + { + "epoch": 1.1223369207116187, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2678349018096924, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7107387781143188, + "num_tokens": 264489951.0, + "step": 10220 + }, + { + "epoch": 1.1224467384142325, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 1.950518012046814, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.7026185989379883, + "num_tokens": 264522420.0, + "step": 10221 + }, + { + "epoch": 1.122556556116846, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.139582395553589, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7111578583717346, + "num_tokens": 264550110.0, + "step": 10222 + }, + { + "epoch": 1.1226663738194598, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.164008855819702, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7035657167434692, + "num_tokens": 264577054.0, + "step": 10223 + }, + { + "epoch": 1.1227761915220733, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2877066135406494, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7148464918136597, + "num_tokens": 264602938.0, + "step": 10224 + }, + { + "epoch": 1.122886009224687, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2192447185516357, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7176034450531006, + "num_tokens": 264629059.0, + "step": 10225 + }, + { + "epoch": 1.1229958269273006, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.1376473903656006, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.729178786277771, + "num_tokens": 264655928.0, + "step": 10226 + }, + { + "epoch": 1.1231056446299144, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2372896671295166, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7286468744277954, + "num_tokens": 264681093.0, + "step": 10227 + }, + { + "epoch": 1.123215462332528, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.245736598968506, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.728840708732605, + "num_tokens": 264705936.0, + "step": 10228 + }, + { + "epoch": 1.1233252800351416, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.215252637863159, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7131434679031372, + "num_tokens": 264731489.0, + "step": 10229 + }, + { + "epoch": 1.1234350977377554, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.334472417831421, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7393700480461121, + "num_tokens": 264753490.0, + "step": 10230 + }, + { + "epoch": 1.123544915440369, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.321622133255005, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.713887631893158, + "num_tokens": 264779633.0, + "step": 10231 + }, + { + "epoch": 1.1236547331429827, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2503647804260254, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.707007646560669, + "num_tokens": 264804496.0, + "step": 10232 + }, + { + "epoch": 1.1237645508455962, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.219205856323242, + "learning_rate": 1e-06, + "loss": 1.0014, + "mean_token_accuracy": 0.6907573342323303, + "num_tokens": 264832883.0, + "step": 10233 + }, + { + "epoch": 1.12387436854821, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.230987548828125, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7315511703491211, + "num_tokens": 264858624.0, + "step": 10234 + }, + { + "epoch": 1.1239841862508237, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.6043272018432617, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7114588618278503, + "num_tokens": 264880895.0, + "step": 10235 + }, + { + "epoch": 1.1240940039534373, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.322674036026001, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7009142637252808, + "num_tokens": 264907135.0, + "step": 10236 + }, + { + "epoch": 1.124203821656051, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.213372230529785, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7119598388671875, + "num_tokens": 264934646.0, + "step": 10237 + }, + { + "epoch": 1.1243136393586646, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.0083749294281006, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7210571765899658, + "num_tokens": 264963994.0, + "step": 10238 + }, + { + "epoch": 1.1244234570612783, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3301963806152344, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7238428592681885, + "num_tokens": 264988809.0, + "step": 10239 + }, + { + "epoch": 1.1245332747638919, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3069818019866943, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7276804447174072, + "num_tokens": 265011918.0, + "step": 10240 + }, + { + "epoch": 1.1246430924665056, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.1380395889282227, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.713546097278595, + "num_tokens": 265040220.0, + "step": 10241 + }, + { + "epoch": 1.1247529101691192, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.18000864982605, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7330808639526367, + "num_tokens": 265066840.0, + "step": 10242 + }, + { + "epoch": 1.124862727871733, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.40354061126709, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7299057841300964, + "num_tokens": 265090591.0, + "step": 10243 + }, + { + "epoch": 1.1249725455743467, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.2329845428466797, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7098222374916077, + "num_tokens": 265117912.0, + "step": 10244 + }, + { + "epoch": 1.1250823632769602, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.2884156703948975, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7149337530136108, + "num_tokens": 265143335.0, + "step": 10245 + }, + { + "epoch": 1.125192180979574, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.3125181198120117, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6946076154708862, + "num_tokens": 265168875.0, + "step": 10246 + }, + { + "epoch": 1.1253019986821875, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.240812301635742, + "learning_rate": 1e-06, + "loss": 1.0163, + "mean_token_accuracy": 0.6962594985961914, + "num_tokens": 265195136.0, + "step": 10247 + }, + { + "epoch": 1.1254118163848013, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.0835182666778564, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7129072546958923, + "num_tokens": 265222875.0, + "step": 10248 + }, + { + "epoch": 1.125521634087415, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.5953407287597656, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7230744361877441, + "num_tokens": 265242173.0, + "step": 10249 + }, + { + "epoch": 1.1256314517900285, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.3573617935180664, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7370972633361816, + "num_tokens": 265265667.0, + "step": 10250 + }, + { + "epoch": 1.1257412694926423, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.213287830352783, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.694668173789978, + "num_tokens": 265295514.0, + "step": 10251 + }, + { + "epoch": 1.1258510871952558, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.445364236831665, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7386994361877441, + "num_tokens": 265317261.0, + "step": 10252 + }, + { + "epoch": 1.1259609048978696, + "ewc_loss": 5.632638931274414e-06, + "grad_norm": 2.319124460220337, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7031425833702087, + "num_tokens": 265342770.0, + "step": 10253 + }, + { + "epoch": 1.1260707226004831, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.5595595836639404, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7189797163009644, + "num_tokens": 265363346.0, + "step": 10254 + }, + { + "epoch": 1.1261805403030969, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.1289029121398926, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.7141937017440796, + "num_tokens": 265391417.0, + "step": 10255 + }, + { + "epoch": 1.1262903580057104, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.5197551250457764, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7442189455032349, + "num_tokens": 265412658.0, + "step": 10256 + }, + { + "epoch": 1.1264001757083242, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.127784252166748, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7237354516983032, + "num_tokens": 265439582.0, + "step": 10257 + }, + { + "epoch": 1.126509993410938, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.2126612663269043, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7171797752380371, + "num_tokens": 265466946.0, + "step": 10258 + }, + { + "epoch": 1.1266198111135515, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.003309488296509, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7135835289955139, + "num_tokens": 265499854.0, + "step": 10259 + }, + { + "epoch": 1.1267296288161652, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.414039134979248, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7306650280952454, + "num_tokens": 265523938.0, + "step": 10260 + }, + { + "epoch": 1.1268394465187788, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 1.8332445621490479, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7189853191375732, + "num_tokens": 265560468.0, + "step": 10261 + }, + { + "epoch": 1.1269492642213925, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.431478977203369, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.707220733165741, + "num_tokens": 265583667.0, + "step": 10262 + }, + { + "epoch": 1.127059081924006, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.2609119415283203, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7393320202827454, + "num_tokens": 265606595.0, + "step": 10263 + }, + { + "epoch": 1.1271688996266198, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.190595865249634, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7097562551498413, + "num_tokens": 265633660.0, + "step": 10264 + }, + { + "epoch": 1.1272787173292333, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.1703944206237793, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7106298804283142, + "num_tokens": 265659487.0, + "step": 10265 + }, + { + "epoch": 1.127388535031847, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.14593505859375, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7081905007362366, + "num_tokens": 265688273.0, + "step": 10266 + }, + { + "epoch": 1.1274983527344609, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2346537113189697, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6967913508415222, + "num_tokens": 265713972.0, + "step": 10267 + }, + { + "epoch": 1.1276081704370744, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.3089959621429443, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6973670721054077, + "num_tokens": 265739755.0, + "step": 10268 + }, + { + "epoch": 1.1277179881396882, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.4081461429595947, + "learning_rate": 1e-06, + "loss": 0.9802, + "mean_token_accuracy": 0.6965003609657288, + "num_tokens": 265764051.0, + "step": 10269 + }, + { + "epoch": 1.1278278058423017, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2819032669067383, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.713248610496521, + "num_tokens": 265789653.0, + "step": 10270 + }, + { + "epoch": 1.1279376235449154, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.1090710163116455, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7291558980941772, + "num_tokens": 265816462.0, + "step": 10271 + }, + { + "epoch": 1.1280474412475292, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.4710285663604736, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.736166775226593, + "num_tokens": 265840306.0, + "step": 10272 + }, + { + "epoch": 1.1281572589501427, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.005114793777466, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7298538684844971, + "num_tokens": 265870347.0, + "step": 10273 + }, + { + "epoch": 1.1282670766527565, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.5025572776794434, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7064182162284851, + "num_tokens": 265894363.0, + "step": 10274 + }, + { + "epoch": 1.12837689435537, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2607173919677734, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7170611619949341, + "num_tokens": 265921549.0, + "step": 10275 + }, + { + "epoch": 1.1284867120579838, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.1563103199005127, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7236875295639038, + "num_tokens": 265948626.0, + "step": 10276 + }, + { + "epoch": 1.1285965297605973, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 1.9303549528121948, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.708588182926178, + "num_tokens": 265983453.0, + "step": 10277 + }, + { + "epoch": 1.128706347463211, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.428865671157837, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7315486073493958, + "num_tokens": 266005444.0, + "step": 10278 + }, + { + "epoch": 1.1288161651658246, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.197288990020752, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7375943660736084, + "num_tokens": 266031241.0, + "step": 10279 + }, + { + "epoch": 1.1289259828684384, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.493215560913086, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7270944118499756, + "num_tokens": 266053070.0, + "step": 10280 + }, + { + "epoch": 1.1290358005710521, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.323904275894165, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7037546634674072, + "num_tokens": 266076940.0, + "step": 10281 + }, + { + "epoch": 1.1291456182736657, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 1.970289707183838, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7043815851211548, + "num_tokens": 266111542.0, + "step": 10282 + }, + { + "epoch": 1.1292554359762794, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.0710153579711914, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7482033967971802, + "num_tokens": 266137534.0, + "step": 10283 + }, + { + "epoch": 1.129365253678893, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2713921070098877, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7332475185394287, + "num_tokens": 266161658.0, + "step": 10284 + }, + { + "epoch": 1.1294750713815067, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.403681516647339, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7171359062194824, + "num_tokens": 266184514.0, + "step": 10285 + }, + { + "epoch": 1.1295848890841205, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2109017372131348, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.6931484341621399, + "num_tokens": 266214113.0, + "step": 10286 + }, + { + "epoch": 1.129694706786734, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.4601035118103027, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.728528618812561, + "num_tokens": 266234590.0, + "step": 10287 + }, + { + "epoch": 1.1298045244893478, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.097609519958496, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7363227605819702, + "num_tokens": 266262165.0, + "step": 10288 + }, + { + "epoch": 1.1299143421919613, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.218441963195801, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7283861637115479, + "num_tokens": 266287184.0, + "step": 10289 + }, + { + "epoch": 1.130024159894575, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 3.1548426151275635, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7287863492965698, + "num_tokens": 266305758.0, + "step": 10290 + }, + { + "epoch": 1.1301339775971886, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.0897135734558105, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7224740386009216, + "num_tokens": 266335294.0, + "step": 10291 + }, + { + "epoch": 1.1302437952998023, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.063499689102173, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7131990194320679, + "num_tokens": 266363544.0, + "step": 10292 + }, + { + "epoch": 1.1303536130024159, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2994015216827393, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6887441277503967, + "num_tokens": 266389150.0, + "step": 10293 + }, + { + "epoch": 1.1304634307050296, + "ewc_loss": 5.662441253662109e-06, + "grad_norm": 2.0063436031341553, + "learning_rate": 1e-06, + "loss": 1.0049, + "mean_token_accuracy": 0.6983554363250732, + "num_tokens": 266418930.0, + "step": 10294 + }, + { + "epoch": 1.1305732484076434, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.2571024894714355, + "learning_rate": 1e-06, + "loss": 1.0034, + "mean_token_accuracy": 0.6983160972595215, + "num_tokens": 266445519.0, + "step": 10295 + }, + { + "epoch": 1.130683066110257, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.6407744884490967, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7230566740036011, + "num_tokens": 266465326.0, + "step": 10296 + }, + { + "epoch": 1.1307928838128707, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.0091450214385986, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7104695439338684, + "num_tokens": 266498034.0, + "step": 10297 + }, + { + "epoch": 1.1309027015154842, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.37365984916687, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7098239660263062, + "num_tokens": 266523244.0, + "step": 10298 + }, + { + "epoch": 1.131012519218098, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.010687828063965, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.6980225443840027, + "num_tokens": 266556348.0, + "step": 10299 + }, + { + "epoch": 1.1311223369207117, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.174595832824707, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7101637125015259, + "num_tokens": 266585163.0, + "step": 10300 + }, + { + "epoch": 1.1312321546233253, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.3392508029937744, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7363883256912231, + "num_tokens": 266608597.0, + "step": 10301 + }, + { + "epoch": 1.131341972325939, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.4255497455596924, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7262057065963745, + "num_tokens": 266631533.0, + "step": 10302 + }, + { + "epoch": 1.1314517900285526, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.236360788345337, + "learning_rate": 1e-06, + "loss": 1.0373, + "mean_token_accuracy": 0.6860999464988708, + "num_tokens": 266662321.0, + "step": 10303 + }, + { + "epoch": 1.1315616077311663, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.763983964920044, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7200784683227539, + "num_tokens": 266681531.0, + "step": 10304 + }, + { + "epoch": 1.1316714254337799, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.106229782104492, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.729453980922699, + "num_tokens": 266709453.0, + "step": 10305 + }, + { + "epoch": 1.1317812431363936, + "ewc_loss": 5.7220458984375e-06, + "grad_norm": 2.205193519592285, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7311632633209229, + "num_tokens": 266735508.0, + "step": 10306 + }, + { + "epoch": 1.1318910608390071, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.4675960540771484, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7058850526809692, + "num_tokens": 266758362.0, + "step": 10307 + }, + { + "epoch": 1.132000878541621, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.1587963104248047, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7081749439239502, + "num_tokens": 266784633.0, + "step": 10308 + }, + { + "epoch": 1.1321106962442347, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.348379373550415, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7139744758605957, + "num_tokens": 266811149.0, + "step": 10309 + }, + { + "epoch": 1.1322205139468482, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.7503621578216553, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7321289777755737, + "num_tokens": 266829030.0, + "step": 10310 + }, + { + "epoch": 1.132330331649462, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.0771849155426025, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7233607172966003, + "num_tokens": 266859845.0, + "step": 10311 + }, + { + "epoch": 1.1324401493520755, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.2323648929595947, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7170849442481995, + "num_tokens": 266884529.0, + "step": 10312 + }, + { + "epoch": 1.1325499670546892, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.245258092880249, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7056698799133301, + "num_tokens": 266910291.0, + "step": 10313 + }, + { + "epoch": 1.132659784757303, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.192011833190918, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7408319711685181, + "num_tokens": 266937059.0, + "step": 10314 + }, + { + "epoch": 1.1327696024599165, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2476718425750732, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.7004099488258362, + "num_tokens": 266964277.0, + "step": 10315 + }, + { + "epoch": 1.1328794201625303, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 1.9943811893463135, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7073708772659302, + "num_tokens": 266996463.0, + "step": 10316 + }, + { + "epoch": 1.1329892378651438, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.524820327758789, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7158547639846802, + "num_tokens": 267019702.0, + "step": 10317 + }, + { + "epoch": 1.1330990555677576, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.430386543273926, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7307806015014648, + "num_tokens": 267040614.0, + "step": 10318 + }, + { + "epoch": 1.1332088732703711, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2235734462738037, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7208827137947083, + "num_tokens": 267067908.0, + "step": 10319 + }, + { + "epoch": 1.1333186909729849, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.232165575027466, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7143499255180359, + "num_tokens": 267093389.0, + "step": 10320 + }, + { + "epoch": 1.1334285086755984, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.0356576442718506, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7120654582977295, + "num_tokens": 267123727.0, + "step": 10321 + }, + { + "epoch": 1.1335383263782122, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.006477117538452, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7051838636398315, + "num_tokens": 267153359.0, + "step": 10322 + }, + { + "epoch": 1.133648144080826, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1257691383361816, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.719920814037323, + "num_tokens": 267180260.0, + "step": 10323 + }, + { + "epoch": 1.1337579617834395, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.030489444732666, + "learning_rate": 1e-06, + "loss": 1.0299, + "mean_token_accuracy": 0.6884584426879883, + "num_tokens": 267214943.0, + "step": 10324 + }, + { + "epoch": 1.1338677794860532, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.366778612136841, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7131824493408203, + "num_tokens": 267238721.0, + "step": 10325 + }, + { + "epoch": 1.1339775971886668, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.072955846786499, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7207654118537903, + "num_tokens": 267269105.0, + "step": 10326 + }, + { + "epoch": 1.1340874148912805, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2461800575256348, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7106248140335083, + "num_tokens": 267295469.0, + "step": 10327 + }, + { + "epoch": 1.134197232593894, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.0715701580047607, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7052577137947083, + "num_tokens": 267323179.0, + "step": 10328 + }, + { + "epoch": 1.1343070502965078, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.88090443611145, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7204478979110718, + "num_tokens": 267347769.0, + "step": 10329 + }, + { + "epoch": 1.1344168679991213, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2394649982452393, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7162579298019409, + "num_tokens": 267373815.0, + "step": 10330 + }, + { + "epoch": 1.134526685701735, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.367741346359253, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7188835740089417, + "num_tokens": 267396441.0, + "step": 10331 + }, + { + "epoch": 1.1346365034043489, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2344226837158203, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.719402015209198, + "num_tokens": 267422653.0, + "step": 10332 + }, + { + "epoch": 1.1347463211069624, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.478565216064453, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7263193726539612, + "num_tokens": 267445598.0, + "step": 10333 + }, + { + "epoch": 1.1348561388095761, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1464226245880127, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7310506105422974, + "num_tokens": 267471406.0, + "step": 10334 + }, + { + "epoch": 1.1349659565121897, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.4046242237091064, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.71642005443573, + "num_tokens": 267494400.0, + "step": 10335 + }, + { + "epoch": 1.1350757742148034, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1381797790527344, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7303899526596069, + "num_tokens": 267524579.0, + "step": 10336 + }, + { + "epoch": 1.1351855919174172, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 1.9527549743652344, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7351176142692566, + "num_tokens": 267555767.0, + "step": 10337 + }, + { + "epoch": 1.1352954096200307, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 1.9075920581817627, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7140007019042969, + "num_tokens": 267589006.0, + "step": 10338 + }, + { + "epoch": 1.1354052273226445, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.3628079891204834, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7242193222045898, + "num_tokens": 267611517.0, + "step": 10339 + }, + { + "epoch": 1.135515045025258, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.013300657272339, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6990581154823303, + "num_tokens": 267644383.0, + "step": 10340 + }, + { + "epoch": 1.1356248627278718, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.315014362335205, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7041600942611694, + "num_tokens": 267667731.0, + "step": 10341 + }, + { + "epoch": 1.1357346804304853, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.374600410461426, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7102478742599487, + "num_tokens": 267692049.0, + "step": 10342 + }, + { + "epoch": 1.135844498133099, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1909725666046143, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7272611260414124, + "num_tokens": 267719834.0, + "step": 10343 + }, + { + "epoch": 1.1359543158357126, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.6450369358062744, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7345521450042725, + "num_tokens": 267741202.0, + "step": 10344 + }, + { + "epoch": 1.1360641335383264, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.0782573223114014, + "learning_rate": 1e-06, + "loss": 1.0152, + "mean_token_accuracy": 0.6996137499809265, + "num_tokens": 267771467.0, + "step": 10345 + }, + { + "epoch": 1.1361739512409401, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.470435619354248, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7316346168518066, + "num_tokens": 267792675.0, + "step": 10346 + }, + { + "epoch": 1.1362837689435537, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1113529205322266, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7055318355560303, + "num_tokens": 267821926.0, + "step": 10347 + }, + { + "epoch": 1.1363935866461674, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.0499658584594727, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.7114391326904297, + "num_tokens": 267852223.0, + "step": 10348 + }, + { + "epoch": 1.136503404348781, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2525618076324463, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7137172222137451, + "num_tokens": 267876621.0, + "step": 10349 + }, + { + "epoch": 1.1366132220513947, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.0098626613616943, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7134435176849365, + "num_tokens": 267908585.0, + "step": 10350 + }, + { + "epoch": 1.1367230397540085, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.047807455062866, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7248480319976807, + "num_tokens": 267940588.0, + "step": 10351 + }, + { + "epoch": 1.136832857456622, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.37884259223938, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7191051840782166, + "num_tokens": 267965779.0, + "step": 10352 + }, + { + "epoch": 1.1369426751592357, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.261899471282959, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.6946799755096436, + "num_tokens": 267992571.0, + "step": 10353 + }, + { + "epoch": 1.1370524928618493, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2804834842681885, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.7125738859176636, + "num_tokens": 268019872.0, + "step": 10354 + }, + { + "epoch": 1.137162310564463, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.351095676422119, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7245162725448608, + "num_tokens": 268043544.0, + "step": 10355 + }, + { + "epoch": 1.1372721282670766, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.1926052570343018, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7136607766151428, + "num_tokens": 268069948.0, + "step": 10356 + }, + { + "epoch": 1.1373819459696903, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.605417490005493, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.713337779045105, + "num_tokens": 268089960.0, + "step": 10357 + }, + { + "epoch": 1.1374917636723039, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.4436776638031006, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.732619047164917, + "num_tokens": 268111988.0, + "step": 10358 + }, + { + "epoch": 1.1376015813749176, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.2112722396850586, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7311444282531738, + "num_tokens": 268137247.0, + "step": 10359 + }, + { + "epoch": 1.1377113990775314, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.3061869144439697, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7456362247467041, + "num_tokens": 268161393.0, + "step": 10360 + }, + { + "epoch": 1.137821216780145, + "ewc_loss": 5.781650543212891e-06, + "grad_norm": 2.137199878692627, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7096878290176392, + "num_tokens": 268190923.0, + "step": 10361 + }, + { + "epoch": 1.1379310344827587, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.4252703189849854, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7031122446060181, + "num_tokens": 268214444.0, + "step": 10362 + }, + { + "epoch": 1.1380408521853722, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.4647324085235596, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7188061475753784, + "num_tokens": 268236428.0, + "step": 10363 + }, + { + "epoch": 1.138150669887986, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.509467840194702, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7446807026863098, + "num_tokens": 268258604.0, + "step": 10364 + }, + { + "epoch": 1.1382604875905997, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.1687631607055664, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7533968687057495, + "num_tokens": 268285344.0, + "step": 10365 + }, + { + "epoch": 1.1383703052932133, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 1.9035189151763916, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7066901922225952, + "num_tokens": 268320968.0, + "step": 10366 + }, + { + "epoch": 1.138480122995827, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.2777364253997803, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7310208082199097, + "num_tokens": 268347096.0, + "step": 10367 + }, + { + "epoch": 1.1385899406984406, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.276887893676758, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7249327301979065, + "num_tokens": 268373691.0, + "step": 10368 + }, + { + "epoch": 1.1386997584010543, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.2889304161071777, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7047739624977112, + "num_tokens": 268398350.0, + "step": 10369 + }, + { + "epoch": 1.1388095761036678, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.3597605228424072, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7271257638931274, + "num_tokens": 268422393.0, + "step": 10370 + }, + { + "epoch": 1.1389193938062816, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.2013325691223145, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7448225021362305, + "num_tokens": 268447765.0, + "step": 10371 + }, + { + "epoch": 1.1390292115088951, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.391615867614746, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7237878441810608, + "num_tokens": 268476834.0, + "step": 10372 + }, + { + "epoch": 1.139139029211509, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.3082830905914307, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7246547341346741, + "num_tokens": 268501276.0, + "step": 10373 + }, + { + "epoch": 1.1392488469141226, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 1.9676529169082642, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7235288619995117, + "num_tokens": 268533575.0, + "step": 10374 + }, + { + "epoch": 1.1393586646167362, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.5083119869232178, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7375843524932861, + "num_tokens": 268554506.0, + "step": 10375 + }, + { + "epoch": 1.13946848231935, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.277020215988159, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7221718430519104, + "num_tokens": 268579127.0, + "step": 10376 + }, + { + "epoch": 1.1395783000219635, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.419344425201416, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7106342315673828, + "num_tokens": 268602082.0, + "step": 10377 + }, + { + "epoch": 1.1396881177245772, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.260805368423462, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.7018426656723022, + "num_tokens": 268629322.0, + "step": 10378 + }, + { + "epoch": 1.1397979354271908, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.2982940673828125, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7399019002914429, + "num_tokens": 268654858.0, + "step": 10379 + }, + { + "epoch": 1.1399077531298045, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.5718741416931152, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7257934808731079, + "num_tokens": 268674996.0, + "step": 10380 + }, + { + "epoch": 1.1400175708324183, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.183863639831543, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7082051038742065, + "num_tokens": 268701396.0, + "step": 10381 + }, + { + "epoch": 1.1401273885350318, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.355058193206787, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7171652913093567, + "num_tokens": 268725730.0, + "step": 10382 + }, + { + "epoch": 1.1402372062376456, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.072078227996826, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7339144945144653, + "num_tokens": 268756266.0, + "step": 10383 + }, + { + "epoch": 1.140347023940259, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.2755541801452637, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7355743646621704, + "num_tokens": 268779905.0, + "step": 10384 + }, + { + "epoch": 1.1404568416428729, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.1667215824127197, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.727468729019165, + "num_tokens": 268805544.0, + "step": 10385 + }, + { + "epoch": 1.1405666593454864, + "ewc_loss": 5.751848220825195e-06, + "grad_norm": 2.279221296310425, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7413212060928345, + "num_tokens": 268828418.0, + "step": 10386 + }, + { + "epoch": 1.1406764770481002, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.7128796577453613, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7193374633789062, + "num_tokens": 268846625.0, + "step": 10387 + }, + { + "epoch": 1.140786294750714, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.185983180999756, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.714679479598999, + "num_tokens": 268871966.0, + "step": 10388 + }, + { + "epoch": 1.1408961124533274, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.1381890773773193, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7406431436538696, + "num_tokens": 268897336.0, + "step": 10389 + }, + { + "epoch": 1.1410059301559412, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.2701382637023926, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7062721252441406, + "num_tokens": 268924918.0, + "step": 10390 + }, + { + "epoch": 1.1411157478585547, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.230350971221924, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7199221253395081, + "num_tokens": 268950848.0, + "step": 10391 + }, + { + "epoch": 1.1412255655611685, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.281834840774536, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7093988656997681, + "num_tokens": 268976844.0, + "step": 10392 + }, + { + "epoch": 1.141335383263782, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.1513447761535645, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7254692912101746, + "num_tokens": 269004861.0, + "step": 10393 + }, + { + "epoch": 1.1414452009663958, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.4766321182250977, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7282184362411499, + "num_tokens": 269025965.0, + "step": 10394 + }, + { + "epoch": 1.1415550186690093, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.3652865886688232, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7295974493026733, + "num_tokens": 269048403.0, + "step": 10395 + }, + { + "epoch": 1.141664836371623, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.1422765254974365, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.6924114227294922, + "num_tokens": 269077585.0, + "step": 10396 + }, + { + "epoch": 1.1417746540742368, + "ewc_loss": 5.841255187988281e-06, + "grad_norm": 2.0445258617401123, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7405721545219421, + "num_tokens": 269108294.0, + "step": 10397 + }, + { + "epoch": 1.1418844717768504, + "ewc_loss": 5.811452865600586e-06, + "grad_norm": 2.215608596801758, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7128760814666748, + "num_tokens": 269135641.0, + "step": 10398 + }, + { + "epoch": 1.1419942894794641, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.9919770956039429, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6953926682472229, + "num_tokens": 269165374.0, + "step": 10399 + }, + { + "epoch": 1.1421041071820777, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4313642978668213, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7150586843490601, + "num_tokens": 269187117.0, + "step": 10400 + }, + { + "epoch": 1.1422139248846914, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.189868211746216, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.730562686920166, + "num_tokens": 269213528.0, + "step": 10401 + }, + { + "epoch": 1.1423237425873052, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.216736316680908, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.721763014793396, + "num_tokens": 269237503.0, + "step": 10402 + }, + { + "epoch": 1.1424335602899187, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.243319511413574, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7344516515731812, + "num_tokens": 269263747.0, + "step": 10403 + }, + { + "epoch": 1.1425433779925325, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4190309047698975, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7173978686332703, + "num_tokens": 269286290.0, + "step": 10404 + }, + { + "epoch": 1.142653195695146, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4524879455566406, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7418100833892822, + "num_tokens": 269306380.0, + "step": 10405 + }, + { + "epoch": 1.1427630133977598, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3428597450256348, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7201268672943115, + "num_tokens": 269330612.0, + "step": 10406 + }, + { + "epoch": 1.1428728311003733, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3522634506225586, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7410444021224976, + "num_tokens": 269352725.0, + "step": 10407 + }, + { + "epoch": 1.142982648802987, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.9387623071670532, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7134048342704773, + "num_tokens": 269385926.0, + "step": 10408 + }, + { + "epoch": 1.1430924665056006, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2748844623565674, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6983184814453125, + "num_tokens": 269411889.0, + "step": 10409 + }, + { + "epoch": 1.1432022842082143, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3549625873565674, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7040669918060303, + "num_tokens": 269436910.0, + "step": 10410 + }, + { + "epoch": 1.143312101910828, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.21069598197937, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7129500508308411, + "num_tokens": 269464438.0, + "step": 10411 + }, + { + "epoch": 1.1434219196134416, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2553179264068604, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7101250290870667, + "num_tokens": 269491963.0, + "step": 10412 + }, + { + "epoch": 1.1435317373160554, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.08878231048584, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7127541303634644, + "num_tokens": 269521966.0, + "step": 10413 + }, + { + "epoch": 1.143641555018669, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.9669548273086548, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.713515043258667, + "num_tokens": 269554796.0, + "step": 10414 + }, + { + "epoch": 1.1437513727212827, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3154489994049072, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7033786177635193, + "num_tokens": 269580105.0, + "step": 10415 + }, + { + "epoch": 1.1438611904238964, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.08449125289917, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6886281967163086, + "num_tokens": 269608531.0, + "step": 10416 + }, + { + "epoch": 1.14397100812651, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.10876727104187, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7074272632598877, + "num_tokens": 269638760.0, + "step": 10417 + }, + { + "epoch": 1.1440808258291237, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.189568042755127, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6975719928741455, + "num_tokens": 269666361.0, + "step": 10418 + }, + { + "epoch": 1.1441906435317373, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1928865909576416, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7138357162475586, + "num_tokens": 269692763.0, + "step": 10419 + }, + { + "epoch": 1.144300461234351, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.9588066339492798, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.6998147368431091, + "num_tokens": 269727061.0, + "step": 10420 + }, + { + "epoch": 1.1444102789369646, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.207310676574707, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7450234889984131, + "num_tokens": 269752838.0, + "step": 10421 + }, + { + "epoch": 1.1445200966395783, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0176467895507812, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7072455883026123, + "num_tokens": 269783618.0, + "step": 10422 + }, + { + "epoch": 1.1446299143421919, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0761852264404297, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7174882292747498, + "num_tokens": 269813083.0, + "step": 10423 + }, + { + "epoch": 1.1447397320448056, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1695778369903564, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7187376618385315, + "num_tokens": 269841125.0, + "step": 10424 + }, + { + "epoch": 1.1448495497474194, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2604854106903076, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7412147521972656, + "num_tokens": 269865835.0, + "step": 10425 + }, + { + "epoch": 1.144959367450033, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.5952060222625732, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7124556303024292, + "num_tokens": 269888757.0, + "step": 10426 + }, + { + "epoch": 1.1450691851526467, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.43233585357666, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7107772827148438, + "num_tokens": 269912248.0, + "step": 10427 + }, + { + "epoch": 1.1451790028552602, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.324486255645752, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7386056780815125, + "num_tokens": 269936794.0, + "step": 10428 + }, + { + "epoch": 1.145288820557874, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1519412994384766, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7177956700325012, + "num_tokens": 269962722.0, + "step": 10429 + }, + { + "epoch": 1.1453986382604877, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.356768846511841, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7239799499511719, + "num_tokens": 269986040.0, + "step": 10430 + }, + { + "epoch": 1.1455084559631012, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0356698036193848, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7151886224746704, + "num_tokens": 270018188.0, + "step": 10431 + }, + { + "epoch": 1.145618273665715, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.16049861907959, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.731658935546875, + "num_tokens": 270045447.0, + "step": 10432 + }, + { + "epoch": 1.1457280913683285, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2047617435455322, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7064127922058105, + "num_tokens": 270073881.0, + "step": 10433 + }, + { + "epoch": 1.1458379090709423, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.419825792312622, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7240806221961975, + "num_tokens": 270097232.0, + "step": 10434 + }, + { + "epoch": 1.1459477267735558, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.348841428756714, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7102241516113281, + "num_tokens": 270122091.0, + "step": 10435 + }, + { + "epoch": 1.1460575444761696, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2801246643066406, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.699579656124115, + "num_tokens": 270147807.0, + "step": 10436 + }, + { + "epoch": 1.1461673621787831, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4034957885742188, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7239739894866943, + "num_tokens": 270169830.0, + "step": 10437 + }, + { + "epoch": 1.1462771798813969, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3701698780059814, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7265834808349609, + "num_tokens": 270194328.0, + "step": 10438 + }, + { + "epoch": 1.1463869975840106, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1742331981658936, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7183219790458679, + "num_tokens": 270219357.0, + "step": 10439 + }, + { + "epoch": 1.1464968152866242, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.415602207183838, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7117981910705566, + "num_tokens": 270241779.0, + "step": 10440 + }, + { + "epoch": 1.146606632989238, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.563753366470337, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7293292284011841, + "num_tokens": 270261733.0, + "step": 10441 + }, + { + "epoch": 1.1467164506918515, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4736483097076416, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7219247817993164, + "num_tokens": 270283333.0, + "step": 10442 + }, + { + "epoch": 1.1468262683944652, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1358144283294678, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7134619355201721, + "num_tokens": 270315335.0, + "step": 10443 + }, + { + "epoch": 1.1469360860970788, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1055097579956055, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7328037619590759, + "num_tokens": 270342964.0, + "step": 10444 + }, + { + "epoch": 1.1470459037996925, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3039395809173584, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6991665363311768, + "num_tokens": 270368653.0, + "step": 10445 + }, + { + "epoch": 1.147155721502306, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3038582801818848, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7147419452667236, + "num_tokens": 270393362.0, + "step": 10446 + }, + { + "epoch": 1.1472655392049198, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4440369606018066, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7347679138183594, + "num_tokens": 270416873.0, + "step": 10447 + }, + { + "epoch": 1.1473753569075336, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3547725677490234, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7268282175064087, + "num_tokens": 270440103.0, + "step": 10448 + }, + { + "epoch": 1.147485174610147, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3779680728912354, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7251843810081482, + "num_tokens": 270462076.0, + "step": 10449 + }, + { + "epoch": 1.1475949923127609, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1869544982910156, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.704766035079956, + "num_tokens": 270488307.0, + "step": 10450 + }, + { + "epoch": 1.1477048100153744, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.038729190826416, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7260003089904785, + "num_tokens": 270517987.0, + "step": 10451 + }, + { + "epoch": 1.1478146277179881, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.353032350540161, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7085144519805908, + "num_tokens": 270543828.0, + "step": 10452 + }, + { + "epoch": 1.147924445420602, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2243335247039795, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7061057686805725, + "num_tokens": 270570706.0, + "step": 10453 + }, + { + "epoch": 1.1480342631232154, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2522361278533936, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7380223274230957, + "num_tokens": 270596739.0, + "step": 10454 + }, + { + "epoch": 1.1481440808258292, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.083181858062744, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7374059557914734, + "num_tokens": 270623788.0, + "step": 10455 + }, + { + "epoch": 1.1482538985284427, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1957337856292725, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7350090742111206, + "num_tokens": 270649744.0, + "step": 10456 + }, + { + "epoch": 1.1483637162310565, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4150829315185547, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7303714752197266, + "num_tokens": 270671889.0, + "step": 10457 + }, + { + "epoch": 1.14847353393367, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.502204179763794, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.758916437625885, + "num_tokens": 270692169.0, + "step": 10458 + }, + { + "epoch": 1.1485833516362838, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2204885482788086, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7305577993392944, + "num_tokens": 270716963.0, + "step": 10459 + }, + { + "epoch": 1.1486931693388973, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1357579231262207, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7183048725128174, + "num_tokens": 270747232.0, + "step": 10460 + }, + { + "epoch": 1.148802987041511, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.024165153503418, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.708229660987854, + "num_tokens": 270777490.0, + "step": 10461 + }, + { + "epoch": 1.1489128047441248, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3638367652893066, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7282198071479797, + "num_tokens": 270799558.0, + "step": 10462 + }, + { + "epoch": 1.1490226224467384, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3163511753082275, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7089424729347229, + "num_tokens": 270824945.0, + "step": 10463 + }, + { + "epoch": 1.1491324401493521, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.954788327217102, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7039568424224854, + "num_tokens": 270859452.0, + "step": 10464 + }, + { + "epoch": 1.1492422578519657, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0578696727752686, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7236202955245972, + "num_tokens": 270889401.0, + "step": 10465 + }, + { + "epoch": 1.1493520755545794, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.608705997467041, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7302045822143555, + "num_tokens": 270911780.0, + "step": 10466 + }, + { + "epoch": 1.1494618932571932, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2353031635284424, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7157854437828064, + "num_tokens": 270939541.0, + "step": 10467 + }, + { + "epoch": 1.1495717109598067, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3164896965026855, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7196396589279175, + "num_tokens": 270961764.0, + "step": 10468 + }, + { + "epoch": 1.1496815286624205, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4915502071380615, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.730237603187561, + "num_tokens": 270982189.0, + "step": 10469 + }, + { + "epoch": 1.149791346365034, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.769725799560547, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7235098481178284, + "num_tokens": 271000362.0, + "step": 10470 + }, + { + "epoch": 1.1499011640676478, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3252205848693848, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7282286882400513, + "num_tokens": 271023383.0, + "step": 10471 + }, + { + "epoch": 1.1500109817702613, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.122648239135742, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7324983477592468, + "num_tokens": 271052828.0, + "step": 10472 + }, + { + "epoch": 1.150120799472875, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1895148754119873, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6995560526847839, + "num_tokens": 271081617.0, + "step": 10473 + }, + { + "epoch": 1.1502306171754886, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.254899501800537, + "learning_rate": 1e-06, + "loss": 1.0377, + "mean_token_accuracy": 0.6809357404708862, + "num_tokens": 271110567.0, + "step": 10474 + }, + { + "epoch": 1.1503404348781023, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2915172576904297, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7138466835021973, + "num_tokens": 271134094.0, + "step": 10475 + }, + { + "epoch": 1.150450252580716, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2614786624908447, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7385199069976807, + "num_tokens": 271159825.0, + "step": 10476 + }, + { + "epoch": 1.1505600702833296, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3863332271575928, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7337253093719482, + "num_tokens": 271179892.0, + "step": 10477 + }, + { + "epoch": 1.1506698879859434, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.200596332550049, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7151583433151245, + "num_tokens": 271206416.0, + "step": 10478 + }, + { + "epoch": 1.150779705688557, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.112992286682129, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6934857964515686, + "num_tokens": 271235255.0, + "step": 10479 + }, + { + "epoch": 1.1508895233911707, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.8683748245239258, + "learning_rate": 1e-06, + "loss": 1.049, + "mean_token_accuracy": 0.6814089417457581, + "num_tokens": 271273431.0, + "step": 10480 + }, + { + "epoch": 1.1509993410937844, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.398693323135376, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7037492990493774, + "num_tokens": 271296536.0, + "step": 10481 + }, + { + "epoch": 1.151109158796398, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 1.9757953882217407, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.6940150260925293, + "num_tokens": 271331104.0, + "step": 10482 + }, + { + "epoch": 1.1512189764990117, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.198345422744751, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7197336554527283, + "num_tokens": 271358864.0, + "step": 10483 + }, + { + "epoch": 1.1513287942016253, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.152108907699585, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.7085585594177246, + "num_tokens": 271385636.0, + "step": 10484 + }, + { + "epoch": 1.151438611904239, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4928014278411865, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7273029088973999, + "num_tokens": 271406751.0, + "step": 10485 + }, + { + "epoch": 1.1515484296068526, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2241406440734863, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7031068205833435, + "num_tokens": 271433759.0, + "step": 10486 + }, + { + "epoch": 1.1516582473094663, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.540104389190674, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7329162955284119, + "num_tokens": 271453057.0, + "step": 10487 + }, + { + "epoch": 1.1517680650120798, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3129818439483643, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7200779914855957, + "num_tokens": 271478297.0, + "step": 10488 + }, + { + "epoch": 1.1518778827146936, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.412233591079712, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.726054310798645, + "num_tokens": 271500563.0, + "step": 10489 + }, + { + "epoch": 1.1519877004173074, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0888407230377197, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7032512426376343, + "num_tokens": 271533148.0, + "step": 10490 + }, + { + "epoch": 1.152097518119921, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.126086711883545, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7139998078346252, + "num_tokens": 271561841.0, + "step": 10491 + }, + { + "epoch": 1.1522073358225347, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.472414016723633, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7341589331626892, + "num_tokens": 271583226.0, + "step": 10492 + }, + { + "epoch": 1.1523171535251482, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.369781970977783, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7291409373283386, + "num_tokens": 271605953.0, + "step": 10493 + }, + { + "epoch": 1.152426971227762, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2359731197357178, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7195961475372314, + "num_tokens": 271632154.0, + "step": 10494 + }, + { + "epoch": 1.1525367889303757, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4538111686706543, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7238160371780396, + "num_tokens": 271653841.0, + "step": 10495 + }, + { + "epoch": 1.1526466066329892, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2395761013031006, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7196393013000488, + "num_tokens": 271677977.0, + "step": 10496 + }, + { + "epoch": 1.152756424335603, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3346245288848877, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7193745374679565, + "num_tokens": 271701221.0, + "step": 10497 + }, + { + "epoch": 1.1528662420382165, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.06854248046875, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.6971051692962646, + "num_tokens": 271732023.0, + "step": 10498 + }, + { + "epoch": 1.1529760597408303, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.209298610687256, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7238218188285828, + "num_tokens": 271757810.0, + "step": 10499 + }, + { + "epoch": 1.1530858774434438, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.8343183994293213, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7278463244438171, + "num_tokens": 271775593.0, + "step": 10500 + }, + { + "epoch": 1.1531956951460576, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.187483549118042, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7293398380279541, + "num_tokens": 271802380.0, + "step": 10501 + }, + { + "epoch": 1.153305512848671, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.127243995666504, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7146687507629395, + "num_tokens": 271829159.0, + "step": 10502 + }, + { + "epoch": 1.1534153305512849, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3536860942840576, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7137698531150818, + "num_tokens": 271853961.0, + "step": 10503 + }, + { + "epoch": 1.1535251482538986, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1547398567199707, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7291054725646973, + "num_tokens": 271880431.0, + "step": 10504 + }, + { + "epoch": 1.1536349659565122, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0043723583221436, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7201881408691406, + "num_tokens": 271908730.0, + "step": 10505 + }, + { + "epoch": 1.153744783659126, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.284430980682373, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.734622597694397, + "num_tokens": 271931465.0, + "step": 10506 + }, + { + "epoch": 1.1538546013617395, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.32509183883667, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7227016687393188, + "num_tokens": 271957167.0, + "step": 10507 + }, + { + "epoch": 1.1539644190643532, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2515335083007812, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7343318462371826, + "num_tokens": 271982253.0, + "step": 10508 + }, + { + "epoch": 1.1540742367669667, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.0470263957977295, + "learning_rate": 1e-06, + "loss": 1.0612, + "mean_token_accuracy": 0.6798255443572998, + "num_tokens": 272014376.0, + "step": 10509 + }, + { + "epoch": 1.1541840544695805, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.1585710048675537, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7198734283447266, + "num_tokens": 272043291.0, + "step": 10510 + }, + { + "epoch": 1.154293872172194, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.115941047668457, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.6975688934326172, + "num_tokens": 272071838.0, + "step": 10511 + }, + { + "epoch": 1.1544036898748078, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.176875352859497, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7139186859130859, + "num_tokens": 272099074.0, + "step": 10512 + }, + { + "epoch": 1.1545135075774215, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.358781576156616, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.729070782661438, + "num_tokens": 272121508.0, + "step": 10513 + }, + { + "epoch": 1.154623325280035, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.4303112030029297, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7333171367645264, + "num_tokens": 272145371.0, + "step": 10514 + }, + { + "epoch": 1.1547331429826488, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.264692544937134, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7081048488616943, + "num_tokens": 272170042.0, + "step": 10515 + }, + { + "epoch": 1.1548429606852624, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.128180980682373, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7348395586013794, + "num_tokens": 272197860.0, + "step": 10516 + }, + { + "epoch": 1.1549527783878761, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.6323184967041016, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7318964004516602, + "num_tokens": 272219986.0, + "step": 10517 + }, + { + "epoch": 1.15506259609049, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.2381322383880615, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7126152515411377, + "num_tokens": 272247393.0, + "step": 10518 + }, + { + "epoch": 1.1551724137931034, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.544184923171997, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7408853769302368, + "num_tokens": 272266856.0, + "step": 10519 + }, + { + "epoch": 1.1552822314957172, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.135039806365967, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7029900550842285, + "num_tokens": 272296842.0, + "step": 10520 + }, + { + "epoch": 1.1553920491983307, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.122244119644165, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6937540173530579, + "num_tokens": 272324349.0, + "step": 10521 + }, + { + "epoch": 1.1555018669009445, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.148907423019409, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7326875925064087, + "num_tokens": 272351169.0, + "step": 10522 + }, + { + "epoch": 1.155611684603558, + "ewc_loss": 5.8710575103759766e-06, + "grad_norm": 2.3339030742645264, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7089366912841797, + "num_tokens": 272375508.0, + "step": 10523 + }, + { + "epoch": 1.1557215023061718, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 1.942072868347168, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7182241678237915, + "num_tokens": 272405700.0, + "step": 10524 + }, + { + "epoch": 1.1558313200087853, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.0919835567474365, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.6918577551841736, + "num_tokens": 272435400.0, + "step": 10525 + }, + { + "epoch": 1.155941137711399, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.272864818572998, + "learning_rate": 1e-06, + "loss": 1.0827, + "mean_token_accuracy": 0.6806528568267822, + "num_tokens": 272462424.0, + "step": 10526 + }, + { + "epoch": 1.1560509554140128, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.1786789894104004, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.6975526809692383, + "num_tokens": 272492794.0, + "step": 10527 + }, + { + "epoch": 1.1561607731166264, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2734994888305664, + "learning_rate": 1e-06, + "loss": 0.731, + "mean_token_accuracy": 0.7668896317481995, + "num_tokens": 272515654.0, + "step": 10528 + }, + { + "epoch": 1.15627059081924, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2791943550109863, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7186745405197144, + "num_tokens": 272540293.0, + "step": 10529 + }, + { + "epoch": 1.1563804085218536, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.0975940227508545, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7233484387397766, + "num_tokens": 272570385.0, + "step": 10530 + }, + { + "epoch": 1.1564902262244674, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.3867061138153076, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7137978076934814, + "num_tokens": 272591512.0, + "step": 10531 + }, + { + "epoch": 1.1566000439270812, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.5725021362304688, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7209357619285583, + "num_tokens": 272614670.0, + "step": 10532 + }, + { + "epoch": 1.1567098616296947, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.211597204208374, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7159172296524048, + "num_tokens": 272640736.0, + "step": 10533 + }, + { + "epoch": 1.1568196793323084, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.245157480239868, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7170614004135132, + "num_tokens": 272665179.0, + "step": 10534 + }, + { + "epoch": 1.156929497034922, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.016169309616089, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7293091416358948, + "num_tokens": 272693745.0, + "step": 10535 + }, + { + "epoch": 1.1570393147375357, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.585829973220825, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.728119969367981, + "num_tokens": 272714068.0, + "step": 10536 + }, + { + "epoch": 1.1571491324401493, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.3354122638702393, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7044211626052856, + "num_tokens": 272737718.0, + "step": 10537 + }, + { + "epoch": 1.157258950142763, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.485501527786255, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7256395816802979, + "num_tokens": 272760243.0, + "step": 10538 + }, + { + "epoch": 1.1573687678453766, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.103745937347412, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7088720202445984, + "num_tokens": 272788577.0, + "step": 10539 + }, + { + "epoch": 1.1574785855479903, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.181410312652588, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.692659854888916, + "num_tokens": 272816076.0, + "step": 10540 + }, + { + "epoch": 1.157588403250604, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.356463670730591, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7197791337966919, + "num_tokens": 272838322.0, + "step": 10541 + }, + { + "epoch": 1.1576982209532176, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.219696044921875, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7400136590003967, + "num_tokens": 272864169.0, + "step": 10542 + }, + { + "epoch": 1.1578080386558314, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2621607780456543, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.727350115776062, + "num_tokens": 272888136.0, + "step": 10543 + }, + { + "epoch": 1.157917856358445, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2632157802581787, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7079750895500183, + "num_tokens": 272912632.0, + "step": 10544 + }, + { + "epoch": 1.1580276740610587, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.275700807571411, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7136853933334351, + "num_tokens": 272939621.0, + "step": 10545 + }, + { + "epoch": 1.1581374917636724, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.364274263381958, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7076671123504639, + "num_tokens": 272963177.0, + "step": 10546 + }, + { + "epoch": 1.158247309466286, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2929790019989014, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7368733882904053, + "num_tokens": 272987430.0, + "step": 10547 + }, + { + "epoch": 1.1583571271688997, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.48504900932312, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7282821536064148, + "num_tokens": 273008216.0, + "step": 10548 + }, + { + "epoch": 1.1584669448715132, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.471717596054077, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7113322615623474, + "num_tokens": 273031222.0, + "step": 10549 + }, + { + "epoch": 1.158576762574127, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.4165079593658447, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7299238443374634, + "num_tokens": 273053142.0, + "step": 10550 + }, + { + "epoch": 1.1586865802767405, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.2014920711517334, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7046597003936768, + "num_tokens": 273080271.0, + "step": 10551 + }, + { + "epoch": 1.1587963979793543, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.315462350845337, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7263233065605164, + "num_tokens": 273103267.0, + "step": 10552 + }, + { + "epoch": 1.1589062156819678, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.234710693359375, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7145907878875732, + "num_tokens": 273129129.0, + "step": 10553 + }, + { + "epoch": 1.1590160333845816, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.047517776489258, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7312228083610535, + "num_tokens": 273159544.0, + "step": 10554 + }, + { + "epoch": 1.1591258510871953, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.0996968746185303, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7203337550163269, + "num_tokens": 273186077.0, + "step": 10555 + }, + { + "epoch": 1.1592356687898089, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.177030324935913, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7187042236328125, + "num_tokens": 273213808.0, + "step": 10556 + }, + { + "epoch": 1.1593454864924226, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.178704261779785, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7255277037620544, + "num_tokens": 273242014.0, + "step": 10557 + }, + { + "epoch": 1.1594553041950362, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.0364346504211426, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7123257517814636, + "num_tokens": 273272011.0, + "step": 10558 + }, + { + "epoch": 1.15956512189765, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.359241008758545, + "learning_rate": 1e-06, + "loss": 0.9879, + "mean_token_accuracy": 0.6950668096542358, + "num_tokens": 273296193.0, + "step": 10559 + }, + { + "epoch": 1.1596749396002635, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.144451379776001, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7193876504898071, + "num_tokens": 273324154.0, + "step": 10560 + }, + { + "epoch": 1.1597847573028772, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 1.99342942237854, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7017147541046143, + "num_tokens": 273356852.0, + "step": 10561 + }, + { + "epoch": 1.159894575005491, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.180487871170044, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7021478414535522, + "num_tokens": 273386090.0, + "step": 10562 + }, + { + "epoch": 1.1600043927081045, + "ewc_loss": 5.930662155151367e-06, + "grad_norm": 2.0549123287200928, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7147192358970642, + "num_tokens": 273416319.0, + "step": 10563 + }, + { + "epoch": 1.1601142104107183, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.446518898010254, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7090257406234741, + "num_tokens": 273439730.0, + "step": 10564 + }, + { + "epoch": 1.1602240281133318, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.2342751026153564, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7009717226028442, + "num_tokens": 273465595.0, + "step": 10565 + }, + { + "epoch": 1.1603338458159456, + "ewc_loss": 5.9604644775390625e-06, + "grad_norm": 2.0790348052978516, + "learning_rate": 1e-06, + "loss": 1.0341, + "mean_token_accuracy": 0.6831567287445068, + "num_tokens": 273496619.0, + "step": 10566 + }, + { + "epoch": 1.160443663518559, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.5497548580169678, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7327948808670044, + "num_tokens": 273517232.0, + "step": 10567 + }, + { + "epoch": 1.1605534812211729, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2648816108703613, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.7106591463088989, + "num_tokens": 273543123.0, + "step": 10568 + }, + { + "epoch": 1.1606632989237866, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0190982818603516, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7172418832778931, + "num_tokens": 273573824.0, + "step": 10569 + }, + { + "epoch": 1.1607731166264001, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.365166664123535, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7334457635879517, + "num_tokens": 273596774.0, + "step": 10570 + }, + { + "epoch": 1.160882934329014, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0340986251831055, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7190255522727966, + "num_tokens": 273625480.0, + "step": 10571 + }, + { + "epoch": 1.1609927520316274, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2450413703918457, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7196458578109741, + "num_tokens": 273650325.0, + "step": 10572 + }, + { + "epoch": 1.1611025697342412, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3251547813415527, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6989774703979492, + "num_tokens": 273673384.0, + "step": 10573 + }, + { + "epoch": 1.1612123874368547, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3440258502960205, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.720481812953949, + "num_tokens": 273696548.0, + "step": 10574 + }, + { + "epoch": 1.1613222051394685, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.056318998336792, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7128506898880005, + "num_tokens": 273725958.0, + "step": 10575 + }, + { + "epoch": 1.161432022842082, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 1.9745303392410278, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7084803581237793, + "num_tokens": 273758568.0, + "step": 10576 + }, + { + "epoch": 1.1615418405446958, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.1735358238220215, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7532178163528442, + "num_tokens": 273786322.0, + "step": 10577 + }, + { + "epoch": 1.1616516582473095, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.22273850440979, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7193402051925659, + "num_tokens": 273812749.0, + "step": 10578 + }, + { + "epoch": 1.161761475949923, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2279891967773438, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7273372411727905, + "num_tokens": 273836812.0, + "step": 10579 + }, + { + "epoch": 1.1618712936525368, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.380455493927002, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.689376175403595, + "num_tokens": 273861639.0, + "step": 10580 + }, + { + "epoch": 1.1619811113551504, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0873172283172607, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7188363671302795, + "num_tokens": 273890785.0, + "step": 10581 + }, + { + "epoch": 1.1620909290577641, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.069929599761963, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7127254009246826, + "num_tokens": 273920468.0, + "step": 10582 + }, + { + "epoch": 1.1622007467603779, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.155914545059204, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7398992776870728, + "num_tokens": 273946197.0, + "step": 10583 + }, + { + "epoch": 1.1623105644629914, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0974578857421875, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7112253904342651, + "num_tokens": 273974399.0, + "step": 10584 + }, + { + "epoch": 1.1624203821656052, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.222395896911621, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7245913147926331, + "num_tokens": 274001424.0, + "step": 10585 + }, + { + "epoch": 1.1625301998682187, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.5408968925476074, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7048224806785583, + "num_tokens": 274024276.0, + "step": 10586 + }, + { + "epoch": 1.1626400175708325, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.29854416847229, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7084094285964966, + "num_tokens": 274048851.0, + "step": 10587 + }, + { + "epoch": 1.162749835273446, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.397712230682373, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7174205780029297, + "num_tokens": 274071445.0, + "step": 10588 + }, + { + "epoch": 1.1628596529760598, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0465736389160156, + "learning_rate": 1e-06, + "loss": 0.742, + "mean_token_accuracy": 0.7590619325637817, + "num_tokens": 274098455.0, + "step": 10589 + }, + { + "epoch": 1.1629694706786733, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.1992478370666504, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7435604333877563, + "num_tokens": 274124178.0, + "step": 10590 + }, + { + "epoch": 1.163079288381287, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.1433188915252686, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.717708170413971, + "num_tokens": 274151069.0, + "step": 10591 + }, + { + "epoch": 1.1631891060839008, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.4490699768066406, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7105296850204468, + "num_tokens": 274173561.0, + "step": 10592 + }, + { + "epoch": 1.1632989237865143, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0433292388916016, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.6843447685241699, + "num_tokens": 274205144.0, + "step": 10593 + }, + { + "epoch": 1.163408741489128, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2944650650024414, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.701633870601654, + "num_tokens": 274231929.0, + "step": 10594 + }, + { + "epoch": 1.1635185591917416, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.4004461765289307, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7373422384262085, + "num_tokens": 274256158.0, + "step": 10595 + }, + { + "epoch": 1.1636283768943554, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3551197052001953, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6996212005615234, + "num_tokens": 274283026.0, + "step": 10596 + }, + { + "epoch": 1.1637381945969691, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2233219146728516, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.703177273273468, + "num_tokens": 274309775.0, + "step": 10597 + }, + { + "epoch": 1.1638480122995827, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3504507541656494, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7175747156143188, + "num_tokens": 274334540.0, + "step": 10598 + }, + { + "epoch": 1.1639578300021964, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.4513888359069824, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7174558639526367, + "num_tokens": 274357535.0, + "step": 10599 + }, + { + "epoch": 1.16406764770481, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.288750171661377, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7207107543945312, + "num_tokens": 274382477.0, + "step": 10600 + }, + { + "epoch": 1.1641774654074237, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3210604190826416, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7297772169113159, + "num_tokens": 274408585.0, + "step": 10601 + }, + { + "epoch": 1.1642872831100373, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0402259826660156, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.7493688464164734, + "num_tokens": 274437220.0, + "step": 10602 + }, + { + "epoch": 1.164397100812651, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.008129596710205, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7182048559188843, + "num_tokens": 274469802.0, + "step": 10603 + }, + { + "epoch": 1.1645069185152646, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 1.980289340019226, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7139872312545776, + "num_tokens": 274499495.0, + "step": 10604 + }, + { + "epoch": 1.1646167362178783, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.346470355987549, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7303224802017212, + "num_tokens": 274522048.0, + "step": 10605 + }, + { + "epoch": 1.164726553920492, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.16821551322937, + "learning_rate": 1e-06, + "loss": 0.9803, + "mean_token_accuracy": 0.699367105960846, + "num_tokens": 274550397.0, + "step": 10606 + }, + { + "epoch": 1.1648363716231056, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.333915948867798, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7264467477798462, + "num_tokens": 274572289.0, + "step": 10607 + }, + { + "epoch": 1.1649461893257194, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.246577739715576, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.70803302526474, + "num_tokens": 274598271.0, + "step": 10608 + }, + { + "epoch": 1.165056007028333, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 1.9904518127441406, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7208088040351868, + "num_tokens": 274626039.0, + "step": 10609 + }, + { + "epoch": 1.1651658247309467, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.1857666969299316, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7059974074363708, + "num_tokens": 274657055.0, + "step": 10610 + }, + { + "epoch": 1.1652756424335604, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.1241090297698975, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7229344844818115, + "num_tokens": 274685455.0, + "step": 10611 + }, + { + "epoch": 1.165385460136174, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.4158709049224854, + "learning_rate": 1e-06, + "loss": 1.0267, + "mean_token_accuracy": 0.6966934204101562, + "num_tokens": 274709761.0, + "step": 10612 + }, + { + "epoch": 1.1654952778387877, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.404863119125366, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7232668995857239, + "num_tokens": 274734396.0, + "step": 10613 + }, + { + "epoch": 1.1656050955414012, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 1.9538233280181885, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7150482535362244, + "num_tokens": 274768069.0, + "step": 10614 + }, + { + "epoch": 1.165714913244015, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 1.9656320810317993, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7247006297111511, + "num_tokens": 274800029.0, + "step": 10615 + }, + { + "epoch": 1.1658247309466285, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.652339220046997, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7274367213249207, + "num_tokens": 274819886.0, + "step": 10616 + }, + { + "epoch": 1.1659345486492423, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2123398780822754, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7085795998573303, + "num_tokens": 274848395.0, + "step": 10617 + }, + { + "epoch": 1.1660443663518558, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.256223201751709, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7192569971084595, + "num_tokens": 274872749.0, + "step": 10618 + }, + { + "epoch": 1.1661541840544696, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2312729358673096, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7317880988121033, + "num_tokens": 274897661.0, + "step": 10619 + }, + { + "epoch": 1.1662640017570833, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.249880313873291, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7059437036514282, + "num_tokens": 274922900.0, + "step": 10620 + }, + { + "epoch": 1.1663738194596969, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0844733715057373, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7184375524520874, + "num_tokens": 274953199.0, + "step": 10621 + }, + { + "epoch": 1.1664836371623106, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.400425672531128, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.724272608757019, + "num_tokens": 274975312.0, + "step": 10622 + }, + { + "epoch": 1.1665934548649242, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.117405414581299, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6971362829208374, + "num_tokens": 275005756.0, + "step": 10623 + }, + { + "epoch": 1.166703272567538, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2870519161224365, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7317845225334167, + "num_tokens": 275029331.0, + "step": 10624 + }, + { + "epoch": 1.1668130902701515, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.4982292652130127, + "learning_rate": 1e-06, + "loss": 1.06, + "mean_token_accuracy": 0.6811704635620117, + "num_tokens": 275054582.0, + "step": 10625 + }, + { + "epoch": 1.1669229079727652, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3205418586730957, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.738099217414856, + "num_tokens": 275079544.0, + "step": 10626 + }, + { + "epoch": 1.167032725675379, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.102111339569092, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.7082728147506714, + "num_tokens": 275109495.0, + "step": 10627 + }, + { + "epoch": 1.1671425433779925, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.3180623054504395, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7155126333236694, + "num_tokens": 275132812.0, + "step": 10628 + }, + { + "epoch": 1.1672523610806063, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.0646681785583496, + "learning_rate": 1e-06, + "loss": 1.0098, + "mean_token_accuracy": 0.6944292783737183, + "num_tokens": 275161256.0, + "step": 10629 + }, + { + "epoch": 1.1673621787832198, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.2397446632385254, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7284279465675354, + "num_tokens": 275186318.0, + "step": 10630 + }, + { + "epoch": 1.1674719964858336, + "ewc_loss": 6.020069122314453e-06, + "grad_norm": 2.5063419342041016, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6967993974685669, + "num_tokens": 275207258.0, + "step": 10631 + }, + { + "epoch": 1.167581814188447, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2585318088531494, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.716050386428833, + "num_tokens": 275233345.0, + "step": 10632 + }, + { + "epoch": 1.1676916318910608, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.278779983520508, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7330062985420227, + "num_tokens": 275258457.0, + "step": 10633 + }, + { + "epoch": 1.1678014495936746, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.6064324378967285, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7341680526733398, + "num_tokens": 275278010.0, + "step": 10634 + }, + { + "epoch": 1.1679112672962881, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.312075614929199, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7245089411735535, + "num_tokens": 275304279.0, + "step": 10635 + }, + { + "epoch": 1.168021084998902, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.428905487060547, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7171016931533813, + "num_tokens": 275326567.0, + "step": 10636 + }, + { + "epoch": 1.1681309027015154, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 1.8681291341781616, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.6884690523147583, + "num_tokens": 275363466.0, + "step": 10637 + }, + { + "epoch": 1.1682407204041292, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.517801523208618, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7361695170402527, + "num_tokens": 275383405.0, + "step": 10638 + }, + { + "epoch": 1.1683505381067427, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.074259042739868, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.728165328502655, + "num_tokens": 275410995.0, + "step": 10639 + }, + { + "epoch": 1.1684603558093565, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0051565170288086, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.6965701580047607, + "num_tokens": 275442348.0, + "step": 10640 + }, + { + "epoch": 1.16857017351197, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 1.9642654657363892, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7075328826904297, + "num_tokens": 275473597.0, + "step": 10641 + }, + { + "epoch": 1.1686799912145838, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.017894744873047, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6985844373703003, + "num_tokens": 275505412.0, + "step": 10642 + }, + { + "epoch": 1.1687898089171975, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.418095111846924, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7161428928375244, + "num_tokens": 275528121.0, + "step": 10643 + }, + { + "epoch": 1.168899626619811, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3352437019348145, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7399104833602905, + "num_tokens": 275549898.0, + "step": 10644 + }, + { + "epoch": 1.1690094443224248, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0401391983032227, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.6945832967758179, + "num_tokens": 275579992.0, + "step": 10645 + }, + { + "epoch": 1.1691192620250384, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.697291135787964, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7103421688079834, + "num_tokens": 275602864.0, + "step": 10646 + }, + { + "epoch": 1.169229079727652, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0010433197021484, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7298009395599365, + "num_tokens": 275631977.0, + "step": 10647 + }, + { + "epoch": 1.1693388974302659, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.176001787185669, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7211666107177734, + "num_tokens": 275661055.0, + "step": 10648 + }, + { + "epoch": 1.1694487151328794, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.1602861881256104, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7307186126708984, + "num_tokens": 275686776.0, + "step": 10649 + }, + { + "epoch": 1.1695585328354932, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.033252716064453, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7426730394363403, + "num_tokens": 275715084.0, + "step": 10650 + }, + { + "epoch": 1.1696683505381067, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 1.949188232421875, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.693525493144989, + "num_tokens": 275748446.0, + "step": 10651 + }, + { + "epoch": 1.1697781682407205, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.463390588760376, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7066171169281006, + "num_tokens": 275770040.0, + "step": 10652 + }, + { + "epoch": 1.169887985943334, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.4748048782348633, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7234342098236084, + "num_tokens": 275793328.0, + "step": 10653 + }, + { + "epoch": 1.1699978036459477, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.027674913406372, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7085059881210327, + "num_tokens": 275823134.0, + "step": 10654 + }, + { + "epoch": 1.1701076213485613, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.182912826538086, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7030289173126221, + "num_tokens": 275851797.0, + "step": 10655 + }, + { + "epoch": 1.170217439051175, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.4312376976013184, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7154556512832642, + "num_tokens": 275874996.0, + "step": 10656 + }, + { + "epoch": 1.1703272567537888, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.211437702178955, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7112928628921509, + "num_tokens": 275901105.0, + "step": 10657 + }, + { + "epoch": 1.1704370744564023, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.1770565509796143, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7157665491104126, + "num_tokens": 275928391.0, + "step": 10658 + }, + { + "epoch": 1.170546892159016, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2213642597198486, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7211298942565918, + "num_tokens": 275955151.0, + "step": 10659 + }, + { + "epoch": 1.1706567098616296, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.4013736248016357, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7171869277954102, + "num_tokens": 275977107.0, + "step": 10660 + }, + { + "epoch": 1.1707665275642434, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0926387310028076, + "learning_rate": 1e-06, + "loss": 1.0223, + "mean_token_accuracy": 0.6861570477485657, + "num_tokens": 276006547.0, + "step": 10661 + }, + { + "epoch": 1.1708763452668571, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0095481872558594, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7185438871383667, + "num_tokens": 276034604.0, + "step": 10662 + }, + { + "epoch": 1.1709861629694707, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.206821918487549, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7020052671432495, + "num_tokens": 276062945.0, + "step": 10663 + }, + { + "epoch": 1.1710959806720844, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.657437801361084, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.743949294090271, + "num_tokens": 276081219.0, + "step": 10664 + }, + { + "epoch": 1.171205798374698, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3278112411499023, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7051049470901489, + "num_tokens": 276105780.0, + "step": 10665 + }, + { + "epoch": 1.1713156160773117, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3861231803894043, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7331706881523132, + "num_tokens": 276127673.0, + "step": 10666 + }, + { + "epoch": 1.1714254337799253, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.148294448852539, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7170410752296448, + "num_tokens": 276155202.0, + "step": 10667 + }, + { + "epoch": 1.171535251482539, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2350425720214844, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7093644142150879, + "num_tokens": 276182164.0, + "step": 10668 + }, + { + "epoch": 1.1716450691851525, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.011868476867676, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.732477068901062, + "num_tokens": 276210888.0, + "step": 10669 + }, + { + "epoch": 1.1717548868877663, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.032144784927368, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7254321575164795, + "num_tokens": 276238745.0, + "step": 10670 + }, + { + "epoch": 1.17186470459038, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.347735643386841, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7253256440162659, + "num_tokens": 276262017.0, + "step": 10671 + }, + { + "epoch": 1.1719745222929936, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.1022377014160156, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7108537554740906, + "num_tokens": 276291389.0, + "step": 10672 + }, + { + "epoch": 1.1720843399956073, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3804211616516113, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7254341840744019, + "num_tokens": 276314507.0, + "step": 10673 + }, + { + "epoch": 1.1721941576982209, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.082090139389038, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7170466184616089, + "num_tokens": 276341334.0, + "step": 10674 + }, + { + "epoch": 1.1723039754008346, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.223128080368042, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.7140935659408569, + "num_tokens": 276366493.0, + "step": 10675 + }, + { + "epoch": 1.1724137931034484, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3144009113311768, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7238420248031616, + "num_tokens": 276388674.0, + "step": 10676 + }, + { + "epoch": 1.172523610806062, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2165415287017822, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7098068594932556, + "num_tokens": 276412824.0, + "step": 10677 + }, + { + "epoch": 1.1726334285086757, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2501418590545654, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7260945439338684, + "num_tokens": 276440248.0, + "step": 10678 + }, + { + "epoch": 1.1727432462112892, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.214834451675415, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7241771817207336, + "num_tokens": 276466961.0, + "step": 10679 + }, + { + "epoch": 1.172853063913903, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.418968677520752, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.7444696426391602, + "num_tokens": 276488568.0, + "step": 10680 + }, + { + "epoch": 1.1729628816165165, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.2845399379730225, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.724096417427063, + "num_tokens": 276514628.0, + "step": 10681 + }, + { + "epoch": 1.1730726993191303, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.0300774574279785, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7147672176361084, + "num_tokens": 276546001.0, + "step": 10682 + }, + { + "epoch": 1.1731825170217438, + "ewc_loss": 6.079673767089844e-06, + "grad_norm": 2.3769781589508057, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7268961668014526, + "num_tokens": 276567144.0, + "step": 10683 + }, + { + "epoch": 1.1732923347243576, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.0058462619781494, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7037249803543091, + "num_tokens": 276600189.0, + "step": 10684 + }, + { + "epoch": 1.1734021524269713, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.285834312438965, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7201581597328186, + "num_tokens": 276626814.0, + "step": 10685 + }, + { + "epoch": 1.1735119701295849, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.0620951652526855, + "learning_rate": 1e-06, + "loss": 1.0468, + "mean_token_accuracy": 0.6853863000869751, + "num_tokens": 276660157.0, + "step": 10686 + }, + { + "epoch": 1.1736217878321986, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.34977126121521, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7100386619567871, + "num_tokens": 276687209.0, + "step": 10687 + }, + { + "epoch": 1.1737316055348122, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.040461778640747, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7236230373382568, + "num_tokens": 276719689.0, + "step": 10688 + }, + { + "epoch": 1.173841423237426, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 1.9789378643035889, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7191793918609619, + "num_tokens": 276750718.0, + "step": 10689 + }, + { + "epoch": 1.1739512409400394, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.2965362071990967, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7155552506446838, + "num_tokens": 276775344.0, + "step": 10690 + }, + { + "epoch": 1.1740610586426532, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.187617063522339, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.6935788989067078, + "num_tokens": 276802480.0, + "step": 10691 + }, + { + "epoch": 1.1741708763452667, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.3669281005859375, + "learning_rate": 1e-06, + "loss": 0.7881, + "mean_token_accuracy": 0.7471227645874023, + "num_tokens": 276825093.0, + "step": 10692 + }, + { + "epoch": 1.1742806940478805, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.1092417240142822, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6903978586196899, + "num_tokens": 276859244.0, + "step": 10693 + }, + { + "epoch": 1.1743905117504942, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.2238709926605225, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7195104360580444, + "num_tokens": 276883396.0, + "step": 10694 + }, + { + "epoch": 1.1745003294531078, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.1255788803100586, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7310751080513, + "num_tokens": 276910340.0, + "step": 10695 + }, + { + "epoch": 1.1746101471557215, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.3478777408599854, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7274609804153442, + "num_tokens": 276935607.0, + "step": 10696 + }, + { + "epoch": 1.174719964858335, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.461137533187866, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.717380702495575, + "num_tokens": 276958407.0, + "step": 10697 + }, + { + "epoch": 1.1748297825609488, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.2584755420684814, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7108359932899475, + "num_tokens": 276984274.0, + "step": 10698 + }, + { + "epoch": 1.1749396002635626, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.488316297531128, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7410387992858887, + "num_tokens": 277003924.0, + "step": 10699 + }, + { + "epoch": 1.1750494179661761, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.2949767112731934, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7079757452011108, + "num_tokens": 277029326.0, + "step": 10700 + }, + { + "epoch": 1.1751592356687899, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.350839614868164, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7462577819824219, + "num_tokens": 277052982.0, + "step": 10701 + }, + { + "epoch": 1.1752690533714034, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.0745365619659424, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7231429815292358, + "num_tokens": 277082856.0, + "step": 10702 + }, + { + "epoch": 1.1753788710740172, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.187305212020874, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.6951489448547363, + "num_tokens": 277110580.0, + "step": 10703 + }, + { + "epoch": 1.1754886887766307, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.400343894958496, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7206757068634033, + "num_tokens": 277133087.0, + "step": 10704 + }, + { + "epoch": 1.1755985064792445, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.244569778442383, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7179878950119019, + "num_tokens": 277157403.0, + "step": 10705 + }, + { + "epoch": 1.175708324181858, + "ewc_loss": 6.139278411865234e-06, + "grad_norm": 2.1414427757263184, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7147684097290039, + "num_tokens": 277183739.0, + "step": 10706 + }, + { + "epoch": 1.1758181418844718, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.106250286102295, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7077455520629883, + "num_tokens": 277210821.0, + "step": 10707 + }, + { + "epoch": 1.1759279595870855, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.087052822113037, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7310138940811157, + "num_tokens": 277236725.0, + "step": 10708 + }, + { + "epoch": 1.176037777289699, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 1.9795514345169067, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7108811736106873, + "num_tokens": 277268483.0, + "step": 10709 + }, + { + "epoch": 1.1761475949923128, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0520081520080566, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7286763787269592, + "num_tokens": 277296117.0, + "step": 10710 + }, + { + "epoch": 1.1762574126949263, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4402942657470703, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7234024405479431, + "num_tokens": 277318543.0, + "step": 10711 + }, + { + "epoch": 1.17636723039754, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.103139638900757, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7109778523445129, + "num_tokens": 277347866.0, + "step": 10712 + }, + { + "epoch": 1.1764770481001539, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1670548915863037, + "learning_rate": 1e-06, + "loss": 0.9935, + "mean_token_accuracy": 0.6936917304992676, + "num_tokens": 277376341.0, + "step": 10713 + }, + { + "epoch": 1.1765868658027674, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1972930431365967, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7295533418655396, + "num_tokens": 277403427.0, + "step": 10714 + }, + { + "epoch": 1.1766966835053811, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1070284843444824, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7244474291801453, + "num_tokens": 277430711.0, + "step": 10715 + }, + { + "epoch": 1.1768065012079947, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0469484329223633, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.6966260075569153, + "num_tokens": 277460092.0, + "step": 10716 + }, + { + "epoch": 1.1769163189106084, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1552207469940186, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.6991459131240845, + "num_tokens": 277486876.0, + "step": 10717 + }, + { + "epoch": 1.177026136613222, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.35318660736084, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7098104953765869, + "num_tokens": 277510917.0, + "step": 10718 + }, + { + "epoch": 1.1771359543158357, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2247390747070312, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.706402063369751, + "num_tokens": 277537077.0, + "step": 10719 + }, + { + "epoch": 1.1772457720184493, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4510254859924316, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7210996747016907, + "num_tokens": 277559338.0, + "step": 10720 + }, + { + "epoch": 1.177355589721063, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.5652542114257812, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7346013784408569, + "num_tokens": 277580893.0, + "step": 10721 + }, + { + "epoch": 1.1774654074236768, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4158213138580322, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7120550274848938, + "num_tokens": 277602487.0, + "step": 10722 + }, + { + "epoch": 1.1775752251262903, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0209803581237793, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7117902636528015, + "num_tokens": 277633982.0, + "step": 10723 + }, + { + "epoch": 1.177685042828904, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.367331027984619, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7267093658447266, + "num_tokens": 277656418.0, + "step": 10724 + }, + { + "epoch": 1.1777948605315176, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.318990707397461, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7026727795600891, + "num_tokens": 277681706.0, + "step": 10725 + }, + { + "epoch": 1.1779046782341314, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.366931676864624, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7312595248222351, + "num_tokens": 277704022.0, + "step": 10726 + }, + { + "epoch": 1.1780144959367451, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.551778554916382, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7512486577033997, + "num_tokens": 277722483.0, + "step": 10727 + }, + { + "epoch": 1.1781243136393587, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3804078102111816, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7242024540901184, + "num_tokens": 277744425.0, + "step": 10728 + }, + { + "epoch": 1.1782341313419724, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.679950475692749, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7285354733467102, + "num_tokens": 277762468.0, + "step": 10729 + }, + { + "epoch": 1.178343949044586, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2178585529327393, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7073729038238525, + "num_tokens": 277789603.0, + "step": 10730 + }, + { + "epoch": 1.1784537667471997, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.211662530899048, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.6983358860015869, + "num_tokens": 277816454.0, + "step": 10731 + }, + { + "epoch": 1.1785635844498132, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.121624231338501, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.735427975654602, + "num_tokens": 277844616.0, + "step": 10732 + }, + { + "epoch": 1.178673402152427, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.288839340209961, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.710633397102356, + "num_tokens": 277870836.0, + "step": 10733 + }, + { + "epoch": 1.1787832198550405, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3100428581237793, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7289490699768066, + "num_tokens": 277894579.0, + "step": 10734 + }, + { + "epoch": 1.1788930375576543, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.125505208969116, + "learning_rate": 1e-06, + "loss": 1.0304, + "mean_token_accuracy": 0.687315046787262, + "num_tokens": 277924162.0, + "step": 10735 + }, + { + "epoch": 1.179002855260268, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.052412986755371, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7185086011886597, + "num_tokens": 277952941.0, + "step": 10736 + }, + { + "epoch": 1.1791126729628816, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.260202407836914, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7041642069816589, + "num_tokens": 277977943.0, + "step": 10737 + }, + { + "epoch": 1.1792224906654953, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.320507287979126, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7067548036575317, + "num_tokens": 278000702.0, + "step": 10738 + }, + { + "epoch": 1.1793323083681089, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.5269343852996826, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7217686176300049, + "num_tokens": 278023195.0, + "step": 10739 + }, + { + "epoch": 1.1794421260707226, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1265461444854736, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7278951406478882, + "num_tokens": 278051203.0, + "step": 10740 + }, + { + "epoch": 1.1795519437733364, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1412012577056885, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7215080261230469, + "num_tokens": 278076421.0, + "step": 10741 + }, + { + "epoch": 1.17966176147595, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3596341609954834, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7093391418457031, + "num_tokens": 278101100.0, + "step": 10742 + }, + { + "epoch": 1.1797715791785637, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3778347969055176, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.704315185546875, + "num_tokens": 278126360.0, + "step": 10743 + }, + { + "epoch": 1.1798813968811772, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1459009647369385, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7219221591949463, + "num_tokens": 278152646.0, + "step": 10744 + }, + { + "epoch": 1.179991214583791, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.43190860748291, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7088395953178406, + "num_tokens": 278174769.0, + "step": 10745 + }, + { + "epoch": 1.1801010322864045, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1486151218414307, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7260252237319946, + "num_tokens": 278203501.0, + "step": 10746 + }, + { + "epoch": 1.1802108499890183, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.1151375770568848, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7321683168411255, + "num_tokens": 278230519.0, + "step": 10747 + }, + { + "epoch": 1.1803206676916318, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2126569747924805, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7299122214317322, + "num_tokens": 278256524.0, + "step": 10748 + }, + { + "epoch": 1.1804304853942456, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.4710373878479004, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.735129714012146, + "num_tokens": 278278805.0, + "step": 10749 + }, + { + "epoch": 1.1805403030968593, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.172785997390747, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7153991460800171, + "num_tokens": 278307389.0, + "step": 10750 + }, + { + "epoch": 1.1806501207994728, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.297750473022461, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7084864377975464, + "num_tokens": 278332455.0, + "step": 10751 + }, + { + "epoch": 1.1807599385020866, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0979316234588623, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7150601744651794, + "num_tokens": 278362515.0, + "step": 10752 + }, + { + "epoch": 1.1808697562047001, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.231539011001587, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7368190288543701, + "num_tokens": 278388616.0, + "step": 10753 + }, + { + "epoch": 1.180979573907314, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3779776096343994, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7076696157455444, + "num_tokens": 278414433.0, + "step": 10754 + }, + { + "epoch": 1.1810893916099274, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.417243480682373, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7292604446411133, + "num_tokens": 278435896.0, + "step": 10755 + }, + { + "epoch": 1.1811992093125412, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.403200387954712, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7426458597183228, + "num_tokens": 278457544.0, + "step": 10756 + }, + { + "epoch": 1.1813090270151547, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.202190399169922, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7022475600242615, + "num_tokens": 278483248.0, + "step": 10757 + }, + { + "epoch": 1.1814188447177685, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3895363807678223, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7114617228507996, + "num_tokens": 278506615.0, + "step": 10758 + }, + { + "epoch": 1.1815286624203822, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2014706134796143, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7128618359565735, + "num_tokens": 278533229.0, + "step": 10759 + }, + { + "epoch": 1.1816384801229958, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.153071641921997, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7253669500350952, + "num_tokens": 278559948.0, + "step": 10760 + }, + { + "epoch": 1.1817482978256095, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2954812049865723, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7150660753250122, + "num_tokens": 278583103.0, + "step": 10761 + }, + { + "epoch": 1.181858115528223, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2162675857543945, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7251261472702026, + "num_tokens": 278610819.0, + "step": 10762 + }, + { + "epoch": 1.1819679332308368, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0882656574249268, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7428250312805176, + "num_tokens": 278639297.0, + "step": 10763 + }, + { + "epoch": 1.1820777509334506, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.261216402053833, + "learning_rate": 1e-06, + "loss": 0.9893, + "mean_token_accuracy": 0.7030038833618164, + "num_tokens": 278668029.0, + "step": 10764 + }, + { + "epoch": 1.1821875686360641, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.9938251972198486, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7190825939178467, + "num_tokens": 278684837.0, + "step": 10765 + }, + { + "epoch": 1.1822973863386779, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.143601417541504, + "learning_rate": 1e-06, + "loss": 1.0573, + "mean_token_accuracy": 0.67881178855896, + "num_tokens": 278716488.0, + "step": 10766 + }, + { + "epoch": 1.1824072040412914, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.5170750617980957, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7144500613212585, + "num_tokens": 278739057.0, + "step": 10767 + }, + { + "epoch": 1.1825170217439052, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.6993331909179688, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7331730723381042, + "num_tokens": 278757143.0, + "step": 10768 + }, + { + "epoch": 1.1826268394465187, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.0236456394195557, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.722236156463623, + "num_tokens": 278788023.0, + "step": 10769 + }, + { + "epoch": 1.1827366571491325, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.2938621044158936, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7189978361129761, + "num_tokens": 278811875.0, + "step": 10770 + }, + { + "epoch": 1.182846474851746, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.3506126403808594, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7087174654006958, + "num_tokens": 278837201.0, + "step": 10771 + }, + { + "epoch": 1.1829562925543597, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.696059226989746, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7367357015609741, + "num_tokens": 278856995.0, + "step": 10772 + }, + { + "epoch": 1.1830661102569735, + "ewc_loss": 6.16908073425293e-06, + "grad_norm": 2.406949043273926, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7156832218170166, + "num_tokens": 278879410.0, + "step": 10773 + }, + { + "epoch": 1.183175927959587, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.343621015548706, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7128326892852783, + "num_tokens": 278902806.0, + "step": 10774 + }, + { + "epoch": 1.1832857456622008, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.311434745788574, + "learning_rate": 1e-06, + "loss": 0.9732, + "mean_token_accuracy": 0.7065913677215576, + "num_tokens": 278928228.0, + "step": 10775 + }, + { + "epoch": 1.1833955633648143, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.2367358207702637, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.6979267597198486, + "num_tokens": 278956107.0, + "step": 10776 + }, + { + "epoch": 1.183505381067428, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.196364402770996, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7393327951431274, + "num_tokens": 278980976.0, + "step": 10777 + }, + { + "epoch": 1.1836151987700418, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.324963331222534, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7187906503677368, + "num_tokens": 279006438.0, + "step": 10778 + }, + { + "epoch": 1.1837250164726554, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.757239580154419, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7142586708068848, + "num_tokens": 279025046.0, + "step": 10779 + }, + { + "epoch": 1.1838348341752691, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.0925261974334717, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7191969156265259, + "num_tokens": 279054000.0, + "step": 10780 + }, + { + "epoch": 1.1839446518778827, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9571020603179932, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.694575309753418, + "num_tokens": 279086191.0, + "step": 10781 + }, + { + "epoch": 1.1840544695804964, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.5543370246887207, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7225965261459351, + "num_tokens": 279105947.0, + "step": 10782 + }, + { + "epoch": 1.18416428728311, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.240447521209717, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7358813881874084, + "num_tokens": 279132323.0, + "step": 10783 + }, + { + "epoch": 1.1842741049857237, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.277787446975708, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7082329988479614, + "num_tokens": 279157452.0, + "step": 10784 + }, + { + "epoch": 1.1843839226883373, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.126321315765381, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7167254686355591, + "num_tokens": 279185069.0, + "step": 10785 + }, + { + "epoch": 1.184493740390951, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.15874981880188, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.730149507522583, + "num_tokens": 279212124.0, + "step": 10786 + }, + { + "epoch": 1.1846035580935648, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.2505195140838623, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7206734418869019, + "num_tokens": 279239591.0, + "step": 10787 + }, + { + "epoch": 1.1847133757961783, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.0600035190582275, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7277705669403076, + "num_tokens": 279269866.0, + "step": 10788 + }, + { + "epoch": 1.184823193498792, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.3381435871124268, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.711108922958374, + "num_tokens": 279293521.0, + "step": 10789 + }, + { + "epoch": 1.1849330112014056, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9263304471969604, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7283138036727905, + "num_tokens": 279326470.0, + "step": 10790 + }, + { + "epoch": 1.1850428289040194, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1712894439697266, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7490614056587219, + "num_tokens": 279351307.0, + "step": 10791 + }, + { + "epoch": 1.185152646606633, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1216347217559814, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7003523111343384, + "num_tokens": 279379656.0, + "step": 10792 + }, + { + "epoch": 1.1852624643092466, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.350386619567871, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7316136360168457, + "num_tokens": 279401365.0, + "step": 10793 + }, + { + "epoch": 1.1853722820118604, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9850192070007324, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7030629515647888, + "num_tokens": 279431184.0, + "step": 10794 + }, + { + "epoch": 1.185482099714474, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.3519740104675293, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7121171951293945, + "num_tokens": 279455386.0, + "step": 10795 + }, + { + "epoch": 1.1855919174170877, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.279893636703491, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7172847986221313, + "num_tokens": 279480157.0, + "step": 10796 + }, + { + "epoch": 1.1857017351197012, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9057211875915527, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7216736674308777, + "num_tokens": 279511184.0, + "step": 10797 + }, + { + "epoch": 1.185811552822315, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.358154296875, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7183937430381775, + "num_tokens": 279536340.0, + "step": 10798 + }, + { + "epoch": 1.1859213705249285, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.0004539489746094, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7279470562934875, + "num_tokens": 279568355.0, + "step": 10799 + }, + { + "epoch": 1.1860311882275423, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.344026565551758, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7169914245605469, + "num_tokens": 279592058.0, + "step": 10800 + }, + { + "epoch": 1.186141005930156, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.311683416366577, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7050131559371948, + "num_tokens": 279616851.0, + "step": 10801 + }, + { + "epoch": 1.1862508236327696, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.24928879737854, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7195248603820801, + "num_tokens": 279643692.0, + "step": 10802 + }, + { + "epoch": 1.1863606413353833, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1863646507263184, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7223156094551086, + "num_tokens": 279668872.0, + "step": 10803 + }, + { + "epoch": 1.1864704590379969, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.2345032691955566, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7201147079467773, + "num_tokens": 279695404.0, + "step": 10804 + }, + { + "epoch": 1.1865802767406106, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.3939332962036133, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.725793719291687, + "num_tokens": 279718073.0, + "step": 10805 + }, + { + "epoch": 1.1866900944432242, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.2232859134674072, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7114245891571045, + "num_tokens": 279747779.0, + "step": 10806 + }, + { + "epoch": 1.186799912145838, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9509129524230957, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.706017255783081, + "num_tokens": 279781480.0, + "step": 10807 + }, + { + "epoch": 1.1869097298484517, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.115513324737549, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7165331840515137, + "num_tokens": 279809082.0, + "step": 10808 + }, + { + "epoch": 1.1870195475510652, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.3927724361419678, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.729297399520874, + "num_tokens": 279832111.0, + "step": 10809 + }, + { + "epoch": 1.187129365253679, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1782431602478027, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7136049270629883, + "num_tokens": 279860701.0, + "step": 10810 + }, + { + "epoch": 1.1872391829562925, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.0370357036590576, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7070726156234741, + "num_tokens": 279891858.0, + "step": 10811 + }, + { + "epoch": 1.1873490006589063, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.568058967590332, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7273225784301758, + "num_tokens": 279911871.0, + "step": 10812 + }, + { + "epoch": 1.1874588183615198, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1485228538513184, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7162668704986572, + "num_tokens": 279937887.0, + "step": 10813 + }, + { + "epoch": 1.1875686360641335, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.1373178958892822, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7123546600341797, + "num_tokens": 279965585.0, + "step": 10814 + }, + { + "epoch": 1.1876784537667473, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 2.6683249473571777, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7089235782623291, + "num_tokens": 279985822.0, + "step": 10815 + }, + { + "epoch": 1.1877882714693608, + "ewc_loss": 6.198883056640625e-06, + "grad_norm": 1.9997742176055908, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7207884788513184, + "num_tokens": 280016291.0, + "step": 10816 + }, + { + "epoch": 1.1878980891719746, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.1275794506073, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7068036794662476, + "num_tokens": 280045789.0, + "step": 10817 + }, + { + "epoch": 1.1880079068745881, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.4442641735076904, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.73836350440979, + "num_tokens": 280068256.0, + "step": 10818 + }, + { + "epoch": 1.1881177245772019, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3155484199523926, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7514628171920776, + "num_tokens": 280092650.0, + "step": 10819 + }, + { + "epoch": 1.1882275422798154, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2892391681671143, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7175110578536987, + "num_tokens": 280118424.0, + "step": 10820 + }, + { + "epoch": 1.1883373599824292, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.305875301361084, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7191394567489624, + "num_tokens": 280142730.0, + "step": 10821 + }, + { + "epoch": 1.1884471776850427, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.562239408493042, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7103086709976196, + "num_tokens": 280163927.0, + "step": 10822 + }, + { + "epoch": 1.1885569953876565, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.5338127613067627, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7220451831817627, + "num_tokens": 280184953.0, + "step": 10823 + }, + { + "epoch": 1.1886668130902702, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.534865379333496, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7108056545257568, + "num_tokens": 280206670.0, + "step": 10824 + }, + { + "epoch": 1.1887766307928838, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.172915458679199, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7230682373046875, + "num_tokens": 280232882.0, + "step": 10825 + }, + { + "epoch": 1.1888864484954975, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.321727991104126, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7286485433578491, + "num_tokens": 280257097.0, + "step": 10826 + }, + { + "epoch": 1.188996266198111, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.159193992614746, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7310160994529724, + "num_tokens": 280283838.0, + "step": 10827 + }, + { + "epoch": 1.1891060839007248, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.557845115661621, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7271660566329956, + "num_tokens": 280305715.0, + "step": 10828 + }, + { + "epoch": 1.1892159016033386, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.6478488445281982, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7342749834060669, + "num_tokens": 280323948.0, + "step": 10829 + }, + { + "epoch": 1.189325719305952, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2133219242095947, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.71600341796875, + "num_tokens": 280352256.0, + "step": 10830 + }, + { + "epoch": 1.1894355370085659, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3315343856811523, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.7011585235595703, + "num_tokens": 280381460.0, + "step": 10831 + }, + { + "epoch": 1.1895453547111794, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.509225368499756, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7288051843643188, + "num_tokens": 280402295.0, + "step": 10832 + }, + { + "epoch": 1.1896551724137931, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.354583501815796, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7523252367973328, + "num_tokens": 280424774.0, + "step": 10833 + }, + { + "epoch": 1.1897649901164067, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.611058473587036, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7037036418914795, + "num_tokens": 280444824.0, + "step": 10834 + }, + { + "epoch": 1.1898748078190204, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.392427921295166, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7122254371643066, + "num_tokens": 280469134.0, + "step": 10835 + }, + { + "epoch": 1.189984625521634, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.0250556468963623, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7126169204711914, + "num_tokens": 280499041.0, + "step": 10836 + }, + { + "epoch": 1.1900944432242477, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.1147844791412354, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7098925113677979, + "num_tokens": 280528314.0, + "step": 10837 + }, + { + "epoch": 1.1902042609268615, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2328786849975586, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7117599248886108, + "num_tokens": 280554192.0, + "step": 10838 + }, + { + "epoch": 1.190314078629475, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.1938018798828125, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7234372496604919, + "num_tokens": 280579201.0, + "step": 10839 + }, + { + "epoch": 1.1904238963320888, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 1.942231297492981, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6965259313583374, + "num_tokens": 280613242.0, + "step": 10840 + }, + { + "epoch": 1.1905337140347023, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.104886770248413, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7246718406677246, + "num_tokens": 280641227.0, + "step": 10841 + }, + { + "epoch": 1.190643531737316, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.0084590911865234, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7126992344856262, + "num_tokens": 280671471.0, + "step": 10842 + }, + { + "epoch": 1.1907533494399298, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3254621028900146, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.7064043879508972, + "num_tokens": 280695950.0, + "step": 10843 + }, + { + "epoch": 1.1908631671425434, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.6953213214874268, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7453843355178833, + "num_tokens": 280714361.0, + "step": 10844 + }, + { + "epoch": 1.1909729848451571, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.6676840782165527, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7462754249572754, + "num_tokens": 280734389.0, + "step": 10845 + }, + { + "epoch": 1.1910828025477707, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.1566216945648193, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.73481285572052, + "num_tokens": 280759798.0, + "step": 10846 + }, + { + "epoch": 1.1911926202503844, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.4345743656158447, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7295414805412292, + "num_tokens": 280782156.0, + "step": 10847 + }, + { + "epoch": 1.191302437952998, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2721574306488037, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7142423391342163, + "num_tokens": 280807703.0, + "step": 10848 + }, + { + "epoch": 1.1914122556556117, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3384509086608887, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7114144563674927, + "num_tokens": 280831959.0, + "step": 10849 + }, + { + "epoch": 1.1915220733582252, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 1.9409993886947632, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7176851034164429, + "num_tokens": 280864614.0, + "step": 10850 + }, + { + "epoch": 1.191631891060839, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.243969678878784, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7053387761116028, + "num_tokens": 280891174.0, + "step": 10851 + }, + { + "epoch": 1.1917417087634528, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.1298210620880127, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7168737053871155, + "num_tokens": 280919838.0, + "step": 10852 + }, + { + "epoch": 1.1918515264660663, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.750993013381958, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7249893546104431, + "num_tokens": 280938891.0, + "step": 10853 + }, + { + "epoch": 1.19196134416868, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2887470722198486, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7193865776062012, + "num_tokens": 280962915.0, + "step": 10854 + }, + { + "epoch": 1.1920711618712936, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.0642166137695312, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7061073780059814, + "num_tokens": 280991722.0, + "step": 10855 + }, + { + "epoch": 1.1921809795739073, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.0440778732299805, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.7069120407104492, + "num_tokens": 281021599.0, + "step": 10856 + }, + { + "epoch": 1.192290797276521, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 1.9964845180511475, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7241369485855103, + "num_tokens": 281051685.0, + "step": 10857 + }, + { + "epoch": 1.1924006149791346, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.380591869354248, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7148940563201904, + "num_tokens": 281074117.0, + "step": 10858 + }, + { + "epoch": 1.1925104326817484, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.213644504547119, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7074616551399231, + "num_tokens": 281102277.0, + "step": 10859 + }, + { + "epoch": 1.192620250384362, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2415928840637207, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7115892767906189, + "num_tokens": 281128923.0, + "step": 10860 + }, + { + "epoch": 1.1927300680869757, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.086174488067627, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7454025149345398, + "num_tokens": 281156396.0, + "step": 10861 + }, + { + "epoch": 1.1928398857895892, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.3226890563964844, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7260291576385498, + "num_tokens": 281181795.0, + "step": 10862 + }, + { + "epoch": 1.192949703492203, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.435518503189087, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7130516171455383, + "num_tokens": 281205956.0, + "step": 10863 + }, + { + "epoch": 1.1930595211948165, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.2149014472961426, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7116421461105347, + "num_tokens": 281231044.0, + "step": 10864 + }, + { + "epoch": 1.1931693388974303, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.224138021469116, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7340139746665955, + "num_tokens": 281256379.0, + "step": 10865 + }, + { + "epoch": 1.193279156600044, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.0530388355255127, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7160138487815857, + "num_tokens": 281286506.0, + "step": 10866 + }, + { + "epoch": 1.1933889743026576, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.218683958053589, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7129122018814087, + "num_tokens": 281315045.0, + "step": 10867 + }, + { + "epoch": 1.1934987920052713, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.344813585281372, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7240950465202332, + "num_tokens": 281338268.0, + "step": 10868 + }, + { + "epoch": 1.1936086097078848, + "ewc_loss": 6.258487701416016e-06, + "grad_norm": 2.600984811782837, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.732337474822998, + "num_tokens": 281358237.0, + "step": 10869 + }, + { + "epoch": 1.1937184274104986, + "ewc_loss": 6.288290023803711e-06, + "grad_norm": 2.4659087657928467, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7376143932342529, + "num_tokens": 281379501.0, + "step": 10870 + }, + { + "epoch": 1.1938282451131121, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5223593711853027, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7280266284942627, + "num_tokens": 281399907.0, + "step": 10871 + }, + { + "epoch": 1.193938062815726, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1643505096435547, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7203652858734131, + "num_tokens": 281426171.0, + "step": 10872 + }, + { + "epoch": 1.1940478805183394, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2922329902648926, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7131698131561279, + "num_tokens": 281452592.0, + "step": 10873 + }, + { + "epoch": 1.1941576982209532, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1701011657714844, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.705842912197113, + "num_tokens": 281479924.0, + "step": 10874 + }, + { + "epoch": 1.194267515923567, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.357117176055908, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7194913029670715, + "num_tokens": 281503398.0, + "step": 10875 + }, + { + "epoch": 1.1943773336261805, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3508758544921875, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7205281257629395, + "num_tokens": 281525676.0, + "step": 10876 + }, + { + "epoch": 1.1944871513287942, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3013434410095215, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7267731428146362, + "num_tokens": 281551631.0, + "step": 10877 + }, + { + "epoch": 1.1945969690314078, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5720677375793457, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7183654308319092, + "num_tokens": 281572796.0, + "step": 10878 + }, + { + "epoch": 1.1947067867340215, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.415536880493164, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.714832603931427, + "num_tokens": 281597170.0, + "step": 10879 + }, + { + "epoch": 1.1948166044366353, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.406759738922119, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.727999210357666, + "num_tokens": 281620614.0, + "step": 10880 + }, + { + "epoch": 1.1949264221392488, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2151033878326416, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7171303033828735, + "num_tokens": 281646288.0, + "step": 10881 + }, + { + "epoch": 1.1950362398418626, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.409846305847168, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7336517572402954, + "num_tokens": 281667782.0, + "step": 10882 + }, + { + "epoch": 1.1951460575444761, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2082936763763428, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.724153995513916, + "num_tokens": 281694308.0, + "step": 10883 + }, + { + "epoch": 1.1952558752470899, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3866240978240967, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7316327095031738, + "num_tokens": 281716188.0, + "step": 10884 + }, + { + "epoch": 1.1953656929497034, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.177661657333374, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.6964062452316284, + "num_tokens": 281743927.0, + "step": 10885 + }, + { + "epoch": 1.1954755106523172, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0854172706604004, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7176265120506287, + "num_tokens": 281770630.0, + "step": 10886 + }, + { + "epoch": 1.1955853283549307, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.612011432647705, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.701027512550354, + "num_tokens": 281793140.0, + "step": 10887 + }, + { + "epoch": 1.1956951460575445, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.497783660888672, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7327994108200073, + "num_tokens": 281814623.0, + "step": 10888 + }, + { + "epoch": 1.1958049637601582, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3607449531555176, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7178773880004883, + "num_tokens": 281838998.0, + "step": 10889 + }, + { + "epoch": 1.1959147814627717, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.069204330444336, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7307765483856201, + "num_tokens": 281867363.0, + "step": 10890 + }, + { + "epoch": 1.1960245991653855, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.391510248184204, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7254619598388672, + "num_tokens": 281889231.0, + "step": 10891 + }, + { + "epoch": 1.196134416867999, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1832737922668457, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7093328237533569, + "num_tokens": 281917109.0, + "step": 10892 + }, + { + "epoch": 1.1962442345706128, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.365354061126709, + "learning_rate": 1e-06, + "loss": 0.9786, + "mean_token_accuracy": 0.7086331844329834, + "num_tokens": 281942238.0, + "step": 10893 + }, + { + "epoch": 1.1963540522732266, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5573787689208984, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.725588321685791, + "num_tokens": 281963200.0, + "step": 10894 + }, + { + "epoch": 1.19646386997584, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.178327798843384, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7081105709075928, + "num_tokens": 281991008.0, + "step": 10895 + }, + { + "epoch": 1.1965736876784538, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4144160747528076, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7354655265808105, + "num_tokens": 282013040.0, + "step": 10896 + }, + { + "epoch": 1.1966835053810674, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4746291637420654, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.728224515914917, + "num_tokens": 282033724.0, + "step": 10897 + }, + { + "epoch": 1.1967933230836811, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0401062965393066, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7016251683235168, + "num_tokens": 282064888.0, + "step": 10898 + }, + { + "epoch": 1.1969031407862947, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1842384338378906, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7207593321800232, + "num_tokens": 282091842.0, + "step": 10899 + }, + { + "epoch": 1.1970129584889084, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.055957794189453, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7296761274337769, + "num_tokens": 282121276.0, + "step": 10900 + }, + { + "epoch": 1.197122776191522, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.873070001602173, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7326589822769165, + "num_tokens": 282138436.0, + "step": 10901 + }, + { + "epoch": 1.1972325938941357, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3473126888275146, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7278581261634827, + "num_tokens": 282161974.0, + "step": 10902 + }, + { + "epoch": 1.1973424115967495, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.280900239944458, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.6970634460449219, + "num_tokens": 282187070.0, + "step": 10903 + }, + { + "epoch": 1.197452229299363, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5660383701324463, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7168686389923096, + "num_tokens": 282210391.0, + "step": 10904 + }, + { + "epoch": 1.1975620470019768, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1774916648864746, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7313200235366821, + "num_tokens": 282236268.0, + "step": 10905 + }, + { + "epoch": 1.1976718647045903, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0004663467407227, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.6924711465835571, + "num_tokens": 282266725.0, + "step": 10906 + }, + { + "epoch": 1.197781682407204, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4809625148773193, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7508876323699951, + "num_tokens": 282286293.0, + "step": 10907 + }, + { + "epoch": 1.1978915001098178, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1426453590393066, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7045948505401611, + "num_tokens": 282317493.0, + "step": 10908 + }, + { + "epoch": 1.1980013178124314, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4278194904327393, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.6947246789932251, + "num_tokens": 282342822.0, + "step": 10909 + }, + { + "epoch": 1.1981111355150451, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.31001353263855, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7427864670753479, + "num_tokens": 282367191.0, + "step": 10910 + }, + { + "epoch": 1.1982209532176586, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0497708320617676, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.6972860097885132, + "num_tokens": 282397156.0, + "step": 10911 + }, + { + "epoch": 1.1983307709202724, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3524601459503174, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7292090654373169, + "num_tokens": 282422169.0, + "step": 10912 + }, + { + "epoch": 1.198440588622886, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 1.9912647008895874, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7211839556694031, + "num_tokens": 282451878.0, + "step": 10913 + }, + { + "epoch": 1.1985504063254997, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.239433526992798, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7065927386283875, + "num_tokens": 282479141.0, + "step": 10914 + }, + { + "epoch": 1.1986602240281132, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0664570331573486, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7081632018089294, + "num_tokens": 282508341.0, + "step": 10915 + }, + { + "epoch": 1.198770041730727, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2335193157196045, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7283610701560974, + "num_tokens": 282533930.0, + "step": 10916 + }, + { + "epoch": 1.1988798594333407, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3474233150482178, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7214187383651733, + "num_tokens": 282556891.0, + "step": 10917 + }, + { + "epoch": 1.1989896771359543, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.560344696044922, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7250379323959351, + "num_tokens": 282578796.0, + "step": 10918 + }, + { + "epoch": 1.199099494838568, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3288066387176514, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7108951807022095, + "num_tokens": 282603968.0, + "step": 10919 + }, + { + "epoch": 1.1992093125411816, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.687357187271118, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7238883972167969, + "num_tokens": 282623436.0, + "step": 10920 + }, + { + "epoch": 1.1993191302437953, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4070959091186523, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7261740565299988, + "num_tokens": 282644872.0, + "step": 10921 + }, + { + "epoch": 1.199428947946409, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0585551261901855, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.6859951615333557, + "num_tokens": 282675436.0, + "step": 10922 + }, + { + "epoch": 1.1995387656490226, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3979029655456543, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7413287162780762, + "num_tokens": 282698546.0, + "step": 10923 + }, + { + "epoch": 1.1996485833516364, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.309480667114258, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.711859941482544, + "num_tokens": 282722349.0, + "step": 10924 + }, + { + "epoch": 1.19975840105425, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4634954929351807, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7019490599632263, + "num_tokens": 282745554.0, + "step": 10925 + }, + { + "epoch": 1.1998682187568637, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3802993297576904, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7236769199371338, + "num_tokens": 282770136.0, + "step": 10926 + }, + { + "epoch": 1.1999780364594772, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4478840827941895, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7146048545837402, + "num_tokens": 282793880.0, + "step": 10927 + }, + { + "epoch": 1.200087854162091, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3311493396759033, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7210386991500854, + "num_tokens": 282819050.0, + "step": 10928 + }, + { + "epoch": 1.2001976718647045, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.601468801498413, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7299285531044006, + "num_tokens": 282838645.0, + "step": 10929 + }, + { + "epoch": 1.2003074895673183, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 1.9791699647903442, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7131210565567017, + "num_tokens": 282870167.0, + "step": 10930 + }, + { + "epoch": 1.200417307269932, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.118870258331299, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7187425494194031, + "num_tokens": 282900252.0, + "step": 10931 + }, + { + "epoch": 1.2005271249725455, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2282466888427734, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.710753321647644, + "num_tokens": 282924793.0, + "step": 10932 + }, + { + "epoch": 1.2006369426751593, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.149980068206787, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7113146185874939, + "num_tokens": 282951368.0, + "step": 10933 + }, + { + "epoch": 1.2007467603777728, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.013111114501953, + "learning_rate": 1e-06, + "loss": 1.0197, + "mean_token_accuracy": 0.689507007598877, + "num_tokens": 282982310.0, + "step": 10934 + }, + { + "epoch": 1.2008565780803866, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4467527866363525, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7100021839141846, + "num_tokens": 283005262.0, + "step": 10935 + }, + { + "epoch": 1.2009663957830001, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4096696376800537, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7352453470230103, + "num_tokens": 283029265.0, + "step": 10936 + }, + { + "epoch": 1.2010762134856139, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.212242603302002, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.737608015537262, + "num_tokens": 283055446.0, + "step": 10937 + }, + { + "epoch": 1.2011860311882274, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.267767906188965, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7497711181640625, + "num_tokens": 283080794.0, + "step": 10938 + }, + { + "epoch": 1.2012958488908412, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4040465354919434, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7171862125396729, + "num_tokens": 283103477.0, + "step": 10939 + }, + { + "epoch": 1.201405666593455, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.345771312713623, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7027865648269653, + "num_tokens": 283128921.0, + "step": 10940 + }, + { + "epoch": 1.2015154842960685, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 1.9730124473571777, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7135908603668213, + "num_tokens": 283159069.0, + "step": 10941 + }, + { + "epoch": 1.2016253019986822, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4639320373535156, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7198055386543274, + "num_tokens": 283182509.0, + "step": 10942 + }, + { + "epoch": 1.2017351197012958, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1418092250823975, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6927887201309204, + "num_tokens": 283211447.0, + "step": 10943 + }, + { + "epoch": 1.2018449374039095, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.120824098587036, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7221218943595886, + "num_tokens": 283239895.0, + "step": 10944 + }, + { + "epoch": 1.2019547551065233, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.8314802646636963, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.712607741355896, + "num_tokens": 283260397.0, + "step": 10945 + }, + { + "epoch": 1.2020645728091368, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.304572820663452, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7329928874969482, + "num_tokens": 283285092.0, + "step": 10946 + }, + { + "epoch": 1.2021743905117506, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.4107935428619385, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7323497533798218, + "num_tokens": 283308409.0, + "step": 10947 + }, + { + "epoch": 1.202284208214364, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.194281578063965, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.7055994272232056, + "num_tokens": 283337028.0, + "step": 10948 + }, + { + "epoch": 1.2023940259169779, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1869935989379883, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7231598496437073, + "num_tokens": 283362535.0, + "step": 10949 + }, + { + "epoch": 1.2025038436195914, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.097933292388916, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.686509370803833, + "num_tokens": 283392480.0, + "step": 10950 + }, + { + "epoch": 1.2026136613222052, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.442533254623413, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7220534682273865, + "num_tokens": 283414290.0, + "step": 10951 + }, + { + "epoch": 1.2027234790248187, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.535649538040161, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7396693229675293, + "num_tokens": 283434515.0, + "step": 10952 + }, + { + "epoch": 1.2028332967274324, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3460023403167725, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7228908538818359, + "num_tokens": 283459691.0, + "step": 10953 + }, + { + "epoch": 1.2029431144300462, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2580790519714355, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7259025573730469, + "num_tokens": 283483621.0, + "step": 10954 + }, + { + "epoch": 1.2030529321326597, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2698185443878174, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7321285009384155, + "num_tokens": 283507753.0, + "step": 10955 + }, + { + "epoch": 1.2031627498352735, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.298311471939087, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7173965573310852, + "num_tokens": 283532007.0, + "step": 10956 + }, + { + "epoch": 1.203272567537887, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.421020746231079, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.7011189460754395, + "num_tokens": 283555781.0, + "step": 10957 + }, + { + "epoch": 1.2033823852405008, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.427960157394409, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7357300519943237, + "num_tokens": 283578730.0, + "step": 10958 + }, + { + "epoch": 1.2034922029431145, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2909677028656006, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7461910247802734, + "num_tokens": 283602621.0, + "step": 10959 + }, + { + "epoch": 1.203602020645728, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 1.8772385120391846, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7176034450531006, + "num_tokens": 283638322.0, + "step": 10960 + }, + { + "epoch": 1.2037118383483418, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.0654046535491943, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.713660717010498, + "num_tokens": 283668495.0, + "step": 10961 + }, + { + "epoch": 1.2038216560509554, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5415091514587402, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7294471859931946, + "num_tokens": 283688969.0, + "step": 10962 + }, + { + "epoch": 1.2039314737535691, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2258362770080566, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7003766298294067, + "num_tokens": 283717810.0, + "step": 10963 + }, + { + "epoch": 1.2040412914561827, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.471559762954712, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7065253257751465, + "num_tokens": 283740254.0, + "step": 10964 + }, + { + "epoch": 1.2041511091587964, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.564666748046875, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7240651845932007, + "num_tokens": 283760259.0, + "step": 10965 + }, + { + "epoch": 1.20426092686141, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3208189010620117, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7153135538101196, + "num_tokens": 283784178.0, + "step": 10966 + }, + { + "epoch": 1.2043707445640237, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.400935649871826, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7441664934158325, + "num_tokens": 283806168.0, + "step": 10967 + }, + { + "epoch": 1.2044805622666375, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3646650314331055, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7292993068695068, + "num_tokens": 283830981.0, + "step": 10968 + }, + { + "epoch": 1.204590379969251, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2037174701690674, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7128937244415283, + "num_tokens": 283856809.0, + "step": 10969 + }, + { + "epoch": 1.2047001976718648, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.439889430999756, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7213819026947021, + "num_tokens": 283880097.0, + "step": 10970 + }, + { + "epoch": 1.2048100153744783, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.2797605991363525, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7330892086029053, + "num_tokens": 283904831.0, + "step": 10971 + }, + { + "epoch": 1.204919833077092, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.215848207473755, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.7097820043563843, + "num_tokens": 283931547.0, + "step": 10972 + }, + { + "epoch": 1.2050296507797058, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5961222648620605, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7245217561721802, + "num_tokens": 283950698.0, + "step": 10973 + }, + { + "epoch": 1.2051394684823193, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.554330825805664, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7261156439781189, + "num_tokens": 283974470.0, + "step": 10974 + }, + { + "epoch": 1.205249286184933, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.3719592094421387, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7065901160240173, + "num_tokens": 284001087.0, + "step": 10975 + }, + { + "epoch": 1.2053591038875466, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.6797573566436768, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7391828298568726, + "num_tokens": 284020322.0, + "step": 10976 + }, + { + "epoch": 1.2054689215901604, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.5810678005218506, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.6998966932296753, + "num_tokens": 284040858.0, + "step": 10977 + }, + { + "epoch": 1.205578739292774, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.198397159576416, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7020537853240967, + "num_tokens": 284067636.0, + "step": 10978 + }, + { + "epoch": 1.2056885569953877, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.1082358360290527, + "learning_rate": 1e-06, + "loss": 1.0262, + "mean_token_accuracy": 0.7038140296936035, + "num_tokens": 284096336.0, + "step": 10979 + }, + { + "epoch": 1.2057983746980012, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.382584810256958, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7100335359573364, + "num_tokens": 284120413.0, + "step": 10980 + }, + { + "epoch": 1.205908192400615, + "ewc_loss": 6.318092346191406e-06, + "grad_norm": 2.162381172180176, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7246192097663879, + "num_tokens": 284145303.0, + "step": 10981 + }, + { + "epoch": 1.2060180101032287, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.278285026550293, + "learning_rate": 1e-06, + "loss": 0.7487, + "mean_token_accuracy": 0.7601539492607117, + "num_tokens": 284167398.0, + "step": 10982 + }, + { + "epoch": 1.2061278278058423, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.363952159881592, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7229921817779541, + "num_tokens": 284191959.0, + "step": 10983 + }, + { + "epoch": 1.206237645508456, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.25990891456604, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7169529795646667, + "num_tokens": 284217135.0, + "step": 10984 + }, + { + "epoch": 1.2063474632110696, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 1.911128282546997, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6943004727363586, + "num_tokens": 284252636.0, + "step": 10985 + }, + { + "epoch": 1.2064572809136833, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2459614276885986, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.728527307510376, + "num_tokens": 284277462.0, + "step": 10986 + }, + { + "epoch": 1.2065670986162969, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.4155564308166504, + "learning_rate": 1e-06, + "loss": 0.7905, + "mean_token_accuracy": 0.7489297986030579, + "num_tokens": 284298181.0, + "step": 10987 + }, + { + "epoch": 1.2066769163189106, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1102776527404785, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7101707458496094, + "num_tokens": 284326972.0, + "step": 10988 + }, + { + "epoch": 1.2067867340215244, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.3011298179626465, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7286512851715088, + "num_tokens": 284352298.0, + "step": 10989 + }, + { + "epoch": 1.206896551724138, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.636451244354248, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7208847999572754, + "num_tokens": 284371893.0, + "step": 10990 + }, + { + "epoch": 1.2070063694267517, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2520902156829834, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7058369517326355, + "num_tokens": 284398281.0, + "step": 10991 + }, + { + "epoch": 1.2071161871293652, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.4103968143463135, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7186281085014343, + "num_tokens": 284421089.0, + "step": 10992 + }, + { + "epoch": 1.207226004831979, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.64025616645813, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7329446077346802, + "num_tokens": 284441192.0, + "step": 10993 + }, + { + "epoch": 1.2073358225345925, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.117105484008789, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.697604238986969, + "num_tokens": 284471296.0, + "step": 10994 + }, + { + "epoch": 1.2074456402372062, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1593523025512695, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7095258831977844, + "num_tokens": 284497238.0, + "step": 10995 + }, + { + "epoch": 1.20755545793982, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1092398166656494, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7017714381217957, + "num_tokens": 284525660.0, + "step": 10996 + }, + { + "epoch": 1.2076652756424335, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.12807035446167, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7250466346740723, + "num_tokens": 284554370.0, + "step": 10997 + }, + { + "epoch": 1.2077750933450473, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.379956007003784, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7278835773468018, + "num_tokens": 284577055.0, + "step": 10998 + }, + { + "epoch": 1.2078849110476608, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.3060195446014404, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7321512699127197, + "num_tokens": 284601534.0, + "step": 10999 + }, + { + "epoch": 1.2079947287502746, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.4844753742218018, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7200765609741211, + "num_tokens": 284624533.0, + "step": 11000 + }, + { + "epoch": 1.2081045464528881, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.0028467178344727, + "learning_rate": 1e-06, + "loss": 1.0355, + "mean_token_accuracy": 0.6802358627319336, + "num_tokens": 284655548.0, + "step": 11001 + }, + { + "epoch": 1.2082143641555019, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2959420680999756, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.716296911239624, + "num_tokens": 284678884.0, + "step": 11002 + }, + { + "epoch": 1.2083241818581154, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.3717050552368164, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7166216373443604, + "num_tokens": 284703846.0, + "step": 11003 + }, + { + "epoch": 1.2084339995607292, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1243457794189453, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7234306335449219, + "num_tokens": 284729345.0, + "step": 11004 + }, + { + "epoch": 1.208543817263343, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.038468837738037, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7201207876205444, + "num_tokens": 284759880.0, + "step": 11005 + }, + { + "epoch": 1.2086536349659565, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.6888508796691895, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.738791823387146, + "num_tokens": 284778553.0, + "step": 11006 + }, + { + "epoch": 1.2087634526685702, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1681008338928223, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7358644008636475, + "num_tokens": 284803349.0, + "step": 11007 + }, + { + "epoch": 1.2088732703711838, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.0989556312561035, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6932665109634399, + "num_tokens": 284836162.0, + "step": 11008 + }, + { + "epoch": 1.2089830880737975, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.538252115249634, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7284060120582581, + "num_tokens": 284855788.0, + "step": 11009 + }, + { + "epoch": 1.2090929057764113, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.359100818634033, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7189968824386597, + "num_tokens": 284880287.0, + "step": 11010 + }, + { + "epoch": 1.2092027234790248, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2466557025909424, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7284850478172302, + "num_tokens": 284904543.0, + "step": 11011 + }, + { + "epoch": 1.2093125411816386, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.123260021209717, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7117133736610413, + "num_tokens": 284933876.0, + "step": 11012 + }, + { + "epoch": 1.209422358884252, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2270398139953613, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7102354764938354, + "num_tokens": 284961257.0, + "step": 11013 + }, + { + "epoch": 1.2095321765868658, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1919994354248047, + "learning_rate": 1e-06, + "loss": 0.9955, + "mean_token_accuracy": 0.6953375935554504, + "num_tokens": 284988656.0, + "step": 11014 + }, + { + "epoch": 1.2096419942894794, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.540493965148926, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7314267754554749, + "num_tokens": 285009556.0, + "step": 11015 + }, + { + "epoch": 1.2097518119920931, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2492990493774414, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7397693395614624, + "num_tokens": 285034371.0, + "step": 11016 + }, + { + "epoch": 1.2098616296947067, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.3654141426086426, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7274656295776367, + "num_tokens": 285056245.0, + "step": 11017 + }, + { + "epoch": 1.2099714473973204, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2430355548858643, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7201846837997437, + "num_tokens": 285081384.0, + "step": 11018 + }, + { + "epoch": 1.2100812650999342, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.5654966831207275, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7272741794586182, + "num_tokens": 285101546.0, + "step": 11019 + }, + { + "epoch": 1.2101910828025477, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1068387031555176, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.735527515411377, + "num_tokens": 285128909.0, + "step": 11020 + }, + { + "epoch": 1.2103009005051615, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.6361749172210693, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7144337892532349, + "num_tokens": 285149217.0, + "step": 11021 + }, + { + "epoch": 1.210410718207775, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.25343656539917, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7421547174453735, + "num_tokens": 285173057.0, + "step": 11022 + }, + { + "epoch": 1.2105205359103888, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.2773399353027344, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7311126589775085, + "num_tokens": 285198193.0, + "step": 11023 + }, + { + "epoch": 1.2106303536130025, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.1931631565093994, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7100380659103394, + "num_tokens": 285226208.0, + "step": 11024 + }, + { + "epoch": 1.210740171315616, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.358203172683716, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7162684202194214, + "num_tokens": 285249441.0, + "step": 11025 + }, + { + "epoch": 1.2108499890182298, + "ewc_loss": 6.3478946685791016e-06, + "grad_norm": 2.198193311691284, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7447409629821777, + "num_tokens": 285274785.0, + "step": 11026 + }, + { + "epoch": 1.2109598067208434, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 1.9291613101959229, + "learning_rate": 1e-06, + "loss": 1.0229, + "mean_token_accuracy": 0.6919176578521729, + "num_tokens": 285311003.0, + "step": 11027 + }, + { + "epoch": 1.2110696244234571, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.3565664291381836, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7244497537612915, + "num_tokens": 285334538.0, + "step": 11028 + }, + { + "epoch": 1.2111794421260706, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.1907739639282227, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7186912298202515, + "num_tokens": 285361596.0, + "step": 11029 + }, + { + "epoch": 1.2112892598286844, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.1764979362487793, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7066952586174011, + "num_tokens": 285389838.0, + "step": 11030 + }, + { + "epoch": 1.211399077531298, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.147716760635376, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.699698805809021, + "num_tokens": 285417739.0, + "step": 11031 + }, + { + "epoch": 1.2115088952339117, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.8740346431732178, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7313810586929321, + "num_tokens": 285435299.0, + "step": 11032 + }, + { + "epoch": 1.2116187129365255, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.3591387271881104, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7149311304092407, + "num_tokens": 285460233.0, + "step": 11033 + }, + { + "epoch": 1.211728530639139, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.439561128616333, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7137969732284546, + "num_tokens": 285485153.0, + "step": 11034 + }, + { + "epoch": 1.2118383483417527, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.2391650676727295, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7313176393508911, + "num_tokens": 285510008.0, + "step": 11035 + }, + { + "epoch": 1.2119481660443663, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.214160680770874, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7155717611312866, + "num_tokens": 285535003.0, + "step": 11036 + }, + { + "epoch": 1.21205798374698, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.0492093563079834, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7114964723587036, + "num_tokens": 285565024.0, + "step": 11037 + }, + { + "epoch": 1.2121678014495938, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 1.9089645147323608, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7239866256713867, + "num_tokens": 285598590.0, + "step": 11038 + }, + { + "epoch": 1.2122776191522073, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.467525005340576, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7243716716766357, + "num_tokens": 285619570.0, + "step": 11039 + }, + { + "epoch": 1.212387436854821, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.2805206775665283, + "learning_rate": 1e-06, + "loss": 0.9858, + "mean_token_accuracy": 0.6980589628219604, + "num_tokens": 285644388.0, + "step": 11040 + }, + { + "epoch": 1.2124972545574346, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.06630539894104, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6955550312995911, + "num_tokens": 285676890.0, + "step": 11041 + }, + { + "epoch": 1.2126070722600484, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.6534101963043213, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7389775514602661, + "num_tokens": 285694848.0, + "step": 11042 + }, + { + "epoch": 1.212716889962662, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.617711067199707, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.727715015411377, + "num_tokens": 285713171.0, + "step": 11043 + }, + { + "epoch": 1.2128267076652757, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.2540180683135986, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.7512717843055725, + "num_tokens": 285736481.0, + "step": 11044 + }, + { + "epoch": 1.2129365253678892, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.0721518993377686, + "learning_rate": 1e-06, + "loss": 0.9963, + "mean_token_accuracy": 0.6913307905197144, + "num_tokens": 285766170.0, + "step": 11045 + }, + { + "epoch": 1.213046343070503, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.9352400302886963, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7406010627746582, + "num_tokens": 285782962.0, + "step": 11046 + }, + { + "epoch": 1.2131561607731167, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.390989065170288, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7385668754577637, + "num_tokens": 285805223.0, + "step": 11047 + }, + { + "epoch": 1.2132659784757303, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.283987045288086, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7235984206199646, + "num_tokens": 285830478.0, + "step": 11048 + }, + { + "epoch": 1.213375796178344, + "ewc_loss": 6.407499313354492e-06, + "grad_norm": 2.7571771144866943, + "learning_rate": 1e-06, + "loss": 0.7622, + "mean_token_accuracy": 0.7570148706436157, + "num_tokens": 285848620.0, + "step": 11049 + }, + { + "epoch": 1.2134856138809575, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1865217685699463, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7099552750587463, + "num_tokens": 285876089.0, + "step": 11050 + }, + { + "epoch": 1.2135954315835713, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2860147953033447, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7423219680786133, + "num_tokens": 285900376.0, + "step": 11051 + }, + { + "epoch": 1.2137052492861848, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.264035224914551, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7347325086593628, + "num_tokens": 285923703.0, + "step": 11052 + }, + { + "epoch": 1.2138150669887986, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.304438352584839, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7017649412155151, + "num_tokens": 285950717.0, + "step": 11053 + }, + { + "epoch": 1.2139248846914121, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.390065908432007, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7037336826324463, + "num_tokens": 285973567.0, + "step": 11054 + }, + { + "epoch": 1.2140347023940259, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4210317134857178, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7095611095428467, + "num_tokens": 285997017.0, + "step": 11055 + }, + { + "epoch": 1.2141445200966396, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.135108232498169, + "learning_rate": 1e-06, + "loss": 1.0478, + "mean_token_accuracy": 0.6821070909500122, + "num_tokens": 286028655.0, + "step": 11056 + }, + { + "epoch": 1.2142543377992532, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2583673000335693, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.713135302066803, + "num_tokens": 286054292.0, + "step": 11057 + }, + { + "epoch": 1.214364155501867, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.256080389022827, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7293789386749268, + "num_tokens": 286079944.0, + "step": 11058 + }, + { + "epoch": 1.2144739732044805, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.204437017440796, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.6991573572158813, + "num_tokens": 286108619.0, + "step": 11059 + }, + { + "epoch": 1.2145837909070942, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0980215072631836, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7318820357322693, + "num_tokens": 286137641.0, + "step": 11060 + }, + { + "epoch": 1.214693608609708, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1705873012542725, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7214372754096985, + "num_tokens": 286163781.0, + "step": 11061 + }, + { + "epoch": 1.2148034263123215, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0896618366241455, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7082334160804749, + "num_tokens": 286193706.0, + "step": 11062 + }, + { + "epoch": 1.2149132440149353, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.071091413497925, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7058730125427246, + "num_tokens": 286224522.0, + "step": 11063 + }, + { + "epoch": 1.2150230617175488, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1716670989990234, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7336008548736572, + "num_tokens": 286249405.0, + "step": 11064 + }, + { + "epoch": 1.2151328794201626, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.114814519882202, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7219691276550293, + "num_tokens": 286278531.0, + "step": 11065 + }, + { + "epoch": 1.215242697122776, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.9792909622192383, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7131445407867432, + "num_tokens": 286308331.0, + "step": 11066 + }, + { + "epoch": 1.2153525148253899, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0049524307250977, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7042115330696106, + "num_tokens": 286339208.0, + "step": 11067 + }, + { + "epoch": 1.2154623325280034, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.331373691558838, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7338579893112183, + "num_tokens": 286361971.0, + "step": 11068 + }, + { + "epoch": 1.2155721502306172, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.980249285697937, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7190483212471008, + "num_tokens": 286392775.0, + "step": 11069 + }, + { + "epoch": 1.215681967933231, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.6367509365081787, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7299838066101074, + "num_tokens": 286411263.0, + "step": 11070 + }, + { + "epoch": 1.2157917856358444, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2028489112854004, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7129286527633667, + "num_tokens": 286437434.0, + "step": 11071 + }, + { + "epoch": 1.2159016033384582, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.275566577911377, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7234295010566711, + "num_tokens": 286461643.0, + "step": 11072 + }, + { + "epoch": 1.2160114210410717, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0318515300750732, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.706896185874939, + "num_tokens": 286493286.0, + "step": 11073 + }, + { + "epoch": 1.2161212387436855, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.5241470336914062, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7183526754379272, + "num_tokens": 286514586.0, + "step": 11074 + }, + { + "epoch": 1.2162310564462993, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.663999319076538, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7379844188690186, + "num_tokens": 286533142.0, + "step": 11075 + }, + { + "epoch": 1.2163408741489128, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2051138877868652, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7163140177726746, + "num_tokens": 286560108.0, + "step": 11076 + }, + { + "epoch": 1.2164506918515265, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.9667471647262573, + "learning_rate": 1e-06, + "loss": 1.0353, + "mean_token_accuracy": 0.6898331642150879, + "num_tokens": 286592101.0, + "step": 11077 + }, + { + "epoch": 1.21656050955414, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.314171314239502, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7138229012489319, + "num_tokens": 286614678.0, + "step": 11078 + }, + { + "epoch": 1.2166703272567538, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.5559263229370117, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7115448713302612, + "num_tokens": 286636211.0, + "step": 11079 + }, + { + "epoch": 1.2167801449593674, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2006070613861084, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7221616506576538, + "num_tokens": 286662911.0, + "step": 11080 + }, + { + "epoch": 1.2168899626619811, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.159728527069092, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7200201749801636, + "num_tokens": 286690482.0, + "step": 11081 + }, + { + "epoch": 1.2169997803645947, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2674121856689453, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.704175591468811, + "num_tokens": 286716439.0, + "step": 11082 + }, + { + "epoch": 1.2171095980672084, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.035508632659912, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7269822955131531, + "num_tokens": 286745981.0, + "step": 11083 + }, + { + "epoch": 1.2172194157698222, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.112560987472534, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7164284586906433, + "num_tokens": 286774086.0, + "step": 11084 + }, + { + "epoch": 1.2173292334724357, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.3363373279571533, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7221912145614624, + "num_tokens": 286797551.0, + "step": 11085 + }, + { + "epoch": 1.2174390511750495, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4431936740875244, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7172160744667053, + "num_tokens": 286821481.0, + "step": 11086 + }, + { + "epoch": 1.217548868877663, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0561442375183105, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.702281653881073, + "num_tokens": 286851130.0, + "step": 11087 + }, + { + "epoch": 1.2176586865802768, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.6072700023651123, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7307137250900269, + "num_tokens": 286870940.0, + "step": 11088 + }, + { + "epoch": 1.2177685042828905, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.386993169784546, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7304813861846924, + "num_tokens": 286893884.0, + "step": 11089 + }, + { + "epoch": 1.217878321985504, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4262149333953857, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7382034659385681, + "num_tokens": 286916003.0, + "step": 11090 + }, + { + "epoch": 1.2179881396881178, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4950878620147705, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7093844413757324, + "num_tokens": 286938420.0, + "step": 11091 + }, + { + "epoch": 1.2180979573907313, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.7280447483062744, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7360742092132568, + "num_tokens": 286958223.0, + "step": 11092 + }, + { + "epoch": 1.218207775093345, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2578816413879395, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7207764387130737, + "num_tokens": 286982199.0, + "step": 11093 + }, + { + "epoch": 1.2183175927959586, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.205401659011841, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7165262699127197, + "num_tokens": 287010430.0, + "step": 11094 + }, + { + "epoch": 1.2184274104985724, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.534154176712036, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7260595560073853, + "num_tokens": 287030970.0, + "step": 11095 + }, + { + "epoch": 1.218537228201186, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1171188354492188, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.706413745880127, + "num_tokens": 287057132.0, + "step": 11096 + }, + { + "epoch": 1.2186470459037997, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.29085636138916, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.732253909111023, + "num_tokens": 287082012.0, + "step": 11097 + }, + { + "epoch": 1.2187568636064134, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2487473487854004, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7014771103858948, + "num_tokens": 287109465.0, + "step": 11098 + }, + { + "epoch": 1.218866681309027, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1883132457733154, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7487369179725647, + "num_tokens": 287135634.0, + "step": 11099 + }, + { + "epoch": 1.2189764990116407, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.835784673690796, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7379801869392395, + "num_tokens": 287151167.0, + "step": 11100 + }, + { + "epoch": 1.2190863167142543, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.306581497192383, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7388643026351929, + "num_tokens": 287176289.0, + "step": 11101 + }, + { + "epoch": 1.219196134416868, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4075236320495605, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.712257981300354, + "num_tokens": 287199749.0, + "step": 11102 + }, + { + "epoch": 1.2193059521194818, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1083755493164062, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7055194973945618, + "num_tokens": 287229177.0, + "step": 11103 + }, + { + "epoch": 1.2194157698220953, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.088888168334961, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7090985774993896, + "num_tokens": 287257362.0, + "step": 11104 + }, + { + "epoch": 1.219525587524709, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0672378540039062, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7246750593185425, + "num_tokens": 287286351.0, + "step": 11105 + }, + { + "epoch": 1.2196354052273226, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.3542520999908447, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7227751016616821, + "num_tokens": 287312114.0, + "step": 11106 + }, + { + "epoch": 1.2197452229299364, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1641337871551514, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7259994745254517, + "num_tokens": 287340424.0, + "step": 11107 + }, + { + "epoch": 1.21985504063255, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.305633306503296, + "learning_rate": 1e-06, + "loss": 0.9781, + "mean_token_accuracy": 0.7060383558273315, + "num_tokens": 287365830.0, + "step": 11108 + }, + { + "epoch": 1.2199648583351637, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1654863357543945, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.732548177242279, + "num_tokens": 287392435.0, + "step": 11109 + }, + { + "epoch": 1.2200746760377772, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.6570754051208496, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7305753231048584, + "num_tokens": 287412571.0, + "step": 11110 + }, + { + "epoch": 1.220184493740391, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.062427520751953, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7162144184112549, + "num_tokens": 287441777.0, + "step": 11111 + }, + { + "epoch": 1.2202943114430047, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.607823371887207, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7155172228813171, + "num_tokens": 287462481.0, + "step": 11112 + }, + { + "epoch": 1.2204041291456182, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1990718841552734, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7152171730995178, + "num_tokens": 287491128.0, + "step": 11113 + }, + { + "epoch": 1.220513946848232, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.8681021928787231, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7250795960426331, + "num_tokens": 287524819.0, + "step": 11114 + }, + { + "epoch": 1.2206237645508455, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.8667378425598145, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7165430188179016, + "num_tokens": 287562937.0, + "step": 11115 + }, + { + "epoch": 1.2207335822534593, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.386906623840332, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7401188015937805, + "num_tokens": 287584404.0, + "step": 11116 + }, + { + "epoch": 1.2208433999560728, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2320332527160645, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7449840903282166, + "num_tokens": 287609866.0, + "step": 11117 + }, + { + "epoch": 1.2209532176586866, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0624473094940186, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.7083951234817505, + "num_tokens": 287639847.0, + "step": 11118 + }, + { + "epoch": 1.2210630353613001, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.9612846374511719, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7175829410552979, + "num_tokens": 287670981.0, + "step": 11119 + }, + { + "epoch": 1.2211728530639139, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.073903799057007, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.709325909614563, + "num_tokens": 287701178.0, + "step": 11120 + }, + { + "epoch": 1.2212826707665276, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.244619607925415, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7026945948600769, + "num_tokens": 287728371.0, + "step": 11121 + }, + { + "epoch": 1.2213924884691412, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.3211512565612793, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7203225493431091, + "num_tokens": 287752359.0, + "step": 11122 + }, + { + "epoch": 1.221502306171755, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4327564239501953, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.731846809387207, + "num_tokens": 287774074.0, + "step": 11123 + }, + { + "epoch": 1.2216121238743685, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.0264041423797607, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6988657116889954, + "num_tokens": 287805557.0, + "step": 11124 + }, + { + "epoch": 1.2217219415769822, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2589354515075684, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7044111490249634, + "num_tokens": 287831463.0, + "step": 11125 + }, + { + "epoch": 1.221831759279596, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.441664457321167, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7422162294387817, + "num_tokens": 287852361.0, + "step": 11126 + }, + { + "epoch": 1.2219415769822095, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.568436861038208, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7258333563804626, + "num_tokens": 287871387.0, + "step": 11127 + }, + { + "epoch": 1.2220513946848233, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.3745439052581787, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7185013294219971, + "num_tokens": 287897123.0, + "step": 11128 + }, + { + "epoch": 1.2221612123874368, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.542128324508667, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7179537415504456, + "num_tokens": 287920282.0, + "step": 11129 + }, + { + "epoch": 1.2222710300900506, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.3176321983337402, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7147709131240845, + "num_tokens": 287945621.0, + "step": 11130 + }, + { + "epoch": 1.222380847792664, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.183573007583618, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7325426936149597, + "num_tokens": 287972594.0, + "step": 11131 + }, + { + "epoch": 1.2224906654952779, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 1.9363354444503784, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7045494914054871, + "num_tokens": 288005889.0, + "step": 11132 + }, + { + "epoch": 1.2226004831978914, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.365941286087036, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7328628301620483, + "num_tokens": 288026620.0, + "step": 11133 + }, + { + "epoch": 1.2227103009005051, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1913113594055176, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7146131992340088, + "num_tokens": 288055937.0, + "step": 11134 + }, + { + "epoch": 1.222820118603119, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2432327270507812, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7213536500930786, + "num_tokens": 288082276.0, + "step": 11135 + }, + { + "epoch": 1.2229299363057324, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.46567440032959, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7152097225189209, + "num_tokens": 288105292.0, + "step": 11136 + }, + { + "epoch": 1.2230397540083462, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1712148189544678, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7250482439994812, + "num_tokens": 288131219.0, + "step": 11137 + }, + { + "epoch": 1.2231495717109597, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.014667510986328, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7120242714881897, + "num_tokens": 288164176.0, + "step": 11138 + }, + { + "epoch": 1.2232593894135735, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.4005444049835205, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7182651162147522, + "num_tokens": 288188719.0, + "step": 11139 + }, + { + "epoch": 1.2233692071161872, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.194406747817993, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7086520195007324, + "num_tokens": 288216109.0, + "step": 11140 + }, + { + "epoch": 1.2234790248188008, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.407777786254883, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7154132127761841, + "num_tokens": 288240464.0, + "step": 11141 + }, + { + "epoch": 1.2235888425214145, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.1330480575561523, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7275576591491699, + "num_tokens": 288266942.0, + "step": 11142 + }, + { + "epoch": 1.223698660224028, + "ewc_loss": 6.4373016357421875e-06, + "grad_norm": 2.2531256675720215, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7035191059112549, + "num_tokens": 288295463.0, + "step": 11143 + }, + { + "epoch": 1.2238084779266418, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.293252468109131, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7291121482849121, + "num_tokens": 288319423.0, + "step": 11144 + }, + { + "epoch": 1.2239182956292554, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.7427964210510254, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7362225651741028, + "num_tokens": 288338142.0, + "step": 11145 + }, + { + "epoch": 1.2240281133318691, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.0230846405029297, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.73134446144104, + "num_tokens": 288365441.0, + "step": 11146 + }, + { + "epoch": 1.2241379310344827, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.3787543773651123, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.701022744178772, + "num_tokens": 288391090.0, + "step": 11147 + }, + { + "epoch": 1.2242477487370964, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 1.980729341506958, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6942225694656372, + "num_tokens": 288423660.0, + "step": 11148 + }, + { + "epoch": 1.2243575664397102, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.1462743282318115, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7217614054679871, + "num_tokens": 288452342.0, + "step": 11149 + }, + { + "epoch": 1.2244673841423237, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.143136978149414, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7095988988876343, + "num_tokens": 288480932.0, + "step": 11150 + }, + { + "epoch": 1.2245772018449375, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.7653043270111084, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7485582828521729, + "num_tokens": 288498077.0, + "step": 11151 + }, + { + "epoch": 1.224687019547551, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.2957699298858643, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7073509693145752, + "num_tokens": 288524421.0, + "step": 11152 + }, + { + "epoch": 1.2247968372501647, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.1881394386291504, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7233539819717407, + "num_tokens": 288551768.0, + "step": 11153 + }, + { + "epoch": 1.2249066549527785, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.3635075092315674, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7268332242965698, + "num_tokens": 288575893.0, + "step": 11154 + }, + { + "epoch": 1.225016472655392, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.349583625793457, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7259862422943115, + "num_tokens": 288598787.0, + "step": 11155 + }, + { + "epoch": 1.2251262903580058, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.4941565990448, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7168096899986267, + "num_tokens": 288620878.0, + "step": 11156 + }, + { + "epoch": 1.2252361080606193, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.140782117843628, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7225186824798584, + "num_tokens": 288648869.0, + "step": 11157 + }, + { + "epoch": 1.225345925763233, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.508218765258789, + "learning_rate": 1e-06, + "loss": 0.7921, + "mean_token_accuracy": 0.7477625608444214, + "num_tokens": 288668783.0, + "step": 11158 + }, + { + "epoch": 1.2254557434658466, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.2119786739349365, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7204205393791199, + "num_tokens": 288696825.0, + "step": 11159 + }, + { + "epoch": 1.2255655611684604, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 1.9812564849853516, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7253602743148804, + "num_tokens": 288729299.0, + "step": 11160 + }, + { + "epoch": 1.225675378871074, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.3536221981048584, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7245447635650635, + "num_tokens": 288754789.0, + "step": 11161 + }, + { + "epoch": 1.2257851965736877, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.3510003089904785, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7016140818595886, + "num_tokens": 288780435.0, + "step": 11162 + }, + { + "epoch": 1.2258950142763014, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.2972464561462402, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7123533487319946, + "num_tokens": 288803995.0, + "step": 11163 + }, + { + "epoch": 1.226004831978915, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.091057777404785, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7036596536636353, + "num_tokens": 288832003.0, + "step": 11164 + }, + { + "epoch": 1.2261146496815287, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.0976905822753906, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7071787714958191, + "num_tokens": 288862135.0, + "step": 11165 + }, + { + "epoch": 1.2262244673841423, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.0777947902679443, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.703637957572937, + "num_tokens": 288890024.0, + "step": 11166 + }, + { + "epoch": 1.226334285086756, + "ewc_loss": 6.467103958129883e-06, + "grad_norm": 2.3193466663360596, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7186863422393799, + "num_tokens": 288913336.0, + "step": 11167 + }, + { + "epoch": 1.2264441027893696, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0315091609954834, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.718449592590332, + "num_tokens": 288944772.0, + "step": 11168 + }, + { + "epoch": 1.2265539204919833, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.087357759475708, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7195778489112854, + "num_tokens": 288972703.0, + "step": 11169 + }, + { + "epoch": 1.226663738194597, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.195089817047119, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7419780492782593, + "num_tokens": 288997926.0, + "step": 11170 + }, + { + "epoch": 1.2267735558972106, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.398843288421631, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7040573358535767, + "num_tokens": 289020373.0, + "step": 11171 + }, + { + "epoch": 1.2268833735998244, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.204296588897705, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6971784234046936, + "num_tokens": 289048954.0, + "step": 11172 + }, + { + "epoch": 1.226993191302438, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3372690677642822, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7199952006340027, + "num_tokens": 289073845.0, + "step": 11173 + }, + { + "epoch": 1.2271030090050516, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.5083155632019043, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7284521460533142, + "num_tokens": 289094320.0, + "step": 11174 + }, + { + "epoch": 1.2272128267076652, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3245692253112793, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7412145137786865, + "num_tokens": 289123281.0, + "step": 11175 + }, + { + "epoch": 1.227322644410279, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.5031771659851074, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7074634432792664, + "num_tokens": 289146395.0, + "step": 11176 + }, + { + "epoch": 1.2274324621128927, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.536339521408081, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7227547764778137, + "num_tokens": 289166789.0, + "step": 11177 + }, + { + "epoch": 1.2275422798155062, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3163132667541504, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7028067111968994, + "num_tokens": 289190811.0, + "step": 11178 + }, + { + "epoch": 1.22765209751812, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2946224212646484, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7220137715339661, + "num_tokens": 289216576.0, + "step": 11179 + }, + { + "epoch": 1.2277619152207335, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.28471302986145, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7213317155838013, + "num_tokens": 289241926.0, + "step": 11180 + }, + { + "epoch": 1.2278717329233473, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4946889877319336, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7452322244644165, + "num_tokens": 289261528.0, + "step": 11181 + }, + { + "epoch": 1.2279815506259608, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.503546714782715, + "learning_rate": 1e-06, + "loss": 1.0329, + "mean_token_accuracy": 0.6933548450469971, + "num_tokens": 289284892.0, + "step": 11182 + }, + { + "epoch": 1.2280913683285746, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2197208404541016, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7083843350410461, + "num_tokens": 289311750.0, + "step": 11183 + }, + { + "epoch": 1.228201186031188, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.293938398361206, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7179730534553528, + "num_tokens": 289336750.0, + "step": 11184 + }, + { + "epoch": 1.2283110037338019, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.8871302604675293, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7069293260574341, + "num_tokens": 289372893.0, + "step": 11185 + }, + { + "epoch": 1.2284208214364156, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1003224849700928, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7071541547775269, + "num_tokens": 289400788.0, + "step": 11186 + }, + { + "epoch": 1.2285306391390292, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2290050983428955, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7163758277893066, + "num_tokens": 289428204.0, + "step": 11187 + }, + { + "epoch": 1.228640456841643, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.5592966079711914, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.739029049873352, + "num_tokens": 289448773.0, + "step": 11188 + }, + { + "epoch": 1.2287502745442564, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3691768646240234, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7269431352615356, + "num_tokens": 289472201.0, + "step": 11189 + }, + { + "epoch": 1.2288600922468702, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.303237199783325, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7420071363449097, + "num_tokens": 289493334.0, + "step": 11190 + }, + { + "epoch": 1.228969909949484, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2089831829071045, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.6995445489883423, + "num_tokens": 289522313.0, + "step": 11191 + }, + { + "epoch": 1.2290797276520975, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.9289569854736328, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.6882482767105103, + "num_tokens": 289555995.0, + "step": 11192 + }, + { + "epoch": 1.2291895453547113, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0572757720947266, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7069061994552612, + "num_tokens": 289588907.0, + "step": 11193 + }, + { + "epoch": 1.2292993630573248, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3154571056365967, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7158661484718323, + "num_tokens": 289613880.0, + "step": 11194 + }, + { + "epoch": 1.2294091807599385, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.474639415740967, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7243432998657227, + "num_tokens": 289635850.0, + "step": 11195 + }, + { + "epoch": 1.229518998462552, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.14858078956604, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7197312712669373, + "num_tokens": 289661786.0, + "step": 11196 + }, + { + "epoch": 1.2296288161651658, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.196542263031006, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7287399768829346, + "num_tokens": 289690312.0, + "step": 11197 + }, + { + "epoch": 1.2297386338677794, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3235175609588623, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7071472406387329, + "num_tokens": 289716247.0, + "step": 11198 + }, + { + "epoch": 1.2298484515703931, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.491300582885742, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.7065894603729248, + "num_tokens": 289737054.0, + "step": 11199 + }, + { + "epoch": 1.2299582692730069, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1824870109558105, + "learning_rate": 1e-06, + "loss": 1.0235, + "mean_token_accuracy": 0.691719651222229, + "num_tokens": 289765541.0, + "step": 11200 + }, + { + "epoch": 1.2300680869756204, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2098608016967773, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7202653288841248, + "num_tokens": 289793721.0, + "step": 11201 + }, + { + "epoch": 1.2301779046782342, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1498606204986572, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7263542413711548, + "num_tokens": 289818886.0, + "step": 11202 + }, + { + "epoch": 1.2302877223808477, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.536527633666992, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.725225031375885, + "num_tokens": 289839342.0, + "step": 11203 + }, + { + "epoch": 1.2303975400834615, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.503041982650757, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7093727588653564, + "num_tokens": 289859828.0, + "step": 11204 + }, + { + "epoch": 1.2305073577860752, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.6109349727630615, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7400912642478943, + "num_tokens": 289878027.0, + "step": 11205 + }, + { + "epoch": 1.2306171754886888, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.342827796936035, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.7013230919837952, + "num_tokens": 289902423.0, + "step": 11206 + }, + { + "epoch": 1.2307269931913025, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.353534698486328, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7075903415679932, + "num_tokens": 289926283.0, + "step": 11207 + }, + { + "epoch": 1.230836810893916, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4851396083831787, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7254906296730042, + "num_tokens": 289947044.0, + "step": 11208 + }, + { + "epoch": 1.2309466285965298, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.166511297225952, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7062737941741943, + "num_tokens": 289973967.0, + "step": 11209 + }, + { + "epoch": 1.2310564462991433, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4243526458740234, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7175115942955017, + "num_tokens": 289997008.0, + "step": 11210 + }, + { + "epoch": 1.231166264001757, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.201373815536499, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.6984653472900391, + "num_tokens": 290025815.0, + "step": 11211 + }, + { + "epoch": 1.2312760817043706, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3636467456817627, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7099500298500061, + "num_tokens": 290049001.0, + "step": 11212 + }, + { + "epoch": 1.2313858994069844, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.451277494430542, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7183316946029663, + "num_tokens": 290072534.0, + "step": 11213 + }, + { + "epoch": 1.2314957171095982, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.417163133621216, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.6999108791351318, + "num_tokens": 290098182.0, + "step": 11214 + }, + { + "epoch": 1.2316055348122117, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4156782627105713, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7277685403823853, + "num_tokens": 290121546.0, + "step": 11215 + }, + { + "epoch": 1.2317153525148254, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.340823173522949, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7395098209381104, + "num_tokens": 290144929.0, + "step": 11216 + }, + { + "epoch": 1.231825170217439, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.335012435913086, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.721362292766571, + "num_tokens": 290169041.0, + "step": 11217 + }, + { + "epoch": 1.2319349879200527, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1152725219726562, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7300498485565186, + "num_tokens": 290196061.0, + "step": 11218 + }, + { + "epoch": 1.2320448056226665, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2349836826324463, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.710185170173645, + "num_tokens": 290222130.0, + "step": 11219 + }, + { + "epoch": 1.23215462332528, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4721317291259766, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7297707796096802, + "num_tokens": 290245566.0, + "step": 11220 + }, + { + "epoch": 1.2322644410278938, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.587904930114746, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7098013758659363, + "num_tokens": 290268633.0, + "step": 11221 + }, + { + "epoch": 1.2323742587305073, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0532543659210205, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7386044263839722, + "num_tokens": 290298052.0, + "step": 11222 + }, + { + "epoch": 1.232484076433121, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.335383176803589, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7337583303451538, + "num_tokens": 290323306.0, + "step": 11223 + }, + { + "epoch": 1.2325938941357346, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.161750078201294, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7219825387001038, + "num_tokens": 290351222.0, + "step": 11224 + }, + { + "epoch": 1.2327037118383484, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1229310035705566, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.691391110420227, + "num_tokens": 290380297.0, + "step": 11225 + }, + { + "epoch": 1.232813529540962, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0815603733062744, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7322435975074768, + "num_tokens": 290410462.0, + "step": 11226 + }, + { + "epoch": 1.2329233472435757, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2659599781036377, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7156811952590942, + "num_tokens": 290438048.0, + "step": 11227 + }, + { + "epoch": 1.2330331649461894, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.9742518663406372, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7054332494735718, + "num_tokens": 290469821.0, + "step": 11228 + }, + { + "epoch": 1.233142982648803, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.162991762161255, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7204407453536987, + "num_tokens": 290495476.0, + "step": 11229 + }, + { + "epoch": 1.2332528003514167, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3389101028442383, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7018007636070251, + "num_tokens": 290519504.0, + "step": 11230 + }, + { + "epoch": 1.2333626180540302, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.087843418121338, + "learning_rate": 1e-06, + "loss": 0.9818, + "mean_token_accuracy": 0.7058852314949036, + "num_tokens": 290547917.0, + "step": 11231 + }, + { + "epoch": 1.233472435756644, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3486416339874268, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7073776721954346, + "num_tokens": 290575889.0, + "step": 11232 + }, + { + "epoch": 1.2335822534592575, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4501090049743652, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7325539588928223, + "num_tokens": 290597692.0, + "step": 11233 + }, + { + "epoch": 1.2336920711618713, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.244157314300537, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.6997315287590027, + "num_tokens": 290624743.0, + "step": 11234 + }, + { + "epoch": 1.2338018888644848, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2490248680114746, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7076253890991211, + "num_tokens": 290651588.0, + "step": 11235 + }, + { + "epoch": 1.2339117065670986, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.213921546936035, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7088093757629395, + "num_tokens": 290679796.0, + "step": 11236 + }, + { + "epoch": 1.2340215242697123, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.7715604305267334, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7220842242240906, + "num_tokens": 290698190.0, + "step": 11237 + }, + { + "epoch": 1.2341313419723259, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.267932415008545, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7285616993904114, + "num_tokens": 290722165.0, + "step": 11238 + }, + { + "epoch": 1.2342411596749396, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.9487823247909546, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.6968404054641724, + "num_tokens": 290755251.0, + "step": 11239 + }, + { + "epoch": 1.2343509773775532, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.059312582015991, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7169522047042847, + "num_tokens": 290782938.0, + "step": 11240 + }, + { + "epoch": 1.234460795080167, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.9416589736938477, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.730314314365387, + "num_tokens": 290814105.0, + "step": 11241 + }, + { + "epoch": 1.2345706127827807, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0522971153259277, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7349948287010193, + "num_tokens": 290841887.0, + "step": 11242 + }, + { + "epoch": 1.2346804304853942, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.130641460418701, + "learning_rate": 1e-06, + "loss": 1.0674, + "mean_token_accuracy": 0.6732008457183838, + "num_tokens": 290872908.0, + "step": 11243 + }, + { + "epoch": 1.234790248188008, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.120508909225464, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7129446268081665, + "num_tokens": 290902229.0, + "step": 11244 + }, + { + "epoch": 1.2349000658906215, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1230244636535645, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6951097249984741, + "num_tokens": 290930274.0, + "step": 11245 + }, + { + "epoch": 1.2350098835932353, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2255046367645264, + "learning_rate": 1e-06, + "loss": 1.0283, + "mean_token_accuracy": 0.6857112646102905, + "num_tokens": 290957237.0, + "step": 11246 + }, + { + "epoch": 1.2351197012958488, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.268171548843384, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7268987894058228, + "num_tokens": 290982019.0, + "step": 11247 + }, + { + "epoch": 1.2352295189984626, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4229090213775635, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7364020943641663, + "num_tokens": 291004914.0, + "step": 11248 + }, + { + "epoch": 1.235339336701076, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0842905044555664, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7147555947303772, + "num_tokens": 291035370.0, + "step": 11249 + }, + { + "epoch": 1.2354491544036899, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1831765174865723, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6974049210548401, + "num_tokens": 291064938.0, + "step": 11250 + }, + { + "epoch": 1.2355589721063036, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.193861246109009, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7100359201431274, + "num_tokens": 291090672.0, + "step": 11251 + }, + { + "epoch": 1.2356687898089171, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4579193592071533, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7195971012115479, + "num_tokens": 291113233.0, + "step": 11252 + }, + { + "epoch": 1.235778607511531, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.108391046524048, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7182880640029907, + "num_tokens": 291142475.0, + "step": 11253 + }, + { + "epoch": 1.2358884252141444, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.08879017829895, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7170779705047607, + "num_tokens": 291169504.0, + "step": 11254 + }, + { + "epoch": 1.2359982429167582, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.0547728538513184, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7110342979431152, + "num_tokens": 291202195.0, + "step": 11255 + }, + { + "epoch": 1.236108060619372, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2561912536621094, + "learning_rate": 1e-06, + "loss": 1.0331, + "mean_token_accuracy": 0.6894267797470093, + "num_tokens": 291228469.0, + "step": 11256 + }, + { + "epoch": 1.2362178783219855, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.1255340576171875, + "learning_rate": 1e-06, + "loss": 1.0007, + "mean_token_accuracy": 0.6892337799072266, + "num_tokens": 291259022.0, + "step": 11257 + }, + { + "epoch": 1.2363276960245992, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.265436887741089, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7281357049942017, + "num_tokens": 291283767.0, + "step": 11258 + }, + { + "epoch": 1.2364375137272128, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.4358699321746826, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7258573770523071, + "num_tokens": 291306139.0, + "step": 11259 + }, + { + "epoch": 1.2365473314298265, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 1.9925562143325806, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.71578049659729, + "num_tokens": 291336887.0, + "step": 11260 + }, + { + "epoch": 1.23665714913244, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.270155191421509, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7020460963249207, + "num_tokens": 291363600.0, + "step": 11261 + }, + { + "epoch": 1.2367669668350538, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.2228050231933594, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7211402654647827, + "num_tokens": 291389067.0, + "step": 11262 + }, + { + "epoch": 1.2368767845376674, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.433392286300659, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.6941883563995361, + "num_tokens": 291413585.0, + "step": 11263 + }, + { + "epoch": 1.2369866022402811, + "ewc_loss": 6.5267086029052734e-06, + "grad_norm": 2.3655107021331787, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.735273003578186, + "num_tokens": 291434867.0, + "step": 11264 + }, + { + "epoch": 1.2370964199428949, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2631075382232666, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7207455635070801, + "num_tokens": 291461764.0, + "step": 11265 + }, + { + "epoch": 1.2372062376455084, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.224241256713867, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7085240483283997, + "num_tokens": 291488950.0, + "step": 11266 + }, + { + "epoch": 1.2373160553481222, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.143871307373047, + "learning_rate": 1e-06, + "loss": 1.0358, + "mean_token_accuracy": 0.6833624839782715, + "num_tokens": 291516738.0, + "step": 11267 + }, + { + "epoch": 1.2374258730507357, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.016118288040161, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7128627300262451, + "num_tokens": 291548428.0, + "step": 11268 + }, + { + "epoch": 1.2375356907533495, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.259798526763916, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7272643446922302, + "num_tokens": 291572552.0, + "step": 11269 + }, + { + "epoch": 1.2376455084559632, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1917436122894287, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7385066747665405, + "num_tokens": 291597888.0, + "step": 11270 + }, + { + "epoch": 1.2377553261585768, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.185744285583496, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7157799601554871, + "num_tokens": 291624458.0, + "step": 11271 + }, + { + "epoch": 1.2378651438611905, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.279324769973755, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7130958437919617, + "num_tokens": 291649258.0, + "step": 11272 + }, + { + "epoch": 1.237974961563804, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2100701332092285, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7367974519729614, + "num_tokens": 291673698.0, + "step": 11273 + }, + { + "epoch": 1.2380847792664178, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3281397819519043, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7256481647491455, + "num_tokens": 291697675.0, + "step": 11274 + }, + { + "epoch": 1.2381945969690313, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.6800284385681152, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7110452651977539, + "num_tokens": 291724573.0, + "step": 11275 + }, + { + "epoch": 1.238304414671645, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.6726584434509277, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7330842018127441, + "num_tokens": 291743913.0, + "step": 11276 + }, + { + "epoch": 1.2384142323742586, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.202382802963257, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7357856631278992, + "num_tokens": 291768326.0, + "step": 11277 + }, + { + "epoch": 1.2385240500768724, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.274174213409424, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.707634687423706, + "num_tokens": 291796062.0, + "step": 11278 + }, + { + "epoch": 1.2386338677794861, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.180495023727417, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7184032201766968, + "num_tokens": 291825094.0, + "step": 11279 + }, + { + "epoch": 1.2387436854820997, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2947561740875244, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7197049856185913, + "num_tokens": 291851180.0, + "step": 11280 + }, + { + "epoch": 1.2388535031847134, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2694408893585205, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7297539710998535, + "num_tokens": 291875864.0, + "step": 11281 + }, + { + "epoch": 1.238963320887327, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2398550510406494, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7215724587440491, + "num_tokens": 291901757.0, + "step": 11282 + }, + { + "epoch": 1.2390731385899407, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.331441879272461, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7360677719116211, + "num_tokens": 291924418.0, + "step": 11283 + }, + { + "epoch": 1.2391829562925545, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.130086660385132, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7116997241973877, + "num_tokens": 291953676.0, + "step": 11284 + }, + { + "epoch": 1.239292773995168, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.150484085083008, + "learning_rate": 1e-06, + "loss": 1.0272, + "mean_token_accuracy": 0.6882328391075134, + "num_tokens": 291982067.0, + "step": 11285 + }, + { + "epoch": 1.2394025916977818, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.5527493953704834, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7251772880554199, + "num_tokens": 292002505.0, + "step": 11286 + }, + { + "epoch": 1.2395124094003953, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1022534370422363, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7153209447860718, + "num_tokens": 292030540.0, + "step": 11287 + }, + { + "epoch": 1.239622227103009, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.343935251235962, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7021665573120117, + "num_tokens": 292055186.0, + "step": 11288 + }, + { + "epoch": 1.2397320448056226, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.449378728866577, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7417843341827393, + "num_tokens": 292075507.0, + "step": 11289 + }, + { + "epoch": 1.2398418625082364, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.163503885269165, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.7008851766586304, + "num_tokens": 292103532.0, + "step": 11290 + }, + { + "epoch": 1.23995168021085, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.127385377883911, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7334960103034973, + "num_tokens": 292131001.0, + "step": 11291 + }, + { + "epoch": 1.2400614979134637, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.0970966815948486, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7173982858657837, + "num_tokens": 292159966.0, + "step": 11292 + }, + { + "epoch": 1.2401713156160774, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.275890827178955, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7280789017677307, + "num_tokens": 292185311.0, + "step": 11293 + }, + { + "epoch": 1.240281133318691, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1830577850341797, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7270593643188477, + "num_tokens": 292210456.0, + "step": 11294 + }, + { + "epoch": 1.2403909510213047, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.144805431365967, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7081732749938965, + "num_tokens": 292238954.0, + "step": 11295 + }, + { + "epoch": 1.2405007687239182, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.0762670040130615, + "learning_rate": 1e-06, + "loss": 0.7896, + "mean_token_accuracy": 0.7513523697853088, + "num_tokens": 292268336.0, + "step": 11296 + }, + { + "epoch": 1.240610586426532, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3699088096618652, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7348313331604004, + "num_tokens": 292290510.0, + "step": 11297 + }, + { + "epoch": 1.2407204041291455, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1411256790161133, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.733045220375061, + "num_tokens": 292319469.0, + "step": 11298 + }, + { + "epoch": 1.2408302218317593, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3171041011810303, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7111436128616333, + "num_tokens": 292345632.0, + "step": 11299 + }, + { + "epoch": 1.2409400395343728, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2206335067749023, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7090567946434021, + "num_tokens": 292372304.0, + "step": 11300 + }, + { + "epoch": 1.2410498572369866, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.275646686553955, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7515302896499634, + "num_tokens": 292395329.0, + "step": 11301 + }, + { + "epoch": 1.2411596749396003, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.36077618598938, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7216134071350098, + "num_tokens": 292420030.0, + "step": 11302 + }, + { + "epoch": 1.2412694926422139, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.228201389312744, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7131313681602478, + "num_tokens": 292444679.0, + "step": 11303 + }, + { + "epoch": 1.2413793103448276, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.177845001220703, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7222409248352051, + "num_tokens": 292471191.0, + "step": 11304 + }, + { + "epoch": 1.2414891280474412, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.532360553741455, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7216435670852661, + "num_tokens": 292493761.0, + "step": 11305 + }, + { + "epoch": 1.241598945750055, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3597922325134277, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7133381366729736, + "num_tokens": 292516844.0, + "step": 11306 + }, + { + "epoch": 1.2417087634526687, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.4815382957458496, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.7025088667869568, + "num_tokens": 292540030.0, + "step": 11307 + }, + { + "epoch": 1.2418185811552822, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1258833408355713, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7215933203697205, + "num_tokens": 292568982.0, + "step": 11308 + }, + { + "epoch": 1.241928398857896, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.5529258251190186, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7202118635177612, + "num_tokens": 292589148.0, + "step": 11309 + }, + { + "epoch": 1.2420382165605095, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.142194986343384, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7051874995231628, + "num_tokens": 292617921.0, + "step": 11310 + }, + { + "epoch": 1.2421480342631233, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1671338081359863, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7383794188499451, + "num_tokens": 292644301.0, + "step": 11311 + }, + { + "epoch": 1.2422578519657368, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.0446627140045166, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7220522165298462, + "num_tokens": 292674415.0, + "step": 11312 + }, + { + "epoch": 1.2423676696683505, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3590288162231445, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7220458984375, + "num_tokens": 292698202.0, + "step": 11313 + }, + { + "epoch": 1.242477487370964, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.084221124649048, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7165681719779968, + "num_tokens": 292727895.0, + "step": 11314 + }, + { + "epoch": 1.2425873050735778, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.4592843055725098, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.6971021294593811, + "num_tokens": 292751726.0, + "step": 11315 + }, + { + "epoch": 1.2426971227761916, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.2294106483459473, + "learning_rate": 1e-06, + "loss": 1.0585, + "mean_token_accuracy": 0.6787283420562744, + "num_tokens": 292782025.0, + "step": 11316 + }, + { + "epoch": 1.2428069404788051, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.1802663803100586, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7188386917114258, + "num_tokens": 292809852.0, + "step": 11317 + }, + { + "epoch": 1.242916758181419, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.031827688217163, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7063695192337036, + "num_tokens": 292840054.0, + "step": 11318 + }, + { + "epoch": 1.2430265758840324, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.6519742012023926, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7441563010215759, + "num_tokens": 292858529.0, + "step": 11319 + }, + { + "epoch": 1.2431363935866462, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.003573417663574, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7071301937103271, + "num_tokens": 292888213.0, + "step": 11320 + }, + { + "epoch": 1.24324621128926, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 1.8912181854248047, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.716474175453186, + "num_tokens": 292921633.0, + "step": 11321 + }, + { + "epoch": 1.2433560289918735, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.0957601070404053, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7161926627159119, + "num_tokens": 292950474.0, + "step": 11322 + }, + { + "epoch": 1.2434658466944872, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.339952230453491, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7007139921188354, + "num_tokens": 292975140.0, + "step": 11323 + }, + { + "epoch": 1.2435756643971008, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.029047727584839, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7246935963630676, + "num_tokens": 293005180.0, + "step": 11324 + }, + { + "epoch": 1.2436854820997145, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.3381564617156982, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7273771166801453, + "num_tokens": 293028606.0, + "step": 11325 + }, + { + "epoch": 1.243795299802328, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 2.482997417449951, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7314819693565369, + "num_tokens": 293049332.0, + "step": 11326 + }, + { + "epoch": 1.2439051175049418, + "ewc_loss": 6.556510925292969e-06, + "grad_norm": 1.9284112453460693, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7158194184303284, + "num_tokens": 293079125.0, + "step": 11327 + }, + { + "epoch": 1.2440149352075554, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.171018362045288, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.6993322372436523, + "num_tokens": 293106284.0, + "step": 11328 + }, + { + "epoch": 1.244124752910169, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 1.885783314704895, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7244938611984253, + "num_tokens": 293137777.0, + "step": 11329 + }, + { + "epoch": 1.2442345706127829, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.6799092292785645, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7410227656364441, + "num_tokens": 293157064.0, + "step": 11330 + }, + { + "epoch": 1.2443443883153964, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.418160915374756, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6970870494842529, + "num_tokens": 293181353.0, + "step": 11331 + }, + { + "epoch": 1.2444542060180102, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1833126544952393, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7062280774116516, + "num_tokens": 293211082.0, + "step": 11332 + }, + { + "epoch": 1.2445640237206237, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.203958749771118, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7293640375137329, + "num_tokens": 293238078.0, + "step": 11333 + }, + { + "epoch": 1.2446738414232374, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.8131933212280273, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7311191558837891, + "num_tokens": 293255608.0, + "step": 11334 + }, + { + "epoch": 1.2447836591258512, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.0497798919677734, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7128169536590576, + "num_tokens": 293285103.0, + "step": 11335 + }, + { + "epoch": 1.2448934768284647, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2705018520355225, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7111589908599854, + "num_tokens": 293309605.0, + "step": 11336 + }, + { + "epoch": 1.2450032945310785, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2215569019317627, + "learning_rate": 1e-06, + "loss": 1.0141, + "mean_token_accuracy": 0.6900884509086609, + "num_tokens": 293336245.0, + "step": 11337 + }, + { + "epoch": 1.245113112233692, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.306009292602539, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7025497555732727, + "num_tokens": 293361734.0, + "step": 11338 + }, + { + "epoch": 1.2452229299363058, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1385796070098877, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7248412370681763, + "num_tokens": 293389235.0, + "step": 11339 + }, + { + "epoch": 1.2453327476389193, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.4966013431549072, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7185143828392029, + "num_tokens": 293410452.0, + "step": 11340 + }, + { + "epoch": 1.245442565341533, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.3659772872924805, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7366536855697632, + "num_tokens": 293435361.0, + "step": 11341 + }, + { + "epoch": 1.2455523830441466, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.4295668601989746, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7370909452438354, + "num_tokens": 293457915.0, + "step": 11342 + }, + { + "epoch": 1.2456622007467604, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.470309257507324, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7458736300468445, + "num_tokens": 293478810.0, + "step": 11343 + }, + { + "epoch": 1.2457720184493741, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.3137712478637695, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7322689890861511, + "num_tokens": 293502615.0, + "step": 11344 + }, + { + "epoch": 1.2458818361519877, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2411646842956543, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7241312265396118, + "num_tokens": 293527291.0, + "step": 11345 + }, + { + "epoch": 1.2459916538546014, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2015187740325928, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7359305620193481, + "num_tokens": 293554129.0, + "step": 11346 + }, + { + "epoch": 1.246101471557215, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1854469776153564, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7249962091445923, + "num_tokens": 293579967.0, + "step": 11347 + }, + { + "epoch": 1.2462112892598287, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1763784885406494, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7337965965270996, + "num_tokens": 293606901.0, + "step": 11348 + }, + { + "epoch": 1.2463211069624422, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2185752391815186, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.702258825302124, + "num_tokens": 293634345.0, + "step": 11349 + }, + { + "epoch": 1.246430924665056, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2059319019317627, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7118874788284302, + "num_tokens": 293660439.0, + "step": 11350 + }, + { + "epoch": 1.2465407423676698, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.207414388656616, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7366325855255127, + "num_tokens": 293685799.0, + "step": 11351 + }, + { + "epoch": 1.2466505600702833, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.340160846710205, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7312915325164795, + "num_tokens": 293710985.0, + "step": 11352 + }, + { + "epoch": 1.246760377772897, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.13110613822937, + "learning_rate": 1e-06, + "loss": 1.0268, + "mean_token_accuracy": 0.6890264749526978, + "num_tokens": 293742124.0, + "step": 11353 + }, + { + "epoch": 1.2468701954755106, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.4410288333892822, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7154226303100586, + "num_tokens": 293764646.0, + "step": 11354 + }, + { + "epoch": 1.2469800131781243, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.435267448425293, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7254325151443481, + "num_tokens": 293788042.0, + "step": 11355 + }, + { + "epoch": 1.2470898308807379, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.4309630393981934, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7337794303894043, + "num_tokens": 293811110.0, + "step": 11356 + }, + { + "epoch": 1.2471996485833516, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1163525581359863, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.7002214193344116, + "num_tokens": 293839656.0, + "step": 11357 + }, + { + "epoch": 1.2473094662859654, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.3363070487976074, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.6978890895843506, + "num_tokens": 293864921.0, + "step": 11358 + }, + { + "epoch": 1.247419283988579, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.2722132205963135, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7139731049537659, + "num_tokens": 293891997.0, + "step": 11359 + }, + { + "epoch": 1.2475291016911927, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 1.9385510683059692, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.6978695392608643, + "num_tokens": 293928415.0, + "step": 11360 + }, + { + "epoch": 1.2476389193938062, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 1.8591868877410889, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.710071325302124, + "num_tokens": 293964945.0, + "step": 11361 + }, + { + "epoch": 1.24774873709642, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1603288650512695, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7087920904159546, + "num_tokens": 293991970.0, + "step": 11362 + }, + { + "epoch": 1.2478585547990335, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.278262138366699, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7214928865432739, + "num_tokens": 294018661.0, + "step": 11363 + }, + { + "epoch": 1.2479683725016473, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1204710006713867, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7489378452301025, + "num_tokens": 294046716.0, + "step": 11364 + }, + { + "epoch": 1.2480781902042608, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.1431241035461426, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7492084503173828, + "num_tokens": 294071987.0, + "step": 11365 + }, + { + "epoch": 1.2481880079068746, + "ewc_loss": 6.586313247680664e-06, + "grad_norm": 2.3429224491119385, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7417682409286499, + "num_tokens": 294095503.0, + "step": 11366 + }, + { + "epoch": 1.2482978256094883, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.308292865753174, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6963600516319275, + "num_tokens": 294122424.0, + "step": 11367 + }, + { + "epoch": 1.2484076433121019, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.128094434738159, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7151652574539185, + "num_tokens": 294150376.0, + "step": 11368 + }, + { + "epoch": 1.2485174610147156, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.189624309539795, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7206498980522156, + "num_tokens": 294177393.0, + "step": 11369 + }, + { + "epoch": 1.2486272787173291, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.248563051223755, + "learning_rate": 1e-06, + "loss": 0.9895, + "mean_token_accuracy": 0.71045982837677, + "num_tokens": 294203511.0, + "step": 11370 + }, + { + "epoch": 1.248737096419943, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.0550198554992676, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7156925201416016, + "num_tokens": 294232230.0, + "step": 11371 + }, + { + "epoch": 1.2488469141225567, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.080209255218506, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7419748306274414, + "num_tokens": 294261107.0, + "step": 11372 + }, + { + "epoch": 1.2489567318251702, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5951757431030273, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7299565672874451, + "num_tokens": 294283492.0, + "step": 11373 + }, + { + "epoch": 1.249066549527784, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2255172729492188, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7244296669960022, + "num_tokens": 294308887.0, + "step": 11374 + }, + { + "epoch": 1.2491763672303975, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2843105792999268, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7107457518577576, + "num_tokens": 294336422.0, + "step": 11375 + }, + { + "epoch": 1.2492861849330112, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2024171352386475, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7009636163711548, + "num_tokens": 294363697.0, + "step": 11376 + }, + { + "epoch": 1.2493960026356248, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2354400157928467, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7319937348365784, + "num_tokens": 294387983.0, + "step": 11377 + }, + { + "epoch": 1.2495058203382385, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1790435314178467, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7100400924682617, + "num_tokens": 294414661.0, + "step": 11378 + }, + { + "epoch": 1.249615638040852, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1450533866882324, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7286738157272339, + "num_tokens": 294440692.0, + "step": 11379 + }, + { + "epoch": 1.2497254557434658, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.4208438396453857, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7163459062576294, + "num_tokens": 294468741.0, + "step": 11380 + }, + { + "epoch": 1.2498352734460796, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.516733169555664, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7264654636383057, + "num_tokens": 294489488.0, + "step": 11381 + }, + { + "epoch": 1.2499450911486931, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3266570568084717, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7149393558502197, + "num_tokens": 294513430.0, + "step": 11382 + }, + { + "epoch": 1.2500549088513069, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.0442910194396973, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7195768356323242, + "num_tokens": 294542188.0, + "step": 11383 + }, + { + "epoch": 1.2501647265539204, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.339099884033203, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7158597707748413, + "num_tokens": 294566002.0, + "step": 11384 + }, + { + "epoch": 1.2502745442565342, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1866025924682617, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7150505781173706, + "num_tokens": 294591308.0, + "step": 11385 + }, + { + "epoch": 1.250384361959148, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2802557945251465, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7025121450424194, + "num_tokens": 294615500.0, + "step": 11386 + }, + { + "epoch": 1.2504941796617615, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.233440399169922, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7184978723526001, + "num_tokens": 294640524.0, + "step": 11387 + }, + { + "epoch": 1.2506039973643752, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.0270602703094482, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7104032635688782, + "num_tokens": 294670595.0, + "step": 11388 + }, + { + "epoch": 1.2507138150669888, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2636704444885254, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7040165662765503, + "num_tokens": 294696095.0, + "step": 11389 + }, + { + "epoch": 1.2508236327696025, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1551311016082764, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7275798320770264, + "num_tokens": 294722676.0, + "step": 11390 + }, + { + "epoch": 1.250933450472216, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.076655387878418, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.718189537525177, + "num_tokens": 294752360.0, + "step": 11391 + }, + { + "epoch": 1.2510432681748298, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2767462730407715, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7103751301765442, + "num_tokens": 294777603.0, + "step": 11392 + }, + { + "epoch": 1.2511530858774433, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.195225477218628, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7326065897941589, + "num_tokens": 294802529.0, + "step": 11393 + }, + { + "epoch": 1.251262903580057, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3093597888946533, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7013297080993652, + "num_tokens": 294827373.0, + "step": 11394 + }, + { + "epoch": 1.2513727212826709, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.107489585876465, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7198543548583984, + "num_tokens": 294855380.0, + "step": 11395 + }, + { + "epoch": 1.2514825389852844, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3247969150543213, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.7493811845779419, + "num_tokens": 294877037.0, + "step": 11396 + }, + { + "epoch": 1.2515923566878981, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.294865131378174, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7141585946083069, + "num_tokens": 294901232.0, + "step": 11397 + }, + { + "epoch": 1.2517021743905117, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.4677088260650635, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7316850423812866, + "num_tokens": 294921628.0, + "step": 11398 + }, + { + "epoch": 1.2518119920931254, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2668845653533936, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7288336753845215, + "num_tokens": 294947376.0, + "step": 11399 + }, + { + "epoch": 1.2519218097957392, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1472997665405273, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7324113845825195, + "num_tokens": 294976345.0, + "step": 11400 + }, + { + "epoch": 1.2520316274983527, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2688748836517334, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7237682342529297, + "num_tokens": 295000527.0, + "step": 11401 + }, + { + "epoch": 1.2521414452009663, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5481863021850586, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.700486421585083, + "num_tokens": 295021857.0, + "step": 11402 + }, + { + "epoch": 1.25225126290358, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3661036491394043, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7489608526229858, + "num_tokens": 295045127.0, + "step": 11403 + }, + { + "epoch": 1.2523610806061938, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.243279457092285, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7379496097564697, + "num_tokens": 295070148.0, + "step": 11404 + }, + { + "epoch": 1.2524708983088073, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5211238861083984, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7171781063079834, + "num_tokens": 295089819.0, + "step": 11405 + }, + { + "epoch": 1.252580716011421, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.810774326324463, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7205424308776855, + "num_tokens": 295107843.0, + "step": 11406 + }, + { + "epoch": 1.2526905337140346, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3078970909118652, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7308253049850464, + "num_tokens": 295133757.0, + "step": 11407 + }, + { + "epoch": 1.2528003514166484, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.256263017654419, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7105907201766968, + "num_tokens": 295161640.0, + "step": 11408 + }, + { + "epoch": 1.2529101691192621, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1039493083953857, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7322425246238708, + "num_tokens": 295189968.0, + "step": 11409 + }, + { + "epoch": 1.2530199868218757, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.4574074745178223, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.734217643737793, + "num_tokens": 295214500.0, + "step": 11410 + }, + { + "epoch": 1.2531298045244894, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.416438102722168, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.709537148475647, + "num_tokens": 295237895.0, + "step": 11411 + }, + { + "epoch": 1.253239622227103, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3208582401275635, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7307202219963074, + "num_tokens": 295260666.0, + "step": 11412 + }, + { + "epoch": 1.2533494399297167, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.01029634475708, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7039165496826172, + "num_tokens": 295293423.0, + "step": 11413 + }, + { + "epoch": 1.2534592576323305, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2540946006774902, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7238918542861938, + "num_tokens": 295318031.0, + "step": 11414 + }, + { + "epoch": 1.253569075334944, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1525216102600098, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6912067532539368, + "num_tokens": 295345764.0, + "step": 11415 + }, + { + "epoch": 1.2536788930375575, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3688597679138184, + "learning_rate": 1e-06, + "loss": 1.0179, + "mean_token_accuracy": 0.6913455128669739, + "num_tokens": 295369222.0, + "step": 11416 + }, + { + "epoch": 1.2537887107401713, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5236871242523193, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7429763078689575, + "num_tokens": 295390231.0, + "step": 11417 + }, + { + "epoch": 1.253898528442785, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5930674076080322, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7259156703948975, + "num_tokens": 295411044.0, + "step": 11418 + }, + { + "epoch": 1.2540083461453986, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.671982765197754, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7325177192687988, + "num_tokens": 295432250.0, + "step": 11419 + }, + { + "epoch": 1.2541181638480123, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.162388324737549, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7204131484031677, + "num_tokens": 295462432.0, + "step": 11420 + }, + { + "epoch": 1.2542279815506259, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.115447521209717, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7023072242736816, + "num_tokens": 295490547.0, + "step": 11421 + }, + { + "epoch": 1.2543377992532396, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.322145700454712, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7184665203094482, + "num_tokens": 295513531.0, + "step": 11422 + }, + { + "epoch": 1.2544476169558534, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.409083127975464, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7399859428405762, + "num_tokens": 295535425.0, + "step": 11423 + }, + { + "epoch": 1.254557434658467, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2873854637145996, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7223207950592041, + "num_tokens": 295561283.0, + "step": 11424 + }, + { + "epoch": 1.2546672523610807, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1318442821502686, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7222077250480652, + "num_tokens": 295589515.0, + "step": 11425 + }, + { + "epoch": 1.2547770700636942, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3277904987335205, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7091891765594482, + "num_tokens": 295615953.0, + "step": 11426 + }, + { + "epoch": 1.254886887766308, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.193331241607666, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7185639142990112, + "num_tokens": 295644770.0, + "step": 11427 + }, + { + "epoch": 1.2549967054689217, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.469355344772339, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7272883653640747, + "num_tokens": 295665904.0, + "step": 11428 + }, + { + "epoch": 1.2551065231715353, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 1.9263771772384644, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7117502689361572, + "num_tokens": 295698262.0, + "step": 11429 + }, + { + "epoch": 1.2552163408741488, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.26804518699646, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6893959045410156, + "num_tokens": 295723447.0, + "step": 11430 + }, + { + "epoch": 1.2553261585767626, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.4346556663513184, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7328776121139526, + "num_tokens": 295744698.0, + "step": 11431 + }, + { + "epoch": 1.2554359762793763, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5969083309173584, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.728710949420929, + "num_tokens": 295764393.0, + "step": 11432 + }, + { + "epoch": 1.2555457939819898, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.321030855178833, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7126718163490295, + "num_tokens": 295790134.0, + "step": 11433 + }, + { + "epoch": 1.2556556116846036, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.495936393737793, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7229191064834595, + "num_tokens": 295811072.0, + "step": 11434 + }, + { + "epoch": 1.2557654293872171, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.1235971450805664, + "learning_rate": 1e-06, + "loss": 0.8235, + "mean_token_accuracy": 0.7464704513549805, + "num_tokens": 295837017.0, + "step": 11435 + }, + { + "epoch": 1.255875247089831, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3734171390533447, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7178865671157837, + "num_tokens": 295859040.0, + "step": 11436 + }, + { + "epoch": 1.2559850647924446, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.7922732830047607, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7463161945343018, + "num_tokens": 295875491.0, + "step": 11437 + }, + { + "epoch": 1.2560948824950582, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2935550212860107, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7332168817520142, + "num_tokens": 295897898.0, + "step": 11438 + }, + { + "epoch": 1.256204700197672, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.030416488647461, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.6969799995422363, + "num_tokens": 295928789.0, + "step": 11439 + }, + { + "epoch": 1.2563145179002855, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.22141432762146, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7232710123062134, + "num_tokens": 295953425.0, + "step": 11440 + }, + { + "epoch": 1.2564243356028992, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.0975472927093506, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.699531078338623, + "num_tokens": 295982609.0, + "step": 11441 + }, + { + "epoch": 1.2565341533055128, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2857837677001953, + "learning_rate": 1e-06, + "loss": 0.999, + "mean_token_accuracy": 0.6930254697799683, + "num_tokens": 296008120.0, + "step": 11442 + }, + { + "epoch": 1.2566439710081265, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.716151714324951, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.735977053642273, + "num_tokens": 296025613.0, + "step": 11443 + }, + { + "epoch": 1.25675378871074, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.118243932723999, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7139503359794617, + "num_tokens": 296053874.0, + "step": 11444 + }, + { + "epoch": 1.2568636064133538, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.647033929824829, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7333402633666992, + "num_tokens": 296074215.0, + "step": 11445 + }, + { + "epoch": 1.2569734241159676, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.143660306930542, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7064695358276367, + "num_tokens": 296102474.0, + "step": 11446 + }, + { + "epoch": 1.257083241818581, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.112396478652954, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.6973479390144348, + "num_tokens": 296131694.0, + "step": 11447 + }, + { + "epoch": 1.2571930595211949, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.421342372894287, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7373632192611694, + "num_tokens": 296153126.0, + "step": 11448 + }, + { + "epoch": 1.2573028772238084, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.3504137992858887, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7020705938339233, + "num_tokens": 296179671.0, + "step": 11449 + }, + { + "epoch": 1.2574126949264222, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.444108486175537, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7027402520179749, + "num_tokens": 296204581.0, + "step": 11450 + }, + { + "epoch": 1.257522512629036, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.6629953384399414, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7193259000778198, + "num_tokens": 296224412.0, + "step": 11451 + }, + { + "epoch": 1.2576323303316495, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5061240196228027, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7183152437210083, + "num_tokens": 296246127.0, + "step": 11452 + }, + { + "epoch": 1.257742148034263, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2879326343536377, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7101004719734192, + "num_tokens": 296269949.0, + "step": 11453 + }, + { + "epoch": 1.2578519657368767, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.5236687660217285, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.6941604614257812, + "num_tokens": 296291955.0, + "step": 11454 + }, + { + "epoch": 1.2579617834394905, + "ewc_loss": 6.645917892456055e-06, + "grad_norm": 2.2377126216888428, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7246436476707458, + "num_tokens": 296319512.0, + "step": 11455 + }, + { + "epoch": 1.258071601142104, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.305746078491211, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7133736610412598, + "num_tokens": 296344515.0, + "step": 11456 + }, + { + "epoch": 1.2581814188447178, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2537877559661865, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7079930305480957, + "num_tokens": 296371510.0, + "step": 11457 + }, + { + "epoch": 1.2582912365473313, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3093669414520264, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7253232598304749, + "num_tokens": 296395037.0, + "step": 11458 + }, + { + "epoch": 1.258401054249945, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.266002655029297, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7215613722801208, + "num_tokens": 296420556.0, + "step": 11459 + }, + { + "epoch": 1.2585108719525588, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1624741554260254, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7014937400817871, + "num_tokens": 296449227.0, + "step": 11460 + }, + { + "epoch": 1.2586206896551724, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.160855531692505, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7220736145973206, + "num_tokens": 296474535.0, + "step": 11461 + }, + { + "epoch": 1.2587305073577861, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3551390171051025, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7059401869773865, + "num_tokens": 296498255.0, + "step": 11462 + }, + { + "epoch": 1.2588403250603997, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.0971436500549316, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7291561365127563, + "num_tokens": 296527472.0, + "step": 11463 + }, + { + "epoch": 1.2589501427630134, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3867034912109375, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7129552364349365, + "num_tokens": 296551741.0, + "step": 11464 + }, + { + "epoch": 1.2590599604656272, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.284080982208252, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7459976077079773, + "num_tokens": 296577777.0, + "step": 11465 + }, + { + "epoch": 1.2591697781682407, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.001736640930176, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7171212434768677, + "num_tokens": 296611448.0, + "step": 11466 + }, + { + "epoch": 1.2592795958708543, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3872628211975098, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7174568176269531, + "num_tokens": 296638612.0, + "step": 11467 + }, + { + "epoch": 1.259389413573468, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2409191131591797, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7190555334091187, + "num_tokens": 296663290.0, + "step": 11468 + }, + { + "epoch": 1.2594992312760818, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2299304008483887, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7314170002937317, + "num_tokens": 296688612.0, + "step": 11469 + }, + { + "epoch": 1.2596090489786953, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2854955196380615, + "learning_rate": 1e-06, + "loss": 1.0084, + "mean_token_accuracy": 0.6908161640167236, + "num_tokens": 296718311.0, + "step": 11470 + }, + { + "epoch": 1.259718866681309, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.354768753051758, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7121857404708862, + "num_tokens": 296741384.0, + "step": 11471 + }, + { + "epoch": 1.2598286843839226, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3193106651306152, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7357429265975952, + "num_tokens": 296764291.0, + "step": 11472 + }, + { + "epoch": 1.2599385020865363, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.245049476623535, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7150503396987915, + "num_tokens": 296790654.0, + "step": 11473 + }, + { + "epoch": 1.26004831978915, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 1.9808372259140015, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7309507131576538, + "num_tokens": 296821205.0, + "step": 11474 + }, + { + "epoch": 1.2601581374917636, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.0454494953155518, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7313015460968018, + "num_tokens": 296849946.0, + "step": 11475 + }, + { + "epoch": 1.2602679551943774, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.17903733253479, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7136490345001221, + "num_tokens": 296879116.0, + "step": 11476 + }, + { + "epoch": 1.260377772896991, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.253446102142334, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.6964013576507568, + "num_tokens": 296905714.0, + "step": 11477 + }, + { + "epoch": 1.2604875905996047, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1938388347625732, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7132883667945862, + "num_tokens": 296932215.0, + "step": 11478 + }, + { + "epoch": 1.2605974083022184, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.0459344387054443, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.7627712488174438, + "num_tokens": 296959360.0, + "step": 11479 + }, + { + "epoch": 1.260707226004832, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3913917541503906, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7049010992050171, + "num_tokens": 296981937.0, + "step": 11480 + }, + { + "epoch": 1.2608170437074455, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 1.941786766052246, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7121188044548035, + "num_tokens": 297013783.0, + "step": 11481 + }, + { + "epoch": 1.2609268614100593, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 1.9869024753570557, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7023407816886902, + "num_tokens": 297045639.0, + "step": 11482 + }, + { + "epoch": 1.261036679112673, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1047840118408203, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7051551938056946, + "num_tokens": 297074810.0, + "step": 11483 + }, + { + "epoch": 1.2611464968152866, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1886308193206787, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7398679256439209, + "num_tokens": 297100410.0, + "step": 11484 + }, + { + "epoch": 1.2612563145179003, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 1.9585466384887695, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7261955738067627, + "num_tokens": 297132999.0, + "step": 11485 + }, + { + "epoch": 1.2613661322205139, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.379643678665161, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7220538854598999, + "num_tokens": 297157221.0, + "step": 11486 + }, + { + "epoch": 1.2614759499231276, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2869577407836914, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7085567116737366, + "num_tokens": 297183062.0, + "step": 11487 + }, + { + "epoch": 1.2615857676257414, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.311518907546997, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7317661643028259, + "num_tokens": 297206154.0, + "step": 11488 + }, + { + "epoch": 1.261695585328355, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.353584051132202, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7331640720367432, + "num_tokens": 297228651.0, + "step": 11489 + }, + { + "epoch": 1.2618054030309687, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.114499807357788, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.715116560459137, + "num_tokens": 297259028.0, + "step": 11490 + }, + { + "epoch": 1.2619152207335822, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.4022390842437744, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.721316397190094, + "num_tokens": 297285857.0, + "step": 11491 + }, + { + "epoch": 1.262025038436196, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3339710235595703, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6953188180923462, + "num_tokens": 297312496.0, + "step": 11492 + }, + { + "epoch": 1.2621348561388097, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2256722450256348, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7008625864982605, + "num_tokens": 297338471.0, + "step": 11493 + }, + { + "epoch": 1.2622446738414232, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.4021904468536377, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.724656343460083, + "num_tokens": 297363122.0, + "step": 11494 + }, + { + "epoch": 1.2623544915440368, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.091930866241455, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7115674018859863, + "num_tokens": 297393466.0, + "step": 11495 + }, + { + "epoch": 1.2624643092466505, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2457618713378906, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.72031569480896, + "num_tokens": 297422391.0, + "step": 11496 + }, + { + "epoch": 1.2625741269492643, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.117130994796753, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6955850720405579, + "num_tokens": 297451099.0, + "step": 11497 + }, + { + "epoch": 1.2626839446518778, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.253479242324829, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7063797116279602, + "num_tokens": 297476115.0, + "step": 11498 + }, + { + "epoch": 1.2627937623544916, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.4657862186431885, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.703259289264679, + "num_tokens": 297498409.0, + "step": 11499 + }, + { + "epoch": 1.2629035800571051, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3622963428497314, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7310582399368286, + "num_tokens": 297521611.0, + "step": 11500 + }, + { + "epoch": 1.2630133977597189, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1293373107910156, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7330319285392761, + "num_tokens": 297548205.0, + "step": 11501 + }, + { + "epoch": 1.2631232154623326, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.324923515319824, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.6983230113983154, + "num_tokens": 297573850.0, + "step": 11502 + }, + { + "epoch": 1.2632330331649462, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.7027807235717773, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7241764664649963, + "num_tokens": 297593424.0, + "step": 11503 + }, + { + "epoch": 1.26334285086756, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3146073818206787, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7039974331855774, + "num_tokens": 297618308.0, + "step": 11504 + }, + { + "epoch": 1.2634526685701735, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.360670328140259, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7201792001724243, + "num_tokens": 297642742.0, + "step": 11505 + }, + { + "epoch": 1.2635624862727872, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.300395965576172, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7278199195861816, + "num_tokens": 297667547.0, + "step": 11506 + }, + { + "epoch": 1.2636723039754008, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.225811004638672, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7467769384384155, + "num_tokens": 297691845.0, + "step": 11507 + }, + { + "epoch": 1.2637821216780145, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3098583221435547, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7258152961730957, + "num_tokens": 297716930.0, + "step": 11508 + }, + { + "epoch": 1.263891939380628, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.0314230918884277, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7068209648132324, + "num_tokens": 297747549.0, + "step": 11509 + }, + { + "epoch": 1.2640017570832418, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.711998701095581, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7347545623779297, + "num_tokens": 297766898.0, + "step": 11510 + }, + { + "epoch": 1.2641115747858556, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3422515392303467, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7041276097297668, + "num_tokens": 297791316.0, + "step": 11511 + }, + { + "epoch": 1.264221392488469, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.591885805130005, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.717870831489563, + "num_tokens": 297811952.0, + "step": 11512 + }, + { + "epoch": 1.2643312101910829, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.4498684406280518, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7241283059120178, + "num_tokens": 297833174.0, + "step": 11513 + }, + { + "epoch": 1.2644410278936964, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.585897922515869, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.714942216873169, + "num_tokens": 297853684.0, + "step": 11514 + }, + { + "epoch": 1.2645508455963101, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.2459969520568848, + "learning_rate": 1e-06, + "loss": 0.9616, + "mean_token_accuracy": 0.7014570832252502, + "num_tokens": 297880080.0, + "step": 11515 + }, + { + "epoch": 1.264660663298924, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.0926101207733154, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7006435990333557, + "num_tokens": 297909560.0, + "step": 11516 + }, + { + "epoch": 1.2647704810015374, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.355006456375122, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7159149050712585, + "num_tokens": 297932492.0, + "step": 11517 + }, + { + "epoch": 1.264880298704151, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.094777822494507, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.7012921571731567, + "num_tokens": 297965319.0, + "step": 11518 + }, + { + "epoch": 1.2649901164067647, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.237628936767578, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7012436389923096, + "num_tokens": 297993903.0, + "step": 11519 + }, + { + "epoch": 1.2650999341093785, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.21285343170166, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.7060292363166809, + "num_tokens": 298022800.0, + "step": 11520 + }, + { + "epoch": 1.265209751811992, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1308770179748535, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7322137355804443, + "num_tokens": 298052026.0, + "step": 11521 + }, + { + "epoch": 1.2653195695146058, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.294938087463379, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7292730808258057, + "num_tokens": 298076619.0, + "step": 11522 + }, + { + "epoch": 1.2654293872172193, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3734638690948486, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7243040800094604, + "num_tokens": 298100965.0, + "step": 11523 + }, + { + "epoch": 1.265539204919833, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.379638433456421, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6899665594100952, + "num_tokens": 298124809.0, + "step": 11524 + }, + { + "epoch": 1.2656490226224468, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1112911701202393, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7056701183319092, + "num_tokens": 298155550.0, + "step": 11525 + }, + { + "epoch": 1.2657588403250604, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.148564338684082, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.711225152015686, + "num_tokens": 298181998.0, + "step": 11526 + }, + { + "epoch": 1.2658686580276741, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.003788709640503, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7281479835510254, + "num_tokens": 298211803.0, + "step": 11527 + }, + { + "epoch": 1.2659784757302877, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.5404276847839355, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7069228291511536, + "num_tokens": 298234110.0, + "step": 11528 + }, + { + "epoch": 1.2660882934329014, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 1.9843512773513794, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.6952953934669495, + "num_tokens": 298265890.0, + "step": 11529 + }, + { + "epoch": 1.2661981111355152, + "ewc_loss": 6.705522537231445e-06, + "grad_norm": 2.2633838653564453, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.725775957107544, + "num_tokens": 298290478.0, + "step": 11530 + }, + { + "epoch": 1.2663079288381287, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.1118433475494385, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7314807176589966, + "num_tokens": 298318961.0, + "step": 11531 + }, + { + "epoch": 1.2664177465407422, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3471107482910156, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7342656254768372, + "num_tokens": 298341171.0, + "step": 11532 + }, + { + "epoch": 1.266527564243356, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.3438539505004883, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7278938293457031, + "num_tokens": 298365083.0, + "step": 11533 + }, + { + "epoch": 1.2666373819459698, + "ewc_loss": 6.67572021484375e-06, + "grad_norm": 2.036954402923584, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7109754085540771, + "num_tokens": 298396270.0, + "step": 11534 + }, + { + "epoch": 1.2667471996485833, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3957343101501465, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7122139930725098, + "num_tokens": 298420344.0, + "step": 11535 + }, + { + "epoch": 1.266857017351197, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.9658854007720947, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7167133688926697, + "num_tokens": 298451386.0, + "step": 11536 + }, + { + "epoch": 1.2669668350538106, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4549410343170166, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7084178924560547, + "num_tokens": 298473767.0, + "step": 11537 + }, + { + "epoch": 1.2670766527564243, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.5011587142944336, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7291074991226196, + "num_tokens": 298495330.0, + "step": 11538 + }, + { + "epoch": 1.267186470459038, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1884443759918213, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7050434350967407, + "num_tokens": 298524502.0, + "step": 11539 + }, + { + "epoch": 1.2672962881616516, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1887407302856445, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7142484188079834, + "num_tokens": 298552364.0, + "step": 11540 + }, + { + "epoch": 1.2674061058642654, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.236051559448242, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7172994017601013, + "num_tokens": 298578078.0, + "step": 11541 + }, + { + "epoch": 1.267515923566879, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1169168949127197, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7190952301025391, + "num_tokens": 298605107.0, + "step": 11542 + }, + { + "epoch": 1.2676257412694927, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4633960723876953, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7321439385414124, + "num_tokens": 298625913.0, + "step": 11543 + }, + { + "epoch": 1.2677355589721064, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.083191394805908, + "learning_rate": 1e-06, + "loss": 0.9805, + "mean_token_accuracy": 0.6932212114334106, + "num_tokens": 298656895.0, + "step": 11544 + }, + { + "epoch": 1.26784537667472, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4773943424224854, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.71793133020401, + "num_tokens": 298680044.0, + "step": 11545 + }, + { + "epoch": 1.2679551943773335, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.223045825958252, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7092258930206299, + "num_tokens": 298707118.0, + "step": 11546 + }, + { + "epoch": 1.2680650120799473, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1657192707061768, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.703799843788147, + "num_tokens": 298731089.0, + "step": 11547 + }, + { + "epoch": 1.268174829782561, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1631176471710205, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.739486813545227, + "num_tokens": 298757668.0, + "step": 11548 + }, + { + "epoch": 1.2682846474851746, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1235034465789795, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7301607131958008, + "num_tokens": 298785147.0, + "step": 11549 + }, + { + "epoch": 1.2683944651877883, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1045749187469482, + "learning_rate": 1e-06, + "loss": 0.9525, + "mean_token_accuracy": 0.7099224925041199, + "num_tokens": 298812858.0, + "step": 11550 + }, + { + "epoch": 1.2685042828904018, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2493577003479004, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7300783395767212, + "num_tokens": 298835848.0, + "step": 11551 + }, + { + "epoch": 1.2686141005930156, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4763143062591553, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.714055061340332, + "num_tokens": 298856536.0, + "step": 11552 + }, + { + "epoch": 1.2687239182956294, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.140197515487671, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7304648160934448, + "num_tokens": 298883334.0, + "step": 11553 + }, + { + "epoch": 1.268833735998243, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.6741085052490234, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7458222508430481, + "num_tokens": 298902690.0, + "step": 11554 + }, + { + "epoch": 1.2689435537008567, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.966999888420105, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.709357738494873, + "num_tokens": 298935723.0, + "step": 11555 + }, + { + "epoch": 1.2690533714034702, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.263150215148926, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7164160013198853, + "num_tokens": 298961129.0, + "step": 11556 + }, + { + "epoch": 1.269163189106084, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.081848382949829, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7288569211959839, + "num_tokens": 298989359.0, + "step": 11557 + }, + { + "epoch": 1.2692730068086975, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.135845422744751, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7291325330734253, + "num_tokens": 299018271.0, + "step": 11558 + }, + { + "epoch": 1.2693828245113112, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4351632595062256, + "learning_rate": 1e-06, + "loss": 1.0143, + "mean_token_accuracy": 0.6960980296134949, + "num_tokens": 299041732.0, + "step": 11559 + }, + { + "epoch": 1.2694926422139248, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3057045936584473, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.727679967880249, + "num_tokens": 299066729.0, + "step": 11560 + }, + { + "epoch": 1.2696024599165385, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.433535575866699, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7075092196464539, + "num_tokens": 299088342.0, + "step": 11561 + }, + { + "epoch": 1.2697122776191523, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3554139137268066, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.6987382173538208, + "num_tokens": 299113587.0, + "step": 11562 + }, + { + "epoch": 1.2698220953217658, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.310933828353882, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7102454304695129, + "num_tokens": 299140214.0, + "step": 11563 + }, + { + "epoch": 1.2699319130243796, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.9460909366607666, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7069562077522278, + "num_tokens": 299174699.0, + "step": 11564 + }, + { + "epoch": 1.2700417307269931, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.9418646097183228, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.72955721616745, + "num_tokens": 299206309.0, + "step": 11565 + }, + { + "epoch": 1.2701515484296069, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.5735855102539062, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7231723070144653, + "num_tokens": 299225998.0, + "step": 11566 + }, + { + "epoch": 1.2702613661322206, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.015930652618408, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7157281637191772, + "num_tokens": 299258633.0, + "step": 11567 + }, + { + "epoch": 1.2703711838348342, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.5026285648345947, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7090862989425659, + "num_tokens": 299280740.0, + "step": 11568 + }, + { + "epoch": 1.270481001537448, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4229929447174072, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.716967761516571, + "num_tokens": 299304369.0, + "step": 11569 + }, + { + "epoch": 1.2705908192400615, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.207056999206543, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7011747360229492, + "num_tokens": 299332152.0, + "step": 11570 + }, + { + "epoch": 1.2707006369426752, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1698200702667236, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.6990243196487427, + "num_tokens": 299361670.0, + "step": 11571 + }, + { + "epoch": 1.2708104546452887, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.286893844604492, + "learning_rate": 1e-06, + "loss": 1.0165, + "mean_token_accuracy": 0.699155867099762, + "num_tokens": 299389803.0, + "step": 11572 + }, + { + "epoch": 1.2709202723479025, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.0069005489349365, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7170165181159973, + "num_tokens": 299420332.0, + "step": 11573 + }, + { + "epoch": 1.271030090050516, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3409347534179688, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.6980348229408264, + "num_tokens": 299444266.0, + "step": 11574 + }, + { + "epoch": 1.2711399077531298, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.9579894542694092, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.712942361831665, + "num_tokens": 299476084.0, + "step": 11575 + }, + { + "epoch": 1.2712497254557436, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2750072479248047, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7203894853591919, + "num_tokens": 299501733.0, + "step": 11576 + }, + { + "epoch": 1.271359543158357, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.402095079421997, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7118636965751648, + "num_tokens": 299524569.0, + "step": 11577 + }, + { + "epoch": 1.2714693608609708, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.291499137878418, + "learning_rate": 1e-06, + "loss": 0.7889, + "mean_token_accuracy": 0.745914876461029, + "num_tokens": 299546818.0, + "step": 11578 + }, + { + "epoch": 1.2715791785635844, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4711177349090576, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7190933227539062, + "num_tokens": 299568630.0, + "step": 11579 + }, + { + "epoch": 1.2716889962661981, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2306485176086426, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7060682773590088, + "num_tokens": 299595091.0, + "step": 11580 + }, + { + "epoch": 1.271798813968812, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.0714497566223145, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7217217683792114, + "num_tokens": 299624124.0, + "step": 11581 + }, + { + "epoch": 1.2719086316714254, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.270242929458618, + "learning_rate": 1e-06, + "loss": 0.9574, + "mean_token_accuracy": 0.7001721262931824, + "num_tokens": 299649454.0, + "step": 11582 + }, + { + "epoch": 1.272018449374039, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3635895252227783, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7056692838668823, + "num_tokens": 299675046.0, + "step": 11583 + }, + { + "epoch": 1.2721282670766527, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.169419765472412, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7008552551269531, + "num_tokens": 299704160.0, + "step": 11584 + }, + { + "epoch": 1.2722380847792665, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2805304527282715, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6877636313438416, + "num_tokens": 299730361.0, + "step": 11585 + }, + { + "epoch": 1.27234790248188, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.166421413421631, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7181607484817505, + "num_tokens": 299757251.0, + "step": 11586 + }, + { + "epoch": 1.2724577201844938, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4841361045837402, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.733709454536438, + "num_tokens": 299777812.0, + "step": 11587 + }, + { + "epoch": 1.2725675378871073, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3970894813537598, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7459766864776611, + "num_tokens": 299799763.0, + "step": 11588 + }, + { + "epoch": 1.272677355589721, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.259280204772949, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7110596299171448, + "num_tokens": 299823944.0, + "step": 11589 + }, + { + "epoch": 1.2727871732923348, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4271488189697266, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7374295592308044, + "num_tokens": 299846384.0, + "step": 11590 + }, + { + "epoch": 1.2728969909949484, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.0501346588134766, + "learning_rate": 1e-06, + "loss": 1.0463, + "mean_token_accuracy": 0.6796886920928955, + "num_tokens": 299877965.0, + "step": 11591 + }, + { + "epoch": 1.273006808697562, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.085292100906372, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.729568600654602, + "num_tokens": 299907103.0, + "step": 11592 + }, + { + "epoch": 1.2731166264001756, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.061675548553467, + "learning_rate": 1e-06, + "loss": 0.9868, + "mean_token_accuracy": 0.7059373259544373, + "num_tokens": 299939433.0, + "step": 11593 + }, + { + "epoch": 1.2732264441027894, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1963906288146973, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7439677119255066, + "num_tokens": 299965674.0, + "step": 11594 + }, + { + "epoch": 1.2733362618054032, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.447068929672241, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7093569040298462, + "num_tokens": 299988125.0, + "step": 11595 + }, + { + "epoch": 1.2734460795080167, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.59302020072937, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7409241199493408, + "num_tokens": 300007609.0, + "step": 11596 + }, + { + "epoch": 1.2735558972106302, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 1.9729044437408447, + "learning_rate": 1e-06, + "loss": 1.0771, + "mean_token_accuracy": 0.6756751537322998, + "num_tokens": 300042810.0, + "step": 11597 + }, + { + "epoch": 1.273665714913244, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.4496397972106934, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7102440595626831, + "num_tokens": 300065000.0, + "step": 11598 + }, + { + "epoch": 1.2737755326158577, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.148930311203003, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7267118692398071, + "num_tokens": 300093292.0, + "step": 11599 + }, + { + "epoch": 1.2738853503184713, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.258382558822632, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7311143279075623, + "num_tokens": 300116894.0, + "step": 11600 + }, + { + "epoch": 1.273995168021085, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.564556837081909, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7065340876579285, + "num_tokens": 300138074.0, + "step": 11601 + }, + { + "epoch": 1.2741049857236986, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.0457348823547363, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.727483868598938, + "num_tokens": 300167547.0, + "step": 11602 + }, + { + "epoch": 1.2742148034263123, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.0470259189605713, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7086684703826904, + "num_tokens": 300197864.0, + "step": 11603 + }, + { + "epoch": 1.274324621128926, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2675607204437256, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7060670852661133, + "num_tokens": 300223913.0, + "step": 11604 + }, + { + "epoch": 1.2744344388315396, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.077941656112671, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7232658863067627, + "num_tokens": 300253707.0, + "step": 11605 + }, + { + "epoch": 1.2745442565341534, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.277048110961914, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.73294997215271, + "num_tokens": 300277603.0, + "step": 11606 + }, + { + "epoch": 1.274654074236767, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.051846981048584, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7152783274650574, + "num_tokens": 300308542.0, + "step": 11607 + }, + { + "epoch": 1.2747638919393807, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.7571158409118652, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.720836877822876, + "num_tokens": 300326984.0, + "step": 11608 + }, + { + "epoch": 1.2748737096419944, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.1605865955352783, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.719455361366272, + "num_tokens": 300353479.0, + "step": 11609 + }, + { + "epoch": 1.274983527344608, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.363814353942871, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6977765560150146, + "num_tokens": 300377842.0, + "step": 11610 + }, + { + "epoch": 1.2750933450472215, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.61145281791687, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7190983891487122, + "num_tokens": 300399800.0, + "step": 11611 + }, + { + "epoch": 1.2752031627498353, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.5878663063049316, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.6996715664863586, + "num_tokens": 300420136.0, + "step": 11612 + }, + { + "epoch": 1.275312980452449, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.3397767543792725, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7375123500823975, + "num_tokens": 300443520.0, + "step": 11613 + }, + { + "epoch": 1.2754227981550625, + "ewc_loss": 6.735324859619141e-06, + "grad_norm": 2.2765305042266846, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7328590154647827, + "num_tokens": 300468267.0, + "step": 11614 + }, + { + "epoch": 1.2755326158576763, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.146981954574585, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7110159397125244, + "num_tokens": 300498103.0, + "step": 11615 + }, + { + "epoch": 1.2756424335602898, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.229234457015991, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7309372425079346, + "num_tokens": 300521745.0, + "step": 11616 + }, + { + "epoch": 1.2757522512629036, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.0821542739868164, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.726393461227417, + "num_tokens": 300550409.0, + "step": 11617 + }, + { + "epoch": 1.2758620689655173, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.3680331707000732, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7298874258995056, + "num_tokens": 300573697.0, + "step": 11618 + }, + { + "epoch": 1.2759718866681309, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.4685354232788086, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7228058576583862, + "num_tokens": 300595692.0, + "step": 11619 + }, + { + "epoch": 1.2760817043707446, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.1771278381347656, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7206132411956787, + "num_tokens": 300621855.0, + "step": 11620 + }, + { + "epoch": 1.2761915220733582, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.2254750728607178, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7434384226799011, + "num_tokens": 300648529.0, + "step": 11621 + }, + { + "epoch": 1.276301339775972, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.1018149852752686, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7065188884735107, + "num_tokens": 300680264.0, + "step": 11622 + }, + { + "epoch": 1.2764111574785855, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.480677843093872, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7151444554328918, + "num_tokens": 300703359.0, + "step": 11623 + }, + { + "epoch": 1.2765209751811992, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.990243911743164, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.708877444267273, + "num_tokens": 300722624.0, + "step": 11624 + }, + { + "epoch": 1.2766307928838128, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.1603283882141113, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7236161231994629, + "num_tokens": 300749852.0, + "step": 11625 + }, + { + "epoch": 1.2767406105864265, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.3136157989501953, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7415871620178223, + "num_tokens": 300772727.0, + "step": 11626 + }, + { + "epoch": 1.2768504282890403, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.3356218338012695, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7154074907302856, + "num_tokens": 300796476.0, + "step": 11627 + }, + { + "epoch": 1.2769602459916538, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.4542810916900635, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7143248915672302, + "num_tokens": 300821612.0, + "step": 11628 + }, + { + "epoch": 1.2770700636942676, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.2353737354278564, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7396978139877319, + "num_tokens": 300847114.0, + "step": 11629 + }, + { + "epoch": 1.277179881396881, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.5828802585601807, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7178634405136108, + "num_tokens": 300867258.0, + "step": 11630 + }, + { + "epoch": 1.2772896990994949, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.0999999046325684, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6884207725524902, + "num_tokens": 300895077.0, + "step": 11631 + }, + { + "epoch": 1.2773995168021086, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.1096978187561035, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7052465081214905, + "num_tokens": 300926950.0, + "step": 11632 + }, + { + "epoch": 1.2775093345047221, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.0700466632843018, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7227632403373718, + "num_tokens": 300954676.0, + "step": 11633 + }, + { + "epoch": 1.2776191522073357, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.3396999835968018, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7223783731460571, + "num_tokens": 300979637.0, + "step": 11634 + }, + { + "epoch": 1.2777289699099494, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.081381320953369, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7073596715927124, + "num_tokens": 301008969.0, + "step": 11635 + }, + { + "epoch": 1.2778387876125632, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.1156980991363525, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.7062098979949951, + "num_tokens": 301039515.0, + "step": 11636 + }, + { + "epoch": 1.2779486053151767, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.172795057296753, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.725966215133667, + "num_tokens": 301066854.0, + "step": 11637 + }, + { + "epoch": 1.2780584230177905, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 2.0443732738494873, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7030978798866272, + "num_tokens": 301097544.0, + "step": 11638 + }, + { + "epoch": 1.278168240720404, + "ewc_loss": 6.794929504394531e-06, + "grad_norm": 1.935025691986084, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7185478210449219, + "num_tokens": 301128404.0, + "step": 11639 + }, + { + "epoch": 1.2782780584230178, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.2824783325195312, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7153024673461914, + "num_tokens": 301154182.0, + "step": 11640 + }, + { + "epoch": 1.2783878761256315, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.5415608882904053, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7299692630767822, + "num_tokens": 301174498.0, + "step": 11641 + }, + { + "epoch": 1.278497693828245, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.351102828979492, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7216343283653259, + "num_tokens": 301198593.0, + "step": 11642 + }, + { + "epoch": 1.2786075115308588, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.403123617172241, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7351276874542236, + "num_tokens": 301219877.0, + "step": 11643 + }, + { + "epoch": 1.2787173292334724, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.706829309463501, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7435035109519958, + "num_tokens": 301237174.0, + "step": 11644 + }, + { + "epoch": 1.2788271469360861, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.2091526985168457, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.726403534412384, + "num_tokens": 301264961.0, + "step": 11645 + }, + { + "epoch": 1.2789369646386999, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.2984089851379395, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7233814001083374, + "num_tokens": 301288854.0, + "step": 11646 + }, + { + "epoch": 1.2790467823413134, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.114917278289795, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.722807765007019, + "num_tokens": 301317681.0, + "step": 11647 + }, + { + "epoch": 1.279156600043927, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.192701578140259, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7321999669075012, + "num_tokens": 301342976.0, + "step": 11648 + }, + { + "epoch": 1.2792664177465407, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.082918882369995, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7144759893417358, + "num_tokens": 301371968.0, + "step": 11649 + }, + { + "epoch": 1.2793762354491545, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.1068263053894043, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.711262047290802, + "num_tokens": 301399863.0, + "step": 11650 + }, + { + "epoch": 1.279486053151768, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.526376485824585, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7150593996047974, + "num_tokens": 301421566.0, + "step": 11651 + }, + { + "epoch": 1.2795958708543818, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.197748899459839, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7306267619132996, + "num_tokens": 301448732.0, + "step": 11652 + }, + { + "epoch": 1.2797056885569953, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.325711727142334, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.6954354643821716, + "num_tokens": 301472964.0, + "step": 11653 + }, + { + "epoch": 1.279815506259609, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.2793140411376953, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7202610969543457, + "num_tokens": 301499186.0, + "step": 11654 + }, + { + "epoch": 1.2799253239622228, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.0657052993774414, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7185429334640503, + "num_tokens": 301526014.0, + "step": 11655 + }, + { + "epoch": 1.2800351416648363, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.257086992263794, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.708153247833252, + "num_tokens": 301551649.0, + "step": 11656 + }, + { + "epoch": 1.28014495936745, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.166842460632324, + "learning_rate": 1e-06, + "loss": 1.0153, + "mean_token_accuracy": 0.6947397589683533, + "num_tokens": 301580729.0, + "step": 11657 + }, + { + "epoch": 1.2802547770700636, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.011420249938965, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.6976947784423828, + "num_tokens": 301610803.0, + "step": 11658 + }, + { + "epoch": 1.2803645947726774, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.327256202697754, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7259116172790527, + "num_tokens": 301634563.0, + "step": 11659 + }, + { + "epoch": 1.2804744124752911, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.4427874088287354, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7340549230575562, + "num_tokens": 301655487.0, + "step": 11660 + }, + { + "epoch": 1.2805842301779047, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 1.859884262084961, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7023299336433411, + "num_tokens": 301692870.0, + "step": 11661 + }, + { + "epoch": 1.2806940478805182, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.309643507003784, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7034960985183716, + "num_tokens": 301718774.0, + "step": 11662 + }, + { + "epoch": 1.280803865583132, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.0143423080444336, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7407887578010559, + "num_tokens": 301745583.0, + "step": 11663 + }, + { + "epoch": 1.2809136832857457, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.71280837059021, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7200922966003418, + "num_tokens": 301766400.0, + "step": 11664 + }, + { + "epoch": 1.2810235009883593, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.3489267826080322, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7351032495498657, + "num_tokens": 301788337.0, + "step": 11665 + }, + { + "epoch": 1.281133318690973, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.3216264247894287, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7107906341552734, + "num_tokens": 301812213.0, + "step": 11666 + }, + { + "epoch": 1.2812431363935866, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.42862868309021, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7410907745361328, + "num_tokens": 301834284.0, + "step": 11667 + }, + { + "epoch": 1.2813529540962003, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.0044894218444824, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.70953768491745, + "num_tokens": 301866714.0, + "step": 11668 + }, + { + "epoch": 1.281462771798814, + "ewc_loss": 6.854534149169922e-06, + "grad_norm": 2.1529343128204346, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7187780141830444, + "num_tokens": 301895602.0, + "step": 11669 + }, + { + "epoch": 1.2815725895014276, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.290799140930176, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7103273868560791, + "num_tokens": 301922347.0, + "step": 11670 + }, + { + "epoch": 1.2816824072040414, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.4767866134643555, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7246953248977661, + "num_tokens": 301948200.0, + "step": 11671 + }, + { + "epoch": 1.281792224906655, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.3900551795959473, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7171481847763062, + "num_tokens": 301972825.0, + "step": 11672 + }, + { + "epoch": 1.2819020426092687, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.6360373497009277, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7288323640823364, + "num_tokens": 301992609.0, + "step": 11673 + }, + { + "epoch": 1.2820118603118824, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.1137990951538086, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7063932418823242, + "num_tokens": 302019194.0, + "step": 11674 + }, + { + "epoch": 1.282121678014496, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.328278064727783, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7063837647438049, + "num_tokens": 302045526.0, + "step": 11675 + }, + { + "epoch": 1.2822314957171095, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.343308687210083, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7084277868270874, + "num_tokens": 302071242.0, + "step": 11676 + }, + { + "epoch": 1.2823413134197232, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.2763659954071045, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7057453393936157, + "num_tokens": 302099235.0, + "step": 11677 + }, + { + "epoch": 1.282451131122337, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.2941126823425293, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7164332866668701, + "num_tokens": 302125416.0, + "step": 11678 + }, + { + "epoch": 1.2825609488249505, + "ewc_loss": 6.884336471557617e-06, + "grad_norm": 2.0577595233917236, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7146516442298889, + "num_tokens": 302155002.0, + "step": 11679 + }, + { + "epoch": 1.2826707665275643, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3798460960388184, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7171577215194702, + "num_tokens": 302177355.0, + "step": 11680 + }, + { + "epoch": 1.2827805842301778, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.339583396911621, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7286458015441895, + "num_tokens": 302199821.0, + "step": 11681 + }, + { + "epoch": 1.2828904019327916, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1650447845458984, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7115038633346558, + "num_tokens": 302228048.0, + "step": 11682 + }, + { + "epoch": 1.2830002196354053, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3117430210113525, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7330535650253296, + "num_tokens": 302250998.0, + "step": 11683 + }, + { + "epoch": 1.2831100373380189, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.309575080871582, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7345090508460999, + "num_tokens": 302275876.0, + "step": 11684 + }, + { + "epoch": 1.2832198550406326, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1236398220062256, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7255114316940308, + "num_tokens": 302303707.0, + "step": 11685 + }, + { + "epoch": 1.2833296727432462, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1306843757629395, + "learning_rate": 1e-06, + "loss": 0.792, + "mean_token_accuracy": 0.7425950169563293, + "num_tokens": 302329954.0, + "step": 11686 + }, + { + "epoch": 1.28343949044586, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.552449941635132, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7254494428634644, + "num_tokens": 302352435.0, + "step": 11687 + }, + { + "epoch": 1.2835493081484735, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.986630916595459, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7250610589981079, + "num_tokens": 302383080.0, + "step": 11688 + }, + { + "epoch": 1.2836591258510872, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.279849052429199, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.6990648508071899, + "num_tokens": 302407828.0, + "step": 11689 + }, + { + "epoch": 1.2837689435537007, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.268724203109741, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7170125246047974, + "num_tokens": 302434324.0, + "step": 11690 + }, + { + "epoch": 1.2838787612563145, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.9553273916244507, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7171475887298584, + "num_tokens": 302466784.0, + "step": 11691 + }, + { + "epoch": 1.2839885789589283, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3672707080841064, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7387083768844604, + "num_tokens": 302490422.0, + "step": 11692 + }, + { + "epoch": 1.2840983966615418, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1412861347198486, + "learning_rate": 1e-06, + "loss": 1.0406, + "mean_token_accuracy": 0.6928070187568665, + "num_tokens": 302521864.0, + "step": 11693 + }, + { + "epoch": 1.2842082143641556, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1254611015319824, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7171273827552795, + "num_tokens": 302550059.0, + "step": 11694 + }, + { + "epoch": 1.284318032066769, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3627309799194336, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7236507534980774, + "num_tokens": 302572257.0, + "step": 11695 + }, + { + "epoch": 1.2844278497693828, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.246250629425049, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7201292514801025, + "num_tokens": 302597184.0, + "step": 11696 + }, + { + "epoch": 1.2845376674719966, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1410765647888184, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7235590815544128, + "num_tokens": 302625370.0, + "step": 11697 + }, + { + "epoch": 1.2846474851746101, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.5100302696228027, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7139061093330383, + "num_tokens": 302647698.0, + "step": 11698 + }, + { + "epoch": 1.2847573028772237, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1935060024261475, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.723156213760376, + "num_tokens": 302675217.0, + "step": 11699 + }, + { + "epoch": 1.2848671205798374, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.239952564239502, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7127562761306763, + "num_tokens": 302701118.0, + "step": 11700 + }, + { + "epoch": 1.2849769382824512, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.6658947467803955, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7072614431381226, + "num_tokens": 302721586.0, + "step": 11701 + }, + { + "epoch": 1.2850867559850647, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.165861129760742, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7202714085578918, + "num_tokens": 302749143.0, + "step": 11702 + }, + { + "epoch": 1.2851965736876785, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.380458116531372, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7214226126670837, + "num_tokens": 302771892.0, + "step": 11703 + }, + { + "epoch": 1.285306391390292, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2530834674835205, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7188841104507446, + "num_tokens": 302795399.0, + "step": 11704 + }, + { + "epoch": 1.2854162090929058, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.242368459701538, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7288819551467896, + "num_tokens": 302820330.0, + "step": 11705 + }, + { + "epoch": 1.2855260267955195, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.315035104751587, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7019323706626892, + "num_tokens": 302844470.0, + "step": 11706 + }, + { + "epoch": 1.285635844498133, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.098287582397461, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6924892067909241, + "num_tokens": 302875201.0, + "step": 11707 + }, + { + "epoch": 1.2857456622007468, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2558352947235107, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7292779684066772, + "num_tokens": 302900116.0, + "step": 11708 + }, + { + "epoch": 1.2858554799033604, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.22830867767334, + "learning_rate": 1e-06, + "loss": 0.9769, + "mean_token_accuracy": 0.7009174823760986, + "num_tokens": 302926066.0, + "step": 11709 + }, + { + "epoch": 1.2859652976059741, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3273956775665283, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7183840274810791, + "num_tokens": 302951977.0, + "step": 11710 + }, + { + "epoch": 1.2860751153085879, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3470077514648438, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7148077487945557, + "num_tokens": 302976941.0, + "step": 11711 + }, + { + "epoch": 1.2861849330112014, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.4855289459228516, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7214958667755127, + "num_tokens": 302999808.0, + "step": 11712 + }, + { + "epoch": 1.286294750713815, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.9795888662338257, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7465128302574158, + "num_tokens": 303030124.0, + "step": 11713 + }, + { + "epoch": 1.2864045684164287, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.9843908548355103, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7303369045257568, + "num_tokens": 303062520.0, + "step": 11714 + }, + { + "epoch": 1.2865143861190425, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.343371868133545, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7175882458686829, + "num_tokens": 303087211.0, + "step": 11715 + }, + { + "epoch": 1.286624203821656, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1283230781555176, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7227777242660522, + "num_tokens": 303114010.0, + "step": 11716 + }, + { + "epoch": 1.2867340215242697, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.8443174362182617, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7388162016868591, + "num_tokens": 303132223.0, + "step": 11717 + }, + { + "epoch": 1.2868438392268833, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.401883125305176, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.727495014667511, + "num_tokens": 303154590.0, + "step": 11718 + }, + { + "epoch": 1.286953656929497, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1085588932037354, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.709848165512085, + "num_tokens": 303181959.0, + "step": 11719 + }, + { + "epoch": 1.2870634746321108, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3863868713378906, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7391170859336853, + "num_tokens": 303206582.0, + "step": 11720 + }, + { + "epoch": 1.2871732923347243, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.35872220993042, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7098770141601562, + "num_tokens": 303230746.0, + "step": 11721 + }, + { + "epoch": 1.287283110037338, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.131812810897827, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7467011213302612, + "num_tokens": 303256952.0, + "step": 11722 + }, + { + "epoch": 1.2873929277399516, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.6382839679718018, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7334806323051453, + "num_tokens": 303276667.0, + "step": 11723 + }, + { + "epoch": 1.2875027454425654, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.108288526535034, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7362715005874634, + "num_tokens": 303302982.0, + "step": 11724 + }, + { + "epoch": 1.2876125631451791, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3786730766296387, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.708789587020874, + "num_tokens": 303330334.0, + "step": 11725 + }, + { + "epoch": 1.2877223808477927, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2303359508514404, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6950374841690063, + "num_tokens": 303358251.0, + "step": 11726 + }, + { + "epoch": 1.2878321985504062, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.217334270477295, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7441819906234741, + "num_tokens": 303381982.0, + "step": 11727 + }, + { + "epoch": 1.28794201625302, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.5843310356140137, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.734996497631073, + "num_tokens": 303401346.0, + "step": 11728 + }, + { + "epoch": 1.2880518339556337, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1740407943725586, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7107190489768982, + "num_tokens": 303426493.0, + "step": 11729 + }, + { + "epoch": 1.2881616516582473, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0848236083984375, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6824115514755249, + "num_tokens": 303458971.0, + "step": 11730 + }, + { + "epoch": 1.288271469360861, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.37211537361145, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7442636489868164, + "num_tokens": 303482654.0, + "step": 11731 + }, + { + "epoch": 1.2883812870634745, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.933476448059082, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7246959209442139, + "num_tokens": 303515731.0, + "step": 11732 + }, + { + "epoch": 1.2884911047660883, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1968326568603516, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7369787693023682, + "num_tokens": 303540606.0, + "step": 11733 + }, + { + "epoch": 1.288600922468702, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.385838508605957, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7249126434326172, + "num_tokens": 303562877.0, + "step": 11734 + }, + { + "epoch": 1.2887107401713156, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.152550458908081, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7035225629806519, + "num_tokens": 303590329.0, + "step": 11735 + }, + { + "epoch": 1.2888205578739294, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.9612795114517212, + "learning_rate": 1e-06, + "loss": 1.0496, + "mean_token_accuracy": 0.6780268549919128, + "num_tokens": 303622275.0, + "step": 11736 + }, + { + "epoch": 1.2889303755765429, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0194284915924072, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7039895057678223, + "num_tokens": 303652835.0, + "step": 11737 + }, + { + "epoch": 1.2890401932791566, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3011281490325928, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7213001251220703, + "num_tokens": 303677543.0, + "step": 11738 + }, + { + "epoch": 1.2891500109817704, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.165771484375, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7155404090881348, + "num_tokens": 303705579.0, + "step": 11739 + }, + { + "epoch": 1.289259828684384, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.310375452041626, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7332751154899597, + "num_tokens": 303727515.0, + "step": 11740 + }, + { + "epoch": 1.2893696463869975, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1002981662750244, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7177308797836304, + "num_tokens": 303757209.0, + "step": 11741 + }, + { + "epoch": 1.2894794640896112, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1649813652038574, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7010778188705444, + "num_tokens": 303785195.0, + "step": 11742 + }, + { + "epoch": 1.289589281792225, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.589205503463745, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7164410352706909, + "num_tokens": 303804918.0, + "step": 11743 + }, + { + "epoch": 1.2896990994948385, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2556710243225098, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.710354208946228, + "num_tokens": 303829718.0, + "step": 11744 + }, + { + "epoch": 1.2898089171974523, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.139620542526245, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7105098962783813, + "num_tokens": 303858466.0, + "step": 11745 + }, + { + "epoch": 1.2899187349000658, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.374781847000122, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7470105886459351, + "num_tokens": 303881926.0, + "step": 11746 + }, + { + "epoch": 1.2900285526026796, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1490108966827393, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7298694849014282, + "num_tokens": 303908711.0, + "step": 11747 + }, + { + "epoch": 1.2901383703052933, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.207031726837158, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7079467177391052, + "num_tokens": 303934977.0, + "step": 11748 + }, + { + "epoch": 1.2902481880079069, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0678608417510986, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7249486446380615, + "num_tokens": 303962166.0, + "step": 11749 + }, + { + "epoch": 1.2903580057105206, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0282163619995117, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7131167650222778, + "num_tokens": 303993221.0, + "step": 11750 + }, + { + "epoch": 1.2904678234131342, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.408172130584717, + "learning_rate": 1e-06, + "loss": 1.0159, + "mean_token_accuracy": 0.6975914239883423, + "num_tokens": 304018005.0, + "step": 11751 + }, + { + "epoch": 1.290577641115748, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.307931661605835, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7252408266067505, + "num_tokens": 304042860.0, + "step": 11752 + }, + { + "epoch": 1.2906874588183614, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0712296962738037, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7110563516616821, + "num_tokens": 304072245.0, + "step": 11753 + }, + { + "epoch": 1.2907972765209752, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2717363834381104, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7198377251625061, + "num_tokens": 304096862.0, + "step": 11754 + }, + { + "epoch": 1.2909070942235887, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.145315170288086, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7195225954055786, + "num_tokens": 304124158.0, + "step": 11755 + }, + { + "epoch": 1.2910169119262025, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2827439308166504, + "learning_rate": 1e-06, + "loss": 1.0046, + "mean_token_accuracy": 0.694083571434021, + "num_tokens": 304152468.0, + "step": 11756 + }, + { + "epoch": 1.2911267296288162, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.395235300064087, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7197223901748657, + "num_tokens": 304175812.0, + "step": 11757 + }, + { + "epoch": 1.2912365473314298, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.332357168197632, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7364988923072815, + "num_tokens": 304200028.0, + "step": 11758 + }, + { + "epoch": 1.2913463650340435, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3741304874420166, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7329323291778564, + "num_tokens": 304224531.0, + "step": 11759 + }, + { + "epoch": 1.291456182736657, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.9724397659301758, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6938583850860596, + "num_tokens": 304256347.0, + "step": 11760 + }, + { + "epoch": 1.2915660004392708, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2565951347351074, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7039690017700195, + "num_tokens": 304281763.0, + "step": 11761 + }, + { + "epoch": 1.2916758181418846, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.4390945434570312, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7558053731918335, + "num_tokens": 304304235.0, + "step": 11762 + }, + { + "epoch": 1.2917856358444981, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.305525302886963, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7256574034690857, + "num_tokens": 304329706.0, + "step": 11763 + }, + { + "epoch": 1.2918954535471117, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0888912677764893, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7340273857116699, + "num_tokens": 304360223.0, + "step": 11764 + }, + { + "epoch": 1.2920052712497254, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.340841054916382, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7297885417938232, + "num_tokens": 304382198.0, + "step": 11765 + }, + { + "epoch": 1.2921150889523392, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 1.936015248298645, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.731863260269165, + "num_tokens": 304415819.0, + "step": 11766 + }, + { + "epoch": 1.2922249066549527, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.211230516433716, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7089606523513794, + "num_tokens": 304443269.0, + "step": 11767 + }, + { + "epoch": 1.2923347243575665, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.376373052597046, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7030688524246216, + "num_tokens": 304467737.0, + "step": 11768 + }, + { + "epoch": 1.29244454206018, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.0275261402130127, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7064494490623474, + "num_tokens": 304498891.0, + "step": 11769 + }, + { + "epoch": 1.2925543597627938, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.5582969188690186, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7195174694061279, + "num_tokens": 304521902.0, + "step": 11770 + }, + { + "epoch": 1.2926641774654075, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.3444511890411377, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7203044295310974, + "num_tokens": 304546070.0, + "step": 11771 + }, + { + "epoch": 1.292773995168021, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.5047783851623535, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.716626763343811, + "num_tokens": 304567669.0, + "step": 11772 + }, + { + "epoch": 1.2928838128706348, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.127892255783081, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7175010442733765, + "num_tokens": 304594245.0, + "step": 11773 + }, + { + "epoch": 1.2929936305732483, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.1802098751068115, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7286738157272339, + "num_tokens": 304622361.0, + "step": 11774 + }, + { + "epoch": 1.293103448275862, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.6592793464660645, + "learning_rate": 1e-06, + "loss": 0.7564, + "mean_token_accuracy": 0.7585793733596802, + "num_tokens": 304641523.0, + "step": 11775 + }, + { + "epoch": 1.2932132659784759, + "ewc_loss": 6.9141387939453125e-06, + "grad_norm": 2.2057313919067383, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7300505638122559, + "num_tokens": 304666763.0, + "step": 11776 + }, + { + "epoch": 1.2933230836810894, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.251466989517212, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7120264768600464, + "num_tokens": 304691674.0, + "step": 11777 + }, + { + "epoch": 1.293432901383703, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.481882095336914, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7251865267753601, + "num_tokens": 304713565.0, + "step": 11778 + }, + { + "epoch": 1.2935427190863167, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2418136596679688, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7117865681648254, + "num_tokens": 304738147.0, + "step": 11779 + }, + { + "epoch": 1.2936525367889304, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.33223295211792, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7049861550331116, + "num_tokens": 304763544.0, + "step": 11780 + }, + { + "epoch": 1.293762354491544, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.140963315963745, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6996176838874817, + "num_tokens": 304793019.0, + "step": 11781 + }, + { + "epoch": 1.2938721721941577, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.0965576171875, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.707460880279541, + "num_tokens": 304821249.0, + "step": 11782 + }, + { + "epoch": 1.2939819898967713, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.133265256881714, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6994014978408813, + "num_tokens": 304849658.0, + "step": 11783 + }, + { + "epoch": 1.294091807599385, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.3953933715820312, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.722905158996582, + "num_tokens": 304872325.0, + "step": 11784 + }, + { + "epoch": 1.2942016253019988, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2404067516326904, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7456820011138916, + "num_tokens": 304896426.0, + "step": 11785 + }, + { + "epoch": 1.2943114430046123, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.391263246536255, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7032224535942078, + "num_tokens": 304920851.0, + "step": 11786 + }, + { + "epoch": 1.294421260707226, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.3520848751068115, + "learning_rate": 1e-06, + "loss": 0.7577, + "mean_token_accuracy": 0.7659602761268616, + "num_tokens": 304943334.0, + "step": 11787 + }, + { + "epoch": 1.2945310784098396, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.278533935546875, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.6989326477050781, + "num_tokens": 304969500.0, + "step": 11788 + }, + { + "epoch": 1.2946408961124534, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.408656597137451, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.723626434803009, + "num_tokens": 304991110.0, + "step": 11789 + }, + { + "epoch": 1.2947507138150671, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.515254497528076, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7185613512992859, + "num_tokens": 305015955.0, + "step": 11790 + }, + { + "epoch": 1.2948605315176807, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.208717107772827, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7163946628570557, + "num_tokens": 305042520.0, + "step": 11791 + }, + { + "epoch": 1.2949703492202942, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2264888286590576, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7082054018974304, + "num_tokens": 305071471.0, + "step": 11792 + }, + { + "epoch": 1.295080166922908, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2989912033081055, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.7492780685424805, + "num_tokens": 305095514.0, + "step": 11793 + }, + { + "epoch": 1.2951899846255217, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4106028079986572, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7067043781280518, + "num_tokens": 305117920.0, + "step": 11794 + }, + { + "epoch": 1.2952998023281352, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4497296810150146, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7461999654769897, + "num_tokens": 305139672.0, + "step": 11795 + }, + { + "epoch": 1.295409620030749, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.255089044570923, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7290534377098083, + "num_tokens": 305164213.0, + "step": 11796 + }, + { + "epoch": 1.2955194377333625, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.359614133834839, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7121936082839966, + "num_tokens": 305188419.0, + "step": 11797 + }, + { + "epoch": 1.2956292554359763, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.131856918334961, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7084704637527466, + "num_tokens": 305217696.0, + "step": 11798 + }, + { + "epoch": 1.29573907313859, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.219724178314209, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7210907936096191, + "num_tokens": 305244888.0, + "step": 11799 + }, + { + "epoch": 1.2958488908412036, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.103564739227295, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7496421933174133, + "num_tokens": 305271407.0, + "step": 11800 + }, + { + "epoch": 1.2959587085438173, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.453200340270996, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.720228374004364, + "num_tokens": 305293901.0, + "step": 11801 + }, + { + "epoch": 1.2960685262464309, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2531888484954834, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7167056798934937, + "num_tokens": 305319982.0, + "step": 11802 + }, + { + "epoch": 1.2961783439490446, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.405461072921753, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.72278892993927, + "num_tokens": 305341512.0, + "step": 11803 + }, + { + "epoch": 1.2962881616516582, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1613476276397705, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7044055461883545, + "num_tokens": 305369414.0, + "step": 11804 + }, + { + "epoch": 1.296397979354272, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4567177295684814, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7370926141738892, + "num_tokens": 305392887.0, + "step": 11805 + }, + { + "epoch": 1.2965077970568855, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 1.989123821258545, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.6933574676513672, + "num_tokens": 305423865.0, + "step": 11806 + }, + { + "epoch": 1.2966176147594992, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.019681453704834, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7051993608474731, + "num_tokens": 305454469.0, + "step": 11807 + }, + { + "epoch": 1.296727432462113, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.082916259765625, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7182896137237549, + "num_tokens": 305483821.0, + "step": 11808 + }, + { + "epoch": 1.2968372501647265, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4196085929870605, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7336602210998535, + "num_tokens": 305506685.0, + "step": 11809 + }, + { + "epoch": 1.2969470678673403, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.357203722000122, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7341772317886353, + "num_tokens": 305530489.0, + "step": 11810 + }, + { + "epoch": 1.2970568855699538, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.374763250350952, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7327772974967957, + "num_tokens": 305552683.0, + "step": 11811 + }, + { + "epoch": 1.2971667032725676, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.298048734664917, + "learning_rate": 1e-06, + "loss": 0.725, + "mean_token_accuracy": 0.7653443813323975, + "num_tokens": 305575005.0, + "step": 11812 + }, + { + "epoch": 1.2972765209751813, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2869441509246826, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7352365851402283, + "num_tokens": 305601169.0, + "step": 11813 + }, + { + "epoch": 1.2973863386777948, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.263679265975952, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7324075698852539, + "num_tokens": 305628213.0, + "step": 11814 + }, + { + "epoch": 1.2974961563804086, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1619465351104736, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7129952907562256, + "num_tokens": 305655998.0, + "step": 11815 + }, + { + "epoch": 1.2976059740830221, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2305009365081787, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7065954208374023, + "num_tokens": 305682178.0, + "step": 11816 + }, + { + "epoch": 1.297715791785636, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.435643434524536, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7088285684585571, + "num_tokens": 305706363.0, + "step": 11817 + }, + { + "epoch": 1.2978256094882494, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 1.902342677116394, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7248905897140503, + "num_tokens": 305736478.0, + "step": 11818 + }, + { + "epoch": 1.2979354271908632, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1488044261932373, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7128229737281799, + "num_tokens": 305765312.0, + "step": 11819 + }, + { + "epoch": 1.2980452448934767, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.297609806060791, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.738448977470398, + "num_tokens": 305788731.0, + "step": 11820 + }, + { + "epoch": 1.2981550625960905, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.238849639892578, + "learning_rate": 1e-06, + "loss": 1.0166, + "mean_token_accuracy": 0.6939057111740112, + "num_tokens": 305814487.0, + "step": 11821 + }, + { + "epoch": 1.2982648802987042, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.248028516769409, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7380905151367188, + "num_tokens": 305840110.0, + "step": 11822 + }, + { + "epoch": 1.2983746980013178, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.042956590652466, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7232837677001953, + "num_tokens": 305870635.0, + "step": 11823 + }, + { + "epoch": 1.2984845157039315, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1079680919647217, + "learning_rate": 1e-06, + "loss": 0.9907, + "mean_token_accuracy": 0.7056642174720764, + "num_tokens": 305900461.0, + "step": 11824 + }, + { + "epoch": 1.298594333406545, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4810614585876465, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7234818935394287, + "num_tokens": 305920842.0, + "step": 11825 + }, + { + "epoch": 1.2987041511091588, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2408218383789062, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.7042835354804993, + "num_tokens": 305947118.0, + "step": 11826 + }, + { + "epoch": 1.2988139688117726, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.0893266201019287, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7242729067802429, + "num_tokens": 305973592.0, + "step": 11827 + }, + { + "epoch": 1.2989237865143861, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.2514820098876953, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7193645238876343, + "num_tokens": 306000339.0, + "step": 11828 + }, + { + "epoch": 1.2990336042169996, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 3.025223970413208, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7139856219291687, + "num_tokens": 306018889.0, + "step": 11829 + }, + { + "epoch": 1.2991434219196134, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4847793579101562, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7231836915016174, + "num_tokens": 306041685.0, + "step": 11830 + }, + { + "epoch": 1.2992532396222272, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.132972002029419, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7103025913238525, + "num_tokens": 306072866.0, + "step": 11831 + }, + { + "epoch": 1.2993630573248407, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1949870586395264, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7172776460647583, + "num_tokens": 306100711.0, + "step": 11832 + }, + { + "epoch": 1.2994728750274545, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.5384087562561035, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7006964087486267, + "num_tokens": 306122537.0, + "step": 11833 + }, + { + "epoch": 1.299582692730068, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.3820321559906006, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7235312461853027, + "num_tokens": 306145470.0, + "step": 11834 + }, + { + "epoch": 1.2996925104326817, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.268090009689331, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7099286317825317, + "num_tokens": 306171805.0, + "step": 11835 + }, + { + "epoch": 1.2998023281352955, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1756041049957275, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7327550649642944, + "num_tokens": 306196865.0, + "step": 11836 + }, + { + "epoch": 1.299912145837909, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.0936851501464844, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7205747365951538, + "num_tokens": 306224739.0, + "step": 11837 + }, + { + "epoch": 1.3000219635405228, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.383236885070801, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7030614614486694, + "num_tokens": 306248787.0, + "step": 11838 + }, + { + "epoch": 1.3001317812431363, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.1882991790771484, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7117215394973755, + "num_tokens": 306274373.0, + "step": 11839 + }, + { + "epoch": 1.30024159894575, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.171865463256836, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.718409538269043, + "num_tokens": 306299504.0, + "step": 11840 + }, + { + "epoch": 1.3003514166483638, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.030912160873413, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7073497772216797, + "num_tokens": 306328299.0, + "step": 11841 + }, + { + "epoch": 1.3004612343509774, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4573044776916504, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7355883121490479, + "num_tokens": 306351122.0, + "step": 11842 + }, + { + "epoch": 1.300571052053591, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.4733529090881348, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7343026399612427, + "num_tokens": 306371039.0, + "step": 11843 + }, + { + "epoch": 1.3006808697562047, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.177642583847046, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7150699496269226, + "num_tokens": 306398171.0, + "step": 11844 + }, + { + "epoch": 1.3007906874588184, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.111549139022827, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7179712057113647, + "num_tokens": 306426726.0, + "step": 11845 + }, + { + "epoch": 1.300900505161432, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.0996804237365723, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.6969321966171265, + "num_tokens": 306459056.0, + "step": 11846 + }, + { + "epoch": 1.3010103228640457, + "ewc_loss": 6.943941116333008e-06, + "grad_norm": 2.051314353942871, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7040998339653015, + "num_tokens": 306488203.0, + "step": 11847 + }, + { + "epoch": 1.3011201405666593, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2374649047851562, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7383934259414673, + "num_tokens": 306514586.0, + "step": 11848 + }, + { + "epoch": 1.301229958269273, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.3179819583892822, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7335273027420044, + "num_tokens": 306541307.0, + "step": 11849 + }, + { + "epoch": 1.3013397759718868, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1739602088928223, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7272635698318481, + "num_tokens": 306568190.0, + "step": 11850 + }, + { + "epoch": 1.3014495936745003, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.0492773056030273, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7229658961296082, + "num_tokens": 306598005.0, + "step": 11851 + }, + { + "epoch": 1.301559411377114, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1380176544189453, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7248846888542175, + "num_tokens": 306626348.0, + "step": 11852 + }, + { + "epoch": 1.3016692290797276, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.5465052127838135, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7145667672157288, + "num_tokens": 306648604.0, + "step": 11853 + }, + { + "epoch": 1.3017790467823414, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.13877272605896, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7076331973075867, + "num_tokens": 306677033.0, + "step": 11854 + }, + { + "epoch": 1.301888864484955, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.115891218185425, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7294461131095886, + "num_tokens": 306704591.0, + "step": 11855 + }, + { + "epoch": 1.3019986821875686, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.563767194747925, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7329162359237671, + "num_tokens": 306722762.0, + "step": 11856 + }, + { + "epoch": 1.3021084998901822, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 1.9730594158172607, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7044359445571899, + "num_tokens": 306755451.0, + "step": 11857 + }, + { + "epoch": 1.302218317592796, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4239237308502197, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7478110194206238, + "num_tokens": 306777213.0, + "step": 11858 + }, + { + "epoch": 1.3023281352954097, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 1.9468361139297485, + "learning_rate": 1e-06, + "loss": 1.0298, + "mean_token_accuracy": 0.6908178329467773, + "num_tokens": 306811483.0, + "step": 11859 + }, + { + "epoch": 1.3024379529980232, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2825331687927246, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7174035310745239, + "num_tokens": 306835767.0, + "step": 11860 + }, + { + "epoch": 1.302547770700637, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.342128038406372, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7338070273399353, + "num_tokens": 306858057.0, + "step": 11861 + }, + { + "epoch": 1.3026575884032505, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4402287006378174, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7381365299224854, + "num_tokens": 306881002.0, + "step": 11862 + }, + { + "epoch": 1.3027674061058643, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2658092975616455, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.714158296585083, + "num_tokens": 306907122.0, + "step": 11863 + }, + { + "epoch": 1.302877223808478, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.300140619277954, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7232879400253296, + "num_tokens": 306931095.0, + "step": 11864 + }, + { + "epoch": 1.3029870415110916, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.3782050609588623, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7286038994789124, + "num_tokens": 306953333.0, + "step": 11865 + }, + { + "epoch": 1.3030968592137053, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.3635094165802, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7132017612457275, + "num_tokens": 306979310.0, + "step": 11866 + }, + { + "epoch": 1.3032066769163189, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.283940553665161, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7260509729385376, + "num_tokens": 307005402.0, + "step": 11867 + }, + { + "epoch": 1.3033164946189326, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.295915126800537, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7177940607070923, + "num_tokens": 307031279.0, + "step": 11868 + }, + { + "epoch": 1.3034263123215462, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4602577686309814, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7254663109779358, + "num_tokens": 307054553.0, + "step": 11869 + }, + { + "epoch": 1.30353613002416, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1803088188171387, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7093846201896667, + "num_tokens": 307081607.0, + "step": 11870 + }, + { + "epoch": 1.3036459477267734, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.14501953125, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.71755051612854, + "num_tokens": 307108952.0, + "step": 11871 + }, + { + "epoch": 1.3037557654293872, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.5761165618896484, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7106470465660095, + "num_tokens": 307129054.0, + "step": 11872 + }, + { + "epoch": 1.303865583132001, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.0604970455169678, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7443090677261353, + "num_tokens": 307155838.0, + "step": 11873 + }, + { + "epoch": 1.3039754008346145, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.143861770629883, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7041780948638916, + "num_tokens": 307182441.0, + "step": 11874 + }, + { + "epoch": 1.3040852185372283, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.541872978210449, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7276443839073181, + "num_tokens": 307204790.0, + "step": 11875 + }, + { + "epoch": 1.3041950362398418, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2401649951934814, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7102404832839966, + "num_tokens": 307230438.0, + "step": 11876 + }, + { + "epoch": 1.3043048539424555, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.0667035579681396, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7076311111450195, + "num_tokens": 307262556.0, + "step": 11877 + }, + { + "epoch": 1.3044146716450693, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.6098411083221436, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7044248580932617, + "num_tokens": 307283816.0, + "step": 11878 + }, + { + "epoch": 1.3045244893476828, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.646758794784546, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7105302214622498, + "num_tokens": 307304357.0, + "step": 11879 + }, + { + "epoch": 1.3046343070502964, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2558631896972656, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7113492488861084, + "num_tokens": 307329642.0, + "step": 11880 + }, + { + "epoch": 1.3047441247529101, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.0225374698638916, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7248477339744568, + "num_tokens": 307357631.0, + "step": 11881 + }, + { + "epoch": 1.3048539424555239, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1753368377685547, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.717574954032898, + "num_tokens": 307384204.0, + "step": 11882 + }, + { + "epoch": 1.3049637601581374, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.126788377761841, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7022815346717834, + "num_tokens": 307411603.0, + "step": 11883 + }, + { + "epoch": 1.3050735778607512, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.099815845489502, + "learning_rate": 1e-06, + "loss": 1.0126, + "mean_token_accuracy": 0.6856801509857178, + "num_tokens": 307440959.0, + "step": 11884 + }, + { + "epoch": 1.3051833955633647, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.0243730545043945, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6969971656799316, + "num_tokens": 307472194.0, + "step": 11885 + }, + { + "epoch": 1.3052932132659785, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.033159017562866, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7200897932052612, + "num_tokens": 307502651.0, + "step": 11886 + }, + { + "epoch": 1.3054030309685922, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4705541133880615, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.728268563747406, + "num_tokens": 307525708.0, + "step": 11887 + }, + { + "epoch": 1.3055128486712058, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.289393901824951, + "learning_rate": 1e-06, + "loss": 0.9869, + "mean_token_accuracy": 0.7053430080413818, + "num_tokens": 307551774.0, + "step": 11888 + }, + { + "epoch": 1.3056226663738195, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.139207363128662, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7288103699684143, + "num_tokens": 307577389.0, + "step": 11889 + }, + { + "epoch": 1.305732484076433, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1111245155334473, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7131295800209045, + "num_tokens": 307605984.0, + "step": 11890 + }, + { + "epoch": 1.3058423017790468, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.5192136764526367, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7234088182449341, + "num_tokens": 307627950.0, + "step": 11891 + }, + { + "epoch": 1.3059521194816606, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.242239236831665, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7163676619529724, + "num_tokens": 307653276.0, + "step": 11892 + }, + { + "epoch": 1.306061937184274, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1223673820495605, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7069846391677856, + "num_tokens": 307681643.0, + "step": 11893 + }, + { + "epoch": 1.3061717548868876, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1362698078155518, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7319962382316589, + "num_tokens": 307707993.0, + "step": 11894 + }, + { + "epoch": 1.3062815725895014, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4351625442504883, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7124123573303223, + "num_tokens": 307730284.0, + "step": 11895 + }, + { + "epoch": 1.3063913902921152, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.5601673126220703, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7379569411277771, + "num_tokens": 307750876.0, + "step": 11896 + }, + { + "epoch": 1.3065012079947287, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4778249263763428, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7363527417182922, + "num_tokens": 307771352.0, + "step": 11897 + }, + { + "epoch": 1.3066110256973424, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.250591278076172, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.6973042488098145, + "num_tokens": 307799490.0, + "step": 11898 + }, + { + "epoch": 1.306720843399956, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.483158588409424, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7419149875640869, + "num_tokens": 307822092.0, + "step": 11899 + }, + { + "epoch": 1.3068306611025697, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.2361936569213867, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7195313572883606, + "num_tokens": 307847578.0, + "step": 11900 + }, + { + "epoch": 1.3069404788051835, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.055258274078369, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7092043161392212, + "num_tokens": 307879184.0, + "step": 11901 + }, + { + "epoch": 1.307050296507797, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.36923885345459, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7068010568618774, + "num_tokens": 307902946.0, + "step": 11902 + }, + { + "epoch": 1.3071601142104108, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.4034998416900635, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7007424831390381, + "num_tokens": 307925745.0, + "step": 11903 + }, + { + "epoch": 1.3072699319130243, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.381073474884033, + "learning_rate": 1e-06, + "loss": 0.972, + "mean_token_accuracy": 0.7046515345573425, + "num_tokens": 307950738.0, + "step": 11904 + }, + { + "epoch": 1.307379749615638, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.083582639694214, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7087458372116089, + "num_tokens": 307980167.0, + "step": 11905 + }, + { + "epoch": 1.3074895673182518, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.5789694786071777, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7206159830093384, + "num_tokens": 308003106.0, + "step": 11906 + }, + { + "epoch": 1.3075993850208654, + "ewc_loss": 6.973743438720703e-06, + "grad_norm": 2.1457314491271973, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7278009653091431, + "num_tokens": 308029924.0, + "step": 11907 + }, + { + "epoch": 1.307709202723479, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.3656203746795654, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7052584886550903, + "num_tokens": 308053551.0, + "step": 11908 + }, + { + "epoch": 1.3078190204260927, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.434673547744751, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7241159677505493, + "num_tokens": 308075839.0, + "step": 11909 + }, + { + "epoch": 1.3079288381287064, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.3768374919891357, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7426738142967224, + "num_tokens": 308098003.0, + "step": 11910 + }, + { + "epoch": 1.30803865583132, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 1.9749815464019775, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7065037488937378, + "num_tokens": 308131484.0, + "step": 11911 + }, + { + "epoch": 1.3081484735339337, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.4428250789642334, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7243232727050781, + "num_tokens": 308152707.0, + "step": 11912 + }, + { + "epoch": 1.3082582912365472, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.1644372940063477, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7168158292770386, + "num_tokens": 308183132.0, + "step": 11913 + }, + { + "epoch": 1.308368108939161, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.6946287155151367, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7482470870018005, + "num_tokens": 308201717.0, + "step": 11914 + }, + { + "epoch": 1.3084779266417748, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.5792369842529297, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7246394157409668, + "num_tokens": 308221777.0, + "step": 11915 + }, + { + "epoch": 1.3085877443443883, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.0327565670013428, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7116546034812927, + "num_tokens": 308252221.0, + "step": 11916 + }, + { + "epoch": 1.308697562047002, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.4002349376678467, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7284189462661743, + "num_tokens": 308274420.0, + "step": 11917 + }, + { + "epoch": 1.3088073797496156, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 1.9371815919876099, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.6917678117752075, + "num_tokens": 308309469.0, + "step": 11918 + }, + { + "epoch": 1.3089171974522293, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.195317029953003, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7382118701934814, + "num_tokens": 308334493.0, + "step": 11919 + }, + { + "epoch": 1.309027015154843, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.298140525817871, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.711784839630127, + "num_tokens": 308362169.0, + "step": 11920 + }, + { + "epoch": 1.3091368328574566, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.4572970867156982, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7327873706817627, + "num_tokens": 308383792.0, + "step": 11921 + }, + { + "epoch": 1.3092466505600702, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.2778236865997314, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.720541775226593, + "num_tokens": 308408941.0, + "step": 11922 + }, + { + "epoch": 1.309356468262684, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.214656352996826, + "learning_rate": 1e-06, + "loss": 0.9897, + "mean_token_accuracy": 0.7033672332763672, + "num_tokens": 308437407.0, + "step": 11923 + }, + { + "epoch": 1.3094662859652977, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.1302099227905273, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7089905738830566, + "num_tokens": 308464422.0, + "step": 11924 + }, + { + "epoch": 1.3095761036679112, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.1488938331604004, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.6964356899261475, + "num_tokens": 308493901.0, + "step": 11925 + }, + { + "epoch": 1.309685921370525, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 1.9695688486099243, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7223833799362183, + "num_tokens": 308527056.0, + "step": 11926 + }, + { + "epoch": 1.3097957390731385, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.0093801021575928, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7142065763473511, + "num_tokens": 308559454.0, + "step": 11927 + }, + { + "epoch": 1.3099055567757523, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.415637969970703, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7244303822517395, + "num_tokens": 308582588.0, + "step": 11928 + }, + { + "epoch": 1.310015374478366, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.4780349731445312, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7149772644042969, + "num_tokens": 308607198.0, + "step": 11929 + }, + { + "epoch": 1.3101251921809796, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.279323101043701, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7136765718460083, + "num_tokens": 308633119.0, + "step": 11930 + }, + { + "epoch": 1.3102350098835933, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.3145291805267334, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7148011326789856, + "num_tokens": 308658039.0, + "step": 11931 + }, + { + "epoch": 1.3103448275862069, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.423837900161743, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7416723966598511, + "num_tokens": 308678716.0, + "step": 11932 + }, + { + "epoch": 1.3104546452888206, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.080587863922119, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7103216648101807, + "num_tokens": 308708883.0, + "step": 11933 + }, + { + "epoch": 1.3105644629914341, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.228675365447998, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7273368835449219, + "num_tokens": 308733671.0, + "step": 11934 + }, + { + "epoch": 1.310674280694048, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.259629726409912, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.7123934626579285, + "num_tokens": 308760280.0, + "step": 11935 + }, + { + "epoch": 1.3107840983966614, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.5219929218292236, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7358754277229309, + "num_tokens": 308780609.0, + "step": 11936 + }, + { + "epoch": 1.3108939160992752, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.37044095993042, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7020834684371948, + "num_tokens": 308804795.0, + "step": 11937 + }, + { + "epoch": 1.311003733801889, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.5549566745758057, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7278550863265991, + "num_tokens": 308827372.0, + "step": 11938 + }, + { + "epoch": 1.3111135515045025, + "ewc_loss": 7.0035457611083984e-06, + "grad_norm": 2.2277660369873047, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7293181419372559, + "num_tokens": 308851157.0, + "step": 11939 + }, + { + "epoch": 1.3112233692071162, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.337961196899414, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7166398763656616, + "num_tokens": 308876515.0, + "step": 11940 + }, + { + "epoch": 1.3113331869097298, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.460331678390503, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7021893262863159, + "num_tokens": 308898561.0, + "step": 11941 + }, + { + "epoch": 1.3114430046123435, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.0208840370178223, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7161076664924622, + "num_tokens": 308928792.0, + "step": 11942 + }, + { + "epoch": 1.3115528223149573, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.1073131561279297, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7364993095397949, + "num_tokens": 308955868.0, + "step": 11943 + }, + { + "epoch": 1.3116626400175708, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.3168654441833496, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7176116108894348, + "num_tokens": 308979800.0, + "step": 11944 + }, + { + "epoch": 1.3117724577201844, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.1253631114959717, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7108772397041321, + "num_tokens": 309009514.0, + "step": 11945 + }, + { + "epoch": 1.3118822754227981, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.5825650691986084, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7222964763641357, + "num_tokens": 309030237.0, + "step": 11946 + }, + { + "epoch": 1.3119920931254119, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.194305419921875, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7287911176681519, + "num_tokens": 309058352.0, + "step": 11947 + }, + { + "epoch": 1.3121019108280254, + "ewc_loss": 7.033348083496094e-06, + "grad_norm": 2.525313138961792, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7217459678649902, + "num_tokens": 309078411.0, + "step": 11948 + }, + { + "epoch": 1.3122117285306392, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.401047706604004, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7393283247947693, + "num_tokens": 309099362.0, + "step": 11949 + }, + { + "epoch": 1.3123215462332527, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.492603063583374, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7395057678222656, + "num_tokens": 309120392.0, + "step": 11950 + }, + { + "epoch": 1.3124313639358665, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.074047803878784, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7259599566459656, + "num_tokens": 309149883.0, + "step": 11951 + }, + { + "epoch": 1.3125411816384802, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.230011224746704, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.712474524974823, + "num_tokens": 309176404.0, + "step": 11952 + }, + { + "epoch": 1.3126509993410937, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.984384298324585, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.703225314617157, + "num_tokens": 309204974.0, + "step": 11953 + }, + { + "epoch": 1.3127608170437075, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0931437015533447, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7219128608703613, + "num_tokens": 309233538.0, + "step": 11954 + }, + { + "epoch": 1.312870634746321, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.9604706764221191, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7085699439048767, + "num_tokens": 309265874.0, + "step": 11955 + }, + { + "epoch": 1.3129804524489348, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3387863636016846, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6941975355148315, + "num_tokens": 309292134.0, + "step": 11956 + }, + { + "epoch": 1.3130902701515486, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.4514052867889404, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7400882244110107, + "num_tokens": 309314202.0, + "step": 11957 + }, + { + "epoch": 1.313200087854162, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.9041227102279663, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7222853899002075, + "num_tokens": 309345545.0, + "step": 11958 + }, + { + "epoch": 1.3133099055567756, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.406102180480957, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7050414681434631, + "num_tokens": 309368933.0, + "step": 11959 + }, + { + "epoch": 1.3134197232593894, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.121002197265625, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7165690660476685, + "num_tokens": 309395945.0, + "step": 11960 + }, + { + "epoch": 1.3135295409620031, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.27274751663208, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7046773433685303, + "num_tokens": 309421344.0, + "step": 11961 + }, + { + "epoch": 1.3136393586646167, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.452308177947998, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7138518691062927, + "num_tokens": 309444483.0, + "step": 11962 + }, + { + "epoch": 1.3137491763672304, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3630223274230957, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7210044860839844, + "num_tokens": 309468167.0, + "step": 11963 + }, + { + "epoch": 1.313858994069844, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.9218217134475708, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7069160342216492, + "num_tokens": 309503149.0, + "step": 11964 + }, + { + "epoch": 1.3139688117724577, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2213566303253174, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7360904216766357, + "num_tokens": 309528160.0, + "step": 11965 + }, + { + "epoch": 1.3140786294750715, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0898330211639404, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7258636951446533, + "num_tokens": 309556572.0, + "step": 11966 + }, + { + "epoch": 1.314188447177685, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.241586208343506, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7243233919143677, + "num_tokens": 309582611.0, + "step": 11967 + }, + { + "epoch": 1.3142982648802988, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2910571098327637, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7408219575881958, + "num_tokens": 309609761.0, + "step": 11968 + }, + { + "epoch": 1.3144080825829123, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1438710689544678, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7165110111236572, + "num_tokens": 309635992.0, + "step": 11969 + }, + { + "epoch": 1.314517900285526, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.4751524925231934, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7178967595100403, + "num_tokens": 309659002.0, + "step": 11970 + }, + { + "epoch": 1.3146277179881398, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1328630447387695, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7078470587730408, + "num_tokens": 309689015.0, + "step": 11971 + }, + { + "epoch": 1.3147375356907534, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2741246223449707, + "learning_rate": 1e-06, + "loss": 0.9937, + "mean_token_accuracy": 0.6963274478912354, + "num_tokens": 309715038.0, + "step": 11972 + }, + { + "epoch": 1.314847353393367, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3990638256073, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7121089696884155, + "num_tokens": 309739071.0, + "step": 11973 + }, + { + "epoch": 1.3149571710959806, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1982643604278564, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6937334537506104, + "num_tokens": 309769186.0, + "step": 11974 + }, + { + "epoch": 1.3150669887985944, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0912976264953613, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.719724178314209, + "num_tokens": 309800157.0, + "step": 11975 + }, + { + "epoch": 1.315176806501208, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.192706823348999, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.6979303359985352, + "num_tokens": 309827522.0, + "step": 11976 + }, + { + "epoch": 1.3152866242038217, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1200122833251953, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6933844685554504, + "num_tokens": 309855405.0, + "step": 11977 + }, + { + "epoch": 1.3153964419064352, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0186288356781006, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7159038782119751, + "num_tokens": 309885220.0, + "step": 11978 + }, + { + "epoch": 1.315506259609049, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.244144916534424, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7290815114974976, + "num_tokens": 309912524.0, + "step": 11979 + }, + { + "epoch": 1.3156160773116627, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0001299381256104, + "learning_rate": 1e-06, + "loss": 1.0296, + "mean_token_accuracy": 0.6897470951080322, + "num_tokens": 309944880.0, + "step": 11980 + }, + { + "epoch": 1.3157258950142763, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3328444957733154, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7066552639007568, + "num_tokens": 309971327.0, + "step": 11981 + }, + { + "epoch": 1.31583571271689, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.4614193439483643, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7046605348587036, + "num_tokens": 309996491.0, + "step": 11982 + }, + { + "epoch": 1.3159455304195036, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.61684513092041, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7385741472244263, + "num_tokens": 310016335.0, + "step": 11983 + }, + { + "epoch": 1.3160553481221173, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.910642385482788, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7033945322036743, + "num_tokens": 310051258.0, + "step": 11984 + }, + { + "epoch": 1.3161651658247309, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.9515823125839233, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7118955254554749, + "num_tokens": 310084879.0, + "step": 11985 + }, + { + "epoch": 1.3162749835273446, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1980466842651367, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7401829957962036, + "num_tokens": 310110181.0, + "step": 11986 + }, + { + "epoch": 1.3163848012299582, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2520244121551514, + "learning_rate": 1e-06, + "loss": 0.776, + "mean_token_accuracy": 0.7504626512527466, + "num_tokens": 310132279.0, + "step": 11987 + }, + { + "epoch": 1.316494618932572, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.275498628616333, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7326023578643799, + "num_tokens": 310154905.0, + "step": 11988 + }, + { + "epoch": 1.3166044366351857, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.247403621673584, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7204481363296509, + "num_tokens": 310182554.0, + "step": 11989 + }, + { + "epoch": 1.3167142543377992, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.032334327697754, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7205770611763, + "num_tokens": 310216138.0, + "step": 11990 + }, + { + "epoch": 1.316824072040413, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.3308262825012207, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7199654579162598, + "num_tokens": 310243173.0, + "step": 11991 + }, + { + "epoch": 1.3169338897430265, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.369199752807617, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.719235897064209, + "num_tokens": 310266769.0, + "step": 11992 + }, + { + "epoch": 1.3170437074456403, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.132401466369629, + "learning_rate": 1e-06, + "loss": 1.022, + "mean_token_accuracy": 0.6950068473815918, + "num_tokens": 310299363.0, + "step": 11993 + }, + { + "epoch": 1.317153525148254, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.0874922275543213, + "learning_rate": 1e-06, + "loss": 1.0004, + "mean_token_accuracy": 0.6946113109588623, + "num_tokens": 310327603.0, + "step": 11994 + }, + { + "epoch": 1.3172633428508675, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.314486026763916, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7224417924880981, + "num_tokens": 310352373.0, + "step": 11995 + }, + { + "epoch": 1.3173731605534813, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.2186434268951416, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7307981252670288, + "num_tokens": 310378023.0, + "step": 11996 + }, + { + "epoch": 1.3174829782560948, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.9572707414627075, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.702057421207428, + "num_tokens": 310408772.0, + "step": 11997 + }, + { + "epoch": 1.3175927959587086, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.1598594188690186, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7137958407402039, + "num_tokens": 310437185.0, + "step": 11998 + }, + { + "epoch": 1.3177026136613221, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 1.856181263923645, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7219069004058838, + "num_tokens": 310472318.0, + "step": 11999 + }, + { + "epoch": 1.3178124313639359, + "ewc_loss": 7.063150405883789e-06, + "grad_norm": 2.104020118713379, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7349300384521484, + "num_tokens": 310501902.0, + "step": 12000 + }, + { + "epoch": 1.3179222490665494, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.455385684967041, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.751711368560791, + "num_tokens": 310523070.0, + "step": 12001 + }, + { + "epoch": 1.3180320667691632, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.1822831630706787, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7177436947822571, + "num_tokens": 310548975.0, + "step": 12002 + }, + { + "epoch": 1.318141884471777, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9547501802444458, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7536982297897339, + "num_tokens": 310578440.0, + "step": 12003 + }, + { + "epoch": 1.3182517021743905, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.1481199264526367, + "learning_rate": 1e-06, + "loss": 0.9795, + "mean_token_accuracy": 0.7018845677375793, + "num_tokens": 310608050.0, + "step": 12004 + }, + { + "epoch": 1.3183615198770042, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2098119258880615, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.718338668346405, + "num_tokens": 310634990.0, + "step": 12005 + }, + { + "epoch": 1.3184713375796178, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.162762403488159, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7191157937049866, + "num_tokens": 310663671.0, + "step": 12006 + }, + { + "epoch": 1.3185811552822315, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.3757426738739014, + "learning_rate": 1e-06, + "loss": 0.7837, + "mean_token_accuracy": 0.7437466382980347, + "num_tokens": 310686032.0, + "step": 12007 + }, + { + "epoch": 1.3186909729848453, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.090471029281616, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7138240337371826, + "num_tokens": 310715174.0, + "step": 12008 + }, + { + "epoch": 1.3188007906874588, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.280623435974121, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.6986352801322937, + "num_tokens": 310741269.0, + "step": 12009 + }, + { + "epoch": 1.3189106083900723, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.592015266418457, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7356868982315063, + "num_tokens": 310761641.0, + "step": 12010 + }, + { + "epoch": 1.319020426092686, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.3129758834838867, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7308871150016785, + "num_tokens": 310784991.0, + "step": 12011 + }, + { + "epoch": 1.3191302437952999, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.5068883895874023, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.700069010257721, + "num_tokens": 310807560.0, + "step": 12012 + }, + { + "epoch": 1.3192400614979134, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.178734540939331, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7241148948669434, + "num_tokens": 310833251.0, + "step": 12013 + }, + { + "epoch": 1.3193498792005272, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9562020301818848, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.715196967124939, + "num_tokens": 310865625.0, + "step": 12014 + }, + { + "epoch": 1.3194596969031407, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.374084949493408, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7200512290000916, + "num_tokens": 310889811.0, + "step": 12015 + }, + { + "epoch": 1.3195695146057544, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.6931490898132324, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7220022678375244, + "num_tokens": 310908529.0, + "step": 12016 + }, + { + "epoch": 1.3196793323083682, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9887745380401611, + "learning_rate": 1e-06, + "loss": 1.0531, + "mean_token_accuracy": 0.6812751293182373, + "num_tokens": 310940143.0, + "step": 12017 + }, + { + "epoch": 1.3197891500109817, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.526198387145996, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7235734462738037, + "num_tokens": 310960261.0, + "step": 12018 + }, + { + "epoch": 1.3198989677135955, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.4592955112457275, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7165300846099854, + "num_tokens": 310983025.0, + "step": 12019 + }, + { + "epoch": 1.320008785416209, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2654902935028076, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7104721069335938, + "num_tokens": 311008978.0, + "step": 12020 + }, + { + "epoch": 1.3201186031188228, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.137519598007202, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7296168804168701, + "num_tokens": 311037576.0, + "step": 12021 + }, + { + "epoch": 1.3202284208214365, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.3761367797851562, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.6996868848800659, + "num_tokens": 311061893.0, + "step": 12022 + }, + { + "epoch": 1.32033823852405, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.551403760910034, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7379508018493652, + "num_tokens": 311084726.0, + "step": 12023 + }, + { + "epoch": 1.3204480562266636, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.1868813037872314, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7242487668991089, + "num_tokens": 311110099.0, + "step": 12024 + }, + { + "epoch": 1.3205578739292774, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.578395128250122, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.737248420715332, + "num_tokens": 311130690.0, + "step": 12025 + }, + { + "epoch": 1.3206676916318911, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.27726411819458, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.6952589750289917, + "num_tokens": 311157366.0, + "step": 12026 + }, + { + "epoch": 1.3207775093345047, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.4771950244903564, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7118024826049805, + "num_tokens": 311179686.0, + "step": 12027 + }, + { + "epoch": 1.3208873270371184, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.338932752609253, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7067289352416992, + "num_tokens": 311203656.0, + "step": 12028 + }, + { + "epoch": 1.320997144739732, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.169059991836548, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7154889702796936, + "num_tokens": 311230369.0, + "step": 12029 + }, + { + "epoch": 1.3211069624423457, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9990837574005127, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7353178262710571, + "num_tokens": 311260256.0, + "step": 12030 + }, + { + "epoch": 1.3212167801449595, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.490910768508911, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7024234533309937, + "num_tokens": 311282552.0, + "step": 12031 + }, + { + "epoch": 1.321326597847573, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.3245368003845215, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7436563372612, + "num_tokens": 311304267.0, + "step": 12032 + }, + { + "epoch": 1.3214364155501868, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2198846340179443, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7031534910202026, + "num_tokens": 311330001.0, + "step": 12033 + }, + { + "epoch": 1.3215462332528003, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.5111002922058105, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7194693088531494, + "num_tokens": 311355827.0, + "step": 12034 + }, + { + "epoch": 1.321656050955414, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.344712495803833, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7223178148269653, + "num_tokens": 311379536.0, + "step": 12035 + }, + { + "epoch": 1.3217658686580278, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.263478994369507, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7249284386634827, + "num_tokens": 311404203.0, + "step": 12036 + }, + { + "epoch": 1.3218756863606413, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2395098209381104, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7281585931777954, + "num_tokens": 311429686.0, + "step": 12037 + }, + { + "epoch": 1.3219855040632549, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.0653467178344727, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.710693895816803, + "num_tokens": 311459298.0, + "step": 12038 + }, + { + "epoch": 1.3220953217658686, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.467498779296875, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7482343316078186, + "num_tokens": 311479998.0, + "step": 12039 + }, + { + "epoch": 1.3222051394684824, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.1317365169525146, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7006272077560425, + "num_tokens": 311508414.0, + "step": 12040 + }, + { + "epoch": 1.322314957171096, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.341878890991211, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7330911159515381, + "num_tokens": 311531813.0, + "step": 12041 + }, + { + "epoch": 1.3224247748737097, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9840112924575806, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7209636569023132, + "num_tokens": 311565499.0, + "step": 12042 + }, + { + "epoch": 1.3225345925763232, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.6167807579040527, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7193742990493774, + "num_tokens": 311585634.0, + "step": 12043 + }, + { + "epoch": 1.322644410278937, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.126406192779541, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.718562126159668, + "num_tokens": 311612145.0, + "step": 12044 + }, + { + "epoch": 1.3227542279815507, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.1852455139160156, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.7010229229927063, + "num_tokens": 311637952.0, + "step": 12045 + }, + { + "epoch": 1.3228640456841643, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 1.9614551067352295, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7302660942077637, + "num_tokens": 311667932.0, + "step": 12046 + }, + { + "epoch": 1.322973863386778, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2798867225646973, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.73103928565979, + "num_tokens": 311691471.0, + "step": 12047 + }, + { + "epoch": 1.3230836810893916, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.6052896976470947, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7284910082817078, + "num_tokens": 311710910.0, + "step": 12048 + }, + { + "epoch": 1.3231934987920053, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.4040334224700928, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.740023672580719, + "num_tokens": 311733474.0, + "step": 12049 + }, + { + "epoch": 1.3233033164946189, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.58915376663208, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7152707576751709, + "num_tokens": 311754050.0, + "step": 12050 + }, + { + "epoch": 1.3234131341972326, + "ewc_loss": 7.092952728271484e-06, + "grad_norm": 2.2944345474243164, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7045257687568665, + "num_tokens": 311782576.0, + "step": 12051 + }, + { + "epoch": 1.3235229518998461, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1127452850341797, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7303831577301025, + "num_tokens": 311810613.0, + "step": 12052 + }, + { + "epoch": 1.32363276960246, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 1.9649913311004639, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.6935872435569763, + "num_tokens": 311845004.0, + "step": 12053 + }, + { + "epoch": 1.3237425873050737, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.470768690109253, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7427794933319092, + "num_tokens": 311865498.0, + "step": 12054 + }, + { + "epoch": 1.3238524050076872, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.010535955429077, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.7109450101852417, + "num_tokens": 311895993.0, + "step": 12055 + }, + { + "epoch": 1.323962222710301, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1877083778381348, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7220847010612488, + "num_tokens": 311922674.0, + "step": 12056 + }, + { + "epoch": 1.3240720404129145, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.451143980026245, + "learning_rate": 1e-06, + "loss": 0.7476, + "mean_token_accuracy": 0.7578153610229492, + "num_tokens": 311944228.0, + "step": 12057 + }, + { + "epoch": 1.3241818581155282, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.502882957458496, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7142077684402466, + "num_tokens": 311967283.0, + "step": 12058 + }, + { + "epoch": 1.324291675818142, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.6506073474884033, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7255551815032959, + "num_tokens": 311986322.0, + "step": 12059 + }, + { + "epoch": 1.3244014935207555, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.275334358215332, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.717644989490509, + "num_tokens": 312011426.0, + "step": 12060 + }, + { + "epoch": 1.324511311223369, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.0567448139190674, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.6973870992660522, + "num_tokens": 312043022.0, + "step": 12061 + }, + { + "epoch": 1.3246211289259828, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.15179443359375, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.696602463722229, + "num_tokens": 312073029.0, + "step": 12062 + }, + { + "epoch": 1.3247309466285966, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.3288660049438477, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.730647623538971, + "num_tokens": 312098938.0, + "step": 12063 + }, + { + "epoch": 1.3248407643312101, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.4882729053497314, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7183088064193726, + "num_tokens": 312122336.0, + "step": 12064 + }, + { + "epoch": 1.3249505820338239, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.378453493118286, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.708264946937561, + "num_tokens": 312145917.0, + "step": 12065 + }, + { + "epoch": 1.3250603997364374, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.5163345336914062, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.705236554145813, + "num_tokens": 312167339.0, + "step": 12066 + }, + { + "epoch": 1.3251702174390512, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.3525497913360596, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7268396615982056, + "num_tokens": 312189994.0, + "step": 12067 + }, + { + "epoch": 1.325280035141665, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.206555128097534, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.71583092212677, + "num_tokens": 312220602.0, + "step": 12068 + }, + { + "epoch": 1.3253898528442785, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.31876540184021, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7147053480148315, + "num_tokens": 312244048.0, + "step": 12069 + }, + { + "epoch": 1.3254996705468922, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1112358570098877, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7138791084289551, + "num_tokens": 312272462.0, + "step": 12070 + }, + { + "epoch": 1.3256094882495058, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.030184507369995, + "learning_rate": 1e-06, + "loss": 1.0294, + "mean_token_accuracy": 0.6926248669624329, + "num_tokens": 312304011.0, + "step": 12071 + }, + { + "epoch": 1.3257193059521195, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 1.9253290891647339, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7077045440673828, + "num_tokens": 312339018.0, + "step": 12072 + }, + { + "epoch": 1.3258291236547333, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.23148512840271, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.709833025932312, + "num_tokens": 312365181.0, + "step": 12073 + }, + { + "epoch": 1.3259389413573468, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.9160892963409424, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7435012459754944, + "num_tokens": 312380926.0, + "step": 12074 + }, + { + "epoch": 1.3260487590599603, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1228904724121094, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7246924042701721, + "num_tokens": 312408282.0, + "step": 12075 + }, + { + "epoch": 1.326158576762574, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.2919692993164062, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7315797805786133, + "num_tokens": 312432842.0, + "step": 12076 + }, + { + "epoch": 1.3262683944651878, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.171617269515991, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.731741189956665, + "num_tokens": 312457932.0, + "step": 12077 + }, + { + "epoch": 1.3263782121678014, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.2016334533691406, + "learning_rate": 1e-06, + "loss": 0.7874, + "mean_token_accuracy": 0.7493623495101929, + "num_tokens": 312482510.0, + "step": 12078 + }, + { + "epoch": 1.3264880298704151, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 1.8515397310256958, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6949671506881714, + "num_tokens": 312522526.0, + "step": 12079 + }, + { + "epoch": 1.3265978475730287, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.267612934112549, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7094799280166626, + "num_tokens": 312547872.0, + "step": 12080 + }, + { + "epoch": 1.3267076652756424, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.541025400161743, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7486684322357178, + "num_tokens": 312567048.0, + "step": 12081 + }, + { + "epoch": 1.3268174829782562, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.356222629547119, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7212837934494019, + "num_tokens": 312590450.0, + "step": 12082 + }, + { + "epoch": 1.3269273006808697, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.0579099655151367, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.702857255935669, + "num_tokens": 312621839.0, + "step": 12083 + }, + { + "epoch": 1.3270371183834835, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1796791553497314, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7002251744270325, + "num_tokens": 312649493.0, + "step": 12084 + }, + { + "epoch": 1.327146936086097, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.195582151412964, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7082025408744812, + "num_tokens": 312678801.0, + "step": 12085 + }, + { + "epoch": 1.3272567537887108, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.2653627395629883, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7322982549667358, + "num_tokens": 312702894.0, + "step": 12086 + }, + { + "epoch": 1.3273665714913245, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.3065080642700195, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7184113264083862, + "num_tokens": 312726396.0, + "step": 12087 + }, + { + "epoch": 1.327476389193938, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.0446999073028564, + "learning_rate": 1e-06, + "loss": 0.9597, + "mean_token_accuracy": 0.7048859596252441, + "num_tokens": 312755864.0, + "step": 12088 + }, + { + "epoch": 1.3275862068965516, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.3491008281707764, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.698164701461792, + "num_tokens": 312780319.0, + "step": 12089 + }, + { + "epoch": 1.3276960245991654, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1331119537353516, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7167013883590698, + "num_tokens": 312808674.0, + "step": 12090 + }, + { + "epoch": 1.3278058423017791, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.2548534870147705, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7035419940948486, + "num_tokens": 312834242.0, + "step": 12091 + }, + { + "epoch": 1.3279156600043927, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.374861001968384, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7335047125816345, + "num_tokens": 312856847.0, + "step": 12092 + }, + { + "epoch": 1.3280254777070064, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.1064510345458984, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7024042010307312, + "num_tokens": 312885152.0, + "step": 12093 + }, + { + "epoch": 1.32813529540962, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.4915645122528076, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7122610807418823, + "num_tokens": 312907134.0, + "step": 12094 + }, + { + "epoch": 1.3282451131122337, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.596137523651123, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7400468587875366, + "num_tokens": 312929375.0, + "step": 12095 + }, + { + "epoch": 1.3283549308148475, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.163696050643921, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6894956827163696, + "num_tokens": 312959592.0, + "step": 12096 + }, + { + "epoch": 1.328464748517461, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.2315638065338135, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7281683683395386, + "num_tokens": 312984791.0, + "step": 12097 + }, + { + "epoch": 1.3285745662200747, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.439335346221924, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7361494302749634, + "num_tokens": 313008554.0, + "step": 12098 + }, + { + "epoch": 1.3286843839226883, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.130714178085327, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7114711403846741, + "num_tokens": 313036100.0, + "step": 12099 + }, + { + "epoch": 1.328794201625302, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.389965057373047, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7206071615219116, + "num_tokens": 313060742.0, + "step": 12100 + }, + { + "epoch": 1.3289040193279158, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.28019642829895, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7277960181236267, + "num_tokens": 313085469.0, + "step": 12101 + }, + { + "epoch": 1.3290138370305293, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.132535219192505, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7074212431907654, + "num_tokens": 313112793.0, + "step": 12102 + }, + { + "epoch": 1.3291236547331429, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.542604923248291, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7358397841453552, + "num_tokens": 313134516.0, + "step": 12103 + }, + { + "epoch": 1.3292334724357566, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 2.139331340789795, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7156072854995728, + "num_tokens": 313163023.0, + "step": 12104 + }, + { + "epoch": 1.3293432901383704, + "ewc_loss": 7.152557373046875e-06, + "grad_norm": 1.9774503707885742, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7212734818458557, + "num_tokens": 313196724.0, + "step": 12105 + }, + { + "epoch": 1.329453107840984, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.989904522895813, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7149198055267334, + "num_tokens": 313227346.0, + "step": 12106 + }, + { + "epoch": 1.3295629255435977, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.140876531600952, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7106453776359558, + "num_tokens": 313255908.0, + "step": 12107 + }, + { + "epoch": 1.3296727432462112, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2764768600463867, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7160879373550415, + "num_tokens": 313282227.0, + "step": 12108 + }, + { + "epoch": 1.329782560948825, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.429622173309326, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7311164736747742, + "num_tokens": 313303207.0, + "step": 12109 + }, + { + "epoch": 1.3298923786514387, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3985955715179443, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7067351341247559, + "num_tokens": 313325047.0, + "step": 12110 + }, + { + "epoch": 1.3300021963540523, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.0521278381347656, + "learning_rate": 1e-06, + "loss": 1.0211, + "mean_token_accuracy": 0.6885781288146973, + "num_tokens": 313356491.0, + "step": 12111 + }, + { + "epoch": 1.330112014056666, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9675029516220093, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7049465179443359, + "num_tokens": 313389447.0, + "step": 12112 + }, + { + "epoch": 1.3302218317592795, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2952656745910645, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7156144380569458, + "num_tokens": 313415681.0, + "step": 12113 + }, + { + "epoch": 1.3303316494618933, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3058536052703857, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7256306409835815, + "num_tokens": 313440534.0, + "step": 12114 + }, + { + "epoch": 1.3304414671645068, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.000744581222534, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7172146439552307, + "num_tokens": 313472238.0, + "step": 12115 + }, + { + "epoch": 1.3305512848671206, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3019299507141113, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7107508778572083, + "num_tokens": 313496901.0, + "step": 12116 + }, + { + "epoch": 1.3306611025697341, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2986373901367188, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.700964093208313, + "num_tokens": 313521690.0, + "step": 12117 + }, + { + "epoch": 1.330770920272348, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.246121883392334, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7213181853294373, + "num_tokens": 313546813.0, + "step": 12118 + }, + { + "epoch": 1.3308807379749616, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.1074180603027344, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7257499098777771, + "num_tokens": 313574997.0, + "step": 12119 + }, + { + "epoch": 1.3309905556775752, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9713441133499146, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7056040167808533, + "num_tokens": 313607098.0, + "step": 12120 + }, + { + "epoch": 1.331100373380189, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.312190055847168, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7475979328155518, + "num_tokens": 313630093.0, + "step": 12121 + }, + { + "epoch": 1.3312101910828025, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.176347494125366, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7198763489723206, + "num_tokens": 313657207.0, + "step": 12122 + }, + { + "epoch": 1.3313200087854162, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.0938260555267334, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7115418314933777, + "num_tokens": 313686686.0, + "step": 12123 + }, + { + "epoch": 1.33142982648803, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9473159313201904, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7105713486671448, + "num_tokens": 313718020.0, + "step": 12124 + }, + { + "epoch": 1.3315396441906435, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.1531951427459717, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7273911237716675, + "num_tokens": 313744954.0, + "step": 12125 + }, + { + "epoch": 1.331649461893257, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3583016395568848, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.703621506690979, + "num_tokens": 313767651.0, + "step": 12126 + }, + { + "epoch": 1.3317592795958708, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.23356556892395, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7200608849525452, + "num_tokens": 313793030.0, + "step": 12127 + }, + { + "epoch": 1.3318690972984846, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2031402587890625, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7155563831329346, + "num_tokens": 313819612.0, + "step": 12128 + }, + { + "epoch": 1.331978915001098, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.0974349975585938, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7279129028320312, + "num_tokens": 313847069.0, + "step": 12129 + }, + { + "epoch": 1.3320887327037119, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.141847848892212, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7300852537155151, + "num_tokens": 313875458.0, + "step": 12130 + }, + { + "epoch": 1.3321985504063254, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.4961307048797607, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7311723232269287, + "num_tokens": 313896527.0, + "step": 12131 + }, + { + "epoch": 1.3323083681089392, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.410506010055542, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7176080942153931, + "num_tokens": 313918697.0, + "step": 12132 + }, + { + "epoch": 1.332418185811553, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.1262874603271484, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7182778120040894, + "num_tokens": 313946077.0, + "step": 12133 + }, + { + "epoch": 1.3325280035141664, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2077279090881348, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7097588777542114, + "num_tokens": 313972542.0, + "step": 12134 + }, + { + "epoch": 1.3326378212167802, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9186378717422485, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.726578950881958, + "num_tokens": 314003809.0, + "step": 12135 + }, + { + "epoch": 1.3327476389193937, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.236436605453491, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6932737231254578, + "num_tokens": 314030491.0, + "step": 12136 + }, + { + "epoch": 1.3328574566220075, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9192484617233276, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7228434085845947, + "num_tokens": 314062339.0, + "step": 12137 + }, + { + "epoch": 1.3329672743246213, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.0947232246398926, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7313511967658997, + "num_tokens": 314089700.0, + "step": 12138 + }, + { + "epoch": 1.3330770920272348, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.129166841506958, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7064404487609863, + "num_tokens": 314120020.0, + "step": 12139 + }, + { + "epoch": 1.3331869097298483, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2959625720977783, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7372473478317261, + "num_tokens": 314143549.0, + "step": 12140 + }, + { + "epoch": 1.333296727432462, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.023226499557495, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7210533618927002, + "num_tokens": 314173193.0, + "step": 12141 + }, + { + "epoch": 1.3334065451350758, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.104443311691284, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7131761908531189, + "num_tokens": 314200701.0, + "step": 12142 + }, + { + "epoch": 1.3335163628376894, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2355563640594482, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7287064790725708, + "num_tokens": 314224618.0, + "step": 12143 + }, + { + "epoch": 1.3336261805403031, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.1628799438476562, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.7031455039978027, + "num_tokens": 314254988.0, + "step": 12144 + }, + { + "epoch": 1.3337359982429167, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.162200689315796, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7057047486305237, + "num_tokens": 314282456.0, + "step": 12145 + }, + { + "epoch": 1.3338458159455304, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.417018413543701, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.699158251285553, + "num_tokens": 314305152.0, + "step": 12146 + }, + { + "epoch": 1.3339556336481442, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2059226036071777, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.713981032371521, + "num_tokens": 314333641.0, + "step": 12147 + }, + { + "epoch": 1.3340654513507577, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 1.9720051288604736, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7021975517272949, + "num_tokens": 314368978.0, + "step": 12148 + }, + { + "epoch": 1.3341752690533715, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.126708745956421, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7313960194587708, + "num_tokens": 314394530.0, + "step": 12149 + }, + { + "epoch": 1.334285086755985, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.4260034561157227, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7174322009086609, + "num_tokens": 314417140.0, + "step": 12150 + }, + { + "epoch": 1.3343949044585988, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.343022346496582, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7139080762863159, + "num_tokens": 314440596.0, + "step": 12151 + }, + { + "epoch": 1.3345047221612125, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.277913808822632, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7279496192932129, + "num_tokens": 314462690.0, + "step": 12152 + }, + { + "epoch": 1.334614539863826, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.4468834400177, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7091459631919861, + "num_tokens": 314485297.0, + "step": 12153 + }, + { + "epoch": 1.3347243575664396, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3239855766296387, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7169348001480103, + "num_tokens": 314511086.0, + "step": 12154 + }, + { + "epoch": 1.3348341752690533, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.355792760848999, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7445814609527588, + "num_tokens": 314534220.0, + "step": 12155 + }, + { + "epoch": 1.334943992971667, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.240929365158081, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7042931914329529, + "num_tokens": 314559951.0, + "step": 12156 + }, + { + "epoch": 1.3350538106742806, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2977006435394287, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7216774225234985, + "num_tokens": 314584385.0, + "step": 12157 + }, + { + "epoch": 1.3351636283768944, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.3579487800598145, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7309576272964478, + "num_tokens": 314608107.0, + "step": 12158 + }, + { + "epoch": 1.335273446079508, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.2057697772979736, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6883649826049805, + "num_tokens": 314636606.0, + "step": 12159 + }, + { + "epoch": 1.3353832637821217, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.304654359817505, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7277116179466248, + "num_tokens": 314660635.0, + "step": 12160 + }, + { + "epoch": 1.3354930814847354, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.1654112339019775, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.724972128868103, + "num_tokens": 314687780.0, + "step": 12161 + }, + { + "epoch": 1.335602899187349, + "ewc_loss": 7.212162017822266e-06, + "grad_norm": 2.244453191757202, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7343356609344482, + "num_tokens": 314713069.0, + "step": 12162 + }, + { + "epoch": 1.3357127168899627, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2527925968170166, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7155417203903198, + "num_tokens": 314737852.0, + "step": 12163 + }, + { + "epoch": 1.3358225345925763, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.184408187866211, + "learning_rate": 1e-06, + "loss": 1.0206, + "mean_token_accuracy": 0.6886714696884155, + "num_tokens": 314764199.0, + "step": 12164 + }, + { + "epoch": 1.33593235229519, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.295722723007202, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7273997664451599, + "num_tokens": 314788466.0, + "step": 12165 + }, + { + "epoch": 1.3360421699978036, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.238802433013916, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7070073485374451, + "num_tokens": 314814228.0, + "step": 12166 + }, + { + "epoch": 1.3361519877004173, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9182220697402954, + "learning_rate": 1e-06, + "loss": 1.0289, + "mean_token_accuracy": 0.6879531741142273, + "num_tokens": 314847990.0, + "step": 12167 + }, + { + "epoch": 1.3362618054030309, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.394063711166382, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7368867993354797, + "num_tokens": 314871778.0, + "step": 12168 + }, + { + "epoch": 1.3363716231056446, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5036780834198, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7179956436157227, + "num_tokens": 314894790.0, + "step": 12169 + }, + { + "epoch": 1.3364814408082584, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4001822471618652, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7143956422805786, + "num_tokens": 314918890.0, + "step": 12170 + }, + { + "epoch": 1.336591258510872, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.237262725830078, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7285819053649902, + "num_tokens": 314944008.0, + "step": 12171 + }, + { + "epoch": 1.3367010762134857, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9183934926986694, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7031456828117371, + "num_tokens": 314979013.0, + "step": 12172 + }, + { + "epoch": 1.3368108939160992, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.23002552986145, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7432222366333008, + "num_tokens": 315003209.0, + "step": 12173 + }, + { + "epoch": 1.336920711618713, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1628987789154053, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7352622747421265, + "num_tokens": 315029803.0, + "step": 12174 + }, + { + "epoch": 1.3370305293213267, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2884137630462646, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7120236158370972, + "num_tokens": 315058104.0, + "step": 12175 + }, + { + "epoch": 1.3371403470239402, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.374539375305176, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7325724363327026, + "num_tokens": 315080546.0, + "step": 12176 + }, + { + "epoch": 1.337250164726554, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3180482387542725, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7253996133804321, + "num_tokens": 315105139.0, + "step": 12177 + }, + { + "epoch": 1.3373599824291675, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3665783405303955, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7101747989654541, + "num_tokens": 315128254.0, + "step": 12178 + }, + { + "epoch": 1.3374698001317813, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2423291206359863, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7156468033790588, + "num_tokens": 315151826.0, + "step": 12179 + }, + { + "epoch": 1.3375796178343948, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0431222915649414, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.711812436580658, + "num_tokens": 315180659.0, + "step": 12180 + }, + { + "epoch": 1.3376894355370086, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.210268497467041, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7195647954940796, + "num_tokens": 315208419.0, + "step": 12181 + }, + { + "epoch": 1.3377992532396221, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.19808292388916, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7255822420120239, + "num_tokens": 315233736.0, + "step": 12182 + }, + { + "epoch": 1.3379090709422359, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.204921245574951, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7487820386886597, + "num_tokens": 315258558.0, + "step": 12183 + }, + { + "epoch": 1.3380188886448496, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2906174659729004, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.715411365032196, + "num_tokens": 315284503.0, + "step": 12184 + }, + { + "epoch": 1.3381287063474632, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1422696113586426, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7129273414611816, + "num_tokens": 315314227.0, + "step": 12185 + }, + { + "epoch": 1.338238524050077, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0389621257781982, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7456688284873962, + "num_tokens": 315341861.0, + "step": 12186 + }, + { + "epoch": 1.3383483417526905, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.6831178665161133, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7357299327850342, + "num_tokens": 315362198.0, + "step": 12187 + }, + { + "epoch": 1.3384581594553042, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4089300632476807, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.713918924331665, + "num_tokens": 315384800.0, + "step": 12188 + }, + { + "epoch": 1.338567977157918, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.152240037918091, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7208855152130127, + "num_tokens": 315412593.0, + "step": 12189 + }, + { + "epoch": 1.3386777948605315, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.261810064315796, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.721102237701416, + "num_tokens": 315437974.0, + "step": 12190 + }, + { + "epoch": 1.338787612563145, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0962576866149902, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6934988498687744, + "num_tokens": 315466659.0, + "step": 12191 + }, + { + "epoch": 1.3388974302657588, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9426509141921997, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7304664254188538, + "num_tokens": 315498402.0, + "step": 12192 + }, + { + "epoch": 1.3390072479683726, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4226033687591553, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7282226085662842, + "num_tokens": 315521133.0, + "step": 12193 + }, + { + "epoch": 1.339117065670986, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.27883243560791, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7419470548629761, + "num_tokens": 315545156.0, + "step": 12194 + }, + { + "epoch": 1.3392268833735999, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.163344383239746, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7165012955665588, + "num_tokens": 315571491.0, + "step": 12195 + }, + { + "epoch": 1.3393367010762134, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3065831661224365, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7271305322647095, + "num_tokens": 315597388.0, + "step": 12196 + }, + { + "epoch": 1.3394465187788271, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5987260341644287, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.726902961730957, + "num_tokens": 315618018.0, + "step": 12197 + }, + { + "epoch": 1.339556336481441, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.199040174484253, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7274141311645508, + "num_tokens": 315644214.0, + "step": 12198 + }, + { + "epoch": 1.3396661541840544, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2215802669525146, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7281512022018433, + "num_tokens": 315670775.0, + "step": 12199 + }, + { + "epoch": 1.3397759718866682, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5241036415100098, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7239646911621094, + "num_tokens": 315692463.0, + "step": 12200 + }, + { + "epoch": 1.3398857895892817, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.301076650619507, + "learning_rate": 1e-06, + "loss": 1.0066, + "mean_token_accuracy": 0.6926090121269226, + "num_tokens": 315719338.0, + "step": 12201 + }, + { + "epoch": 1.3399956072918955, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1638922691345215, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.693936288356781, + "num_tokens": 315748717.0, + "step": 12202 + }, + { + "epoch": 1.3401054249945092, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.32139253616333, + "learning_rate": 1e-06, + "loss": 0.7717, + "mean_token_accuracy": 0.7487839460372925, + "num_tokens": 315771304.0, + "step": 12203 + }, + { + "epoch": 1.3402152426971228, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1635875701904297, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7104600071907043, + "num_tokens": 315801733.0, + "step": 12204 + }, + { + "epoch": 1.3403250603997363, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3097965717315674, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7161840200424194, + "num_tokens": 315825516.0, + "step": 12205 + }, + { + "epoch": 1.34043487810235, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1255455017089844, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7083626985549927, + "num_tokens": 315854523.0, + "step": 12206 + }, + { + "epoch": 1.3405446958049638, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.213245391845703, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7230079770088196, + "num_tokens": 315881088.0, + "step": 12207 + }, + { + "epoch": 1.3406545135075774, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1722512245178223, + "learning_rate": 1e-06, + "loss": 1.0533, + "mean_token_accuracy": 0.6828586459159851, + "num_tokens": 315913033.0, + "step": 12208 + }, + { + "epoch": 1.3407643312101911, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.223332643508911, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7096562385559082, + "num_tokens": 315939853.0, + "step": 12209 + }, + { + "epoch": 1.3408741489128047, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2270421981811523, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.710913360118866, + "num_tokens": 315964538.0, + "step": 12210 + }, + { + "epoch": 1.3409839666154184, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.862740159034729, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7090798020362854, + "num_tokens": 316001696.0, + "step": 12211 + }, + { + "epoch": 1.3410937843180322, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.096445322036743, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.715485692024231, + "num_tokens": 316027487.0, + "step": 12212 + }, + { + "epoch": 1.3412036020206457, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2352683544158936, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7274057269096375, + "num_tokens": 316052502.0, + "step": 12213 + }, + { + "epoch": 1.3413134197232595, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3375608921051025, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7412259578704834, + "num_tokens": 316074290.0, + "step": 12214 + }, + { + "epoch": 1.341423237425873, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.594295024871826, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.713775634765625, + "num_tokens": 316095713.0, + "step": 12215 + }, + { + "epoch": 1.3415330551284868, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5268397331237793, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7134671211242676, + "num_tokens": 316119090.0, + "step": 12216 + }, + { + "epoch": 1.3416428728311005, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1858670711517334, + "learning_rate": 1e-06, + "loss": 1.0638, + "mean_token_accuracy": 0.6792072057723999, + "num_tokens": 316151908.0, + "step": 12217 + }, + { + "epoch": 1.341752690533714, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.7842624187469482, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7406681776046753, + "num_tokens": 316173112.0, + "step": 12218 + }, + { + "epoch": 1.3418625082363276, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4580068588256836, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7369984984397888, + "num_tokens": 316195721.0, + "step": 12219 + }, + { + "epoch": 1.3419723259389413, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0508506298065186, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7127330303192139, + "num_tokens": 316224989.0, + "step": 12220 + }, + { + "epoch": 1.342082143641555, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9301369190216064, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7087728977203369, + "num_tokens": 316259046.0, + "step": 12221 + }, + { + "epoch": 1.3421919613441686, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3167202472686768, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7345830202102661, + "num_tokens": 316281973.0, + "step": 12222 + }, + { + "epoch": 1.3423017790467824, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3162646293640137, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7129154205322266, + "num_tokens": 316307353.0, + "step": 12223 + }, + { + "epoch": 1.342411596749396, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3498494625091553, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7160162329673767, + "num_tokens": 316332042.0, + "step": 12224 + }, + { + "epoch": 1.3425214144520097, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0497758388519287, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7266504168510437, + "num_tokens": 316361949.0, + "step": 12225 + }, + { + "epoch": 1.3426312321546234, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.0889530181884766, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7425007820129395, + "num_tokens": 316389388.0, + "step": 12226 + }, + { + "epoch": 1.342741049857237, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4948060512542725, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7283294200897217, + "num_tokens": 316411616.0, + "step": 12227 + }, + { + "epoch": 1.3428508675598507, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.305117130279541, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.723964273929596, + "num_tokens": 316436210.0, + "step": 12228 + }, + { + "epoch": 1.3429606852624643, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3338732719421387, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.7029064297676086, + "num_tokens": 316460338.0, + "step": 12229 + }, + { + "epoch": 1.343070502965078, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9409676790237427, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7238045930862427, + "num_tokens": 316492222.0, + "step": 12230 + }, + { + "epoch": 1.3431803206676916, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.119579553604126, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7300752401351929, + "num_tokens": 316520060.0, + "step": 12231 + }, + { + "epoch": 1.3432901383703053, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 1.9615142345428467, + "learning_rate": 1e-06, + "loss": 1.0107, + "mean_token_accuracy": 0.6917272210121155, + "num_tokens": 316557495.0, + "step": 12232 + }, + { + "epoch": 1.3433999560729188, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3234164714813232, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7101253271102905, + "num_tokens": 316580479.0, + "step": 12233 + }, + { + "epoch": 1.3435097737755326, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2785747051239014, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7254502773284912, + "num_tokens": 316604763.0, + "step": 12234 + }, + { + "epoch": 1.3436195914781464, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2467198371887207, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.742445707321167, + "num_tokens": 316628819.0, + "step": 12235 + }, + { + "epoch": 1.34372940918076, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5015034675598145, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7276566028594971, + "num_tokens": 316649717.0, + "step": 12236 + }, + { + "epoch": 1.3438392268833736, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2919914722442627, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7173999547958374, + "num_tokens": 316673641.0, + "step": 12237 + }, + { + "epoch": 1.3439490445859872, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1848607063293457, + "learning_rate": 1e-06, + "loss": 0.9446, + "mean_token_accuracy": 0.7120497822761536, + "num_tokens": 316703302.0, + "step": 12238 + }, + { + "epoch": 1.344058862288601, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.113619327545166, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7206096649169922, + "num_tokens": 316731974.0, + "step": 12239 + }, + { + "epoch": 1.3441686799912147, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3968732357025146, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7281344532966614, + "num_tokens": 316753735.0, + "step": 12240 + }, + { + "epoch": 1.3442784976938282, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3367888927459717, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7351512908935547, + "num_tokens": 316777376.0, + "step": 12241 + }, + { + "epoch": 1.3443883153964418, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.261456251144409, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.737850546836853, + "num_tokens": 316803200.0, + "step": 12242 + }, + { + "epoch": 1.3444981330990555, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.3796026706695557, + "learning_rate": 1e-06, + "loss": 1.0594, + "mean_token_accuracy": 0.6721251606941223, + "num_tokens": 316830502.0, + "step": 12243 + }, + { + "epoch": 1.3446079508016693, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.200282096862793, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7290566563606262, + "num_tokens": 316858731.0, + "step": 12244 + }, + { + "epoch": 1.3447177685042828, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.047877073287964, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.702800452709198, + "num_tokens": 316890304.0, + "step": 12245 + }, + { + "epoch": 1.3448275862068966, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2102720737457275, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7232186198234558, + "num_tokens": 316914171.0, + "step": 12246 + }, + { + "epoch": 1.34493740390951, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.037353277206421, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7254408597946167, + "num_tokens": 316945920.0, + "step": 12247 + }, + { + "epoch": 1.3450472216121239, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.5473954677581787, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7301018834114075, + "num_tokens": 316967233.0, + "step": 12248 + }, + { + "epoch": 1.3451570393147376, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.027284622192383, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7088841795921326, + "num_tokens": 316996454.0, + "step": 12249 + }, + { + "epoch": 1.3452668570173512, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.299682378768921, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7166461944580078, + "num_tokens": 317021126.0, + "step": 12250 + }, + { + "epoch": 1.345376674719965, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1142830848693848, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7019534111022949, + "num_tokens": 317049288.0, + "step": 12251 + }, + { + "epoch": 1.3454864924225785, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.2653229236602783, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.707173228263855, + "num_tokens": 317075763.0, + "step": 12252 + }, + { + "epoch": 1.3455963101251922, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.1053366661071777, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7247464656829834, + "num_tokens": 317105490.0, + "step": 12253 + }, + { + "epoch": 1.345706127827806, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.317582845687866, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7233569025993347, + "num_tokens": 317128532.0, + "step": 12254 + }, + { + "epoch": 1.3458159455304195, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.4790072441101074, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.725020706653595, + "num_tokens": 317149149.0, + "step": 12255 + }, + { + "epoch": 1.345925763233033, + "ewc_loss": 7.271766662597656e-06, + "grad_norm": 2.064138412475586, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7302610278129578, + "num_tokens": 317179995.0, + "step": 12256 + }, + { + "epoch": 1.3460355809356468, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1651015281677246, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7210869789123535, + "num_tokens": 317208166.0, + "step": 12257 + }, + { + "epoch": 1.3461453986382605, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3591349124908447, + "learning_rate": 1e-06, + "loss": 1.0022, + "mean_token_accuracy": 0.6904942989349365, + "num_tokens": 317235731.0, + "step": 12258 + }, + { + "epoch": 1.346255216340874, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2886412143707275, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7345706224441528, + "num_tokens": 317259161.0, + "step": 12259 + }, + { + "epoch": 1.3463650340434878, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.446643829345703, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7161891460418701, + "num_tokens": 317280911.0, + "step": 12260 + }, + { + "epoch": 1.3464748517461014, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.510057210922241, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7183850407600403, + "num_tokens": 317307150.0, + "step": 12261 + }, + { + "epoch": 1.3465846694487151, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0980162620544434, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7184368968009949, + "num_tokens": 317336466.0, + "step": 12262 + }, + { + "epoch": 1.346694487151329, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.512190341949463, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7447277307510376, + "num_tokens": 317358547.0, + "step": 12263 + }, + { + "epoch": 1.3468043048539424, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 1.8498640060424805, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7280840277671814, + "num_tokens": 317395401.0, + "step": 12264 + }, + { + "epoch": 1.3469141225565562, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0926547050476074, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7174652814865112, + "num_tokens": 317423197.0, + "step": 12265 + }, + { + "epoch": 1.3470239402591697, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2746429443359375, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7078924179077148, + "num_tokens": 317446824.0, + "step": 12266 + }, + { + "epoch": 1.3471337579617835, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.296957492828369, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7191143035888672, + "num_tokens": 317471013.0, + "step": 12267 + }, + { + "epoch": 1.3472435756643972, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3629648685455322, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7339136600494385, + "num_tokens": 317494213.0, + "step": 12268 + }, + { + "epoch": 1.3473533933670108, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1388072967529297, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7246631979942322, + "num_tokens": 317521586.0, + "step": 12269 + }, + { + "epoch": 1.3474632110696243, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.053353786468506, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7239037752151489, + "num_tokens": 317550203.0, + "step": 12270 + }, + { + "epoch": 1.347573028772238, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.054064989089966, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.725426197052002, + "num_tokens": 317577786.0, + "step": 12271 + }, + { + "epoch": 1.3476828464748518, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.136134147644043, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7057482004165649, + "num_tokens": 317605820.0, + "step": 12272 + }, + { + "epoch": 1.3477926641774653, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.203547716140747, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.694586992263794, + "num_tokens": 317632019.0, + "step": 12273 + }, + { + "epoch": 1.347902481880079, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.165863037109375, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.715194821357727, + "num_tokens": 317660099.0, + "step": 12274 + }, + { + "epoch": 1.3480122995826926, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2811830043792725, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7103411555290222, + "num_tokens": 317686365.0, + "step": 12275 + }, + { + "epoch": 1.3481221172853064, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.5025901794433594, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7222370505332947, + "num_tokens": 317707757.0, + "step": 12276 + }, + { + "epoch": 1.3482319349879202, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3495445251464844, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7379844784736633, + "num_tokens": 317731141.0, + "step": 12277 + }, + { + "epoch": 1.3483417526905337, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.23256516456604, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7365379333496094, + "num_tokens": 317758618.0, + "step": 12278 + }, + { + "epoch": 1.3484515703931474, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3593735694885254, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7169733047485352, + "num_tokens": 317780188.0, + "step": 12279 + }, + { + "epoch": 1.348561388095761, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3119640350341797, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7446147203445435, + "num_tokens": 317803852.0, + "step": 12280 + }, + { + "epoch": 1.3486712057983747, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 1.9653120040893555, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7122454643249512, + "num_tokens": 317834475.0, + "step": 12281 + }, + { + "epoch": 1.3487810235009885, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.136321544647217, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7440668344497681, + "num_tokens": 317860962.0, + "step": 12282 + }, + { + "epoch": 1.348890841203602, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1728925704956055, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.723853588104248, + "num_tokens": 317887167.0, + "step": 12283 + }, + { + "epoch": 1.3490006589062156, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.667797327041626, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7381671667098999, + "num_tokens": 317907342.0, + "step": 12284 + }, + { + "epoch": 1.3491104766088293, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.6275265216827393, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7174376249313354, + "num_tokens": 317929542.0, + "step": 12285 + }, + { + "epoch": 1.349220294311443, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.5824267864227295, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7156962752342224, + "num_tokens": 317950297.0, + "step": 12286 + }, + { + "epoch": 1.3493301120140566, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1141538619995117, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7046456933021545, + "num_tokens": 317978347.0, + "step": 12287 + }, + { + "epoch": 1.3494399297166704, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2368357181549072, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7376725077629089, + "num_tokens": 318003315.0, + "step": 12288 + }, + { + "epoch": 1.349549747419284, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.299154043197632, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7287031412124634, + "num_tokens": 318027089.0, + "step": 12289 + }, + { + "epoch": 1.3496595651218977, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.413363456726074, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7239499092102051, + "num_tokens": 318049983.0, + "step": 12290 + }, + { + "epoch": 1.3497693828245114, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.267656087875366, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7114416360855103, + "num_tokens": 318076446.0, + "step": 12291 + }, + { + "epoch": 1.349879200527125, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.387791633605957, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.726714015007019, + "num_tokens": 318099282.0, + "step": 12292 + }, + { + "epoch": 1.3499890182297387, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2825348377227783, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7140483856201172, + "num_tokens": 318125169.0, + "step": 12293 + }, + { + "epoch": 1.3500988359323522, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 1.9292975664138794, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7414647340774536, + "num_tokens": 318157779.0, + "step": 12294 + }, + { + "epoch": 1.350208653634966, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 1.9829306602478027, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7466832399368286, + "num_tokens": 318187301.0, + "step": 12295 + }, + { + "epoch": 1.3503184713375795, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2081644535064697, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7150972485542297, + "num_tokens": 318212475.0, + "step": 12296 + }, + { + "epoch": 1.3504282890401933, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0185976028442383, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7080105543136597, + "num_tokens": 318245545.0, + "step": 12297 + }, + { + "epoch": 1.3505381067428068, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.398283004760742, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7335188388824463, + "num_tokens": 318268347.0, + "step": 12298 + }, + { + "epoch": 1.3506479244454206, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0102720260620117, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7255169749259949, + "num_tokens": 318296588.0, + "step": 12299 + }, + { + "epoch": 1.3507577421480343, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.37176775932312, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7254174947738647, + "num_tokens": 318318387.0, + "step": 12300 + }, + { + "epoch": 1.3508675598506479, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.258371353149414, + "learning_rate": 1e-06, + "loss": 0.7755, + "mean_token_accuracy": 0.7557049989700317, + "num_tokens": 318340938.0, + "step": 12301 + }, + { + "epoch": 1.3509773775532616, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3527474403381348, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7246100306510925, + "num_tokens": 318362842.0, + "step": 12302 + }, + { + "epoch": 1.3510871952558752, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.208744525909424, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7344454526901245, + "num_tokens": 318388176.0, + "step": 12303 + }, + { + "epoch": 1.351197012958489, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.370619058609009, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7347128987312317, + "num_tokens": 318411027.0, + "step": 12304 + }, + { + "epoch": 1.3513068306611027, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.461538076400757, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7314761877059937, + "num_tokens": 318431454.0, + "step": 12305 + }, + { + "epoch": 1.3514166483637162, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.348020315170288, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7089352607727051, + "num_tokens": 318456081.0, + "step": 12306 + }, + { + "epoch": 1.3515264660663298, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.132230043411255, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7161421179771423, + "num_tokens": 318484595.0, + "step": 12307 + }, + { + "epoch": 1.3516362837689435, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.5197458267211914, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7069382071495056, + "num_tokens": 318507267.0, + "step": 12308 + }, + { + "epoch": 1.3517461014715573, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.320462465286255, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7343087196350098, + "num_tokens": 318530460.0, + "step": 12309 + }, + { + "epoch": 1.3518559191741708, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1392714977264404, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7153820395469666, + "num_tokens": 318557169.0, + "step": 12310 + }, + { + "epoch": 1.3519657368767846, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1212387084960938, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7049921154975891, + "num_tokens": 318585845.0, + "step": 12311 + }, + { + "epoch": 1.352075554579398, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.467050790786743, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7045407295227051, + "num_tokens": 318607694.0, + "step": 12312 + }, + { + "epoch": 1.3521853722820119, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.413980007171631, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7142972350120544, + "num_tokens": 318630886.0, + "step": 12313 + }, + { + "epoch": 1.3522951899846256, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.5340843200683594, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7352207899093628, + "num_tokens": 318651318.0, + "step": 12314 + }, + { + "epoch": 1.3524050076872391, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0821077823638916, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7096035480499268, + "num_tokens": 318681369.0, + "step": 12315 + }, + { + "epoch": 1.352514825389853, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2624735832214355, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7070544958114624, + "num_tokens": 318708443.0, + "step": 12316 + }, + { + "epoch": 1.3526246430924664, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.5067248344421387, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7220532298088074, + "num_tokens": 318728753.0, + "step": 12317 + }, + { + "epoch": 1.3527344607950802, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.253726005554199, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7113680243492126, + "num_tokens": 318752398.0, + "step": 12318 + }, + { + "epoch": 1.352844278497694, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3072991371154785, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.6999655365943909, + "num_tokens": 318777523.0, + "step": 12319 + }, + { + "epoch": 1.3529540962003075, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1690878868103027, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.715813398361206, + "num_tokens": 318804092.0, + "step": 12320 + }, + { + "epoch": 1.353063913902921, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.263308525085449, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7271812558174133, + "num_tokens": 318830073.0, + "step": 12321 + }, + { + "epoch": 1.3531737316055348, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.231309652328491, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7212752103805542, + "num_tokens": 318855429.0, + "step": 12322 + }, + { + "epoch": 1.3532835493081485, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.356146812438965, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7319620847702026, + "num_tokens": 318878567.0, + "step": 12323 + }, + { + "epoch": 1.353393367010762, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2323708534240723, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7379075288772583, + "num_tokens": 318905612.0, + "step": 12324 + }, + { + "epoch": 1.3535031847133758, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.2001678943634033, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7151466012001038, + "num_tokens": 318932808.0, + "step": 12325 + }, + { + "epoch": 1.3536130024159894, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.240334987640381, + "learning_rate": 1e-06, + "loss": 0.9691, + "mean_token_accuracy": 0.71352219581604, + "num_tokens": 318958431.0, + "step": 12326 + }, + { + "epoch": 1.3537228201186031, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.184805154800415, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.732996940612793, + "num_tokens": 318981995.0, + "step": 12327 + }, + { + "epoch": 1.3538326378212169, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.1900625228881836, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.6985447406768799, + "num_tokens": 319011387.0, + "step": 12328 + }, + { + "epoch": 1.3539424555238304, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.0412333011627197, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6916078329086304, + "num_tokens": 319044796.0, + "step": 12329 + }, + { + "epoch": 1.3540522732264442, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.021622657775879, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.714877724647522, + "num_tokens": 319074717.0, + "step": 12330 + }, + { + "epoch": 1.3541620909290577, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3551154136657715, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6996012926101685, + "num_tokens": 319099557.0, + "step": 12331 + }, + { + "epoch": 1.3542719086316715, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.45607590675354, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7313529253005981, + "num_tokens": 319123101.0, + "step": 12332 + }, + { + "epoch": 1.3543817263342852, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.507899284362793, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7225354313850403, + "num_tokens": 319145466.0, + "step": 12333 + }, + { + "epoch": 1.3544915440368988, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 1.7660489082336426, + "learning_rate": 1e-06, + "loss": 1.034, + "mean_token_accuracy": 0.6889606714248657, + "num_tokens": 319184665.0, + "step": 12334 + }, + { + "epoch": 1.3546013617395123, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.4920573234558105, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7075064778327942, + "num_tokens": 319205907.0, + "step": 12335 + }, + { + "epoch": 1.354711179442126, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.4321482181549072, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7225728034973145, + "num_tokens": 319227066.0, + "step": 12336 + }, + { + "epoch": 1.3548209971447398, + "ewc_loss": 7.331371307373047e-06, + "grad_norm": 2.3068783283233643, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7030782699584961, + "num_tokens": 319252101.0, + "step": 12337 + }, + { + "epoch": 1.3549308148473533, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.3809680938720703, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7209465503692627, + "num_tokens": 319275052.0, + "step": 12338 + }, + { + "epoch": 1.355040632549967, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.425354242324829, + "learning_rate": 1e-06, + "loss": 0.7843, + "mean_token_accuracy": 0.7509127855300903, + "num_tokens": 319295313.0, + "step": 12339 + }, + { + "epoch": 1.3551504502525806, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.2429568767547607, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7109442949295044, + "num_tokens": 319321658.0, + "step": 12340 + }, + { + "epoch": 1.3552602679551944, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.397127151489258, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7209937572479248, + "num_tokens": 319344828.0, + "step": 12341 + }, + { + "epoch": 1.3553700856578081, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.1551339626312256, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7205032706260681, + "num_tokens": 319372964.0, + "step": 12342 + }, + { + "epoch": 1.3554799033604217, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.379457712173462, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7229177355766296, + "num_tokens": 319397019.0, + "step": 12343 + }, + { + "epoch": 1.3555897210630354, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.253880500793457, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.695783257484436, + "num_tokens": 319423150.0, + "step": 12344 + }, + { + "epoch": 1.355699538765649, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.4787611961364746, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.740350604057312, + "num_tokens": 319444028.0, + "step": 12345 + }, + { + "epoch": 1.3558093564682627, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.239880084991455, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7070283889770508, + "num_tokens": 319471726.0, + "step": 12346 + }, + { + "epoch": 1.3559191741708763, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.2657310962677, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7126197814941406, + "num_tokens": 319496920.0, + "step": 12347 + }, + { + "epoch": 1.35602899187349, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.1470515727996826, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7429729700088501, + "num_tokens": 319522394.0, + "step": 12348 + }, + { + "epoch": 1.3561388095761036, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.049700975418091, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7364816665649414, + "num_tokens": 319549366.0, + "step": 12349 + }, + { + "epoch": 1.3562486272787173, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.2652454376220703, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7396904230117798, + "num_tokens": 319573807.0, + "step": 12350 + }, + { + "epoch": 1.356358444981331, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.3853347301483154, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7187554836273193, + "num_tokens": 319597163.0, + "step": 12351 + }, + { + "epoch": 1.3564682626839446, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.100024461746216, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.733411431312561, + "num_tokens": 319624851.0, + "step": 12352 + }, + { + "epoch": 1.3565780803865584, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.3662617206573486, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7055427432060242, + "num_tokens": 319648659.0, + "step": 12353 + }, + { + "epoch": 1.356687898089172, + "ewc_loss": 7.361173629760742e-06, + "grad_norm": 2.6509368419647217, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7121597528457642, + "num_tokens": 319668633.0, + "step": 12354 + }, + { + "epoch": 1.3567977157917857, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0283095836639404, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7182523608207703, + "num_tokens": 319697321.0, + "step": 12355 + }, + { + "epoch": 1.3569075334943994, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.055656671524048, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6942712664604187, + "num_tokens": 319727386.0, + "step": 12356 + }, + { + "epoch": 1.357017351197013, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.2905967235565186, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7214122414588928, + "num_tokens": 319751705.0, + "step": 12357 + }, + { + "epoch": 1.3571271688996267, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3229939937591553, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7296156287193298, + "num_tokens": 319777094.0, + "step": 12358 + }, + { + "epoch": 1.3572369866022402, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0254218578338623, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7156581878662109, + "num_tokens": 319806021.0, + "step": 12359 + }, + { + "epoch": 1.357346804304854, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1849441528320312, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7112252712249756, + "num_tokens": 319834341.0, + "step": 12360 + }, + { + "epoch": 1.3574566220074675, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.122317314147949, + "learning_rate": 1e-06, + "loss": 1.0112, + "mean_token_accuracy": 0.6867746114730835, + "num_tokens": 319863895.0, + "step": 12361 + }, + { + "epoch": 1.3575664397100813, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.085299015045166, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7059369087219238, + "num_tokens": 319893533.0, + "step": 12362 + }, + { + "epoch": 1.3576762574126948, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.046424388885498, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6995273232460022, + "num_tokens": 319925237.0, + "step": 12363 + }, + { + "epoch": 1.3577860751153086, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.2365481853485107, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7116481065750122, + "num_tokens": 319953740.0, + "step": 12364 + }, + { + "epoch": 1.3578958928179223, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.4593374729156494, + "learning_rate": 1e-06, + "loss": 0.7708, + "mean_token_accuracy": 0.7408367395401001, + "num_tokens": 319975003.0, + "step": 12365 + }, + { + "epoch": 1.3580057105205359, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0290801525115967, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7171280980110168, + "num_tokens": 320005624.0, + "step": 12366 + }, + { + "epoch": 1.3581155282231496, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.4600629806518555, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7188857197761536, + "num_tokens": 320030292.0, + "step": 12367 + }, + { + "epoch": 1.3582253459257632, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.30690336227417, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7087328433990479, + "num_tokens": 320054812.0, + "step": 12368 + }, + { + "epoch": 1.358335163628377, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.091158866882324, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7179875373840332, + "num_tokens": 320083315.0, + "step": 12369 + }, + { + "epoch": 1.3584449813309907, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3401174545288086, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.7107002139091492, + "num_tokens": 320107574.0, + "step": 12370 + }, + { + "epoch": 1.3585547990336042, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.061675548553467, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.702552080154419, + "num_tokens": 320137930.0, + "step": 12371 + }, + { + "epoch": 1.3586646167362177, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.519105911254883, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7152639031410217, + "num_tokens": 320160452.0, + "step": 12372 + }, + { + "epoch": 1.3587744344388315, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.368558883666992, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7075550556182861, + "num_tokens": 320186376.0, + "step": 12373 + }, + { + "epoch": 1.3588842521414453, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3338208198547363, + "learning_rate": 1e-06, + "loss": 1.0164, + "mean_token_accuracy": 0.7060067057609558, + "num_tokens": 320212668.0, + "step": 12374 + }, + { + "epoch": 1.3589940698440588, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.033064603805542, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.723240852355957, + "num_tokens": 320242462.0, + "step": 12375 + }, + { + "epoch": 1.3591038875466726, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0807907581329346, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.730858325958252, + "num_tokens": 320270761.0, + "step": 12376 + }, + { + "epoch": 1.359213705249286, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.14326548576355, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7313529849052429, + "num_tokens": 320299388.0, + "step": 12377 + }, + { + "epoch": 1.3593235229518998, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 1.9675562381744385, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7242361307144165, + "num_tokens": 320329246.0, + "step": 12378 + }, + { + "epoch": 1.3594333406545136, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.180433988571167, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7153888940811157, + "num_tokens": 320357784.0, + "step": 12379 + }, + { + "epoch": 1.3595431583571271, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.0871963500976562, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7206439971923828, + "num_tokens": 320388559.0, + "step": 12380 + }, + { + "epoch": 1.359652976059741, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.37937068939209, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7083360552787781, + "num_tokens": 320413902.0, + "step": 12381 + }, + { + "epoch": 1.3597627937623544, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.6131772994995117, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7205262184143066, + "num_tokens": 320433528.0, + "step": 12382 + }, + { + "epoch": 1.3598726114649682, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.236988067626953, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7001381516456604, + "num_tokens": 320461910.0, + "step": 12383 + }, + { + "epoch": 1.359982429167582, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3974905014038086, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7149083614349365, + "num_tokens": 320488243.0, + "step": 12384 + }, + { + "epoch": 1.3600922468701955, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.34275221824646, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7014005184173584, + "num_tokens": 320513492.0, + "step": 12385 + }, + { + "epoch": 1.360202064572809, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.196532964706421, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7039071321487427, + "num_tokens": 320541935.0, + "step": 12386 + }, + { + "epoch": 1.3603118822754228, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.7057714462280273, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.7478561401367188, + "num_tokens": 320559125.0, + "step": 12387 + }, + { + "epoch": 1.3604216999780365, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.279054641723633, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.718393862247467, + "num_tokens": 320583098.0, + "step": 12388 + }, + { + "epoch": 1.36053151768065, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.4652507305145264, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7185924649238586, + "num_tokens": 320603576.0, + "step": 12389 + }, + { + "epoch": 1.3606413353832638, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 1.7781606912612915, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.6962971687316895, + "num_tokens": 320644237.0, + "step": 12390 + }, + { + "epoch": 1.3607511530858774, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1698923110961914, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7003292441368103, + "num_tokens": 320673042.0, + "step": 12391 + }, + { + "epoch": 1.360860970788491, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.7469446659088135, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.720426619052887, + "num_tokens": 320696520.0, + "step": 12392 + }, + { + "epoch": 1.3609707884911049, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3017873764038086, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7421095371246338, + "num_tokens": 320719902.0, + "step": 12393 + }, + { + "epoch": 1.3610806061937184, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 1.938207983970642, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7193970084190369, + "num_tokens": 320753425.0, + "step": 12394 + }, + { + "epoch": 1.3611904238963322, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1559102535247803, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.717517614364624, + "num_tokens": 320780438.0, + "step": 12395 + }, + { + "epoch": 1.3613002415989457, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.424842596054077, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7352888584136963, + "num_tokens": 320801567.0, + "step": 12396 + }, + { + "epoch": 1.3614100593015594, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.380023717880249, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7251833081245422, + "num_tokens": 320825226.0, + "step": 12397 + }, + { + "epoch": 1.3615198770041732, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.467393636703491, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.7374981641769409, + "num_tokens": 320845317.0, + "step": 12398 + }, + { + "epoch": 1.3616296947067867, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.7532761096954346, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7265913486480713, + "num_tokens": 320863226.0, + "step": 12399 + }, + { + "epoch": 1.3617395124094003, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.5617189407348633, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7327265739440918, + "num_tokens": 320884355.0, + "step": 12400 + }, + { + "epoch": 1.361849330112014, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.3625950813293457, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.717435359954834, + "num_tokens": 320909532.0, + "step": 12401 + }, + { + "epoch": 1.3619591478146278, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.322079658508301, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7107325792312622, + "num_tokens": 320937048.0, + "step": 12402 + }, + { + "epoch": 1.3620689655172413, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.273102283477783, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7211431264877319, + "num_tokens": 320960287.0, + "step": 12403 + }, + { + "epoch": 1.362178783219855, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.7065317630767822, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.7078988552093506, + "num_tokens": 320979821.0, + "step": 12404 + }, + { + "epoch": 1.3622886009224686, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.4297759532928467, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7038608193397522, + "num_tokens": 321002289.0, + "step": 12405 + }, + { + "epoch": 1.3623984186250824, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1219449043273926, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.6992522478103638, + "num_tokens": 321031301.0, + "step": 12406 + }, + { + "epoch": 1.3625082363276961, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.431519031524658, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7082220315933228, + "num_tokens": 321054918.0, + "step": 12407 + }, + { + "epoch": 1.3626180540303097, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.1830148696899414, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7308824062347412, + "num_tokens": 321081261.0, + "step": 12408 + }, + { + "epoch": 1.3627278717329234, + "ewc_loss": 7.3909759521484375e-06, + "grad_norm": 2.119213104248047, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7322178483009338, + "num_tokens": 321109810.0, + "step": 12409 + }, + { + "epoch": 1.362837689435537, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4264731407165527, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7235568761825562, + "num_tokens": 321132982.0, + "step": 12410 + }, + { + "epoch": 1.3629475071381507, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.9768708944320679, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7245767116546631, + "num_tokens": 321162779.0, + "step": 12411 + }, + { + "epoch": 1.3630573248407643, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.5654568672180176, + "learning_rate": 1e-06, + "loss": 0.7654, + "mean_token_accuracy": 0.7564693689346313, + "num_tokens": 321181944.0, + "step": 12412 + }, + { + "epoch": 1.363167142543378, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.056934118270874, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7077521085739136, + "num_tokens": 321210380.0, + "step": 12413 + }, + { + "epoch": 1.3632769602459915, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2349817752838135, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7189759016036987, + "num_tokens": 321235617.0, + "step": 12414 + }, + { + "epoch": 1.3633867779486053, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.302271604537964, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7219526171684265, + "num_tokens": 321260493.0, + "step": 12415 + }, + { + "epoch": 1.363496595651219, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4307353496551514, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7383537292480469, + "num_tokens": 321282094.0, + "step": 12416 + }, + { + "epoch": 1.3636064133538326, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.0075716972351074, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7107566595077515, + "num_tokens": 321313841.0, + "step": 12417 + }, + { + "epoch": 1.3637162310564463, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4298198223114014, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.723270058631897, + "num_tokens": 321336344.0, + "step": 12418 + }, + { + "epoch": 1.3638260487590599, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.023484230041504, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.7023485898971558, + "num_tokens": 321369387.0, + "step": 12419 + }, + { + "epoch": 1.3639358664616736, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.268791675567627, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7290878891944885, + "num_tokens": 321395461.0, + "step": 12420 + }, + { + "epoch": 1.3640456841642874, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1885035037994385, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7332836985588074, + "num_tokens": 321420665.0, + "step": 12421 + }, + { + "epoch": 1.364155501866901, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.187994956970215, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7279886603355408, + "num_tokens": 321446602.0, + "step": 12422 + }, + { + "epoch": 1.3642653195695145, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2484848499298096, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.709015429019928, + "num_tokens": 321472037.0, + "step": 12423 + }, + { + "epoch": 1.3643751372721282, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1643471717834473, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7203245162963867, + "num_tokens": 321498650.0, + "step": 12424 + }, + { + "epoch": 1.364484954974742, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2464706897735596, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7005641460418701, + "num_tokens": 321524056.0, + "step": 12425 + }, + { + "epoch": 1.3645947726773555, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.3535478115081787, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7004716992378235, + "num_tokens": 321548443.0, + "step": 12426 + }, + { + "epoch": 1.3647045903799693, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1509127616882324, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7048820853233337, + "num_tokens": 321578386.0, + "step": 12427 + }, + { + "epoch": 1.3648144080825828, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.044102907180786, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7209585309028625, + "num_tokens": 321607798.0, + "step": 12428 + }, + { + "epoch": 1.3649242257851966, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1248600482940674, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6906664371490479, + "num_tokens": 321637720.0, + "step": 12429 + }, + { + "epoch": 1.3650340434878103, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2935869693756104, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.6981478333473206, + "num_tokens": 321665072.0, + "step": 12430 + }, + { + "epoch": 1.3651438611904239, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2141222953796387, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7298256754875183, + "num_tokens": 321690054.0, + "step": 12431 + }, + { + "epoch": 1.3652536788930376, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.9511053562164307, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7301164865493774, + "num_tokens": 321722282.0, + "step": 12432 + }, + { + "epoch": 1.3653634965956511, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.0494508743286133, + "learning_rate": 1e-06, + "loss": 0.9816, + "mean_token_accuracy": 0.6974862813949585, + "num_tokens": 321754184.0, + "step": 12433 + }, + { + "epoch": 1.365473314298265, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.584456443786621, + "learning_rate": 1e-06, + "loss": 0.7423, + "mean_token_accuracy": 0.7614837884902954, + "num_tokens": 321774303.0, + "step": 12434 + }, + { + "epoch": 1.3655831320008787, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2316131591796875, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7192065715789795, + "num_tokens": 321799656.0, + "step": 12435 + }, + { + "epoch": 1.3656929497034922, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.7788445949554443, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7188888788223267, + "num_tokens": 321818801.0, + "step": 12436 + }, + { + "epoch": 1.3658027674061057, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2674739360809326, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7071921825408936, + "num_tokens": 321844992.0, + "step": 12437 + }, + { + "epoch": 1.3659125851087195, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.5214383602142334, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7204718589782715, + "num_tokens": 321866970.0, + "step": 12438 + }, + { + "epoch": 1.3660224028113332, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2313520908355713, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7531686425209045, + "num_tokens": 321891675.0, + "step": 12439 + }, + { + "epoch": 1.3661322205139468, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.8496379852294922, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7232609391212463, + "num_tokens": 321926398.0, + "step": 12440 + }, + { + "epoch": 1.3662420382165605, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.282381772994995, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7321993112564087, + "num_tokens": 321950293.0, + "step": 12441 + }, + { + "epoch": 1.366351855919174, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.104116916656494, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7228749990463257, + "num_tokens": 321979849.0, + "step": 12442 + }, + { + "epoch": 1.3664616736217878, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.503485679626465, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7217728495597839, + "num_tokens": 322001928.0, + "step": 12443 + }, + { + "epoch": 1.3665714913244016, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1955676078796387, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7251471281051636, + "num_tokens": 322029905.0, + "step": 12444 + }, + { + "epoch": 1.3666813090270151, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.283168077468872, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7124736309051514, + "num_tokens": 322055755.0, + "step": 12445 + }, + { + "epoch": 1.3667911267296289, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.3122122287750244, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.6988445520401001, + "num_tokens": 322081185.0, + "step": 12446 + }, + { + "epoch": 1.3669009444322424, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.378896713256836, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7142205238342285, + "num_tokens": 322107051.0, + "step": 12447 + }, + { + "epoch": 1.3670107621348562, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.3707706928253174, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7359032034873962, + "num_tokens": 322130687.0, + "step": 12448 + }, + { + "epoch": 1.36712057983747, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.017102003097534, + "learning_rate": 1e-06, + "loss": 1.0596, + "mean_token_accuracy": 0.6773961782455444, + "num_tokens": 322164725.0, + "step": 12449 + }, + { + "epoch": 1.3672303975400835, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.9863920211791992, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7210988998413086, + "num_tokens": 322193166.0, + "step": 12450 + }, + { + "epoch": 1.367340215242697, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4491353034973145, + "learning_rate": 1e-06, + "loss": 0.7855, + "mean_token_accuracy": 0.7490574717521667, + "num_tokens": 322212005.0, + "step": 12451 + }, + { + "epoch": 1.3674500329453108, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.247840166091919, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7007104158401489, + "num_tokens": 322239553.0, + "step": 12452 + }, + { + "epoch": 1.3675598506479245, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.096919298171997, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7195172309875488, + "num_tokens": 322269202.0, + "step": 12453 + }, + { + "epoch": 1.367669668350538, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2106826305389404, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7031821012496948, + "num_tokens": 322297129.0, + "step": 12454 + }, + { + "epoch": 1.3677794860531518, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.972036361694336, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7041698694229126, + "num_tokens": 322331302.0, + "step": 12455 + }, + { + "epoch": 1.3678893037557653, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.304253578186035, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7160728573799133, + "num_tokens": 322356109.0, + "step": 12456 + }, + { + "epoch": 1.367999121458379, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2372419834136963, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.6998312473297119, + "num_tokens": 322382372.0, + "step": 12457 + }, + { + "epoch": 1.3681089391609929, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.078519105911255, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.739222526550293, + "num_tokens": 322408429.0, + "step": 12458 + }, + { + "epoch": 1.3682187568636064, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.286515235900879, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7241497039794922, + "num_tokens": 322433048.0, + "step": 12459 + }, + { + "epoch": 1.3683285745662201, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.239248037338257, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7296230792999268, + "num_tokens": 322458545.0, + "step": 12460 + }, + { + "epoch": 1.3684383922688337, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.708861827850342, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7407984733581543, + "num_tokens": 322478183.0, + "step": 12461 + }, + { + "epoch": 1.3685482099714474, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4827799797058105, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7168376445770264, + "num_tokens": 322499421.0, + "step": 12462 + }, + { + "epoch": 1.3686580276740612, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.101525068283081, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7222058773040771, + "num_tokens": 322528571.0, + "step": 12463 + }, + { + "epoch": 1.3687678453766747, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.423736572265625, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7058970928192139, + "num_tokens": 322552969.0, + "step": 12464 + }, + { + "epoch": 1.3688776630792883, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4683680534362793, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7188988924026489, + "num_tokens": 322576081.0, + "step": 12465 + }, + { + "epoch": 1.368987480781902, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.1774871349334717, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7204864025115967, + "num_tokens": 322605259.0, + "step": 12466 + }, + { + "epoch": 1.3690972984845158, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.186523199081421, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7287510633468628, + "num_tokens": 322632366.0, + "step": 12467 + }, + { + "epoch": 1.3692071161871293, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.312091827392578, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6871466040611267, + "num_tokens": 322658014.0, + "step": 12468 + }, + { + "epoch": 1.369316933889743, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.21217679977417, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7189788222312927, + "num_tokens": 322685064.0, + "step": 12469 + }, + { + "epoch": 1.3694267515923566, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.218355417251587, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7310729622840881, + "num_tokens": 322711501.0, + "step": 12470 + }, + { + "epoch": 1.3695365692949704, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.2140395641326904, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7349724769592285, + "num_tokens": 322738897.0, + "step": 12471 + }, + { + "epoch": 1.3696463869975841, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.3696341514587402, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7251456379890442, + "num_tokens": 322762890.0, + "step": 12472 + }, + { + "epoch": 1.3697562047001977, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.244939088821411, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7063766717910767, + "num_tokens": 322788541.0, + "step": 12473 + }, + { + "epoch": 1.3698660224028114, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.3128676414489746, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7143234014511108, + "num_tokens": 322812278.0, + "step": 12474 + }, + { + "epoch": 1.369975840105425, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.0846033096313477, + "learning_rate": 1e-06, + "loss": 1.0469, + "mean_token_accuracy": 0.687543511390686, + "num_tokens": 322844266.0, + "step": 12475 + }, + { + "epoch": 1.3700856578080387, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.36906361579895, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6983852982521057, + "num_tokens": 322867518.0, + "step": 12476 + }, + { + "epoch": 1.3701954755106522, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.4154512882232666, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7240502834320068, + "num_tokens": 322891856.0, + "step": 12477 + }, + { + "epoch": 1.370305293213266, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.350123882293701, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.6977939605712891, + "num_tokens": 322916582.0, + "step": 12478 + }, + { + "epoch": 1.3704151109158795, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.5907795429229736, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7173101902008057, + "num_tokens": 322938281.0, + "step": 12479 + }, + { + "epoch": 1.3705249286184933, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.353027582168579, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7203589677810669, + "num_tokens": 322961516.0, + "step": 12480 + }, + { + "epoch": 1.370634746321107, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.736196994781494, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7497154474258423, + "num_tokens": 322979819.0, + "step": 12481 + }, + { + "epoch": 1.3707445640237206, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 1.8913136720657349, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7094913721084595, + "num_tokens": 323013915.0, + "step": 12482 + }, + { + "epoch": 1.3708543817263343, + "ewc_loss": 7.420778274536133e-06, + "grad_norm": 2.456481695175171, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7382051944732666, + "num_tokens": 323036826.0, + "step": 12483 + }, + { + "epoch": 1.3709641994289479, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.041947603225708, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.708891749382019, + "num_tokens": 323067041.0, + "step": 12484 + }, + { + "epoch": 1.3710740171315616, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3732728958129883, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7197573184967041, + "num_tokens": 323090266.0, + "step": 12485 + }, + { + "epoch": 1.3711838348341754, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3401505947113037, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7030237913131714, + "num_tokens": 323116084.0, + "step": 12486 + }, + { + "epoch": 1.371293652536789, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.077202081680298, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.72147136926651, + "num_tokens": 323145061.0, + "step": 12487 + }, + { + "epoch": 1.3714034702394025, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4210939407348633, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.723173201084137, + "num_tokens": 323167555.0, + "step": 12488 + }, + { + "epoch": 1.3715132879420162, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.435058116912842, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7311230301856995, + "num_tokens": 323189204.0, + "step": 12489 + }, + { + "epoch": 1.37162310564463, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.33847975730896, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7305558919906616, + "num_tokens": 323212801.0, + "step": 12490 + }, + { + "epoch": 1.3717329233472435, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0198957920074463, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7316635847091675, + "num_tokens": 323239307.0, + "step": 12491 + }, + { + "epoch": 1.3718427410498573, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.323817729949951, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7109580636024475, + "num_tokens": 323264180.0, + "step": 12492 + }, + { + "epoch": 1.3719525587524708, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1299588680267334, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7040822505950928, + "num_tokens": 323292103.0, + "step": 12493 + }, + { + "epoch": 1.3720623764550846, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.330231189727783, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7165728807449341, + "num_tokens": 323315127.0, + "step": 12494 + }, + { + "epoch": 1.3721721941576983, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.322624444961548, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6965365409851074, + "num_tokens": 323340495.0, + "step": 12495 + }, + { + "epoch": 1.3722820118603118, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3596298694610596, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6956844329833984, + "num_tokens": 323366720.0, + "step": 12496 + }, + { + "epoch": 1.3723918295629256, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3648269176483154, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7283294200897217, + "num_tokens": 323390323.0, + "step": 12497 + }, + { + "epoch": 1.3725016472655391, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3551557064056396, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7293027639389038, + "num_tokens": 323412899.0, + "step": 12498 + }, + { + "epoch": 1.372611464968153, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.8640403747558594, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7339572310447693, + "num_tokens": 323430627.0, + "step": 12499 + }, + { + "epoch": 1.3727212826707667, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.5715341567993164, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7339591383934021, + "num_tokens": 323451115.0, + "step": 12500 + }, + { + "epoch": 1.3728311003733802, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2983181476593018, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.713729739189148, + "num_tokens": 323477932.0, + "step": 12501 + }, + { + "epoch": 1.3729409180759937, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0481247901916504, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.711434543132782, + "num_tokens": 323507554.0, + "step": 12502 + }, + { + "epoch": 1.3730507357786075, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.189688205718994, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7046924829483032, + "num_tokens": 323536096.0, + "step": 12503 + }, + { + "epoch": 1.3731605534812212, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0642993450164795, + "learning_rate": 1e-06, + "loss": 1.0207, + "mean_token_accuracy": 0.6856963038444519, + "num_tokens": 323568211.0, + "step": 12504 + }, + { + "epoch": 1.3732703711838348, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.143261194229126, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7036085724830627, + "num_tokens": 323597033.0, + "step": 12505 + }, + { + "epoch": 1.3733801888864485, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.16719388961792, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7352455854415894, + "num_tokens": 323623987.0, + "step": 12506 + }, + { + "epoch": 1.373490006589062, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.090792417526245, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7203030586242676, + "num_tokens": 323651896.0, + "step": 12507 + }, + { + "epoch": 1.3735998242916758, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1668434143066406, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7033016085624695, + "num_tokens": 323679655.0, + "step": 12508 + }, + { + "epoch": 1.3737096419942896, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.292025327682495, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7120395302772522, + "num_tokens": 323704418.0, + "step": 12509 + }, + { + "epoch": 1.3738194596969031, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.5775933265686035, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7124649882316589, + "num_tokens": 323727290.0, + "step": 12510 + }, + { + "epoch": 1.3739292773995169, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1425435543060303, + "learning_rate": 1e-06, + "loss": 0.9749, + "mean_token_accuracy": 0.699864387512207, + "num_tokens": 323756261.0, + "step": 12511 + }, + { + "epoch": 1.3740390951021304, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 1.9906402826309204, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7009541988372803, + "num_tokens": 323785724.0, + "step": 12512 + }, + { + "epoch": 1.3741489128047442, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1830718517303467, + "learning_rate": 1e-06, + "loss": 0.9854, + "mean_token_accuracy": 0.6994473934173584, + "num_tokens": 323812965.0, + "step": 12513 + }, + { + "epoch": 1.374258730507358, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.188037633895874, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7397688627243042, + "num_tokens": 323838244.0, + "step": 12514 + }, + { + "epoch": 1.3743685482099715, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4861044883728027, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.747052788734436, + "num_tokens": 323858223.0, + "step": 12515 + }, + { + "epoch": 1.374478365912585, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 1.9906727075576782, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7077581882476807, + "num_tokens": 323890650.0, + "step": 12516 + }, + { + "epoch": 1.3745881836151987, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1917407512664795, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7172776460647583, + "num_tokens": 323915138.0, + "step": 12517 + }, + { + "epoch": 1.3746980013178125, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.397852659225464, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.7014876008033752, + "num_tokens": 323939118.0, + "step": 12518 + }, + { + "epoch": 1.374807819020426, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 1.9966316223144531, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7158282995223999, + "num_tokens": 323967959.0, + "step": 12519 + }, + { + "epoch": 1.3749176367230398, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.450979471206665, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7217528820037842, + "num_tokens": 323989323.0, + "step": 12520 + }, + { + "epoch": 1.3750274544256533, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.5120396614074707, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7279471158981323, + "num_tokens": 324010929.0, + "step": 12521 + }, + { + "epoch": 1.375137272128267, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.6334426403045654, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7484827041625977, + "num_tokens": 324029764.0, + "step": 12522 + }, + { + "epoch": 1.3752470898308808, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.473069667816162, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7376614809036255, + "num_tokens": 324051254.0, + "step": 12523 + }, + { + "epoch": 1.3753569075334944, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4480087757110596, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7244972586631775, + "num_tokens": 324073161.0, + "step": 12524 + }, + { + "epoch": 1.3754667252361081, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2605600357055664, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7231699228286743, + "num_tokens": 324096249.0, + "step": 12525 + }, + { + "epoch": 1.3755765429387217, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3757054805755615, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7166190147399902, + "num_tokens": 324119618.0, + "step": 12526 + }, + { + "epoch": 1.3756863606413354, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1403517723083496, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7162290811538696, + "num_tokens": 324149285.0, + "step": 12527 + }, + { + "epoch": 1.3757961783439492, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.5296871662139893, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7173284292221069, + "num_tokens": 324170406.0, + "step": 12528 + }, + { + "epoch": 1.3759059960465627, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1490418910980225, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7015510201454163, + "num_tokens": 324198163.0, + "step": 12529 + }, + { + "epoch": 1.3760158137491763, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1600229740142822, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7153798341751099, + "num_tokens": 324226223.0, + "step": 12530 + }, + { + "epoch": 1.37612563145179, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.258408784866333, + "learning_rate": 1e-06, + "loss": 0.97, + "mean_token_accuracy": 0.7036449313163757, + "num_tokens": 324252547.0, + "step": 12531 + }, + { + "epoch": 1.3762354491544038, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.876622200012207, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7303814888000488, + "num_tokens": 324269247.0, + "step": 12532 + }, + { + "epoch": 1.3763452668570173, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.155052900314331, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.707987904548645, + "num_tokens": 324294598.0, + "step": 12533 + }, + { + "epoch": 1.376455084559631, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.144451379776001, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7087450623512268, + "num_tokens": 324322960.0, + "step": 12534 + }, + { + "epoch": 1.3765649022622446, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.422917604446411, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7469973564147949, + "num_tokens": 324345051.0, + "step": 12535 + }, + { + "epoch": 1.3766747199648584, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2683820724487305, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7058672308921814, + "num_tokens": 324370134.0, + "step": 12536 + }, + { + "epoch": 1.376784537667472, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1769938468933105, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6939712762832642, + "num_tokens": 324396827.0, + "step": 12537 + }, + { + "epoch": 1.3768943553700856, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.386780261993408, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7133674025535583, + "num_tokens": 324419711.0, + "step": 12538 + }, + { + "epoch": 1.3770041730726994, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3590986728668213, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7187805771827698, + "num_tokens": 324442608.0, + "step": 12539 + }, + { + "epoch": 1.377113990775313, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.402730941772461, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7196514010429382, + "num_tokens": 324463862.0, + "step": 12540 + }, + { + "epoch": 1.3772238084779267, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.443333864212036, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7250792980194092, + "num_tokens": 324485779.0, + "step": 12541 + }, + { + "epoch": 1.3773336261805402, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.206547975540161, + "learning_rate": 1e-06, + "loss": 1.0053, + "mean_token_accuracy": 0.7002017498016357, + "num_tokens": 324513413.0, + "step": 12542 + }, + { + "epoch": 1.377443443883154, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 1.963546872138977, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7268761396408081, + "num_tokens": 324543901.0, + "step": 12543 + }, + { + "epoch": 1.3775532615857675, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.454216718673706, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7073367834091187, + "num_tokens": 324566134.0, + "step": 12544 + }, + { + "epoch": 1.3776630792883813, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.144874334335327, + "learning_rate": 1e-06, + "loss": 1.0221, + "mean_token_accuracy": 0.6939265131950378, + "num_tokens": 324594770.0, + "step": 12545 + }, + { + "epoch": 1.377772896990995, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2057547569274902, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7097858190536499, + "num_tokens": 324619443.0, + "step": 12546 + }, + { + "epoch": 1.3778827146936086, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.5419867038726807, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7271895408630371, + "num_tokens": 324639903.0, + "step": 12547 + }, + { + "epoch": 1.3779925323962223, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3203413486480713, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7312696576118469, + "num_tokens": 324661797.0, + "step": 12548 + }, + { + "epoch": 1.3781023500988359, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.172427177429199, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7368022203445435, + "num_tokens": 324686076.0, + "step": 12549 + }, + { + "epoch": 1.3782121678014496, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3109564781188965, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7379631996154785, + "num_tokens": 324709315.0, + "step": 12550 + }, + { + "epoch": 1.3783219855040634, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2210910320281982, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7327669858932495, + "num_tokens": 324733142.0, + "step": 12551 + }, + { + "epoch": 1.378431803206677, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.342014789581299, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7386950254440308, + "num_tokens": 324756808.0, + "step": 12552 + }, + { + "epoch": 1.3785416209092904, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.632424831390381, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7155437469482422, + "num_tokens": 324777993.0, + "step": 12553 + }, + { + "epoch": 1.3786514386119042, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0629587173461914, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.743177056312561, + "num_tokens": 324806322.0, + "step": 12554 + }, + { + "epoch": 1.378761256314518, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 1.9905675649642944, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7030125260353088, + "num_tokens": 324837557.0, + "step": 12555 + }, + { + "epoch": 1.3788710740171315, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.6067864894866943, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7031541466712952, + "num_tokens": 324857980.0, + "step": 12556 + }, + { + "epoch": 1.3789808917197452, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.232050895690918, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.707683801651001, + "num_tokens": 324887487.0, + "step": 12557 + }, + { + "epoch": 1.3790907094223588, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.469536781311035, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7451095581054688, + "num_tokens": 324908476.0, + "step": 12558 + }, + { + "epoch": 1.3792005271249725, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2570807933807373, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7253748178482056, + "num_tokens": 324934794.0, + "step": 12559 + }, + { + "epoch": 1.3793103448275863, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0311899185180664, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7358458042144775, + "num_tokens": 324965560.0, + "step": 12560 + }, + { + "epoch": 1.3794201625301998, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2019217014312744, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7183140516281128, + "num_tokens": 324992511.0, + "step": 12561 + }, + { + "epoch": 1.3795299802328136, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.227660655975342, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7263239026069641, + "num_tokens": 325019487.0, + "step": 12562 + }, + { + "epoch": 1.3796397979354271, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.360368251800537, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7202672362327576, + "num_tokens": 325043901.0, + "step": 12563 + }, + { + "epoch": 1.3797496156380409, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.213069200515747, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7300176620483398, + "num_tokens": 325068013.0, + "step": 12564 + }, + { + "epoch": 1.3798594333406546, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2902657985687256, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7082683444023132, + "num_tokens": 325094329.0, + "step": 12565 + }, + { + "epoch": 1.3799692510432682, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0884006023406982, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7229270935058594, + "num_tokens": 325121449.0, + "step": 12566 + }, + { + "epoch": 1.3800790687458817, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1540865898132324, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7245528697967529, + "num_tokens": 325146118.0, + "step": 12567 + }, + { + "epoch": 1.3801888864484955, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0674164295196533, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7105360627174377, + "num_tokens": 325177571.0, + "step": 12568 + }, + { + "epoch": 1.3802987041511092, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.0709235668182373, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6956166625022888, + "num_tokens": 325207045.0, + "step": 12569 + }, + { + "epoch": 1.3804085218537228, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.205881118774414, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7147495746612549, + "num_tokens": 325233060.0, + "step": 12570 + }, + { + "epoch": 1.3805183395563365, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3694562911987305, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7316762208938599, + "num_tokens": 325255279.0, + "step": 12571 + }, + { + "epoch": 1.38062815725895, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3482000827789307, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7171146869659424, + "num_tokens": 325279494.0, + "step": 12572 + }, + { + "epoch": 1.3807379749615638, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.150578498840332, + "learning_rate": 1e-06, + "loss": 0.7921, + "mean_token_accuracy": 0.7504935264587402, + "num_tokens": 325303432.0, + "step": 12573 + }, + { + "epoch": 1.3808477926641776, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.263317584991455, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7011178135871887, + "num_tokens": 325329242.0, + "step": 12574 + }, + { + "epoch": 1.380957610366791, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.177116870880127, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7139086723327637, + "num_tokens": 325355446.0, + "step": 12575 + }, + { + "epoch": 1.3810674280694049, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.163402795791626, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7115322351455688, + "num_tokens": 325382954.0, + "step": 12576 + }, + { + "epoch": 1.3811772457720184, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2614176273345947, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7056170701980591, + "num_tokens": 325407603.0, + "step": 12577 + }, + { + "epoch": 1.3812870634746321, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.136307716369629, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7015587091445923, + "num_tokens": 325434185.0, + "step": 12578 + }, + { + "epoch": 1.381396881177246, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.1273810863494873, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7197948694229126, + "num_tokens": 325461343.0, + "step": 12579 + }, + { + "epoch": 1.3815066988798594, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4382853507995605, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7105208039283752, + "num_tokens": 325483556.0, + "step": 12580 + }, + { + "epoch": 1.381616516582473, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.2310237884521484, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.710129976272583, + "num_tokens": 325510162.0, + "step": 12581 + }, + { + "epoch": 1.3817263342850867, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.3570282459259033, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.751182496547699, + "num_tokens": 325533463.0, + "step": 12582 + }, + { + "epoch": 1.3818361519877005, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.239276647567749, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7470675110816956, + "num_tokens": 325558235.0, + "step": 12583 + }, + { + "epoch": 1.381945969690314, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.4511220455169678, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7238539457321167, + "num_tokens": 325579757.0, + "step": 12584 + }, + { + "epoch": 1.3820557873929278, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.286233425140381, + "learning_rate": 1e-06, + "loss": 0.9946, + "mean_token_accuracy": 0.6983600854873657, + "num_tokens": 325605573.0, + "step": 12585 + }, + { + "epoch": 1.3821656050955413, + "ewc_loss": 7.450580596923828e-06, + "grad_norm": 2.324997901916504, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7179459929466248, + "num_tokens": 325629148.0, + "step": 12586 + }, + { + "epoch": 1.382275422798155, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1985068321228027, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7162314653396606, + "num_tokens": 325655848.0, + "step": 12587 + }, + { + "epoch": 1.3823852405007688, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.696716785430908, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7215704321861267, + "num_tokens": 325679168.0, + "step": 12588 + }, + { + "epoch": 1.3824950582033824, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.081047534942627, + "learning_rate": 1e-06, + "loss": 1.0061, + "mean_token_accuracy": 0.6998959183692932, + "num_tokens": 325711489.0, + "step": 12589 + }, + { + "epoch": 1.3826048759059961, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1873996257781982, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6908668279647827, + "num_tokens": 325740922.0, + "step": 12590 + }, + { + "epoch": 1.3827146936086097, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.086681365966797, + "learning_rate": 1e-06, + "loss": 0.9725, + "mean_token_accuracy": 0.7046098709106445, + "num_tokens": 325770207.0, + "step": 12591 + }, + { + "epoch": 1.3828245113112234, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.036508560180664, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7434606552124023, + "num_tokens": 325798524.0, + "step": 12592 + }, + { + "epoch": 1.382934329013837, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.283304452896118, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7009850144386292, + "num_tokens": 325825687.0, + "step": 12593 + }, + { + "epoch": 1.3830441467164507, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.146599531173706, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7180942296981812, + "num_tokens": 325854967.0, + "step": 12594 + }, + { + "epoch": 1.3831539644190642, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 1.991931676864624, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6941112279891968, + "num_tokens": 325887490.0, + "step": 12595 + }, + { + "epoch": 1.383263782121678, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.5859692096710205, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7056447863578796, + "num_tokens": 325908570.0, + "step": 12596 + }, + { + "epoch": 1.3833735998242918, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.598625421524048, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7219290733337402, + "num_tokens": 325928528.0, + "step": 12597 + }, + { + "epoch": 1.3834834175269053, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.171552896499634, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7065690755844116, + "num_tokens": 325956713.0, + "step": 12598 + }, + { + "epoch": 1.383593235229519, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.128730535507202, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7145456075668335, + "num_tokens": 325984839.0, + "step": 12599 + }, + { + "epoch": 1.3837030529321326, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2614123821258545, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7279125452041626, + "num_tokens": 326011963.0, + "step": 12600 + }, + { + "epoch": 1.3838128706347463, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2565956115722656, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7255451083183289, + "num_tokens": 326037627.0, + "step": 12601 + }, + { + "epoch": 1.38392268833736, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 1.950534701347351, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7161600589752197, + "num_tokens": 326068689.0, + "step": 12602 + }, + { + "epoch": 1.3840325060399736, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.119612216949463, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7111968398094177, + "num_tokens": 326099103.0, + "step": 12603 + }, + { + "epoch": 1.3841423237425874, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2719099521636963, + "learning_rate": 1e-06, + "loss": 1.0178, + "mean_token_accuracy": 0.6887979507446289, + "num_tokens": 326125592.0, + "step": 12604 + }, + { + "epoch": 1.384252141445201, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.05159592628479, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.717685341835022, + "num_tokens": 326154108.0, + "step": 12605 + }, + { + "epoch": 1.3843619591478147, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.30732798576355, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7256898283958435, + "num_tokens": 326178614.0, + "step": 12606 + }, + { + "epoch": 1.3844717768504282, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.128917932510376, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7034664154052734, + "num_tokens": 326205864.0, + "step": 12607 + }, + { + "epoch": 1.384581594553042, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.40269136428833, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7110268473625183, + "num_tokens": 326229546.0, + "step": 12608 + }, + { + "epoch": 1.3846914122556555, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.0827462673187256, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7050498723983765, + "num_tokens": 326257447.0, + "step": 12609 + }, + { + "epoch": 1.3848012299582693, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.3464980125427246, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7234572768211365, + "num_tokens": 326279636.0, + "step": 12610 + }, + { + "epoch": 1.384911047660883, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1375765800476074, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7055643200874329, + "num_tokens": 326307847.0, + "step": 12611 + }, + { + "epoch": 1.3850208653634966, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.3103630542755127, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7112946510314941, + "num_tokens": 326334739.0, + "step": 12612 + }, + { + "epoch": 1.3851306830661103, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.6403279304504395, + "learning_rate": 1e-06, + "loss": 0.8274, + "mean_token_accuracy": 0.7420547008514404, + "num_tokens": 326354685.0, + "step": 12613 + }, + { + "epoch": 1.3852405007687238, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1226601600646973, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.6994773745536804, + "num_tokens": 326384519.0, + "step": 12614 + }, + { + "epoch": 1.3853503184713376, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.4758355617523193, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7387402653694153, + "num_tokens": 326407163.0, + "step": 12615 + }, + { + "epoch": 1.3854601361739514, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2334370613098145, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7121724486351013, + "num_tokens": 326432569.0, + "step": 12616 + }, + { + "epoch": 1.385569953876565, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.137929677963257, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.7015427350997925, + "num_tokens": 326462060.0, + "step": 12617 + }, + { + "epoch": 1.3856797715791784, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.259355306625366, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7309813499450684, + "num_tokens": 326487727.0, + "step": 12618 + }, + { + "epoch": 1.3857895892817922, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.4644429683685303, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7326624393463135, + "num_tokens": 326510858.0, + "step": 12619 + }, + { + "epoch": 1.385899406984406, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.4194867610931396, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7286665439605713, + "num_tokens": 326532694.0, + "step": 12620 + }, + { + "epoch": 1.3860092246870195, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.188264846801758, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7277523875236511, + "num_tokens": 326560597.0, + "step": 12621 + }, + { + "epoch": 1.3861190423896332, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.5219597816467285, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6910154819488525, + "num_tokens": 326582964.0, + "step": 12622 + }, + { + "epoch": 1.3862288600922468, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1954400539398193, + "learning_rate": 1e-06, + "loss": 1.023, + "mean_token_accuracy": 0.701921820640564, + "num_tokens": 326609534.0, + "step": 12623 + }, + { + "epoch": 1.3863386777948605, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.256500720977783, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7357667684555054, + "num_tokens": 326634252.0, + "step": 12624 + }, + { + "epoch": 1.3864484954974743, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.2991185188293457, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.732474148273468, + "num_tokens": 326658361.0, + "step": 12625 + }, + { + "epoch": 1.3865583132000878, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.4705512523651123, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7356791496276855, + "num_tokens": 326678597.0, + "step": 12626 + }, + { + "epoch": 1.3866681309027016, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.1745176315307617, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.707044780254364, + "num_tokens": 326707463.0, + "step": 12627 + }, + { + "epoch": 1.3867779486053151, + "ewc_loss": 7.4803829193115234e-06, + "grad_norm": 2.03106951713562, + "learning_rate": 1e-06, + "loss": 1.0336, + "mean_token_accuracy": 0.6861054301261902, + "num_tokens": 326740808.0, + "step": 12628 + }, + { + "epoch": 1.3868877663079289, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.034754753112793, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.7118821144104004, + "num_tokens": 326770351.0, + "step": 12629 + }, + { + "epoch": 1.3869975840105426, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.201385736465454, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7274723052978516, + "num_tokens": 326795628.0, + "step": 12630 + }, + { + "epoch": 1.3871074017131562, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.026031970977783, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7106737494468689, + "num_tokens": 326828612.0, + "step": 12631 + }, + { + "epoch": 1.3872172194157697, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3242037296295166, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.728558361530304, + "num_tokens": 326852073.0, + "step": 12632 + }, + { + "epoch": 1.3873270371183835, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.202648878097534, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7431055307388306, + "num_tokens": 326877643.0, + "step": 12633 + }, + { + "epoch": 1.3874368548209972, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.289440393447876, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7091261744499207, + "num_tokens": 326902042.0, + "step": 12634 + }, + { + "epoch": 1.3875466725236107, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.2442336082458496, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7179123163223267, + "num_tokens": 326927492.0, + "step": 12635 + }, + { + "epoch": 1.3876564902262245, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.173251152038574, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7172969579696655, + "num_tokens": 326955640.0, + "step": 12636 + }, + { + "epoch": 1.387766307928838, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.124425172805786, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7135610580444336, + "num_tokens": 326982934.0, + "step": 12637 + }, + { + "epoch": 1.3878761256314518, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.664417028427124, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7110487222671509, + "num_tokens": 327003284.0, + "step": 12638 + }, + { + "epoch": 1.3879859433340656, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 1.9224187135696411, + "learning_rate": 1e-06, + "loss": 1.0279, + "mean_token_accuracy": 0.6853572130203247, + "num_tokens": 327039297.0, + "step": 12639 + }, + { + "epoch": 1.388095761036679, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.1303460597991943, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6926937103271484, + "num_tokens": 327068485.0, + "step": 12640 + }, + { + "epoch": 1.3882055787392928, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.2133893966674805, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7146717309951782, + "num_tokens": 327094322.0, + "step": 12641 + }, + { + "epoch": 1.3883153964419064, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3214480876922607, + "learning_rate": 1e-06, + "loss": 0.7901, + "mean_token_accuracy": 0.7484524846076965, + "num_tokens": 327118567.0, + "step": 12642 + }, + { + "epoch": 1.3884252141445201, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.1819708347320557, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7199476361274719, + "num_tokens": 327144865.0, + "step": 12643 + }, + { + "epoch": 1.388535031847134, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.017087459564209, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7103108763694763, + "num_tokens": 327174998.0, + "step": 12644 + }, + { + "epoch": 1.3886448495497474, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.223595142364502, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7034125924110413, + "num_tokens": 327200509.0, + "step": 12645 + }, + { + "epoch": 1.388754667252361, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.16555118560791, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7291569113731384, + "num_tokens": 327225745.0, + "step": 12646 + }, + { + "epoch": 1.3888644849549747, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.0885567665100098, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.695936918258667, + "num_tokens": 327254984.0, + "step": 12647 + }, + { + "epoch": 1.3889743026575885, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.1776366233825684, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7274868488311768, + "num_tokens": 327279223.0, + "step": 12648 + }, + { + "epoch": 1.389084120360202, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.4818038940429688, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7198374271392822, + "num_tokens": 327303319.0, + "step": 12649 + }, + { + "epoch": 1.3891939380628158, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.078834295272827, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7125826478004456, + "num_tokens": 327331675.0, + "step": 12650 + }, + { + "epoch": 1.3893037557654293, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3035240173339844, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7203311920166016, + "num_tokens": 327355990.0, + "step": 12651 + }, + { + "epoch": 1.389413573468043, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3912394046783447, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7309150099754333, + "num_tokens": 327379324.0, + "step": 12652 + }, + { + "epoch": 1.3895233911706568, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.3963022232055664, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7089727520942688, + "num_tokens": 327402498.0, + "step": 12653 + }, + { + "epoch": 1.3896332088732704, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.360243082046509, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7104948163032532, + "num_tokens": 327426181.0, + "step": 12654 + }, + { + "epoch": 1.389743026575884, + "ewc_loss": 7.510185241699219e-06, + "grad_norm": 2.019766092300415, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7123489379882812, + "num_tokens": 327455846.0, + "step": 12655 + }, + { + "epoch": 1.3898528442784976, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1862499713897705, + "learning_rate": 1e-06, + "loss": 1.0154, + "mean_token_accuracy": 0.6863913536071777, + "num_tokens": 327486203.0, + "step": 12656 + }, + { + "epoch": 1.3899626619811114, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.2500483989715576, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7413268089294434, + "num_tokens": 327510687.0, + "step": 12657 + }, + { + "epoch": 1.390072479683725, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.226922035217285, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7200208306312561, + "num_tokens": 327536849.0, + "step": 12658 + }, + { + "epoch": 1.3901822973863387, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.338073968887329, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7134124040603638, + "num_tokens": 327559433.0, + "step": 12659 + }, + { + "epoch": 1.3902921150889522, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.4323201179504395, + "learning_rate": 1e-06, + "loss": 0.7289, + "mean_token_accuracy": 0.7635947465896606, + "num_tokens": 327580154.0, + "step": 12660 + }, + { + "epoch": 1.390401932791566, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.4041025638580322, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.736199140548706, + "num_tokens": 327602813.0, + "step": 12661 + }, + { + "epoch": 1.3905117504941797, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.280503034591675, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7322938442230225, + "num_tokens": 327629603.0, + "step": 12662 + }, + { + "epoch": 1.3906215681967933, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.391942262649536, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7103133201599121, + "num_tokens": 327654648.0, + "step": 12663 + }, + { + "epoch": 1.390731385899407, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.107936382293701, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7035099864006042, + "num_tokens": 327684120.0, + "step": 12664 + }, + { + "epoch": 1.3908412036020206, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.454768657684326, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7135568857192993, + "num_tokens": 327707176.0, + "step": 12665 + }, + { + "epoch": 1.3909510213046343, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1677029132843018, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.701033353805542, + "num_tokens": 327734720.0, + "step": 12666 + }, + { + "epoch": 1.391060839007248, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.244854688644409, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7201387286186218, + "num_tokens": 327760150.0, + "step": 12667 + }, + { + "epoch": 1.3911706567098616, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.2254867553710938, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7201969623565674, + "num_tokens": 327784149.0, + "step": 12668 + }, + { + "epoch": 1.3912804744124752, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3258025646209717, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7314995527267456, + "num_tokens": 327808125.0, + "step": 12669 + }, + { + "epoch": 1.391390292115089, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.376380681991577, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7363281846046448, + "num_tokens": 327830077.0, + "step": 12670 + }, + { + "epoch": 1.3915001098177027, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.195361375808716, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7167476415634155, + "num_tokens": 327857461.0, + "step": 12671 + }, + { + "epoch": 1.3916099275203162, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.6014175415039062, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.725472092628479, + "num_tokens": 327877176.0, + "step": 12672 + }, + { + "epoch": 1.39171974522293, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.9480855464935303, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7439725399017334, + "num_tokens": 327895395.0, + "step": 12673 + }, + { + "epoch": 1.3918295629255435, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 1.9643261432647705, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7254996299743652, + "num_tokens": 327927128.0, + "step": 12674 + }, + { + "epoch": 1.3919393806281573, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.2787599563598633, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7430264353752136, + "num_tokens": 327951863.0, + "step": 12675 + }, + { + "epoch": 1.392049198330771, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1896071434020996, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.733767032623291, + "num_tokens": 327975457.0, + "step": 12676 + }, + { + "epoch": 1.3921590160333845, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.2165277004241943, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7101341485977173, + "num_tokens": 328001139.0, + "step": 12677 + }, + { + "epoch": 1.3922688337359983, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.2056005001068115, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.707173228263855, + "num_tokens": 328027431.0, + "step": 12678 + }, + { + "epoch": 1.3923786514386118, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.196455478668213, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7006672620773315, + "num_tokens": 328054469.0, + "step": 12679 + }, + { + "epoch": 1.3924884691412256, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.005875587463379, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7095970511436462, + "num_tokens": 328086474.0, + "step": 12680 + }, + { + "epoch": 1.3925982868438394, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1766042709350586, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7064148187637329, + "num_tokens": 328114272.0, + "step": 12681 + }, + { + "epoch": 1.3927081045464529, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3569517135620117, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7147117853164673, + "num_tokens": 328139879.0, + "step": 12682 + }, + { + "epoch": 1.3928179222490664, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.573641777038574, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7070220708847046, + "num_tokens": 328163458.0, + "step": 12683 + }, + { + "epoch": 1.3929277399516802, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.220651626586914, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7183598279953003, + "num_tokens": 328189557.0, + "step": 12684 + }, + { + "epoch": 1.393037557654294, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1582789421081543, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7111166715621948, + "num_tokens": 328216781.0, + "step": 12685 + }, + { + "epoch": 1.3931473753569075, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.0734164714813232, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7276164293289185, + "num_tokens": 328248094.0, + "step": 12686 + }, + { + "epoch": 1.3932571930595212, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 1.9765924215316772, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.6991456747055054, + "num_tokens": 328282358.0, + "step": 12687 + }, + { + "epoch": 1.3933670107621348, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 1.9872721433639526, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.6990525126457214, + "num_tokens": 328318452.0, + "step": 12688 + }, + { + "epoch": 1.3934768284647485, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.262471914291382, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.7041624188423157, + "num_tokens": 328344819.0, + "step": 12689 + }, + { + "epoch": 1.3935866461673623, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.5756707191467285, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.6995000839233398, + "num_tokens": 328366894.0, + "step": 12690 + }, + { + "epoch": 1.3936964638699758, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3194267749786377, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7123671174049377, + "num_tokens": 328393601.0, + "step": 12691 + }, + { + "epoch": 1.3938062815725896, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1831836700439453, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7254253625869751, + "num_tokens": 328419230.0, + "step": 12692 + }, + { + "epoch": 1.393916099275203, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.4555349349975586, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7109197974205017, + "num_tokens": 328443372.0, + "step": 12693 + }, + { + "epoch": 1.3940259169778169, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3584370613098145, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7293006777763367, + "num_tokens": 328465375.0, + "step": 12694 + }, + { + "epoch": 1.3941357346804306, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.028712272644043, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7151349782943726, + "num_tokens": 328497099.0, + "step": 12695 + }, + { + "epoch": 1.3942455523830442, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3691484928131104, + "learning_rate": 1e-06, + "loss": 0.7775, + "mean_token_accuracy": 0.7438279986381531, + "num_tokens": 328517455.0, + "step": 12696 + }, + { + "epoch": 1.3943553700856577, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.233874559402466, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7034257650375366, + "num_tokens": 328542932.0, + "step": 12697 + }, + { + "epoch": 1.3944651877882714, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.3954660892486572, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7296502590179443, + "num_tokens": 328566209.0, + "step": 12698 + }, + { + "epoch": 1.3945750054908852, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.1339964866638184, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.720360279083252, + "num_tokens": 328591339.0, + "step": 12699 + }, + { + "epoch": 1.3946848231934987, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 1.9624462127685547, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.692699134349823, + "num_tokens": 328624959.0, + "step": 12700 + }, + { + "epoch": 1.3947946408961125, + "ewc_loss": 7.539987564086914e-06, + "grad_norm": 2.254809856414795, + "learning_rate": 1e-06, + "loss": 1.0017, + "mean_token_accuracy": 0.6922183036804199, + "num_tokens": 328653922.0, + "step": 12701 + }, + { + "epoch": 1.394904458598726, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4731099605560303, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7300785183906555, + "num_tokens": 328674331.0, + "step": 12702 + }, + { + "epoch": 1.3950142763013398, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.224374532699585, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7337751388549805, + "num_tokens": 328698819.0, + "step": 12703 + }, + { + "epoch": 1.3951240940039535, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.0901718139648438, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7163547277450562, + "num_tokens": 328727751.0, + "step": 12704 + }, + { + "epoch": 1.395233911706567, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4576430320739746, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.73609459400177, + "num_tokens": 328749952.0, + "step": 12705 + }, + { + "epoch": 1.3953437294091808, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.215815305709839, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7419422268867493, + "num_tokens": 328773941.0, + "step": 12706 + }, + { + "epoch": 1.3954535471117944, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3431804180145264, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.695319414138794, + "num_tokens": 328799129.0, + "step": 12707 + }, + { + "epoch": 1.3955633648144081, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1122350692749023, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7179368734359741, + "num_tokens": 328826607.0, + "step": 12708 + }, + { + "epoch": 1.3956731825170219, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.271594762802124, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7420210242271423, + "num_tokens": 328851609.0, + "step": 12709 + }, + { + "epoch": 1.3957830002196354, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.144164562225342, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7172891497612, + "num_tokens": 328878492.0, + "step": 12710 + }, + { + "epoch": 1.395892817922249, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3354134559631348, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7227759957313538, + "num_tokens": 328903482.0, + "step": 12711 + }, + { + "epoch": 1.3960026356248627, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3900644779205322, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7266409397125244, + "num_tokens": 328925219.0, + "step": 12712 + }, + { + "epoch": 1.3961124533274765, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2417845726013184, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.715258777141571, + "num_tokens": 328949992.0, + "step": 12713 + }, + { + "epoch": 1.39622227103009, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.508009433746338, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7312561273574829, + "num_tokens": 328972712.0, + "step": 12714 + }, + { + "epoch": 1.3963320887327038, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.24074387550354, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.6997096538543701, + "num_tokens": 329000792.0, + "step": 12715 + }, + { + "epoch": 1.3964419064353173, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1124391555786133, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7273292541503906, + "num_tokens": 329028808.0, + "step": 12716 + }, + { + "epoch": 1.396551724137931, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.241398811340332, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.723042368888855, + "num_tokens": 329056023.0, + "step": 12717 + }, + { + "epoch": 1.3966615418405448, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.0344960689544678, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.7010995149612427, + "num_tokens": 329087466.0, + "step": 12718 + }, + { + "epoch": 1.3967713595431583, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2721776962280273, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7181581854820251, + "num_tokens": 329113202.0, + "step": 12719 + }, + { + "epoch": 1.396881177245772, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.292721748352051, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7054972648620605, + "num_tokens": 329138724.0, + "step": 12720 + }, + { + "epoch": 1.3969909949483856, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2390406131744385, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7017740607261658, + "num_tokens": 329165168.0, + "step": 12721 + }, + { + "epoch": 1.3971008126509994, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.335284471511841, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7086009383201599, + "num_tokens": 329189652.0, + "step": 12722 + }, + { + "epoch": 1.397210630353613, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4217371940612793, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7523002028465271, + "num_tokens": 329211107.0, + "step": 12723 + }, + { + "epoch": 1.3973204480562267, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4839930534362793, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7115086317062378, + "num_tokens": 329233763.0, + "step": 12724 + }, + { + "epoch": 1.3974302657588402, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4407973289489746, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7173824310302734, + "num_tokens": 329256232.0, + "step": 12725 + }, + { + "epoch": 1.397540083461454, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4712612628936768, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7203739881515503, + "num_tokens": 329278118.0, + "step": 12726 + }, + { + "epoch": 1.3976499011640677, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1666829586029053, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7424740791320801, + "num_tokens": 329306441.0, + "step": 12727 + }, + { + "epoch": 1.3977597188666813, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.096482992172241, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7069779634475708, + "num_tokens": 329338047.0, + "step": 12728 + }, + { + "epoch": 1.397869536569295, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4424691200256348, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7329494953155518, + "num_tokens": 329361361.0, + "step": 12729 + }, + { + "epoch": 1.3979793542719086, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.5388402938842773, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.729736328125, + "num_tokens": 329382056.0, + "step": 12730 + }, + { + "epoch": 1.3980891719745223, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.509777307510376, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7280566096305847, + "num_tokens": 329404022.0, + "step": 12731 + }, + { + "epoch": 1.398198989677136, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2004852294921875, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7119485139846802, + "num_tokens": 329430905.0, + "step": 12732 + }, + { + "epoch": 1.3983088073797496, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.375261068344116, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.724156379699707, + "num_tokens": 329453502.0, + "step": 12733 + }, + { + "epoch": 1.3984186250823631, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1776397228240967, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7053639888763428, + "num_tokens": 329482453.0, + "step": 12734 + }, + { + "epoch": 1.398528442784977, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.047175407409668, + "learning_rate": 1e-06, + "loss": 0.9982, + "mean_token_accuracy": 0.6921032071113586, + "num_tokens": 329514440.0, + "step": 12735 + }, + { + "epoch": 1.3986382604875907, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.5938079357147217, + "learning_rate": 1e-06, + "loss": 0.7671, + "mean_token_accuracy": 0.7621643543243408, + "num_tokens": 329534295.0, + "step": 12736 + }, + { + "epoch": 1.3987480781902042, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.5551540851593018, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7108518481254578, + "num_tokens": 329555527.0, + "step": 12737 + }, + { + "epoch": 1.398857895892818, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.130899667739868, + "learning_rate": 1e-06, + "loss": 1.0243, + "mean_token_accuracy": 0.6827698945999146, + "num_tokens": 329586322.0, + "step": 12738 + }, + { + "epoch": 1.3989677135954315, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2728707790374756, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7179820537567139, + "num_tokens": 329611497.0, + "step": 12739 + }, + { + "epoch": 1.3990775312980452, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2861592769622803, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7226020097732544, + "num_tokens": 329636836.0, + "step": 12740 + }, + { + "epoch": 1.399187349000659, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.4108314514160156, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7208728790283203, + "num_tokens": 329661010.0, + "step": 12741 + }, + { + "epoch": 1.3992971667032725, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.515244722366333, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7213050127029419, + "num_tokens": 329682644.0, + "step": 12742 + }, + { + "epoch": 1.3994069844058863, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.281562328338623, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.727328360080719, + "num_tokens": 329707774.0, + "step": 12743 + }, + { + "epoch": 1.3995168021084998, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.085235118865967, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7208622694015503, + "num_tokens": 329735999.0, + "step": 12744 + }, + { + "epoch": 1.3996266198111136, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1944849491119385, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6980041265487671, + "num_tokens": 329765151.0, + "step": 12745 + }, + { + "epoch": 1.3997364375137273, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1787922382354736, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7395972013473511, + "num_tokens": 329791205.0, + "step": 12746 + }, + { + "epoch": 1.3998462552163409, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.31168794631958, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.6954379081726074, + "num_tokens": 329815303.0, + "step": 12747 + }, + { + "epoch": 1.3999560729189544, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.332576274871826, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.733619213104248, + "num_tokens": 329838161.0, + "step": 12748 + }, + { + "epoch": 1.4000658906215682, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1060476303100586, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6965361833572388, + "num_tokens": 329869874.0, + "step": 12749 + }, + { + "epoch": 1.400175708324182, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.503925323486328, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7424266338348389, + "num_tokens": 329890223.0, + "step": 12750 + }, + { + "epoch": 1.4002855260267955, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.106309652328491, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7164249420166016, + "num_tokens": 329918639.0, + "step": 12751 + }, + { + "epoch": 1.4003953437294092, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.51015043258667, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7102384567260742, + "num_tokens": 329943018.0, + "step": 12752 + }, + { + "epoch": 1.4005051614320227, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.270233631134033, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7243521213531494, + "num_tokens": 329968118.0, + "step": 12753 + }, + { + "epoch": 1.4006149791346365, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.092942714691162, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7282150387763977, + "num_tokens": 329996407.0, + "step": 12754 + }, + { + "epoch": 1.4007247968372503, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.235713481903076, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7042165994644165, + "num_tokens": 330020797.0, + "step": 12755 + }, + { + "epoch": 1.4008346145398638, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2971701622009277, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7130038738250732, + "num_tokens": 330046151.0, + "step": 12756 + }, + { + "epoch": 1.4009444322424776, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.503772258758545, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7381112575531006, + "num_tokens": 330067459.0, + "step": 12757 + }, + { + "epoch": 1.401054249945091, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3515048027038574, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7346360087394714, + "num_tokens": 330090836.0, + "step": 12758 + }, + { + "epoch": 1.4011640676477048, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.5816516876220703, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.6948962211608887, + "num_tokens": 330112581.0, + "step": 12759 + }, + { + "epoch": 1.4012738853503186, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.402498722076416, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.718192458152771, + "num_tokens": 330134770.0, + "step": 12760 + }, + { + "epoch": 1.4013837030529321, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2450578212738037, + "learning_rate": 1e-06, + "loss": 1.0366, + "mean_token_accuracy": 0.6885552406311035, + "num_tokens": 330162305.0, + "step": 12761 + }, + { + "epoch": 1.4014935207555457, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 1.9337292909622192, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.6956449747085571, + "num_tokens": 330195513.0, + "step": 12762 + }, + { + "epoch": 1.4016033384581594, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.147428512573242, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7016493082046509, + "num_tokens": 330224612.0, + "step": 12763 + }, + { + "epoch": 1.4017131561607732, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.428884744644165, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7020201086997986, + "num_tokens": 330245872.0, + "step": 12764 + }, + { + "epoch": 1.4018229738633867, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.8721559047698975, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7301585078239441, + "num_tokens": 330263478.0, + "step": 12765 + }, + { + "epoch": 1.4019327915660005, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1648011207580566, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7202393412590027, + "num_tokens": 330292579.0, + "step": 12766 + }, + { + "epoch": 1.402042609268614, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.43377423286438, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7158462405204773, + "num_tokens": 330315212.0, + "step": 12767 + }, + { + "epoch": 1.4021524269712278, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.220510244369507, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7135451436042786, + "num_tokens": 330343511.0, + "step": 12768 + }, + { + "epoch": 1.4022622446738415, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1756591796875, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7294108867645264, + "num_tokens": 330371013.0, + "step": 12769 + }, + { + "epoch": 1.402372062376455, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.484044075012207, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7505652904510498, + "num_tokens": 330390565.0, + "step": 12770 + }, + { + "epoch": 1.4024818800790688, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3297252655029297, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7410482168197632, + "num_tokens": 330414047.0, + "step": 12771 + }, + { + "epoch": 1.4025916977816824, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2059202194213867, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7227137088775635, + "num_tokens": 330440831.0, + "step": 12772 + }, + { + "epoch": 1.4027015154842961, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.519771099090576, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7466719150543213, + "num_tokens": 330460607.0, + "step": 12773 + }, + { + "epoch": 1.4028113331869096, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 1.9781190156936646, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6969549059867859, + "num_tokens": 330493053.0, + "step": 12774 + }, + { + "epoch": 1.4029211508895234, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 1.880408525466919, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7188198566436768, + "num_tokens": 330525499.0, + "step": 12775 + }, + { + "epoch": 1.403030968592137, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.245215654373169, + "learning_rate": 1e-06, + "loss": 1.016, + "mean_token_accuracy": 0.6881787776947021, + "num_tokens": 330552529.0, + "step": 12776 + }, + { + "epoch": 1.4031407862947507, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.035957098007202, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.6970087289810181, + "num_tokens": 330584224.0, + "step": 12777 + }, + { + "epoch": 1.4032506039973645, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3083701133728027, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7197596430778503, + "num_tokens": 330608743.0, + "step": 12778 + }, + { + "epoch": 1.403360421699978, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1344635486602783, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7159103751182556, + "num_tokens": 330637403.0, + "step": 12779 + }, + { + "epoch": 1.4034702394025917, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.1115236282348633, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.71579909324646, + "num_tokens": 330669501.0, + "step": 12780 + }, + { + "epoch": 1.4035800571052053, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.0476653575897217, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.708540141582489, + "num_tokens": 330701549.0, + "step": 12781 + }, + { + "epoch": 1.403689874807819, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.153920888900757, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.7034770846366882, + "num_tokens": 330729181.0, + "step": 12782 + }, + { + "epoch": 1.4037996925104328, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.15609073638916, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.7006363272666931, + "num_tokens": 330757862.0, + "step": 12783 + }, + { + "epoch": 1.4039095102130463, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.274137496948242, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7149560451507568, + "num_tokens": 330783695.0, + "step": 12784 + }, + { + "epoch": 1.40401932791566, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 1.8615062236785889, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7075003385543823, + "num_tokens": 330817814.0, + "step": 12785 + }, + { + "epoch": 1.4041291456182736, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2208974361419678, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7270292043685913, + "num_tokens": 330844176.0, + "step": 12786 + }, + { + "epoch": 1.4042389633208874, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.2438669204711914, + "learning_rate": 1e-06, + "loss": 0.9561, + "mean_token_accuracy": 0.7073254585266113, + "num_tokens": 330870796.0, + "step": 12787 + }, + { + "epoch": 1.404348781023501, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3436102867126465, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.718256950378418, + "num_tokens": 330894438.0, + "step": 12788 + }, + { + "epoch": 1.4044585987261147, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3283979892730713, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7064152956008911, + "num_tokens": 330921480.0, + "step": 12789 + }, + { + "epoch": 1.4045684164287282, + "ewc_loss": 7.599592208862305e-06, + "grad_norm": 2.3197250366210938, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7080084681510925, + "num_tokens": 330946772.0, + "step": 12790 + }, + { + "epoch": 1.404678234131342, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0196077823638916, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7184640169143677, + "num_tokens": 330979370.0, + "step": 12791 + }, + { + "epoch": 1.4047880518339557, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3015825748443604, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7026327848434448, + "num_tokens": 331002954.0, + "step": 12792 + }, + { + "epoch": 1.4048978695365693, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 1.95564603805542, + "learning_rate": 1e-06, + "loss": 0.9992, + "mean_token_accuracy": 0.6945523619651794, + "num_tokens": 331035129.0, + "step": 12793 + }, + { + "epoch": 1.405007687239183, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3590757846832275, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7118252515792847, + "num_tokens": 331059434.0, + "step": 12794 + }, + { + "epoch": 1.4051175049417965, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.573864459991455, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7050046920776367, + "num_tokens": 331080140.0, + "step": 12795 + }, + { + "epoch": 1.4052273226444103, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1706185340881348, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7276386022567749, + "num_tokens": 331106893.0, + "step": 12796 + }, + { + "epoch": 1.405337140347024, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.582021474838257, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7490912079811096, + "num_tokens": 331126529.0, + "step": 12797 + }, + { + "epoch": 1.4054469580496376, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0312092304229736, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7056890726089478, + "num_tokens": 331158566.0, + "step": 12798 + }, + { + "epoch": 1.4055567757522511, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0864198207855225, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7164348363876343, + "num_tokens": 331186311.0, + "step": 12799 + }, + { + "epoch": 1.4056665934548649, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0014452934265137, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7405003905296326, + "num_tokens": 331215102.0, + "step": 12800 + }, + { + "epoch": 1.4057764111574786, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.5072882175445557, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.710816502571106, + "num_tokens": 331236980.0, + "step": 12801 + }, + { + "epoch": 1.4058862288600922, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.385988712310791, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.698932409286499, + "num_tokens": 331261948.0, + "step": 12802 + }, + { + "epoch": 1.405996046562706, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2637243270874023, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7110815048217773, + "num_tokens": 331288192.0, + "step": 12803 + }, + { + "epoch": 1.4061058642653195, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.047819137573242, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.71934574842453, + "num_tokens": 331317084.0, + "step": 12804 + }, + { + "epoch": 1.4062156819679332, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0067214965820312, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7364040017127991, + "num_tokens": 331345302.0, + "step": 12805 + }, + { + "epoch": 1.406325499670547, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1677377223968506, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7203152179718018, + "num_tokens": 331372262.0, + "step": 12806 + }, + { + "epoch": 1.4064353173731605, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1775004863739014, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7281785011291504, + "num_tokens": 331398490.0, + "step": 12807 + }, + { + "epoch": 1.4065451350757743, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.196075677871704, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6956436038017273, + "num_tokens": 331424470.0, + "step": 12808 + }, + { + "epoch": 1.4066549527783878, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3501224517822266, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7106989622116089, + "num_tokens": 331447838.0, + "step": 12809 + }, + { + "epoch": 1.4067647704810016, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.528278112411499, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7143279314041138, + "num_tokens": 331470010.0, + "step": 12810 + }, + { + "epoch": 1.4068745881836153, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1524877548217773, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7019535303115845, + "num_tokens": 331496230.0, + "step": 12811 + }, + { + "epoch": 1.4069844058862289, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.154376745223999, + "learning_rate": 1e-06, + "loss": 0.9889, + "mean_token_accuracy": 0.6929444074630737, + "num_tokens": 331525604.0, + "step": 12812 + }, + { + "epoch": 1.4070942235888424, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.111255168914795, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7152369022369385, + "num_tokens": 331553471.0, + "step": 12813 + }, + { + "epoch": 1.4072040412914562, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.4725921154022217, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7365929484367371, + "num_tokens": 331575373.0, + "step": 12814 + }, + { + "epoch": 1.40731385899407, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0536530017852783, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7132748961448669, + "num_tokens": 331604945.0, + "step": 12815 + }, + { + "epoch": 1.4074236766966834, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1454436779022217, + "learning_rate": 1e-06, + "loss": 0.9891, + "mean_token_accuracy": 0.700104296207428, + "num_tokens": 331632112.0, + "step": 12816 + }, + { + "epoch": 1.4075334943992972, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0410828590393066, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.722759485244751, + "num_tokens": 331659631.0, + "step": 12817 + }, + { + "epoch": 1.4076433121019107, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.392502546310425, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7388851642608643, + "num_tokens": 331681471.0, + "step": 12818 + }, + { + "epoch": 1.4077531298045245, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.336760997772217, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7376480102539062, + "num_tokens": 331704106.0, + "step": 12819 + }, + { + "epoch": 1.4078629475071383, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0459728240966797, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6960595846176147, + "num_tokens": 331736036.0, + "step": 12820 + }, + { + "epoch": 1.4079727652097518, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3151793479919434, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7159627079963684, + "num_tokens": 331761772.0, + "step": 12821 + }, + { + "epoch": 1.4080825829123655, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3564202785491943, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7238444685935974, + "num_tokens": 331785723.0, + "step": 12822 + }, + { + "epoch": 1.408192400614979, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3621256351470947, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7234869003295898, + "num_tokens": 331809006.0, + "step": 12823 + }, + { + "epoch": 1.4083022183175928, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.5238826274871826, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7384084463119507, + "num_tokens": 331830095.0, + "step": 12824 + }, + { + "epoch": 1.4084120360202066, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.377546787261963, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7454519271850586, + "num_tokens": 331852970.0, + "step": 12825 + }, + { + "epoch": 1.4085218537228201, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0501816272735596, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7329879999160767, + "num_tokens": 331880021.0, + "step": 12826 + }, + { + "epoch": 1.4086316714254337, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.077599048614502, + "learning_rate": 1e-06, + "loss": 1.0148, + "mean_token_accuracy": 0.6931936740875244, + "num_tokens": 331909539.0, + "step": 12827 + }, + { + "epoch": 1.4087414891280474, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.089036226272583, + "learning_rate": 1e-06, + "loss": 1.0176, + "mean_token_accuracy": 0.690697431564331, + "num_tokens": 331938496.0, + "step": 12828 + }, + { + "epoch": 1.4088513068306612, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 1.9879801273345947, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7312659621238708, + "num_tokens": 331970350.0, + "step": 12829 + }, + { + "epoch": 1.4089611245332747, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.464702844619751, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.718870997428894, + "num_tokens": 331991869.0, + "step": 12830 + }, + { + "epoch": 1.4090709422358885, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.517622709274292, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7225325107574463, + "num_tokens": 332013233.0, + "step": 12831 + }, + { + "epoch": 1.409180759938502, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1912078857421875, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.711709201335907, + "num_tokens": 332039007.0, + "step": 12832 + }, + { + "epoch": 1.4092905776411158, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2069106101989746, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7143304944038391, + "num_tokens": 332066044.0, + "step": 12833 + }, + { + "epoch": 1.4094003953437295, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1635899543762207, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7447066307067871, + "num_tokens": 332089319.0, + "step": 12834 + }, + { + "epoch": 1.409510213046343, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.601046085357666, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7214525938034058, + "num_tokens": 332109374.0, + "step": 12835 + }, + { + "epoch": 1.4096200307489568, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.083122968673706, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7117314338684082, + "num_tokens": 332137601.0, + "step": 12836 + }, + { + "epoch": 1.4097298484515703, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3329312801361084, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.713603675365448, + "num_tokens": 332162374.0, + "step": 12837 + }, + { + "epoch": 1.409839666154184, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0587563514709473, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7121380567550659, + "num_tokens": 332191028.0, + "step": 12838 + }, + { + "epoch": 1.4099494838567976, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.459804058074951, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7254244089126587, + "num_tokens": 332212579.0, + "step": 12839 + }, + { + "epoch": 1.4100593015594114, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.211644172668457, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7259863615036011, + "num_tokens": 332240660.0, + "step": 12840 + }, + { + "epoch": 1.410169119262025, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0802221298217773, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7087631821632385, + "num_tokens": 332271878.0, + "step": 12841 + }, + { + "epoch": 1.4102789369646387, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.054471015930176, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.724504828453064, + "num_tokens": 332300385.0, + "step": 12842 + }, + { + "epoch": 1.4103887546672524, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2823140621185303, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7225639820098877, + "num_tokens": 332324927.0, + "step": 12843 + }, + { + "epoch": 1.410498572369866, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.493410348892212, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7216967344284058, + "num_tokens": 332347441.0, + "step": 12844 + }, + { + "epoch": 1.4106083900724797, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.632817029953003, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7225459218025208, + "num_tokens": 332367849.0, + "step": 12845 + }, + { + "epoch": 1.4107182077750933, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2582716941833496, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.711432158946991, + "num_tokens": 332394262.0, + "step": 12846 + }, + { + "epoch": 1.410828025477707, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.396294593811035, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7006958723068237, + "num_tokens": 332417902.0, + "step": 12847 + }, + { + "epoch": 1.4109378431803208, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 1.956040859222412, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.708297848701477, + "num_tokens": 332450973.0, + "step": 12848 + }, + { + "epoch": 1.4110476608829343, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3200056552886963, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7188159823417664, + "num_tokens": 332474891.0, + "step": 12849 + }, + { + "epoch": 1.4111574785855479, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.177537679672241, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7164508700370789, + "num_tokens": 332501341.0, + "step": 12850 + }, + { + "epoch": 1.4112672962881616, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1980068683624268, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7107710242271423, + "num_tokens": 332529082.0, + "step": 12851 + }, + { + "epoch": 1.4113771139907754, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.266655921936035, + "learning_rate": 1e-06, + "loss": 1.0324, + "mean_token_accuracy": 0.690293550491333, + "num_tokens": 332556574.0, + "step": 12852 + }, + { + "epoch": 1.411486931693389, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.351161003112793, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7150872349739075, + "num_tokens": 332579569.0, + "step": 12853 + }, + { + "epoch": 1.4115967493960027, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.053323268890381, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.704376757144928, + "num_tokens": 332611878.0, + "step": 12854 + }, + { + "epoch": 1.4117065670986162, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2846944332122803, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7055650949478149, + "num_tokens": 332638058.0, + "step": 12855 + }, + { + "epoch": 1.41181638480123, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.155139684677124, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.6989579200744629, + "num_tokens": 332666000.0, + "step": 12856 + }, + { + "epoch": 1.4119262025038437, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0344176292419434, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7264524698257446, + "num_tokens": 332694149.0, + "step": 12857 + }, + { + "epoch": 1.4120360202064572, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2121095657348633, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.686659574508667, + "num_tokens": 332724188.0, + "step": 12858 + }, + { + "epoch": 1.412145837909071, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0711581707000732, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.6996583342552185, + "num_tokens": 332752306.0, + "step": 12859 + }, + { + "epoch": 1.4122556556116845, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3214733600616455, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7085140943527222, + "num_tokens": 332780115.0, + "step": 12860 + }, + { + "epoch": 1.4123654733142983, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.5563693046569824, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7158158421516418, + "num_tokens": 332800506.0, + "step": 12861 + }, + { + "epoch": 1.412475291016912, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.143364191055298, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.712936282157898, + "num_tokens": 332830867.0, + "step": 12862 + }, + { + "epoch": 1.4125851087195256, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1829473972320557, + "learning_rate": 1e-06, + "loss": 1.0149, + "mean_token_accuracy": 0.6834800243377686, + "num_tokens": 332860462.0, + "step": 12863 + }, + { + "epoch": 1.4126949264221391, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.4754626750946045, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7290488481521606, + "num_tokens": 332880760.0, + "step": 12864 + }, + { + "epoch": 1.4128047441247529, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0616238117218018, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7036298513412476, + "num_tokens": 332910473.0, + "step": 12865 + }, + { + "epoch": 1.4129145618273666, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.4524013996124268, + "learning_rate": 1e-06, + "loss": 0.9957, + "mean_token_accuracy": 0.6970179080963135, + "num_tokens": 332936802.0, + "step": 12866 + }, + { + "epoch": 1.4130243795299802, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3968920707702637, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7366706132888794, + "num_tokens": 332959432.0, + "step": 12867 + }, + { + "epoch": 1.413134197232594, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.0992226600646973, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6972987651824951, + "num_tokens": 332987546.0, + "step": 12868 + }, + { + "epoch": 1.4132440149352075, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.249307632446289, + "learning_rate": 1e-06, + "loss": 0.9707, + "mean_token_accuracy": 0.6985987424850464, + "num_tokens": 333014776.0, + "step": 12869 + }, + { + "epoch": 1.4133538326378212, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2945926189422607, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7153740525245667, + "num_tokens": 333040984.0, + "step": 12870 + }, + { + "epoch": 1.413463650340435, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.3356692790985107, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7282469868659973, + "num_tokens": 333064260.0, + "step": 12871 + }, + { + "epoch": 1.4135734680430485, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.5325100421905518, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7295845150947571, + "num_tokens": 333085230.0, + "step": 12872 + }, + { + "epoch": 1.4136832857456623, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2095909118652344, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7056741714477539, + "num_tokens": 333111202.0, + "step": 12873 + }, + { + "epoch": 1.4137931034482758, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2292394638061523, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7108556032180786, + "num_tokens": 333135562.0, + "step": 12874 + }, + { + "epoch": 1.4139029211508896, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2192530632019043, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.716131329536438, + "num_tokens": 333162233.0, + "step": 12875 + }, + { + "epoch": 1.4140127388535033, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2753636837005615, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7191252708435059, + "num_tokens": 333187389.0, + "step": 12876 + }, + { + "epoch": 1.4141225565561168, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.1879191398620605, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7191130518913269, + "num_tokens": 333216193.0, + "step": 12877 + }, + { + "epoch": 1.4142323742587304, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.4953558444976807, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7284237146377563, + "num_tokens": 333237036.0, + "step": 12878 + }, + { + "epoch": 1.4143421919613441, + "ewc_loss": 7.62939453125e-06, + "grad_norm": 2.2054271697998047, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7278381586074829, + "num_tokens": 333261687.0, + "step": 12879 + }, + { + "epoch": 1.414452009663958, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.4627654552459717, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7153247594833374, + "num_tokens": 333285475.0, + "step": 12880 + }, + { + "epoch": 1.4145618273665714, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.4204061031341553, + "learning_rate": 1e-06, + "loss": 1.0027, + "mean_token_accuracy": 0.7083641290664673, + "num_tokens": 333308813.0, + "step": 12881 + }, + { + "epoch": 1.4146716450691852, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.18959903717041, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7167237997055054, + "num_tokens": 333335279.0, + "step": 12882 + }, + { + "epoch": 1.4147814627717987, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0928361415863037, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7088829278945923, + "num_tokens": 333364811.0, + "step": 12883 + }, + { + "epoch": 1.4148912804744125, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.6217799186706543, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7271072864532471, + "num_tokens": 333385365.0, + "step": 12884 + }, + { + "epoch": 1.4150010981770262, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5454440116882324, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7279131412506104, + "num_tokens": 333405477.0, + "step": 12885 + }, + { + "epoch": 1.4151109158796398, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.328723907470703, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7166103720664978, + "num_tokens": 333429113.0, + "step": 12886 + }, + { + "epoch": 1.4152207335822535, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3677759170532227, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7312458753585815, + "num_tokens": 333451628.0, + "step": 12887 + }, + { + "epoch": 1.415330551284867, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.663942813873291, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.7435190081596375, + "num_tokens": 333469094.0, + "step": 12888 + }, + { + "epoch": 1.4154403689874808, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.264784574508667, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7467325329780579, + "num_tokens": 333492280.0, + "step": 12889 + }, + { + "epoch": 1.4155501866900946, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2013914585113525, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.70304274559021, + "num_tokens": 333518127.0, + "step": 12890 + }, + { + "epoch": 1.4156600043927081, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1542270183563232, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6817470788955688, + "num_tokens": 333546384.0, + "step": 12891 + }, + { + "epoch": 1.4157698220953217, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.341026544570923, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7141308784484863, + "num_tokens": 333571273.0, + "step": 12892 + }, + { + "epoch": 1.4158796397979354, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3351714611053467, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7273388504981995, + "num_tokens": 333594588.0, + "step": 12893 + }, + { + "epoch": 1.4159894575005492, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0553243160247803, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7386040091514587, + "num_tokens": 333622666.0, + "step": 12894 + }, + { + "epoch": 1.4160992752031627, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3581461906433105, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7324908971786499, + "num_tokens": 333644976.0, + "step": 12895 + }, + { + "epoch": 1.4162090929057765, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1943068504333496, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7137539386749268, + "num_tokens": 333673260.0, + "step": 12896 + }, + { + "epoch": 1.41631891060839, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2649073600769043, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6898036003112793, + "num_tokens": 333701124.0, + "step": 12897 + }, + { + "epoch": 1.4164287283110037, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.173658609390259, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7158514261245728, + "num_tokens": 333728031.0, + "step": 12898 + }, + { + "epoch": 1.4165385460136175, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2501134872436523, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7343387603759766, + "num_tokens": 333753218.0, + "step": 12899 + }, + { + "epoch": 1.416648363716231, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2615814208984375, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7193467020988464, + "num_tokens": 333780133.0, + "step": 12900 + }, + { + "epoch": 1.4167581814188448, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2421748638153076, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7087050676345825, + "num_tokens": 333805585.0, + "step": 12901 + }, + { + "epoch": 1.4168679991214583, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1758127212524414, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7079029083251953, + "num_tokens": 333832788.0, + "step": 12902 + }, + { + "epoch": 1.416977816824072, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.089482545852661, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7246555089950562, + "num_tokens": 333861074.0, + "step": 12903 + }, + { + "epoch": 1.4170876345266856, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5510189533233643, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7152248620986938, + "num_tokens": 333883421.0, + "step": 12904 + }, + { + "epoch": 1.4171974522292994, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5412304401397705, + "learning_rate": 1e-06, + "loss": 0.7433, + "mean_token_accuracy": 0.7561379671096802, + "num_tokens": 333902876.0, + "step": 12905 + }, + { + "epoch": 1.417307269931913, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0735385417938232, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7164953351020813, + "num_tokens": 333932878.0, + "step": 12906 + }, + { + "epoch": 1.4174170876345267, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.031066656112671, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7141468524932861, + "num_tokens": 333964610.0, + "step": 12907 + }, + { + "epoch": 1.4175269053371404, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.330171585083008, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7047958374023438, + "num_tokens": 333990179.0, + "step": 12908 + }, + { + "epoch": 1.417636723039754, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.4000868797302246, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7299612760543823, + "num_tokens": 334014098.0, + "step": 12909 + }, + { + "epoch": 1.4177465407423677, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5261106491088867, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7208821177482605, + "num_tokens": 334035745.0, + "step": 12910 + }, + { + "epoch": 1.4178563584449813, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0874600410461426, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.719524621963501, + "num_tokens": 334063259.0, + "step": 12911 + }, + { + "epoch": 1.417966176147595, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0458667278289795, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7159625291824341, + "num_tokens": 334092798.0, + "step": 12912 + }, + { + "epoch": 1.4180759938502088, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.06199049949646, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7256021499633789, + "num_tokens": 334121983.0, + "step": 12913 + }, + { + "epoch": 1.4181858115528223, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1768903732299805, + "learning_rate": 1e-06, + "loss": 0.9987, + "mean_token_accuracy": 0.6982198357582092, + "num_tokens": 334150366.0, + "step": 12914 + }, + { + "epoch": 1.4182956292554358, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3395164012908936, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.729955792427063, + "num_tokens": 334173217.0, + "step": 12915 + }, + { + "epoch": 1.4184054469580496, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0529937744140625, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6937524080276489, + "num_tokens": 334203261.0, + "step": 12916 + }, + { + "epoch": 1.4185152646606634, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3497939109802246, + "learning_rate": 1e-06, + "loss": 0.9582, + "mean_token_accuracy": 0.7014092206954956, + "num_tokens": 334229166.0, + "step": 12917 + }, + { + "epoch": 1.418625082363277, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.108319044113159, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6924898624420166, + "num_tokens": 334258769.0, + "step": 12918 + }, + { + "epoch": 1.4187349000658906, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0950608253479004, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.7133612632751465, + "num_tokens": 334287469.0, + "step": 12919 + }, + { + "epoch": 1.4188447177685042, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.175703525543213, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7183499336242676, + "num_tokens": 334314416.0, + "step": 12920 + }, + { + "epoch": 1.418954535471118, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1457200050354004, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7278831005096436, + "num_tokens": 334339914.0, + "step": 12921 + }, + { + "epoch": 1.4190643531737317, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5221824645996094, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7204241752624512, + "num_tokens": 334362130.0, + "step": 12922 + }, + { + "epoch": 1.4191741708763452, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 1.9093914031982422, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7051014304161072, + "num_tokens": 334394745.0, + "step": 12923 + }, + { + "epoch": 1.419283988578959, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.115248441696167, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7225502729415894, + "num_tokens": 334422706.0, + "step": 12924 + }, + { + "epoch": 1.4193938062815725, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.259748935699463, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7093634605407715, + "num_tokens": 334448475.0, + "step": 12925 + }, + { + "epoch": 1.4195036239841863, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2779576778411865, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7297677397727966, + "num_tokens": 334473179.0, + "step": 12926 + }, + { + "epoch": 1.4196134416868, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1452114582061768, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7176228761672974, + "num_tokens": 334499427.0, + "step": 12927 + }, + { + "epoch": 1.4197232593894136, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.0571365356445312, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7042255401611328, + "num_tokens": 334531301.0, + "step": 12928 + }, + { + "epoch": 1.419833077092027, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.2688751220703125, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7181988954544067, + "num_tokens": 334555507.0, + "step": 12929 + }, + { + "epoch": 1.4199428947946409, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 1.9317388534545898, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7135576605796814, + "num_tokens": 334587652.0, + "step": 12930 + }, + { + "epoch": 1.4200527124972546, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.492128372192383, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7205432653427124, + "num_tokens": 334610022.0, + "step": 12931 + }, + { + "epoch": 1.4201625301998682, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.051208019256592, + "learning_rate": 1e-06, + "loss": 0.9695, + "mean_token_accuracy": 0.7034978270530701, + "num_tokens": 334639176.0, + "step": 12932 + }, + { + "epoch": 1.420272347902482, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.5588529109954834, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7376694679260254, + "num_tokens": 334658933.0, + "step": 12933 + }, + { + "epoch": 1.4203821656050954, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.246983289718628, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7036775946617126, + "num_tokens": 334687502.0, + "step": 12934 + }, + { + "epoch": 1.4204919833077092, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 1.9625873565673828, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7056897282600403, + "num_tokens": 334719409.0, + "step": 12935 + }, + { + "epoch": 1.420601801010323, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 1.9643844366073608, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7052638530731201, + "num_tokens": 334749363.0, + "step": 12936 + }, + { + "epoch": 1.4207116187129365, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.059109687805176, + "learning_rate": 1e-06, + "loss": 1.006, + "mean_token_accuracy": 0.6921119689941406, + "num_tokens": 334780552.0, + "step": 12937 + }, + { + "epoch": 1.4208214364155503, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.197056293487549, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7011988162994385, + "num_tokens": 334808206.0, + "step": 12938 + }, + { + "epoch": 1.4209312541181638, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.3359827995300293, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7177559733390808, + "num_tokens": 334831561.0, + "step": 12939 + }, + { + "epoch": 1.4210410718207775, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1735715866088867, + "learning_rate": 1e-06, + "loss": 1.0416, + "mean_token_accuracy": 0.685411274433136, + "num_tokens": 334859188.0, + "step": 12940 + }, + { + "epoch": 1.4211508895233913, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.274174928665161, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7255098819732666, + "num_tokens": 334884809.0, + "step": 12941 + }, + { + "epoch": 1.4212607072260048, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.260399103164673, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7146693468093872, + "num_tokens": 334910764.0, + "step": 12942 + }, + { + "epoch": 1.4213705249286184, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.440699338912964, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.721434473991394, + "num_tokens": 334935069.0, + "step": 12943 + }, + { + "epoch": 1.4214803426312321, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.19040846824646, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7143474817276001, + "num_tokens": 334961762.0, + "step": 12944 + }, + { + "epoch": 1.4215901603338459, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.279763698577881, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.724941074848175, + "num_tokens": 334988095.0, + "step": 12945 + }, + { + "epoch": 1.4216999780364594, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.1091697216033936, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7306768894195557, + "num_tokens": 335015639.0, + "step": 12946 + }, + { + "epoch": 1.4218097957390732, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.444672107696533, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7270070314407349, + "num_tokens": 335038331.0, + "step": 12947 + }, + { + "epoch": 1.4219196134416867, + "ewc_loss": 7.68899917602539e-06, + "grad_norm": 2.320207118988037, + "learning_rate": 1e-06, + "loss": 1.0139, + "mean_token_accuracy": 0.6993260383605957, + "num_tokens": 335064612.0, + "step": 12948 + }, + { + "epoch": 1.4220294311443005, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3223278522491455, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7237305641174316, + "num_tokens": 335088936.0, + "step": 12949 + }, + { + "epoch": 1.4221392488469142, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3245880603790283, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7385641932487488, + "num_tokens": 335111531.0, + "step": 12950 + }, + { + "epoch": 1.4222490665495278, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.183500051498413, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7157479524612427, + "num_tokens": 335138758.0, + "step": 12951 + }, + { + "epoch": 1.4223588842521415, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2092745304107666, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7230595350265503, + "num_tokens": 335165080.0, + "step": 12952 + }, + { + "epoch": 1.422468701954755, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.397935390472412, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7053416967391968, + "num_tokens": 335189641.0, + "step": 12953 + }, + { + "epoch": 1.4225785196573688, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.03518009185791, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.707482099533081, + "num_tokens": 335219361.0, + "step": 12954 + }, + { + "epoch": 1.4226883373599823, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.336278200149536, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7172211408615112, + "num_tokens": 335242980.0, + "step": 12955 + }, + { + "epoch": 1.422798155062596, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.485853433609009, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.705119252204895, + "num_tokens": 335266402.0, + "step": 12956 + }, + { + "epoch": 1.4229079727652096, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.453172206878662, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7284589409828186, + "num_tokens": 335287588.0, + "step": 12957 + }, + { + "epoch": 1.4230177904678234, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2713966369628906, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7283896803855896, + "num_tokens": 335312230.0, + "step": 12958 + }, + { + "epoch": 1.4231276081704372, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0432794094085693, + "learning_rate": 1e-06, + "loss": 0.7659, + "mean_token_accuracy": 0.7518274188041687, + "num_tokens": 335339674.0, + "step": 12959 + }, + { + "epoch": 1.4232374258730507, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3096418380737305, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7147315740585327, + "num_tokens": 335366323.0, + "step": 12960 + }, + { + "epoch": 1.4233472435756644, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.400459051132202, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7366329431533813, + "num_tokens": 335388086.0, + "step": 12961 + }, + { + "epoch": 1.423457061278278, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0532948970794678, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7331682443618774, + "num_tokens": 335416717.0, + "step": 12962 + }, + { + "epoch": 1.4235668789808917, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.073775291442871, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7105515599250793, + "num_tokens": 335448787.0, + "step": 12963 + }, + { + "epoch": 1.4236766966835055, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.238409996032715, + "learning_rate": 1e-06, + "loss": 0.7468, + "mean_token_accuracy": 0.7591485381126404, + "num_tokens": 335471561.0, + "step": 12964 + }, + { + "epoch": 1.423786514386119, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2190170288085938, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7196815609931946, + "num_tokens": 335498861.0, + "step": 12965 + }, + { + "epoch": 1.4238963320887328, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.6944079399108887, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7274544835090637, + "num_tokens": 335518377.0, + "step": 12966 + }, + { + "epoch": 1.4240061497913463, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4678685665130615, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7323955297470093, + "num_tokens": 335539438.0, + "step": 12967 + }, + { + "epoch": 1.42411596749396, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2014219760894775, + "learning_rate": 1e-06, + "loss": 0.7642, + "mean_token_accuracy": 0.7573902010917664, + "num_tokens": 335563992.0, + "step": 12968 + }, + { + "epoch": 1.4242257851965736, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1198527812957764, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6974968314170837, + "num_tokens": 335595342.0, + "step": 12969 + }, + { + "epoch": 1.4243356028991874, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0910098552703857, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.6999843120574951, + "num_tokens": 335623167.0, + "step": 12970 + }, + { + "epoch": 1.424445420601801, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4669671058654785, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7352142333984375, + "num_tokens": 335645301.0, + "step": 12971 + }, + { + "epoch": 1.4245552383044147, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.540032386779785, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7317718267440796, + "num_tokens": 335668265.0, + "step": 12972 + }, + { + "epoch": 1.4246650560070284, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9614536762237549, + "learning_rate": 1e-06, + "loss": 1.0394, + "mean_token_accuracy": 0.6839929819107056, + "num_tokens": 335702380.0, + "step": 12973 + }, + { + "epoch": 1.424774873709642, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.269542932510376, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7099344730377197, + "num_tokens": 335728746.0, + "step": 12974 + }, + { + "epoch": 1.4248846914122557, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.044696807861328, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7141436338424683, + "num_tokens": 335760911.0, + "step": 12975 + }, + { + "epoch": 1.4249945091148692, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1971688270568848, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7339394092559814, + "num_tokens": 335787964.0, + "step": 12976 + }, + { + "epoch": 1.425104326817483, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4846997261047363, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7211979627609253, + "num_tokens": 335808830.0, + "step": 12977 + }, + { + "epoch": 1.4252141445200968, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.016030788421631, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.70620197057724, + "num_tokens": 335838235.0, + "step": 12978 + }, + { + "epoch": 1.4253239622227103, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.324183464050293, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7201095819473267, + "num_tokens": 335861638.0, + "step": 12979 + }, + { + "epoch": 1.4254337799253238, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2674601078033447, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7107905149459839, + "num_tokens": 335887560.0, + "step": 12980 + }, + { + "epoch": 1.4255435976279376, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0441157817840576, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7120630741119385, + "num_tokens": 335916147.0, + "step": 12981 + }, + { + "epoch": 1.4256534153305513, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.098057270050049, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.6960277557373047, + "num_tokens": 335945961.0, + "step": 12982 + }, + { + "epoch": 1.4257632330331649, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3840811252593994, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7276594638824463, + "num_tokens": 335968366.0, + "step": 12983 + }, + { + "epoch": 1.4258730507357786, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4093658924102783, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.737951397895813, + "num_tokens": 335989642.0, + "step": 12984 + }, + { + "epoch": 1.4259828684383922, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1314659118652344, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7111982107162476, + "num_tokens": 336017457.0, + "step": 12985 + }, + { + "epoch": 1.426092686141006, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2463409900665283, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7165162563323975, + "num_tokens": 336043602.0, + "step": 12986 + }, + { + "epoch": 1.4262025038436197, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.176518678665161, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7262941002845764, + "num_tokens": 336070210.0, + "step": 12987 + }, + { + "epoch": 1.4263123215462332, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.223210334777832, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7223338484764099, + "num_tokens": 336097055.0, + "step": 12988 + }, + { + "epoch": 1.426422139248847, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2162587642669678, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7097862362861633, + "num_tokens": 336124255.0, + "step": 12989 + }, + { + "epoch": 1.4265319569514605, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2338979244232178, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7147246599197388, + "num_tokens": 336151343.0, + "step": 12990 + }, + { + "epoch": 1.4266417746540743, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9848157167434692, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7153984904289246, + "num_tokens": 336183647.0, + "step": 12991 + }, + { + "epoch": 1.426751592356688, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4950313568115234, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7277641296386719, + "num_tokens": 336204677.0, + "step": 12992 + }, + { + "epoch": 1.4268614100593016, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0853726863861084, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7106747627258301, + "num_tokens": 336232008.0, + "step": 12993 + }, + { + "epoch": 1.426971227761915, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9761011600494385, + "learning_rate": 1e-06, + "loss": 0.9904, + "mean_token_accuracy": 0.6991623044013977, + "num_tokens": 336265650.0, + "step": 12994 + }, + { + "epoch": 1.4270810454645289, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.182011127471924, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.6986874341964722, + "num_tokens": 336293078.0, + "step": 12995 + }, + { + "epoch": 1.4271908631671426, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.014840841293335, + "learning_rate": 1e-06, + "loss": 0.9951, + "mean_token_accuracy": 0.700766921043396, + "num_tokens": 336325403.0, + "step": 12996 + }, + { + "epoch": 1.4273006808697561, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3253111839294434, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7142244577407837, + "num_tokens": 336350773.0, + "step": 12997 + }, + { + "epoch": 1.42741049857237, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2450802326202393, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7233024835586548, + "num_tokens": 336375432.0, + "step": 12998 + }, + { + "epoch": 1.4275203162749834, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1659152507781982, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7230590581893921, + "num_tokens": 336402467.0, + "step": 12999 + }, + { + "epoch": 1.4276301339775972, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2485134601593018, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7242108583450317, + "num_tokens": 336426745.0, + "step": 13000 + }, + { + "epoch": 1.427739951680211, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.93905770778656, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.717747688293457, + "num_tokens": 336456691.0, + "step": 13001 + }, + { + "epoch": 1.4278497693828245, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.134185314178467, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.737669825553894, + "num_tokens": 336481532.0, + "step": 13002 + }, + { + "epoch": 1.4279595870854382, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.365665912628174, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7471811771392822, + "num_tokens": 336502282.0, + "step": 13003 + }, + { + "epoch": 1.4280694047880518, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9333206415176392, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.7053565979003906, + "num_tokens": 336536836.0, + "step": 13004 + }, + { + "epoch": 1.4281792224906655, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3706979751586914, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7315549254417419, + "num_tokens": 336559622.0, + "step": 13005 + }, + { + "epoch": 1.4282890401932793, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.170823812484741, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7308739423751831, + "num_tokens": 336586090.0, + "step": 13006 + }, + { + "epoch": 1.4283988578958928, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2828245162963867, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7260268926620483, + "num_tokens": 336610788.0, + "step": 13007 + }, + { + "epoch": 1.4285086755985064, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.12141752243042, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7248103618621826, + "num_tokens": 336637114.0, + "step": 13008 + }, + { + "epoch": 1.4286184933011201, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.26580810546875, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7165862321853638, + "num_tokens": 336662254.0, + "step": 13009 + }, + { + "epoch": 1.4287283110037339, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0143210887908936, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7046077251434326, + "num_tokens": 336692324.0, + "step": 13010 + }, + { + "epoch": 1.4288381287063474, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.308203935623169, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7074181437492371, + "num_tokens": 336719362.0, + "step": 13011 + }, + { + "epoch": 1.4289479464089612, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5748088359832764, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7206581830978394, + "num_tokens": 336739122.0, + "step": 13012 + }, + { + "epoch": 1.4290577641115747, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.021355390548706, + "learning_rate": 1e-06, + "loss": 0.982, + "mean_token_accuracy": 0.6983636021614075, + "num_tokens": 336772121.0, + "step": 13013 + }, + { + "epoch": 1.4291675818141885, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.258544921875, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7059672474861145, + "num_tokens": 336796925.0, + "step": 13014 + }, + { + "epoch": 1.4292773995168022, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.34384822845459, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7116150856018066, + "num_tokens": 336821568.0, + "step": 13015 + }, + { + "epoch": 1.4293872172194158, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2882473468780518, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7158417701721191, + "num_tokens": 336847377.0, + "step": 13016 + }, + { + "epoch": 1.4294970349220295, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1940762996673584, + "learning_rate": 1e-06, + "loss": 1.0045, + "mean_token_accuracy": 0.6925270557403564, + "num_tokens": 336872968.0, + "step": 13017 + }, + { + "epoch": 1.429606852624643, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.325603485107422, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.7032825946807861, + "num_tokens": 336897484.0, + "step": 13018 + }, + { + "epoch": 1.4297166703272568, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.033698081970215, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7386766076087952, + "num_tokens": 336925818.0, + "step": 13019 + }, + { + "epoch": 1.4298264880298703, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3004164695739746, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7317495346069336, + "num_tokens": 336949216.0, + "step": 13020 + }, + { + "epoch": 1.429936305732484, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5151453018188477, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7089966535568237, + "num_tokens": 336970982.0, + "step": 13021 + }, + { + "epoch": 1.4300461234350976, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9496504068374634, + "learning_rate": 1e-06, + "loss": 1.0202, + "mean_token_accuracy": 0.704653799533844, + "num_tokens": 337003995.0, + "step": 13022 + }, + { + "epoch": 1.4301559411377114, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.7885849475860596, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7433619499206543, + "num_tokens": 337022070.0, + "step": 13023 + }, + { + "epoch": 1.4302657588403251, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.278757095336914, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7238600850105286, + "num_tokens": 337044754.0, + "step": 13024 + }, + { + "epoch": 1.4303755765429387, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2009763717651367, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.701857328414917, + "num_tokens": 337071397.0, + "step": 13025 + }, + { + "epoch": 1.4304853942455524, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.214186429977417, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.7010162472724915, + "num_tokens": 337097440.0, + "step": 13026 + }, + { + "epoch": 1.430595211948166, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3845276832580566, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7188396453857422, + "num_tokens": 337122037.0, + "step": 13027 + }, + { + "epoch": 1.4307050296507797, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.212599039077759, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7171480059623718, + "num_tokens": 337147695.0, + "step": 13028 + }, + { + "epoch": 1.4308148473533935, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2626285552978516, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7250111699104309, + "num_tokens": 337174018.0, + "step": 13029 + }, + { + "epoch": 1.430924665056007, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.203073501586914, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7117002010345459, + "num_tokens": 337201036.0, + "step": 13030 + }, + { + "epoch": 1.4310344827586206, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1825270652770996, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7242103815078735, + "num_tokens": 337227358.0, + "step": 13031 + }, + { + "epoch": 1.4311443004612343, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.298069953918457, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7156785726547241, + "num_tokens": 337251693.0, + "step": 13032 + }, + { + "epoch": 1.431254118163848, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1683685779571533, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7259254455566406, + "num_tokens": 337278719.0, + "step": 13033 + }, + { + "epoch": 1.4313639358664616, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0671916007995605, + "learning_rate": 1e-06, + "loss": 1.0023, + "mean_token_accuracy": 0.6944564580917358, + "num_tokens": 337310203.0, + "step": 13034 + }, + { + "epoch": 1.4314737535690754, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4346487522125244, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7183820009231567, + "num_tokens": 337332944.0, + "step": 13035 + }, + { + "epoch": 1.431583571271689, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5022480487823486, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7171545028686523, + "num_tokens": 337353080.0, + "step": 13036 + }, + { + "epoch": 1.4316933889743026, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9097222089767456, + "learning_rate": 1e-06, + "loss": 1.0282, + "mean_token_accuracy": 0.6856637001037598, + "num_tokens": 337388005.0, + "step": 13037 + }, + { + "epoch": 1.4318032066769164, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.332944869995117, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7162586450576782, + "num_tokens": 337412486.0, + "step": 13038 + }, + { + "epoch": 1.43191302437953, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.074347734451294, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.704634428024292, + "num_tokens": 337442562.0, + "step": 13039 + }, + { + "epoch": 1.4320228420821437, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.626904010772705, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7395806312561035, + "num_tokens": 337462022.0, + "step": 13040 + }, + { + "epoch": 1.4321326597847572, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1434683799743652, + "learning_rate": 1e-06, + "loss": 1.0269, + "mean_token_accuracy": 0.6869908571243286, + "num_tokens": 337489573.0, + "step": 13041 + }, + { + "epoch": 1.432242477487371, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2291150093078613, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7065879702568054, + "num_tokens": 337517170.0, + "step": 13042 + }, + { + "epoch": 1.4323522951899847, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9140198230743408, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7121137976646423, + "num_tokens": 337550404.0, + "step": 13043 + }, + { + "epoch": 1.4324621128925983, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2699246406555176, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7249144315719604, + "num_tokens": 337573747.0, + "step": 13044 + }, + { + "epoch": 1.4325719305952118, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3996739387512207, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7296468019485474, + "num_tokens": 337596783.0, + "step": 13045 + }, + { + "epoch": 1.4326817482978256, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4128470420837402, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7227227687835693, + "num_tokens": 337619950.0, + "step": 13046 + }, + { + "epoch": 1.4327915660004393, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.958612084388733, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7358547449111938, + "num_tokens": 337650293.0, + "step": 13047 + }, + { + "epoch": 1.4329013837030529, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1936991214752197, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7178963422775269, + "num_tokens": 337680476.0, + "step": 13048 + }, + { + "epoch": 1.4330112014056666, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3825740814208984, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7204312086105347, + "num_tokens": 337702464.0, + "step": 13049 + }, + { + "epoch": 1.4331210191082802, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1467556953430176, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7182186841964722, + "num_tokens": 337729715.0, + "step": 13050 + }, + { + "epoch": 1.433230836810894, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.112718105316162, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.7012925148010254, + "num_tokens": 337757375.0, + "step": 13051 + }, + { + "epoch": 1.4333406545135077, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1943187713623047, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7282648682594299, + "num_tokens": 337784525.0, + "step": 13052 + }, + { + "epoch": 1.4334504722161212, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5516273975372314, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7321248054504395, + "num_tokens": 337807150.0, + "step": 13053 + }, + { + "epoch": 1.433560289918735, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4531912803649902, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7084087133407593, + "num_tokens": 337830400.0, + "step": 13054 + }, + { + "epoch": 1.4336701076213485, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2010421752929688, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7270599007606506, + "num_tokens": 337854411.0, + "step": 13055 + }, + { + "epoch": 1.4337799253239623, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1288719177246094, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7307354807853699, + "num_tokens": 337880853.0, + "step": 13056 + }, + { + "epoch": 1.433889743026576, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2257044315338135, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7229880690574646, + "num_tokens": 337906239.0, + "step": 13057 + }, + { + "epoch": 1.4339995607291895, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.352457046508789, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.74844890832901, + "num_tokens": 337928118.0, + "step": 13058 + }, + { + "epoch": 1.434109378431803, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.162766933441162, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7092026472091675, + "num_tokens": 337957483.0, + "step": 13059 + }, + { + "epoch": 1.4342191961344168, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2637884616851807, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7227590680122375, + "num_tokens": 337986256.0, + "step": 13060 + }, + { + "epoch": 1.4343290138370306, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.070244550704956, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7294741272926331, + "num_tokens": 338016447.0, + "step": 13061 + }, + { + "epoch": 1.4344388315396441, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3982865810394287, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.6958537101745605, + "num_tokens": 338041049.0, + "step": 13062 + }, + { + "epoch": 1.434548649242258, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.027989149093628, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.740010142326355, + "num_tokens": 338069754.0, + "step": 13063 + }, + { + "epoch": 1.4346584669448714, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.433387041091919, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7291520237922668, + "num_tokens": 338090336.0, + "step": 13064 + }, + { + "epoch": 1.4347682846474852, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.127365827560425, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7302733659744263, + "num_tokens": 338118398.0, + "step": 13065 + }, + { + "epoch": 1.434878102350099, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.144437551498413, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.732474148273468, + "num_tokens": 338144635.0, + "step": 13066 + }, + { + "epoch": 1.4349879200527125, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1383488178253174, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7056243419647217, + "num_tokens": 338172175.0, + "step": 13067 + }, + { + "epoch": 1.4350977377553262, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.132141351699829, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7140909433364868, + "num_tokens": 338200321.0, + "step": 13068 + }, + { + "epoch": 1.4352075554579398, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3925673961639404, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7250382304191589, + "num_tokens": 338224147.0, + "step": 13069 + }, + { + "epoch": 1.4353173731605535, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9990414381027222, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7272014617919922, + "num_tokens": 338253884.0, + "step": 13070 + }, + { + "epoch": 1.4354271908631673, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.145440101623535, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7129004001617432, + "num_tokens": 338281900.0, + "step": 13071 + }, + { + "epoch": 1.4355370085657808, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.409165620803833, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7210314273834229, + "num_tokens": 338305899.0, + "step": 13072 + }, + { + "epoch": 1.4356468262683943, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.7641514539718628, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7066131830215454, + "num_tokens": 338344227.0, + "step": 13073 + }, + { + "epoch": 1.435756643971008, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3829896450042725, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7137970924377441, + "num_tokens": 338369871.0, + "step": 13074 + }, + { + "epoch": 1.4358664616736219, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.393989086151123, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7147217392921448, + "num_tokens": 338396378.0, + "step": 13075 + }, + { + "epoch": 1.4359762793762354, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3590540885925293, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7168002724647522, + "num_tokens": 338418917.0, + "step": 13076 + }, + { + "epoch": 1.4360860970788492, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.338900089263916, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7220266461372375, + "num_tokens": 338443366.0, + "step": 13077 + }, + { + "epoch": 1.4361959147814627, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4539222717285156, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7446572184562683, + "num_tokens": 338464714.0, + "step": 13078 + }, + { + "epoch": 1.4363057324840764, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1133456230163574, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7076139450073242, + "num_tokens": 338496171.0, + "step": 13079 + }, + { + "epoch": 1.4364155501866902, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1746089458465576, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.7382839918136597, + "num_tokens": 338521962.0, + "step": 13080 + }, + { + "epoch": 1.4365253678893037, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3593149185180664, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.717665433883667, + "num_tokens": 338547278.0, + "step": 13081 + }, + { + "epoch": 1.4366351855919175, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2434792518615723, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7187131643295288, + "num_tokens": 338574124.0, + "step": 13082 + }, + { + "epoch": 1.436745003294531, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.380794048309326, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7263212203979492, + "num_tokens": 338598037.0, + "step": 13083 + }, + { + "epoch": 1.4368548209971448, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.049391269683838, + "learning_rate": 1e-06, + "loss": 0.9792, + "mean_token_accuracy": 0.7031540274620056, + "num_tokens": 338629611.0, + "step": 13084 + }, + { + "epoch": 1.4369646386997583, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1985557079315186, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7573845386505127, + "num_tokens": 338652735.0, + "step": 13085 + }, + { + "epoch": 1.437074456402372, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.503000020980835, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.728187084197998, + "num_tokens": 338673571.0, + "step": 13086 + }, + { + "epoch": 1.4371842741049856, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.18501615524292, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7080100774765015, + "num_tokens": 338700594.0, + "step": 13087 + }, + { + "epoch": 1.4372940918075994, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5759310722351074, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7304071187973022, + "num_tokens": 338721880.0, + "step": 13088 + }, + { + "epoch": 1.4374039095102131, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.342954158782959, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7353349924087524, + "num_tokens": 338745508.0, + "step": 13089 + }, + { + "epoch": 1.4375137272128267, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.367267370223999, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7212129831314087, + "num_tokens": 338769394.0, + "step": 13090 + }, + { + "epoch": 1.4376235449154404, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0975348949432373, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.713772177696228, + "num_tokens": 338799700.0, + "step": 13091 + }, + { + "epoch": 1.437733362618054, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.421987295150757, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7173018455505371, + "num_tokens": 338823255.0, + "step": 13092 + }, + { + "epoch": 1.4378431803206677, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9558049440383911, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7103745937347412, + "num_tokens": 338855495.0, + "step": 13093 + }, + { + "epoch": 1.4379529980232815, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.545581340789795, + "learning_rate": 1e-06, + "loss": 0.7187, + "mean_token_accuracy": 0.7681536674499512, + "num_tokens": 338873672.0, + "step": 13094 + }, + { + "epoch": 1.438062815725895, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1478936672210693, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.7000137567520142, + "num_tokens": 338900914.0, + "step": 13095 + }, + { + "epoch": 1.4381726334285085, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3163373470306396, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7144614458084106, + "num_tokens": 338926093.0, + "step": 13096 + }, + { + "epoch": 1.4382824511311223, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.8503395318984985, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.7044981718063354, + "num_tokens": 338959557.0, + "step": 13097 + }, + { + "epoch": 1.438392268833736, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9437005519866943, + "learning_rate": 1e-06, + "loss": 0.9999, + "mean_token_accuracy": 0.6943787336349487, + "num_tokens": 338995283.0, + "step": 13098 + }, + { + "epoch": 1.4385020865363496, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.377318859100342, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7036076784133911, + "num_tokens": 339017549.0, + "step": 13099 + }, + { + "epoch": 1.4386119042389633, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0630576610565186, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7141636610031128, + "num_tokens": 339045339.0, + "step": 13100 + }, + { + "epoch": 1.4387217219415769, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.0739622116088867, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7119186520576477, + "num_tokens": 339076105.0, + "step": 13101 + }, + { + "epoch": 1.4388315396441906, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2331478595733643, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.7001992464065552, + "num_tokens": 339101917.0, + "step": 13102 + }, + { + "epoch": 1.4389413573468044, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3708176612854004, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7233858704566956, + "num_tokens": 339125277.0, + "step": 13103 + }, + { + "epoch": 1.439051175049418, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.631772994995117, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.7404525279998779, + "num_tokens": 339142988.0, + "step": 13104 + }, + { + "epoch": 1.4391609927520317, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.5033986568450928, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7106823921203613, + "num_tokens": 339164220.0, + "step": 13105 + }, + { + "epoch": 1.4392708104546452, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1256275177001953, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7089378833770752, + "num_tokens": 339194424.0, + "step": 13106 + }, + { + "epoch": 1.439380628157259, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.288486957550049, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.705101728439331, + "num_tokens": 339219591.0, + "step": 13107 + }, + { + "epoch": 1.4394904458598727, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.036468744277954, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7082929611206055, + "num_tokens": 339250645.0, + "step": 13108 + }, + { + "epoch": 1.4396002635624863, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1817290782928467, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.708957314491272, + "num_tokens": 339276743.0, + "step": 13109 + }, + { + "epoch": 1.4397100812650998, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4291956424713135, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7344995737075806, + "num_tokens": 339297691.0, + "step": 13110 + }, + { + "epoch": 1.4398198989677136, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3611605167388916, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.721787691116333, + "num_tokens": 339320684.0, + "step": 13111 + }, + { + "epoch": 1.4399297166703273, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.323147773742676, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7228208184242249, + "num_tokens": 339345478.0, + "step": 13112 + }, + { + "epoch": 1.4400395343729409, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4506218433380127, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7220368385314941, + "num_tokens": 339368782.0, + "step": 13113 + }, + { + "epoch": 1.4401493520755546, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2589571475982666, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7046802043914795, + "num_tokens": 339396266.0, + "step": 13114 + }, + { + "epoch": 1.4402591697781681, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.118175506591797, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7295759916305542, + "num_tokens": 339425471.0, + "step": 13115 + }, + { + "epoch": 1.440368987480782, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3150112628936768, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7415529489517212, + "num_tokens": 339448977.0, + "step": 13116 + }, + { + "epoch": 1.4404788051833957, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3277969360351562, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7292643785476685, + "num_tokens": 339471164.0, + "step": 13117 + }, + { + "epoch": 1.4405886228860092, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1611106395721436, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7193827629089355, + "num_tokens": 339497571.0, + "step": 13118 + }, + { + "epoch": 1.440698440588623, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.225738048553467, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7264842987060547, + "num_tokens": 339522326.0, + "step": 13119 + }, + { + "epoch": 1.4408082582912365, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.335955858230591, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7124608755111694, + "num_tokens": 339547243.0, + "step": 13120 + }, + { + "epoch": 1.4409180759938502, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3631348609924316, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.716742753982544, + "num_tokens": 339570558.0, + "step": 13121 + }, + { + "epoch": 1.441027893696464, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3960800170898438, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7400510311126709, + "num_tokens": 339592445.0, + "step": 13122 + }, + { + "epoch": 1.4411377113990775, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.294747829437256, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7107561826705933, + "num_tokens": 339617239.0, + "step": 13123 + }, + { + "epoch": 1.441247529101691, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.503938913345337, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.715535044670105, + "num_tokens": 339639633.0, + "step": 13124 + }, + { + "epoch": 1.4413573468043048, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.341979742050171, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7257046103477478, + "num_tokens": 339663745.0, + "step": 13125 + }, + { + "epoch": 1.4414671645069186, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.49599289894104, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7259626984596252, + "num_tokens": 339683723.0, + "step": 13126 + }, + { + "epoch": 1.4415769822095321, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.678497552871704, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7443910837173462, + "num_tokens": 339702235.0, + "step": 13127 + }, + { + "epoch": 1.4416867999121459, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1922550201416016, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7431729435920715, + "num_tokens": 339727218.0, + "step": 13128 + }, + { + "epoch": 1.4417966176147594, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.141964912414551, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7243595123291016, + "num_tokens": 339755696.0, + "step": 13129 + }, + { + "epoch": 1.4419064353173732, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.138692855834961, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7126132249832153, + "num_tokens": 339781899.0, + "step": 13130 + }, + { + "epoch": 1.442016253019987, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.332412004470825, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7409707307815552, + "num_tokens": 339805352.0, + "step": 13131 + }, + { + "epoch": 1.4421260707226005, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.295970916748047, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7374129891395569, + "num_tokens": 339829741.0, + "step": 13132 + }, + { + "epoch": 1.4422358884252142, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.42153000831604, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7290990948677063, + "num_tokens": 339852460.0, + "step": 13133 + }, + { + "epoch": 1.4423457061278278, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2884581089019775, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7084446549415588, + "num_tokens": 339877706.0, + "step": 13134 + }, + { + "epoch": 1.4424555238304415, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.2982494831085205, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7358648777008057, + "num_tokens": 339901952.0, + "step": 13135 + }, + { + "epoch": 1.442565341533055, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.454242706298828, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7256896495819092, + "num_tokens": 339924641.0, + "step": 13136 + }, + { + "epoch": 1.4426751592356688, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 1.9218062162399292, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7214156985282898, + "num_tokens": 339955650.0, + "step": 13137 + }, + { + "epoch": 1.4427849769382823, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1855509281158447, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7153767347335815, + "num_tokens": 339982083.0, + "step": 13138 + }, + { + "epoch": 1.442894794640896, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3975768089294434, + "learning_rate": 1e-06, + "loss": 1.0132, + "mean_token_accuracy": 0.6923438310623169, + "num_tokens": 340003890.0, + "step": 13139 + }, + { + "epoch": 1.4430046123435099, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.3066916465759277, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7140398621559143, + "num_tokens": 340028979.0, + "step": 13140 + }, + { + "epoch": 1.4431144300461234, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.4062886238098145, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7217204570770264, + "num_tokens": 340051821.0, + "step": 13141 + }, + { + "epoch": 1.4432242477487371, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.218705654144287, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7295625805854797, + "num_tokens": 340077118.0, + "step": 13142 + }, + { + "epoch": 1.4433340654513507, + "ewc_loss": 7.748603820800781e-06, + "grad_norm": 2.1831071376800537, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7088789343833923, + "num_tokens": 340103599.0, + "step": 13143 + }, + { + "epoch": 1.4434438831539644, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.4195055961608887, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7388267517089844, + "num_tokens": 340125652.0, + "step": 13144 + }, + { + "epoch": 1.4435537008565782, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2758476734161377, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7277190685272217, + "num_tokens": 340150073.0, + "step": 13145 + }, + { + "epoch": 1.4436635185591917, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.415546417236328, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7364273071289062, + "num_tokens": 340172018.0, + "step": 13146 + }, + { + "epoch": 1.4437733362618055, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2464423179626465, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7120524048805237, + "num_tokens": 340197922.0, + "step": 13147 + }, + { + "epoch": 1.443883153964419, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.9394092559814453, + "learning_rate": 1e-06, + "loss": 1.0086, + "mean_token_accuracy": 0.6886656284332275, + "num_tokens": 340232043.0, + "step": 13148 + }, + { + "epoch": 1.4439929716670328, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.9830617904663086, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7075425982475281, + "num_tokens": 340263950.0, + "step": 13149 + }, + { + "epoch": 1.4441027893696463, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.245218276977539, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7209810018539429, + "num_tokens": 340290039.0, + "step": 13150 + }, + { + "epoch": 1.44421260707226, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.447347640991211, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7075127363204956, + "num_tokens": 340312673.0, + "step": 13151 + }, + { + "epoch": 1.4443224247748736, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2184088230133057, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7358945608139038, + "num_tokens": 340339078.0, + "step": 13152 + }, + { + "epoch": 1.4444322424774874, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.259916067123413, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7154267430305481, + "num_tokens": 340364148.0, + "step": 13153 + }, + { + "epoch": 1.4445420601801011, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.172959089279175, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7283499240875244, + "num_tokens": 340390357.0, + "step": 13154 + }, + { + "epoch": 1.4446518778827147, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3583590984344482, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7286527752876282, + "num_tokens": 340414598.0, + "step": 13155 + }, + { + "epoch": 1.4447616955853284, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.389986515045166, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7151573300361633, + "num_tokens": 340436416.0, + "step": 13156 + }, + { + "epoch": 1.444871513287942, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.012152671813965, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7137038111686707, + "num_tokens": 340468002.0, + "step": 13157 + }, + { + "epoch": 1.4449813309905557, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.6425671577453613, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7122689485549927, + "num_tokens": 340490276.0, + "step": 13158 + }, + { + "epoch": 1.4450911486931695, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.5355193614959717, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7295918464660645, + "num_tokens": 340511926.0, + "step": 13159 + }, + { + "epoch": 1.445200966395783, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1590635776519775, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7285701036453247, + "num_tokens": 340537993.0, + "step": 13160 + }, + { + "epoch": 1.4453107840983965, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.249293565750122, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7117253541946411, + "num_tokens": 340562051.0, + "step": 13161 + }, + { + "epoch": 1.4454206018010103, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3697409629821777, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7104672193527222, + "num_tokens": 340584541.0, + "step": 13162 + }, + { + "epoch": 1.445530419503624, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.4434163570404053, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7299647331237793, + "num_tokens": 340607199.0, + "step": 13163 + }, + { + "epoch": 1.4456402372062376, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3492977619171143, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7154158353805542, + "num_tokens": 340631608.0, + "step": 13164 + }, + { + "epoch": 1.4457500549088513, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1127545833587646, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7004395723342896, + "num_tokens": 340659819.0, + "step": 13165 + }, + { + "epoch": 1.4458598726114649, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.6446328163146973, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7374966144561768, + "num_tokens": 340678889.0, + "step": 13166 + }, + { + "epoch": 1.4459696903140786, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.373535394668579, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.718769371509552, + "num_tokens": 340704348.0, + "step": 13167 + }, + { + "epoch": 1.4460795080166924, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.198498249053955, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7272648811340332, + "num_tokens": 340728379.0, + "step": 13168 + }, + { + "epoch": 1.446189325719306, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.279398202896118, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7062127590179443, + "num_tokens": 340753423.0, + "step": 13169 + }, + { + "epoch": 1.4462991434219197, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1545913219451904, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7087054252624512, + "num_tokens": 340779801.0, + "step": 13170 + }, + { + "epoch": 1.4464089611245332, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.388700008392334, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.705833911895752, + "num_tokens": 340805614.0, + "step": 13171 + }, + { + "epoch": 1.446518778827147, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.8910404443740845, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.71254962682724, + "num_tokens": 340840393.0, + "step": 13172 + }, + { + "epoch": 1.4466285965297607, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.031665802001953, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.705923855304718, + "num_tokens": 340871054.0, + "step": 13173 + }, + { + "epoch": 1.4467384142323743, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1960458755493164, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6923477649688721, + "num_tokens": 340898566.0, + "step": 13174 + }, + { + "epoch": 1.4468482319349878, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.323065757751465, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7420669794082642, + "num_tokens": 340921348.0, + "step": 13175 + }, + { + "epoch": 1.4469580496376016, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.237910509109497, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7171854972839355, + "num_tokens": 340948757.0, + "step": 13176 + }, + { + "epoch": 1.4470678673402153, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3688466548919678, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.712357759475708, + "num_tokens": 340975774.0, + "step": 13177 + }, + { + "epoch": 1.4471776850428288, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.5200040340423584, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7412996292114258, + "num_tokens": 340996031.0, + "step": 13178 + }, + { + "epoch": 1.4472875027454426, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.4062886238098145, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7139474749565125, + "num_tokens": 341020197.0, + "step": 13179 + }, + { + "epoch": 1.4473973204480561, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1674911975860596, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7226740121841431, + "num_tokens": 341047284.0, + "step": 13180 + }, + { + "epoch": 1.44750713815067, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.320106267929077, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7067252993583679, + "num_tokens": 341074120.0, + "step": 13181 + }, + { + "epoch": 1.4476169558532836, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.0420618057250977, + "learning_rate": 1e-06, + "loss": 1.0259, + "mean_token_accuracy": 0.6991996765136719, + "num_tokens": 341106844.0, + "step": 13182 + }, + { + "epoch": 1.4477267735558972, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.119403123855591, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6907512545585632, + "num_tokens": 341133552.0, + "step": 13183 + }, + { + "epoch": 1.447836591258511, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3125712871551514, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7421442866325378, + "num_tokens": 341159404.0, + "step": 13184 + }, + { + "epoch": 1.4479464089611245, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.06882381439209, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7127875685691833, + "num_tokens": 341188828.0, + "step": 13185 + }, + { + "epoch": 1.4480562266637382, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.267982244491577, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7305311560630798, + "num_tokens": 341212905.0, + "step": 13186 + }, + { + "epoch": 1.448166044366352, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1956582069396973, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.746762216091156, + "num_tokens": 341237002.0, + "step": 13187 + }, + { + "epoch": 1.4482758620689655, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.0166420936584473, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7127968668937683, + "num_tokens": 341268766.0, + "step": 13188 + }, + { + "epoch": 1.448385679771579, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.6051769256591797, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7284075617790222, + "num_tokens": 341290040.0, + "step": 13189 + }, + { + "epoch": 1.4484954974741928, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.957029104232788, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6916276216506958, + "num_tokens": 341322524.0, + "step": 13190 + }, + { + "epoch": 1.4486053151768066, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.368846893310547, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7069505453109741, + "num_tokens": 341345369.0, + "step": 13191 + }, + { + "epoch": 1.44871513287942, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.277371406555176, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7227233648300171, + "num_tokens": 341370538.0, + "step": 13192 + }, + { + "epoch": 1.4488249505820339, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.9638864994049072, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7073478698730469, + "num_tokens": 341402489.0, + "step": 13193 + }, + { + "epoch": 1.4489347682846474, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2744202613830566, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.718200147151947, + "num_tokens": 341428296.0, + "step": 13194 + }, + { + "epoch": 1.4490445859872612, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.209569215774536, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.7061551809310913, + "num_tokens": 341455598.0, + "step": 13195 + }, + { + "epoch": 1.449154403689875, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1250503063201904, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7153983116149902, + "num_tokens": 341485051.0, + "step": 13196 + }, + { + "epoch": 1.4492642213924884, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2478108406066895, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7171887159347534, + "num_tokens": 341508687.0, + "step": 13197 + }, + { + "epoch": 1.4493740390951022, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1407699584960938, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7483025193214417, + "num_tokens": 341536130.0, + "step": 13198 + }, + { + "epoch": 1.4494838567977157, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.25905704498291, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7225189208984375, + "num_tokens": 341560952.0, + "step": 13199 + }, + { + "epoch": 1.4495936745003295, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.0741841793060303, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7097176909446716, + "num_tokens": 341591777.0, + "step": 13200 + }, + { + "epoch": 1.449703492202943, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2289938926696777, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7120347023010254, + "num_tokens": 341621448.0, + "step": 13201 + }, + { + "epoch": 1.4498133099055568, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3609259128570557, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.6998685598373413, + "num_tokens": 341647764.0, + "step": 13202 + }, + { + "epoch": 1.4499231276081703, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.179394483566284, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6853004097938538, + "num_tokens": 341677722.0, + "step": 13203 + }, + { + "epoch": 1.450032945310784, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.477505683898926, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.709077775478363, + "num_tokens": 341699976.0, + "step": 13204 + }, + { + "epoch": 1.4501427630133978, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.202030658721924, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.733318567276001, + "num_tokens": 341727137.0, + "step": 13205 + }, + { + "epoch": 1.4502525807160114, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.0291998386383057, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7163957357406616, + "num_tokens": 341756721.0, + "step": 13206 + }, + { + "epoch": 1.4503623984186251, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3699088096618652, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.703109622001648, + "num_tokens": 341781790.0, + "step": 13207 + }, + { + "epoch": 1.4504722161212387, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3719122409820557, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7261283993721008, + "num_tokens": 341804509.0, + "step": 13208 + }, + { + "epoch": 1.4505820338238524, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.519240617752075, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7029621601104736, + "num_tokens": 341828413.0, + "step": 13209 + }, + { + "epoch": 1.4506918515264662, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.25710391998291, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7319386005401611, + "num_tokens": 341851612.0, + "step": 13210 + }, + { + "epoch": 1.4508016692290797, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.430647373199463, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7540183067321777, + "num_tokens": 341873403.0, + "step": 13211 + }, + { + "epoch": 1.4509114869316933, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.0181591510772705, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7207372188568115, + "num_tokens": 341903244.0, + "step": 13212 + }, + { + "epoch": 1.451021304634307, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.777545213699341, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7239317893981934, + "num_tokens": 341921413.0, + "step": 13213 + }, + { + "epoch": 1.4511311223369208, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.255361318588257, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7267125844955444, + "num_tokens": 341946752.0, + "step": 13214 + }, + { + "epoch": 1.4512409400395343, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1797714233398438, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7239747643470764, + "num_tokens": 341974994.0, + "step": 13215 + }, + { + "epoch": 1.451350757742148, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2667298316955566, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7099925875663757, + "num_tokens": 342001400.0, + "step": 13216 + }, + { + "epoch": 1.4514605754447616, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.436014413833618, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7087363004684448, + "num_tokens": 342025161.0, + "step": 13217 + }, + { + "epoch": 1.4515703931473753, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.1919190883636475, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7270039319992065, + "num_tokens": 342052036.0, + "step": 13218 + }, + { + "epoch": 1.451680210849989, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.01953125, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7100733518600464, + "num_tokens": 342081345.0, + "step": 13219 + }, + { + "epoch": 1.4517900285526026, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.4587271213531494, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7174525856971741, + "num_tokens": 342103168.0, + "step": 13220 + }, + { + "epoch": 1.4518998462552164, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2090587615966797, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7166332006454468, + "num_tokens": 342128722.0, + "step": 13221 + }, + { + "epoch": 1.45200966395783, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.225525140762329, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7127649784088135, + "num_tokens": 342153962.0, + "step": 13222 + }, + { + "epoch": 1.4521194816604437, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.434558868408203, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7011940479278564, + "num_tokens": 342177007.0, + "step": 13223 + }, + { + "epoch": 1.4522292993630574, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2229549884796143, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7194896936416626, + "num_tokens": 342204681.0, + "step": 13224 + }, + { + "epoch": 1.452339117065671, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.5428972244262695, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7441223859786987, + "num_tokens": 342223924.0, + "step": 13225 + }, + { + "epoch": 1.4524489347682845, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.20302414894104, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7334146499633789, + "num_tokens": 342250807.0, + "step": 13226 + }, + { + "epoch": 1.4525587524708983, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.145709991455078, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7260851860046387, + "num_tokens": 342275965.0, + "step": 13227 + }, + { + "epoch": 1.452668570173512, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.021519660949707, + "learning_rate": 1e-06, + "loss": 1.0309, + "mean_token_accuracy": 0.6866799592971802, + "num_tokens": 342307194.0, + "step": 13228 + }, + { + "epoch": 1.4527783878761256, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.159695625305176, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7152338624000549, + "num_tokens": 342335970.0, + "step": 13229 + }, + { + "epoch": 1.4528882055787393, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.3339967727661133, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7206462621688843, + "num_tokens": 342359907.0, + "step": 13230 + }, + { + "epoch": 1.4529980232813529, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.258662700653076, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7258054614067078, + "num_tokens": 342384495.0, + "step": 13231 + }, + { + "epoch": 1.4531078409839666, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 1.9515665769577026, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7000764608383179, + "num_tokens": 342419582.0, + "step": 13232 + }, + { + "epoch": 1.4532176586865804, + "ewc_loss": 7.808208465576172e-06, + "grad_norm": 2.2146692276000977, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6968820095062256, + "num_tokens": 342446851.0, + "step": 13233 + }, + { + "epoch": 1.453327476389194, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3505592346191406, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7350565195083618, + "num_tokens": 342470904.0, + "step": 13234 + }, + { + "epoch": 1.4534372940918077, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.481862783432007, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7298677563667297, + "num_tokens": 342491848.0, + "step": 13235 + }, + { + "epoch": 1.4535471117944212, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.104687452316284, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7141416072845459, + "num_tokens": 342521773.0, + "step": 13236 + }, + { + "epoch": 1.453656929497035, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2370386123657227, + "learning_rate": 1e-06, + "loss": 0.9912, + "mean_token_accuracy": 0.6939729452133179, + "num_tokens": 342548782.0, + "step": 13237 + }, + { + "epoch": 1.4537667471996487, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.0882363319396973, + "learning_rate": 1e-06, + "loss": 0.9966, + "mean_token_accuracy": 0.6880444288253784, + "num_tokens": 342581713.0, + "step": 13238 + }, + { + "epoch": 1.4538765649022622, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.119575262069702, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7301868200302124, + "num_tokens": 342608817.0, + "step": 13239 + }, + { + "epoch": 1.4539863826048758, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.0144567489624023, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7240252494812012, + "num_tokens": 342635871.0, + "step": 13240 + }, + { + "epoch": 1.4540962003074895, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.253002882003784, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7197532057762146, + "num_tokens": 342660502.0, + "step": 13241 + }, + { + "epoch": 1.4542060180101033, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.460222005844116, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7310659885406494, + "num_tokens": 342681603.0, + "step": 13242 + }, + { + "epoch": 1.4543158357127168, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.1586124897003174, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7318347692489624, + "num_tokens": 342707872.0, + "step": 13243 + }, + { + "epoch": 1.4544256534153306, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.203596353530884, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6947520971298218, + "num_tokens": 342738552.0, + "step": 13244 + }, + { + "epoch": 1.4545354711179441, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3006961345672607, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7241355180740356, + "num_tokens": 342762795.0, + "step": 13245 + }, + { + "epoch": 1.4546452888205579, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.348629951477051, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7278968095779419, + "num_tokens": 342787005.0, + "step": 13246 + }, + { + "epoch": 1.4547551065231716, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.4620108604431152, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7124117612838745, + "num_tokens": 342807908.0, + "step": 13247 + }, + { + "epoch": 1.4548649242257852, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.0710203647613525, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7013893127441406, + "num_tokens": 342836125.0, + "step": 13248 + }, + { + "epoch": 1.454974741928399, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 1.911935567855835, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7345728874206543, + "num_tokens": 342868122.0, + "step": 13249 + }, + { + "epoch": 1.4550845596310125, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.11452317237854, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7281458377838135, + "num_tokens": 342895977.0, + "step": 13250 + }, + { + "epoch": 1.4551943773336262, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.2188355922698975, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.7068326473236084, + "num_tokens": 342921265.0, + "step": 13251 + }, + { + "epoch": 1.45530419503624, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.208972215652466, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7231007814407349, + "num_tokens": 342948013.0, + "step": 13252 + }, + { + "epoch": 1.4554140127388535, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 1.972171425819397, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7189222574234009, + "num_tokens": 342979030.0, + "step": 13253 + }, + { + "epoch": 1.455523830441467, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3619256019592285, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7272589802742004, + "num_tokens": 343001334.0, + "step": 13254 + }, + { + "epoch": 1.4556336481440808, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.150942802429199, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7091938257217407, + "num_tokens": 343030971.0, + "step": 13255 + }, + { + "epoch": 1.4557434658466946, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3020448684692383, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7044659852981567, + "num_tokens": 343057286.0, + "step": 13256 + }, + { + "epoch": 1.455853283549308, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.0849063396453857, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.708632230758667, + "num_tokens": 343087943.0, + "step": 13257 + }, + { + "epoch": 1.4559631012519219, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3843367099761963, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7312437891960144, + "num_tokens": 343110953.0, + "step": 13258 + }, + { + "epoch": 1.4560729189545354, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.4358584880828857, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7193629741668701, + "num_tokens": 343133106.0, + "step": 13259 + }, + { + "epoch": 1.4561827366571491, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.305982828140259, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7270404696464539, + "num_tokens": 343157036.0, + "step": 13260 + }, + { + "epoch": 1.456292554359763, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.1390717029571533, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6909434199333191, + "num_tokens": 343186413.0, + "step": 13261 + }, + { + "epoch": 1.4564023720623764, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 1.9672250747680664, + "learning_rate": 1e-06, + "loss": 0.9831, + "mean_token_accuracy": 0.7014166116714478, + "num_tokens": 343218981.0, + "step": 13262 + }, + { + "epoch": 1.4565121897649902, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.1402885913848877, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7070035934448242, + "num_tokens": 343245759.0, + "step": 13263 + }, + { + "epoch": 1.4566220074676037, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.4985640048980713, + "learning_rate": 1e-06, + "loss": 0.731, + "mean_token_accuracy": 0.7623140811920166, + "num_tokens": 343265382.0, + "step": 13264 + }, + { + "epoch": 1.4567318251702175, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.528698205947876, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7050594091415405, + "num_tokens": 343287786.0, + "step": 13265 + }, + { + "epoch": 1.456841642872831, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.22790265083313, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7267985343933105, + "num_tokens": 343312571.0, + "step": 13266 + }, + { + "epoch": 1.4569514605754448, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.177483320236206, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7008078694343567, + "num_tokens": 343338845.0, + "step": 13267 + }, + { + "epoch": 1.4570612782780583, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.089839220046997, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7102140188217163, + "num_tokens": 343366259.0, + "step": 13268 + }, + { + "epoch": 1.457171095980672, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.3266496658325195, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7071387767791748, + "num_tokens": 343389709.0, + "step": 13269 + }, + { + "epoch": 1.4572809136832858, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.506093740463257, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7538028359413147, + "num_tokens": 343409779.0, + "step": 13270 + }, + { + "epoch": 1.4573907313858994, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.429597854614258, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7265766859054565, + "num_tokens": 343433056.0, + "step": 13271 + }, + { + "epoch": 1.4575005490885131, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.427337408065796, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7198657393455505, + "num_tokens": 343455858.0, + "step": 13272 + }, + { + "epoch": 1.4576103667911267, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.681305408477783, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7252863049507141, + "num_tokens": 343474974.0, + "step": 13273 + }, + { + "epoch": 1.4577201844937404, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.540501356124878, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7186070680618286, + "num_tokens": 343494711.0, + "step": 13274 + }, + { + "epoch": 1.4578300021963542, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.220324754714966, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7129366993904114, + "num_tokens": 343523387.0, + "step": 13275 + }, + { + "epoch": 1.4579398198989677, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.5120041370391846, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7369588017463684, + "num_tokens": 343544079.0, + "step": 13276 + }, + { + "epoch": 1.4580496376015812, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.0549826622009277, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7209339737892151, + "num_tokens": 343571845.0, + "step": 13277 + }, + { + "epoch": 1.458159455304195, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.1847286224365234, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7265830636024475, + "num_tokens": 343596591.0, + "step": 13278 + }, + { + "epoch": 1.4582692730068088, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.138589859008789, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7302209138870239, + "num_tokens": 343622319.0, + "step": 13279 + }, + { + "epoch": 1.4583790907094223, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.280932664871216, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7199327945709229, + "num_tokens": 343645477.0, + "step": 13280 + }, + { + "epoch": 1.458488908412036, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.791358470916748, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7164697647094727, + "num_tokens": 343664693.0, + "step": 13281 + }, + { + "epoch": 1.4585987261146496, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.5585882663726807, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7468397617340088, + "num_tokens": 343683758.0, + "step": 13282 + }, + { + "epoch": 1.4587085438172633, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.293576240539551, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7096433043479919, + "num_tokens": 343708334.0, + "step": 13283 + }, + { + "epoch": 1.458818361519877, + "ewc_loss": 7.867813110351562e-06, + "grad_norm": 2.018468141555786, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7131012678146362, + "num_tokens": 343740386.0, + "step": 13284 + }, + { + "epoch": 1.4589281792224906, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2732644081115723, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7313694357872009, + "num_tokens": 343765478.0, + "step": 13285 + }, + { + "epoch": 1.4590379969251044, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3581721782684326, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7305280566215515, + "num_tokens": 343790678.0, + "step": 13286 + }, + { + "epoch": 1.459147814627718, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.075608253479004, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7154816389083862, + "num_tokens": 343818608.0, + "step": 13287 + }, + { + "epoch": 1.4592576323303317, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.103456735610962, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7238982915878296, + "num_tokens": 343844894.0, + "step": 13288 + }, + { + "epoch": 1.4593674500329454, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.4011168479919434, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7211670875549316, + "num_tokens": 343866536.0, + "step": 13289 + }, + { + "epoch": 1.459477267735559, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.0058934688568115, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.717441201210022, + "num_tokens": 343894798.0, + "step": 13290 + }, + { + "epoch": 1.4595870854381725, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2121241092681885, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7013352513313293, + "num_tokens": 343921430.0, + "step": 13291 + }, + { + "epoch": 1.4596969031407863, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.0775575637817383, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7341585159301758, + "num_tokens": 343949715.0, + "step": 13292 + }, + { + "epoch": 1.4598067208434, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2852320671081543, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7354220151901245, + "num_tokens": 343974416.0, + "step": 13293 + }, + { + "epoch": 1.4599165385460136, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.1444931030273438, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.6974364519119263, + "num_tokens": 344004519.0, + "step": 13294 + }, + { + "epoch": 1.4600263562486273, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2475314140319824, + "learning_rate": 1e-06, + "loss": 0.7483, + "mean_token_accuracy": 0.7633467316627502, + "num_tokens": 344028304.0, + "step": 13295 + }, + { + "epoch": 1.4601361739512408, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.277003049850464, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7356764078140259, + "num_tokens": 344053256.0, + "step": 13296 + }, + { + "epoch": 1.4602459916538546, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.716891050338745, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7246326208114624, + "num_tokens": 344073480.0, + "step": 13297 + }, + { + "epoch": 1.4603558093564684, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.042233943939209, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7008948922157288, + "num_tokens": 344103404.0, + "step": 13298 + }, + { + "epoch": 1.460465627059082, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.1491284370422363, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7242143154144287, + "num_tokens": 344129978.0, + "step": 13299 + }, + { + "epoch": 1.4605754447616957, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2548789978027344, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.741606593132019, + "num_tokens": 344153913.0, + "step": 13300 + }, + { + "epoch": 1.4606852624643092, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.358067035675049, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7191590666770935, + "num_tokens": 344179762.0, + "step": 13301 + }, + { + "epoch": 1.460795080166923, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.380582809448242, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7244549989700317, + "num_tokens": 344203479.0, + "step": 13302 + }, + { + "epoch": 1.4609048978695367, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3874287605285645, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7129616141319275, + "num_tokens": 344227702.0, + "step": 13303 + }, + { + "epoch": 1.4610147155721502, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.496933698654175, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7235611081123352, + "num_tokens": 344250615.0, + "step": 13304 + }, + { + "epoch": 1.4611245332747638, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.067854642868042, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.6998986005783081, + "num_tokens": 344279535.0, + "step": 13305 + }, + { + "epoch": 1.4612343509773775, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 1.8756266832351685, + "learning_rate": 1e-06, + "loss": 1.1082, + "mean_token_accuracy": 0.6678178906440735, + "num_tokens": 344315144.0, + "step": 13306 + }, + { + "epoch": 1.4613441686799913, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.32295298576355, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7141441702842712, + "num_tokens": 344338533.0, + "step": 13307 + }, + { + "epoch": 1.4614539863826048, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.241011142730713, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7092622518539429, + "num_tokens": 344364248.0, + "step": 13308 + }, + { + "epoch": 1.4615638040852186, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.508753776550293, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7306044101715088, + "num_tokens": 344386105.0, + "step": 13309 + }, + { + "epoch": 1.4616736217878321, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.459158420562744, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7230749130249023, + "num_tokens": 344407088.0, + "step": 13310 + }, + { + "epoch": 1.4617834394904459, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.442432403564453, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.715539813041687, + "num_tokens": 344428868.0, + "step": 13311 + }, + { + "epoch": 1.4618932571930596, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2627546787261963, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7064812183380127, + "num_tokens": 344454809.0, + "step": 13312 + }, + { + "epoch": 1.4620030748956732, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.1702237129211426, + "learning_rate": 1e-06, + "loss": 1.0029, + "mean_token_accuracy": 0.6919169425964355, + "num_tokens": 344487596.0, + "step": 13313 + }, + { + "epoch": 1.462112892598287, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.257131338119507, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.6969588398933411, + "num_tokens": 344515855.0, + "step": 13314 + }, + { + "epoch": 1.4622227103009005, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2386298179626465, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7281311750411987, + "num_tokens": 344540736.0, + "step": 13315 + }, + { + "epoch": 1.4623325280035142, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2047417163848877, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7008241415023804, + "num_tokens": 344569990.0, + "step": 13316 + }, + { + "epoch": 1.462442345706128, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.066437005996704, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7305735349655151, + "num_tokens": 344598249.0, + "step": 13317 + }, + { + "epoch": 1.4625521634087415, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2917723655700684, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.6984778046607971, + "num_tokens": 344623371.0, + "step": 13318 + }, + { + "epoch": 1.462661981111355, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.5757253170013428, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.73031085729599, + "num_tokens": 344645429.0, + "step": 13319 + }, + { + "epoch": 1.4627717988139688, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.2580151557922363, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7297046184539795, + "num_tokens": 344668941.0, + "step": 13320 + }, + { + "epoch": 1.4628816165165826, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.615562915802002, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7505187392234802, + "num_tokens": 344687947.0, + "step": 13321 + }, + { + "epoch": 1.462991434219196, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.1829802989959717, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7470599412918091, + "num_tokens": 344714531.0, + "step": 13322 + }, + { + "epoch": 1.4631012519218098, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 1.9646179676055908, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7204447388648987, + "num_tokens": 344745952.0, + "step": 13323 + }, + { + "epoch": 1.4632110696244234, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.30660080909729, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7226535081863403, + "num_tokens": 344769489.0, + "step": 13324 + }, + { + "epoch": 1.4633208873270371, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.1802585124969482, + "learning_rate": 1e-06, + "loss": 1.011, + "mean_token_accuracy": 0.6888723373413086, + "num_tokens": 344798325.0, + "step": 13325 + }, + { + "epoch": 1.463430705029651, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.074732542037964, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7158489227294922, + "num_tokens": 344826816.0, + "step": 13326 + }, + { + "epoch": 1.4635405227322644, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 1.9742294549942017, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.7001996040344238, + "num_tokens": 344857874.0, + "step": 13327 + }, + { + "epoch": 1.4636503404348782, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.251394271850586, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7064294815063477, + "num_tokens": 344883472.0, + "step": 13328 + }, + { + "epoch": 1.4637601581374917, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.220909595489502, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7166897058486938, + "num_tokens": 344911717.0, + "step": 13329 + }, + { + "epoch": 1.4638699758401055, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3386781215667725, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7080776691436768, + "num_tokens": 344935917.0, + "step": 13330 + }, + { + "epoch": 1.463979793542719, + "ewc_loss": 7.927417755126953e-06, + "grad_norm": 2.3046419620513916, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7262511253356934, + "num_tokens": 344960203.0, + "step": 13331 + }, + { + "epoch": 1.4640896112453328, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2468855381011963, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7293816804885864, + "num_tokens": 344983619.0, + "step": 13332 + }, + { + "epoch": 1.4641994289479463, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9775041341781616, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7139996886253357, + "num_tokens": 345015122.0, + "step": 13333 + }, + { + "epoch": 1.46430924665056, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.000905752182007, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7143734097480774, + "num_tokens": 345045631.0, + "step": 13334 + }, + { + "epoch": 1.4644190643531738, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.072704315185547, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7182016372680664, + "num_tokens": 345076630.0, + "step": 13335 + }, + { + "epoch": 1.4645288820557874, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1198368072509766, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7401611804962158, + "num_tokens": 345101823.0, + "step": 13336 + }, + { + "epoch": 1.464638699758401, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.30759859085083, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7198390960693359, + "num_tokens": 345128795.0, + "step": 13337 + }, + { + "epoch": 1.4647485174610146, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.163853406906128, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7165069580078125, + "num_tokens": 345156852.0, + "step": 13338 + }, + { + "epoch": 1.4648583351636284, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.307422399520874, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7291038632392883, + "num_tokens": 345182349.0, + "step": 13339 + }, + { + "epoch": 1.4649681528662422, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.10402512550354, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7330788373947144, + "num_tokens": 345210182.0, + "step": 13340 + }, + { + "epoch": 1.4650779705688557, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.487517833709717, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7545644044876099, + "num_tokens": 345230677.0, + "step": 13341 + }, + { + "epoch": 1.4651877882714692, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2092878818511963, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7243467569351196, + "num_tokens": 345256864.0, + "step": 13342 + }, + { + "epoch": 1.465297605974083, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1879091262817383, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7391948699951172, + "num_tokens": 345283478.0, + "step": 13343 + }, + { + "epoch": 1.4654074236766967, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.9861433506011963, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7099566459655762, + "num_tokens": 345307144.0, + "step": 13344 + }, + { + "epoch": 1.4655172413793103, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.019465684890747, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.7024425268173218, + "num_tokens": 345336359.0, + "step": 13345 + }, + { + "epoch": 1.465627059081924, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2569897174835205, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7334759831428528, + "num_tokens": 345359927.0, + "step": 13346 + }, + { + "epoch": 1.4657368767845376, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5718677043914795, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.735202431678772, + "num_tokens": 345380985.0, + "step": 13347 + }, + { + "epoch": 1.4658466944871513, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.137576103210449, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.6931419968605042, + "num_tokens": 345409928.0, + "step": 13348 + }, + { + "epoch": 1.465956512189765, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.148587942123413, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7194592952728271, + "num_tokens": 345435622.0, + "step": 13349 + }, + { + "epoch": 1.4660663298923786, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.39312744140625, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7177799940109253, + "num_tokens": 345457381.0, + "step": 13350 + }, + { + "epoch": 1.4661761475949924, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.297215461730957, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7302250862121582, + "num_tokens": 345480374.0, + "step": 13351 + }, + { + "epoch": 1.466285965297606, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.262179374694824, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7525230050086975, + "num_tokens": 345505238.0, + "step": 13352 + }, + { + "epoch": 1.4663957830002197, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2606303691864014, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6992605924606323, + "num_tokens": 345531950.0, + "step": 13353 + }, + { + "epoch": 1.4665056007028334, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3686280250549316, + "learning_rate": 1e-06, + "loss": 0.9944, + "mean_token_accuracy": 0.6963577270507812, + "num_tokens": 345556257.0, + "step": 13354 + }, + { + "epoch": 1.466615418405447, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.283477306365967, + "learning_rate": 1e-06, + "loss": 1.0526, + "mean_token_accuracy": 0.6927952766418457, + "num_tokens": 345586451.0, + "step": 13355 + }, + { + "epoch": 1.4667252361080605, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4951086044311523, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7177953720092773, + "num_tokens": 345610752.0, + "step": 13356 + }, + { + "epoch": 1.4668350538106742, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6541082859039307, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7322326898574829, + "num_tokens": 345631872.0, + "step": 13357 + }, + { + "epoch": 1.466944871513288, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1234023571014404, + "learning_rate": 1e-06, + "loss": 0.983, + "mean_token_accuracy": 0.7020893096923828, + "num_tokens": 345661342.0, + "step": 13358 + }, + { + "epoch": 1.4670546892159015, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.382380723953247, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7567501664161682, + "num_tokens": 345683908.0, + "step": 13359 + }, + { + "epoch": 1.4671645069185153, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.239013195037842, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7086979746818542, + "num_tokens": 345709061.0, + "step": 13360 + }, + { + "epoch": 1.4672743246211288, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2748053073883057, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7205506563186646, + "num_tokens": 345734093.0, + "step": 13361 + }, + { + "epoch": 1.4673841423237426, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.179837942123413, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.705319881439209, + "num_tokens": 345762128.0, + "step": 13362 + }, + { + "epoch": 1.4674939600263563, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2634687423706055, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7226384878158569, + "num_tokens": 345785034.0, + "step": 13363 + }, + { + "epoch": 1.4676037777289699, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2863943576812744, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7198873162269592, + "num_tokens": 345808752.0, + "step": 13364 + }, + { + "epoch": 1.4677135954315836, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1703238487243652, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7108187079429626, + "num_tokens": 345838267.0, + "step": 13365 + }, + { + "epoch": 1.4678234131341972, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.303600311279297, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7370648384094238, + "num_tokens": 345861573.0, + "step": 13366 + }, + { + "epoch": 1.467933230836811, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.143031597137451, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.718230128288269, + "num_tokens": 345888270.0, + "step": 13367 + }, + { + "epoch": 1.4680430485394247, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5368564128875732, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7333307266235352, + "num_tokens": 345909122.0, + "step": 13368 + }, + { + "epoch": 1.4681528662420382, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9442623853683472, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7065613865852356, + "num_tokens": 345941407.0, + "step": 13369 + }, + { + "epoch": 1.4682626839446518, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1421701908111572, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7114272117614746, + "num_tokens": 345967749.0, + "step": 13370 + }, + { + "epoch": 1.4683725016472655, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6103339195251465, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7423380613327026, + "num_tokens": 345986427.0, + "step": 13371 + }, + { + "epoch": 1.4684823193498793, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.267584800720215, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7201070785522461, + "num_tokens": 346014159.0, + "step": 13372 + }, + { + "epoch": 1.4685921370524928, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1169509887695312, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7240254878997803, + "num_tokens": 346043607.0, + "step": 13373 + }, + { + "epoch": 1.4687019547551066, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3334269523620605, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7245678901672363, + "num_tokens": 346068148.0, + "step": 13374 + }, + { + "epoch": 1.46881177245772, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.512338399887085, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7183806896209717, + "num_tokens": 346089553.0, + "step": 13375 + }, + { + "epoch": 1.4689215901603339, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.139248847961426, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7101132869720459, + "num_tokens": 346119179.0, + "step": 13376 + }, + { + "epoch": 1.4690314078629476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5172760486602783, + "learning_rate": 1e-06, + "loss": 0.7802, + "mean_token_accuracy": 0.7494328618049622, + "num_tokens": 346137783.0, + "step": 13377 + }, + { + "epoch": 1.4691412255655611, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.306194543838501, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7125709652900696, + "num_tokens": 346163726.0, + "step": 13378 + }, + { + "epoch": 1.469251043268175, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0638580322265625, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7104527354240417, + "num_tokens": 346192312.0, + "step": 13379 + }, + { + "epoch": 1.4693608609707884, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.125887632369995, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7067481279373169, + "num_tokens": 346221468.0, + "step": 13380 + }, + { + "epoch": 1.4694706786734022, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1342849731445312, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7173594236373901, + "num_tokens": 346247660.0, + "step": 13381 + }, + { + "epoch": 1.4695804963760157, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.196901559829712, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7145382761955261, + "num_tokens": 346275419.0, + "step": 13382 + }, + { + "epoch": 1.4696903140786295, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3212788105010986, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7302924394607544, + "num_tokens": 346299583.0, + "step": 13383 + }, + { + "epoch": 1.469800131781243, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.666084051132202, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7352752089500427, + "num_tokens": 346319314.0, + "step": 13384 + }, + { + "epoch": 1.4699099494838568, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2297446727752686, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6994078159332275, + "num_tokens": 346344487.0, + "step": 13385 + }, + { + "epoch": 1.4700197671864705, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1439414024353027, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7277613878250122, + "num_tokens": 346371288.0, + "step": 13386 + }, + { + "epoch": 1.470129584889084, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.27990984916687, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7301685810089111, + "num_tokens": 346397949.0, + "step": 13387 + }, + { + "epoch": 1.4702394025916978, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5709855556488037, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.727180540561676, + "num_tokens": 346418085.0, + "step": 13388 + }, + { + "epoch": 1.4703492202943114, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1123578548431396, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7123937606811523, + "num_tokens": 346445774.0, + "step": 13389 + }, + { + "epoch": 1.4704590379969251, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9154033660888672, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7119121551513672, + "num_tokens": 346479092.0, + "step": 13390 + }, + { + "epoch": 1.4705688556995389, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.639756441116333, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7087218761444092, + "num_tokens": 346500975.0, + "step": 13391 + }, + { + "epoch": 1.4706786734021524, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.321350574493408, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7202096581459045, + "num_tokens": 346524780.0, + "step": 13392 + }, + { + "epoch": 1.4707884911047662, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1761221885681152, + "learning_rate": 1e-06, + "loss": 0.9838, + "mean_token_accuracy": 0.7016117572784424, + "num_tokens": 346554229.0, + "step": 13393 + }, + { + "epoch": 1.4708983088073797, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.097269058227539, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7347757816314697, + "num_tokens": 346581471.0, + "step": 13394 + }, + { + "epoch": 1.4710081265099935, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2712182998657227, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7289518117904663, + "num_tokens": 346606506.0, + "step": 13395 + }, + { + "epoch": 1.471117944212607, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2387778759002686, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7063137292861938, + "num_tokens": 346632543.0, + "step": 13396 + }, + { + "epoch": 1.4712277619152208, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5179498195648193, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7256923317909241, + "num_tokens": 346654268.0, + "step": 13397 + }, + { + "epoch": 1.4713375796178343, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.034766674041748, + "learning_rate": 1e-06, + "loss": 0.7861, + "mean_token_accuracy": 0.7560040950775146, + "num_tokens": 346681670.0, + "step": 13398 + }, + { + "epoch": 1.471447397320448, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2427589893341064, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.743358850479126, + "num_tokens": 346706343.0, + "step": 13399 + }, + { + "epoch": 1.4715572150230618, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2033047676086426, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7255433797836304, + "num_tokens": 346731063.0, + "step": 13400 + }, + { + "epoch": 1.4716670327256753, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3100929260253906, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7242457866668701, + "num_tokens": 346755843.0, + "step": 13401 + }, + { + "epoch": 1.471776850428289, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2037250995635986, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.710205078125, + "num_tokens": 346782314.0, + "step": 13402 + }, + { + "epoch": 1.4718866681309026, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.444317579269409, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7362507581710815, + "num_tokens": 346804881.0, + "step": 13403 + }, + { + "epoch": 1.4719964858335164, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1944642066955566, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.6945786476135254, + "num_tokens": 346833038.0, + "step": 13404 + }, + { + "epoch": 1.4721063035361301, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4404149055480957, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7376090884208679, + "num_tokens": 346854806.0, + "step": 13405 + }, + { + "epoch": 1.4722161212387437, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0661776065826416, + "learning_rate": 1e-06, + "loss": 0.9837, + "mean_token_accuracy": 0.7004377841949463, + "num_tokens": 346886827.0, + "step": 13406 + }, + { + "epoch": 1.4723259389413572, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2874014377593994, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7190259695053101, + "num_tokens": 346909719.0, + "step": 13407 + }, + { + "epoch": 1.472435756643971, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.139280319213867, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7217610478401184, + "num_tokens": 346936087.0, + "step": 13408 + }, + { + "epoch": 1.4725455743465847, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.327927827835083, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7029301524162292, + "num_tokens": 346962253.0, + "step": 13409 + }, + { + "epoch": 1.4726553920491983, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.244152545928955, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7157661318778992, + "num_tokens": 346987083.0, + "step": 13410 + }, + { + "epoch": 1.472765209751812, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4794883728027344, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7356197834014893, + "num_tokens": 347008954.0, + "step": 13411 + }, + { + "epoch": 1.4728750274544256, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.228937864303589, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7082497477531433, + "num_tokens": 347037526.0, + "step": 13412 + }, + { + "epoch": 1.4729848451570393, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2606377601623535, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7314007878303528, + "num_tokens": 347060196.0, + "step": 13413 + }, + { + "epoch": 1.473094662859653, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6603353023529053, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7354952096939087, + "num_tokens": 347079279.0, + "step": 13414 + }, + { + "epoch": 1.4732044805622666, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.393294095993042, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7218597531318665, + "num_tokens": 347101296.0, + "step": 13415 + }, + { + "epoch": 1.4733142982648804, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2849221229553223, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7290564775466919, + "num_tokens": 347124779.0, + "step": 13416 + }, + { + "epoch": 1.473424115967494, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.591628313064575, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7285906672477722, + "num_tokens": 347148769.0, + "step": 13417 + }, + { + "epoch": 1.4735339336701077, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.07320499420166, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.731775164604187, + "num_tokens": 347176103.0, + "step": 13418 + }, + { + "epoch": 1.4736437513727214, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2626640796661377, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7189915180206299, + "num_tokens": 347201415.0, + "step": 13419 + }, + { + "epoch": 1.473753569075335, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.166964054107666, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7375897765159607, + "num_tokens": 347228122.0, + "step": 13420 + }, + { + "epoch": 1.4738633867779485, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1128790378570557, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7194620370864868, + "num_tokens": 347254149.0, + "step": 13421 + }, + { + "epoch": 1.4739732044805622, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3863770961761475, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7293908596038818, + "num_tokens": 347278195.0, + "step": 13422 + }, + { + "epoch": 1.474083022183176, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3186042308807373, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7468495965003967, + "num_tokens": 347301112.0, + "step": 13423 + }, + { + "epoch": 1.4741928398857895, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1966071128845215, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7168850898742676, + "num_tokens": 347326955.0, + "step": 13424 + }, + { + "epoch": 1.4743026575884033, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.284473419189453, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7183163166046143, + "num_tokens": 347352680.0, + "step": 13425 + }, + { + "epoch": 1.4744124752910168, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2427263259887695, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7166174650192261, + "num_tokens": 347379559.0, + "step": 13426 + }, + { + "epoch": 1.4745222929936306, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.25368332862854, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7352025508880615, + "num_tokens": 347404330.0, + "step": 13427 + }, + { + "epoch": 1.4746321106962443, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1251134872436523, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7170611619949341, + "num_tokens": 347432882.0, + "step": 13428 + }, + { + "epoch": 1.4747419283988579, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.139198064804077, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7232456207275391, + "num_tokens": 347460572.0, + "step": 13429 + }, + { + "epoch": 1.4748517461014716, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.310129404067993, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7105697393417358, + "num_tokens": 347486650.0, + "step": 13430 + }, + { + "epoch": 1.4749615638040852, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1862027645111084, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7294224500656128, + "num_tokens": 347514758.0, + "step": 13431 + }, + { + "epoch": 1.475071381506699, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2521848678588867, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7181382775306702, + "num_tokens": 347539986.0, + "step": 13432 + }, + { + "epoch": 1.4751811992093127, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.227691650390625, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7134863138198853, + "num_tokens": 347563513.0, + "step": 13433 + }, + { + "epoch": 1.4752910169119262, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0857903957366943, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7285960912704468, + "num_tokens": 347591159.0, + "step": 13434 + }, + { + "epoch": 1.4754008346145397, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1824748516082764, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7230023741722107, + "num_tokens": 347618018.0, + "step": 13435 + }, + { + "epoch": 1.4755106523171535, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.146127462387085, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7023625373840332, + "num_tokens": 347645281.0, + "step": 13436 + }, + { + "epoch": 1.4756204700197673, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3163211345672607, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7205683588981628, + "num_tokens": 347669321.0, + "step": 13437 + }, + { + "epoch": 1.4757302877223808, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.183830738067627, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7094330787658691, + "num_tokens": 347696628.0, + "step": 13438 + }, + { + "epoch": 1.4758401054249946, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.320326089859009, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.7099796533584595, + "num_tokens": 347720666.0, + "step": 13439 + }, + { + "epoch": 1.475949923127608, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2854127883911133, + "learning_rate": 1e-06, + "loss": 1.0321, + "mean_token_accuracy": 0.6903607249259949, + "num_tokens": 347748398.0, + "step": 13440 + }, + { + "epoch": 1.4760597408302218, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.046409845352173, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6989130973815918, + "num_tokens": 347779219.0, + "step": 13441 + }, + { + "epoch": 1.4761695585328356, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.485347032546997, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7206168174743652, + "num_tokens": 347800401.0, + "step": 13442 + }, + { + "epoch": 1.4762793762354491, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2927565574645996, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7359510064125061, + "num_tokens": 347824024.0, + "step": 13443 + }, + { + "epoch": 1.476389193938063, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.228572368621826, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7339723110198975, + "num_tokens": 347850139.0, + "step": 13444 + }, + { + "epoch": 1.4764990116406764, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.153759717941284, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7185037136077881, + "num_tokens": 347875516.0, + "step": 13445 + }, + { + "epoch": 1.4766088293432902, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9898879528045654, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.6986622214317322, + "num_tokens": 347906133.0, + "step": 13446 + }, + { + "epoch": 1.4767186470459037, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0558619499206543, + "learning_rate": 1e-06, + "loss": 1.0011, + "mean_token_accuracy": 0.6952092051506042, + "num_tokens": 347938177.0, + "step": 13447 + }, + { + "epoch": 1.4768284647485175, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.197443962097168, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7065591812133789, + "num_tokens": 347964338.0, + "step": 13448 + }, + { + "epoch": 1.476938282451131, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6237130165100098, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7233736515045166, + "num_tokens": 347983881.0, + "step": 13449 + }, + { + "epoch": 1.4770481001537448, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0913960933685303, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7124413847923279, + "num_tokens": 348011842.0, + "step": 13450 + }, + { + "epoch": 1.4771579178563585, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3273510932922363, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7206231355667114, + "num_tokens": 348037002.0, + "step": 13451 + }, + { + "epoch": 1.477267735558972, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2527341842651367, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7208333015441895, + "num_tokens": 348063768.0, + "step": 13452 + }, + { + "epoch": 1.4773775532615858, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0488131046295166, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.742611825466156, + "num_tokens": 348092857.0, + "step": 13453 + }, + { + "epoch": 1.4774873709641994, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.313582181930542, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7174216508865356, + "num_tokens": 348116796.0, + "step": 13454 + }, + { + "epoch": 1.477597188666813, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2224156856536865, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7290758490562439, + "num_tokens": 348142158.0, + "step": 13455 + }, + { + "epoch": 1.4777070063694269, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1246328353881836, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7107326984405518, + "num_tokens": 348170191.0, + "step": 13456 + }, + { + "epoch": 1.4778168240720404, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1847963333129883, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7361880540847778, + "num_tokens": 348194238.0, + "step": 13457 + }, + { + "epoch": 1.477926641774654, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9245314598083496, + "learning_rate": 1e-06, + "loss": 0.828, + "mean_token_accuracy": 0.7426337003707886, + "num_tokens": 348226443.0, + "step": 13458 + }, + { + "epoch": 1.4780364594772677, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.20478892326355, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7175199389457703, + "num_tokens": 348253733.0, + "step": 13459 + }, + { + "epoch": 1.4781462771798815, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.295381546020508, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7170658111572266, + "num_tokens": 348278012.0, + "step": 13460 + }, + { + "epoch": 1.478256094882495, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5549428462982178, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7282004356384277, + "num_tokens": 348298231.0, + "step": 13461 + }, + { + "epoch": 1.4783659125851087, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9284169673919678, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7263239026069641, + "num_tokens": 348330234.0, + "step": 13462 + }, + { + "epoch": 1.4784757302877223, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.093262195587158, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7061880826950073, + "num_tokens": 348359195.0, + "step": 13463 + }, + { + "epoch": 1.478585547990336, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1183454990386963, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7189962267875671, + "num_tokens": 348385778.0, + "step": 13464 + }, + { + "epoch": 1.4786953656929498, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.401005744934082, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6901823282241821, + "num_tokens": 348409555.0, + "step": 13465 + }, + { + "epoch": 1.4788051833955633, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1686692237854004, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7256067991256714, + "num_tokens": 348436878.0, + "step": 13466 + }, + { + "epoch": 1.478915001098177, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2119975090026855, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7247713804244995, + "num_tokens": 348460782.0, + "step": 13467 + }, + { + "epoch": 1.4790248188007906, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9060484170913696, + "learning_rate": 1e-06, + "loss": 1.0057, + "mean_token_accuracy": 0.6884187459945679, + "num_tokens": 348493364.0, + "step": 13468 + }, + { + "epoch": 1.4791346365034044, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9971030950546265, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7271822094917297, + "num_tokens": 348522052.0, + "step": 13469 + }, + { + "epoch": 1.4792444542060181, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.154181718826294, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7126060128211975, + "num_tokens": 348550787.0, + "step": 13470 + }, + { + "epoch": 1.4793542719086317, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.534414291381836, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7473764419555664, + "num_tokens": 348569893.0, + "step": 13471 + }, + { + "epoch": 1.4794640896112452, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5767018795013428, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7401065230369568, + "num_tokens": 348589231.0, + "step": 13472 + }, + { + "epoch": 1.479573907313859, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3927063941955566, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.74434494972229, + "num_tokens": 348612204.0, + "step": 13473 + }, + { + "epoch": 1.4796837250164727, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3420052528381348, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7076040506362915, + "num_tokens": 348635785.0, + "step": 13474 + }, + { + "epoch": 1.4797935427190863, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1049201488494873, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.6998250484466553, + "num_tokens": 348667896.0, + "step": 13475 + }, + { + "epoch": 1.4799033604217, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.321725606918335, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7266990542411804, + "num_tokens": 348693407.0, + "step": 13476 + }, + { + "epoch": 1.4800131781243135, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4907875061035156, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7381006479263306, + "num_tokens": 348716369.0, + "step": 13477 + }, + { + "epoch": 1.4801229958269273, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.120042085647583, + "learning_rate": 1e-06, + "loss": 0.967, + "mean_token_accuracy": 0.7068273425102234, + "num_tokens": 348746705.0, + "step": 13478 + }, + { + "epoch": 1.480232813529541, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.535686492919922, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7502649426460266, + "num_tokens": 348766027.0, + "step": 13479 + }, + { + "epoch": 1.4803426312321546, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0658912658691406, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7213901281356812, + "num_tokens": 348793646.0, + "step": 13480 + }, + { + "epoch": 1.4804524489347684, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1985957622528076, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7237060070037842, + "num_tokens": 348820525.0, + "step": 13481 + }, + { + "epoch": 1.4805622666373819, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.21850323677063, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7116779685020447, + "num_tokens": 348847130.0, + "step": 13482 + }, + { + "epoch": 1.4806720843399956, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.388176441192627, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.727240800857544, + "num_tokens": 348870238.0, + "step": 13483 + }, + { + "epoch": 1.4807819020426094, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2117321491241455, + "learning_rate": 1e-06, + "loss": 0.985, + "mean_token_accuracy": 0.6985815763473511, + "num_tokens": 348898170.0, + "step": 13484 + }, + { + "epoch": 1.480891719745223, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2001049518585205, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.7004927396774292, + "num_tokens": 348928003.0, + "step": 13485 + }, + { + "epoch": 1.4810015374478365, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3554022312164307, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7230402231216431, + "num_tokens": 348953028.0, + "step": 13486 + }, + { + "epoch": 1.4811113551504502, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.145653009414673, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7388565540313721, + "num_tokens": 348981766.0, + "step": 13487 + }, + { + "epoch": 1.481221172853064, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.098933219909668, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7114432454109192, + "num_tokens": 349010130.0, + "step": 13488 + }, + { + "epoch": 1.4813309905556775, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2780168056488037, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6944878101348877, + "num_tokens": 349037423.0, + "step": 13489 + }, + { + "epoch": 1.4814408082582913, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.273599147796631, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7024203538894653, + "num_tokens": 349063582.0, + "step": 13490 + }, + { + "epoch": 1.4815506259609048, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9247022867202759, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7058144211769104, + "num_tokens": 349097443.0, + "step": 13491 + }, + { + "epoch": 1.4816604436635186, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1764118671417236, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7035966515541077, + "num_tokens": 349124836.0, + "step": 13492 + }, + { + "epoch": 1.4817702613661323, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4409406185150146, + "learning_rate": 1e-06, + "loss": 0.9878, + "mean_token_accuracy": 0.6989817023277283, + "num_tokens": 349150018.0, + "step": 13493 + }, + { + "epoch": 1.4818800790687459, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1058967113494873, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7174752354621887, + "num_tokens": 349178029.0, + "step": 13494 + }, + { + "epoch": 1.4819898967713596, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.428746223449707, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7068643569946289, + "num_tokens": 349201132.0, + "step": 13495 + }, + { + "epoch": 1.4820997144739732, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4490365982055664, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7314603924751282, + "num_tokens": 349222121.0, + "step": 13496 + }, + { + "epoch": 1.482209532176587, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.154356002807617, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7377279996871948, + "num_tokens": 349250114.0, + "step": 13497 + }, + { + "epoch": 1.4823193498792007, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.038484811782837, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7079793810844421, + "num_tokens": 349278670.0, + "step": 13498 + }, + { + "epoch": 1.4824291675818142, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.269404649734497, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7084038257598877, + "num_tokens": 349304750.0, + "step": 13499 + }, + { + "epoch": 1.4825389852844277, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1879749298095703, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7341818809509277, + "num_tokens": 349332424.0, + "step": 13500 + }, + { + "epoch": 1.4826488029870415, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2044074535369873, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7065759897232056, + "num_tokens": 349359251.0, + "step": 13501 + }, + { + "epoch": 1.4827586206896552, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.078577756881714, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.7014796733856201, + "num_tokens": 349390802.0, + "step": 13502 + }, + { + "epoch": 1.4828684383922688, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4185802936553955, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7371413111686707, + "num_tokens": 349411462.0, + "step": 13503 + }, + { + "epoch": 1.4829782560948825, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0993220806121826, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7512726783752441, + "num_tokens": 349437804.0, + "step": 13504 + }, + { + "epoch": 1.483088073797496, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4057552814483643, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7278881669044495, + "num_tokens": 349461433.0, + "step": 13505 + }, + { + "epoch": 1.4831978915001098, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.09828782081604, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6950995922088623, + "num_tokens": 349492785.0, + "step": 13506 + }, + { + "epoch": 1.4833077092027236, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.507902145385742, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.720687985420227, + "num_tokens": 349514255.0, + "step": 13507 + }, + { + "epoch": 1.4834175269053371, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.232501983642578, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.72160404920578, + "num_tokens": 349540250.0, + "step": 13508 + }, + { + "epoch": 1.4835273446079509, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.9328322410583496, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.734939694404602, + "num_tokens": 349555718.0, + "step": 13509 + }, + { + "epoch": 1.4836371623105644, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.260878324508667, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7245269417762756, + "num_tokens": 349581064.0, + "step": 13510 + }, + { + "epoch": 1.4837469800131782, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.147343158721924, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7037882804870605, + "num_tokens": 349610750.0, + "step": 13511 + }, + { + "epoch": 1.4838567977157917, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.475045919418335, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7235815525054932, + "num_tokens": 349631750.0, + "step": 13512 + }, + { + "epoch": 1.4839666154184055, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2098917961120605, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7047758102416992, + "num_tokens": 349660878.0, + "step": 13513 + }, + { + "epoch": 1.484076433121019, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.320786237716675, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7322249412536621, + "num_tokens": 349684386.0, + "step": 13514 + }, + { + "epoch": 1.4841862508236328, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0120315551757812, + "learning_rate": 1e-06, + "loss": 1.0031, + "mean_token_accuracy": 0.6928355097770691, + "num_tokens": 349713907.0, + "step": 13515 + }, + { + "epoch": 1.4842960685262465, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9977591037750244, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7164069414138794, + "num_tokens": 349744778.0, + "step": 13516 + }, + { + "epoch": 1.48440588622886, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3944952487945557, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7535091638565063, + "num_tokens": 349767661.0, + "step": 13517 + }, + { + "epoch": 1.4845157039314738, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2618045806884766, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7060382962226868, + "num_tokens": 349794478.0, + "step": 13518 + }, + { + "epoch": 1.4846255216340873, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3419620990753174, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7088415622711182, + "num_tokens": 349819844.0, + "step": 13519 + }, + { + "epoch": 1.484735339336701, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.18684458732605, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7229533791542053, + "num_tokens": 349847165.0, + "step": 13520 + }, + { + "epoch": 1.4848451570393149, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1726624965667725, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7023576498031616, + "num_tokens": 349874515.0, + "step": 13521 + }, + { + "epoch": 1.4849549747419284, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.241352081298828, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7229398488998413, + "num_tokens": 349899695.0, + "step": 13522 + }, + { + "epoch": 1.485064792444542, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.965547800064087, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7059421539306641, + "num_tokens": 349930611.0, + "step": 13523 + }, + { + "epoch": 1.4851746101471557, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.162262201309204, + "learning_rate": 1e-06, + "loss": 0.962, + "mean_token_accuracy": 0.7047992944717407, + "num_tokens": 349958000.0, + "step": 13524 + }, + { + "epoch": 1.4852844278497694, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1493520736694336, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7148083448410034, + "num_tokens": 349984827.0, + "step": 13525 + }, + { + "epoch": 1.485394245552383, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3539438247680664, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7202811241149902, + "num_tokens": 350008778.0, + "step": 13526 + }, + { + "epoch": 1.4855040632549967, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.238569736480713, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7187721729278564, + "num_tokens": 350034827.0, + "step": 13527 + }, + { + "epoch": 1.4856138809576103, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.107879638671875, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7313485145568848, + "num_tokens": 350062547.0, + "step": 13528 + }, + { + "epoch": 1.485723698660224, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.241199493408203, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7226572036743164, + "num_tokens": 350089143.0, + "step": 13529 + }, + { + "epoch": 1.4858335163628378, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2699153423309326, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7065325975418091, + "num_tokens": 350113252.0, + "step": 13530 + }, + { + "epoch": 1.4859433340654513, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9594544172286987, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.715793251991272, + "num_tokens": 350145085.0, + "step": 13531 + }, + { + "epoch": 1.486053151768065, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.489453077316284, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7296491861343384, + "num_tokens": 350167183.0, + "step": 13532 + }, + { + "epoch": 1.4861629694706786, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3359546661376953, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7276252508163452, + "num_tokens": 350189984.0, + "step": 13533 + }, + { + "epoch": 1.4862727871732924, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.229635238647461, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.713948130607605, + "num_tokens": 350216411.0, + "step": 13534 + }, + { + "epoch": 1.4863826048759061, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9600343704223633, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7242927551269531, + "num_tokens": 350247603.0, + "step": 13535 + }, + { + "epoch": 1.4864924225785197, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.365828037261963, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7311328053474426, + "num_tokens": 350273919.0, + "step": 13536 + }, + { + "epoch": 1.4866022402811332, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.176715612411499, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.718163251876831, + "num_tokens": 350300025.0, + "step": 13537 + }, + { + "epoch": 1.486712057983747, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.542019844055176, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7370738983154297, + "num_tokens": 350323919.0, + "step": 13538 + }, + { + "epoch": 1.4868218756863607, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3927812576293945, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7197522521018982, + "num_tokens": 350349168.0, + "step": 13539 + }, + { + "epoch": 1.4869316933889742, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5751023292541504, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7277601957321167, + "num_tokens": 350370466.0, + "step": 13540 + }, + { + "epoch": 1.487041511091588, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.400040626525879, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7311096787452698, + "num_tokens": 350392292.0, + "step": 13541 + }, + { + "epoch": 1.4871513287942015, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.274779796600342, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.6932750344276428, + "num_tokens": 350418165.0, + "step": 13542 + }, + { + "epoch": 1.4872611464968153, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.476543426513672, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7205883264541626, + "num_tokens": 350438677.0, + "step": 13543 + }, + { + "epoch": 1.487370964199429, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4104301929473877, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7188609838485718, + "num_tokens": 350463147.0, + "step": 13544 + }, + { + "epoch": 1.4874807819020426, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9999088048934937, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7033977508544922, + "num_tokens": 350493784.0, + "step": 13545 + }, + { + "epoch": 1.4875905996046563, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5099456310272217, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7157921195030212, + "num_tokens": 350515693.0, + "step": 13546 + }, + { + "epoch": 1.4877004173072699, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.999687671661377, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7227375507354736, + "num_tokens": 350549384.0, + "step": 13547 + }, + { + "epoch": 1.4878102350098836, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0954947471618652, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.6971270442008972, + "num_tokens": 350580496.0, + "step": 13548 + }, + { + "epoch": 1.4879200527124974, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4930200576782227, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7215962409973145, + "num_tokens": 350602962.0, + "step": 13549 + }, + { + "epoch": 1.488029870415111, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.220390558242798, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7157514095306396, + "num_tokens": 350628895.0, + "step": 13550 + }, + { + "epoch": 1.4881396881177245, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6702964305877686, + "learning_rate": 1e-06, + "loss": 0.7746, + "mean_token_accuracy": 0.7509374618530273, + "num_tokens": 350646645.0, + "step": 13551 + }, + { + "epoch": 1.4882495058203382, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3635623455047607, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7092527151107788, + "num_tokens": 350671701.0, + "step": 13552 + }, + { + "epoch": 1.488359323522952, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.918338418006897, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7108397483825684, + "num_tokens": 350703873.0, + "step": 13553 + }, + { + "epoch": 1.4884691412255655, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.059568405151367, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7129961252212524, + "num_tokens": 350733475.0, + "step": 13554 + }, + { + "epoch": 1.4885789589281793, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1884305477142334, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7455195188522339, + "num_tokens": 350758833.0, + "step": 13555 + }, + { + "epoch": 1.4886887766307928, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4820749759674072, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7196270227432251, + "num_tokens": 350780400.0, + "step": 13556 + }, + { + "epoch": 1.4887985943334066, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0308380126953125, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7357439994812012, + "num_tokens": 350808141.0, + "step": 13557 + }, + { + "epoch": 1.4889084120360203, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2614684104919434, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7207240462303162, + "num_tokens": 350835002.0, + "step": 13558 + }, + { + "epoch": 1.4890182297386338, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.393995523452759, + "learning_rate": 1e-06, + "loss": 0.7667, + "mean_token_accuracy": 0.7604525089263916, + "num_tokens": 350856990.0, + "step": 13559 + }, + { + "epoch": 1.4891280474412476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.121453285217285, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.709854006767273, + "num_tokens": 350886291.0, + "step": 13560 + }, + { + "epoch": 1.4892378651438611, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0710086822509766, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7180749773979187, + "num_tokens": 350914005.0, + "step": 13561 + }, + { + "epoch": 1.489347682846475, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2788052558898926, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6930617094039917, + "num_tokens": 350941507.0, + "step": 13562 + }, + { + "epoch": 1.4894575005490884, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.66719651222229, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7370555996894836, + "num_tokens": 350960208.0, + "step": 13563 + }, + { + "epoch": 1.4895673182517022, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1627163887023926, + "learning_rate": 1e-06, + "loss": 0.9952, + "mean_token_accuracy": 0.6913368701934814, + "num_tokens": 350988155.0, + "step": 13564 + }, + { + "epoch": 1.4896771359543157, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2293264865875244, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.727706253528595, + "num_tokens": 351013036.0, + "step": 13565 + }, + { + "epoch": 1.4897869536569295, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3566737174987793, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7279558181762695, + "num_tokens": 351036133.0, + "step": 13566 + }, + { + "epoch": 1.4898967713595432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9615423679351807, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7222737669944763, + "num_tokens": 351069401.0, + "step": 13567 + }, + { + "epoch": 1.4900065890621568, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0209972858428955, + "learning_rate": 1e-06, + "loss": 1.0242, + "mean_token_accuracy": 0.6968474388122559, + "num_tokens": 351102336.0, + "step": 13568 + }, + { + "epoch": 1.4901164067647705, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7473413944244385, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7344506978988647, + "num_tokens": 351120748.0, + "step": 13569 + }, + { + "epoch": 1.490226224467384, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5883114337921143, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7203295230865479, + "num_tokens": 351140821.0, + "step": 13570 + }, + { + "epoch": 1.4903360421699978, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9587316513061523, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.6974382400512695, + "num_tokens": 351170688.0, + "step": 13571 + }, + { + "epoch": 1.4904458598726116, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2396442890167236, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7130125761032104, + "num_tokens": 351195904.0, + "step": 13572 + }, + { + "epoch": 1.4905556775752251, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.427140235900879, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7205758094787598, + "num_tokens": 351218093.0, + "step": 13573 + }, + { + "epoch": 1.4906654952778389, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.197726249694824, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.71124267578125, + "num_tokens": 351243442.0, + "step": 13574 + }, + { + "epoch": 1.4907753129804524, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.597119092941284, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6990737915039062, + "num_tokens": 351264376.0, + "step": 13575 + }, + { + "epoch": 1.4908851306830662, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6582746505737305, + "learning_rate": 1e-06, + "loss": 0.7749, + "mean_token_accuracy": 0.7522123456001282, + "num_tokens": 351282166.0, + "step": 13576 + }, + { + "epoch": 1.4909949483856797, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1517562866210938, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7040688991546631, + "num_tokens": 351310966.0, + "step": 13577 + }, + { + "epoch": 1.4911047660882935, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.203524112701416, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7300673127174377, + "num_tokens": 351338849.0, + "step": 13578 + }, + { + "epoch": 1.491214583790907, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.30920147895813, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7201402187347412, + "num_tokens": 351364478.0, + "step": 13579 + }, + { + "epoch": 1.4913244014935207, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.313474416732788, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7171563506126404, + "num_tokens": 351389319.0, + "step": 13580 + }, + { + "epoch": 1.4914342191961345, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2007763385772705, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7170453071594238, + "num_tokens": 351413848.0, + "step": 13581 + }, + { + "epoch": 1.491544036898748, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0372328758239746, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.6994351148605347, + "num_tokens": 351443166.0, + "step": 13582 + }, + { + "epoch": 1.4916538546013618, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.715811014175415, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7219615578651428, + "num_tokens": 351462274.0, + "step": 13583 + }, + { + "epoch": 1.4917636723039753, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4295880794525146, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7254260778427124, + "num_tokens": 351487010.0, + "step": 13584 + }, + { + "epoch": 1.491873490006589, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1442980766296387, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7424229383468628, + "num_tokens": 351514148.0, + "step": 13585 + }, + { + "epoch": 1.4919833077092028, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.984924554824829, + "learning_rate": 1e-06, + "loss": 1.019, + "mean_token_accuracy": 0.6898484230041504, + "num_tokens": 351548859.0, + "step": 13586 + }, + { + "epoch": 1.4920931254118164, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5308094024658203, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7486157417297363, + "num_tokens": 351568450.0, + "step": 13587 + }, + { + "epoch": 1.49220294311443, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4193873405456543, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7210421562194824, + "num_tokens": 351590179.0, + "step": 13588 + }, + { + "epoch": 1.4923127608170437, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.041515827178955, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7348855137825012, + "num_tokens": 351617740.0, + "step": 13589 + }, + { + "epoch": 1.4924225785196574, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1223902702331543, + "learning_rate": 1e-06, + "loss": 1.0477, + "mean_token_accuracy": 0.688512921333313, + "num_tokens": 351649879.0, + "step": 13590 + }, + { + "epoch": 1.492532396222271, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.531757354736328, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7135918140411377, + "num_tokens": 351671227.0, + "step": 13591 + }, + { + "epoch": 1.4926422139248847, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.47745418548584, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7175296545028687, + "num_tokens": 351693296.0, + "step": 13592 + }, + { + "epoch": 1.4927520316274983, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.629567861557007, + "learning_rate": 1e-06, + "loss": 0.67, + "mean_token_accuracy": 0.7756332159042358, + "num_tokens": 351711924.0, + "step": 13593 + }, + { + "epoch": 1.492861849330112, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2672464847564697, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7316736578941345, + "num_tokens": 351737262.0, + "step": 13594 + }, + { + "epoch": 1.4929716670327258, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2914888858795166, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7311705350875854, + "num_tokens": 351760464.0, + "step": 13595 + }, + { + "epoch": 1.4930814847353393, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2981603145599365, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.6976520419120789, + "num_tokens": 351784671.0, + "step": 13596 + }, + { + "epoch": 1.493191302437953, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2287001609802246, + "learning_rate": 1e-06, + "loss": 0.9516, + "mean_token_accuracy": 0.705715000629425, + "num_tokens": 351813130.0, + "step": 13597 + }, + { + "epoch": 1.4933011201405666, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5196597576141357, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.7107666730880737, + "num_tokens": 351836354.0, + "step": 13598 + }, + { + "epoch": 1.4934109378431804, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.164652109146118, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7336942553520203, + "num_tokens": 351864756.0, + "step": 13599 + }, + { + "epoch": 1.493520755545794, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.8925502300262451, + "learning_rate": 1e-06, + "loss": 0.9705, + "mean_token_accuracy": 0.7010983824729919, + "num_tokens": 351900574.0, + "step": 13600 + }, + { + "epoch": 1.4936305732484076, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.383882522583008, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7330143451690674, + "num_tokens": 351922710.0, + "step": 13601 + }, + { + "epoch": 1.4937403909510212, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3232028484344482, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7246110439300537, + "num_tokens": 351949225.0, + "step": 13602 + }, + { + "epoch": 1.493850208653635, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2396481037139893, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7135242819786072, + "num_tokens": 351974234.0, + "step": 13603 + }, + { + "epoch": 1.4939600263562487, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.368262767791748, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7125183343887329, + "num_tokens": 351996602.0, + "step": 13604 + }, + { + "epoch": 1.4940698440588622, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3464040756225586, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7114593982696533, + "num_tokens": 352020711.0, + "step": 13605 + }, + { + "epoch": 1.494179661761476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.078098773956299, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.6937934756278992, + "num_tokens": 352052927.0, + "step": 13606 + }, + { + "epoch": 1.4942894794640895, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3584344387054443, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7178981304168701, + "num_tokens": 352077195.0, + "step": 13607 + }, + { + "epoch": 1.4943992971667033, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0131850242614746, + "learning_rate": 1e-06, + "loss": 0.9754, + "mean_token_accuracy": 0.7064816355705261, + "num_tokens": 352106228.0, + "step": 13608 + }, + { + "epoch": 1.494509114869317, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0681726932525635, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7148685455322266, + "num_tokens": 352133965.0, + "step": 13609 + }, + { + "epoch": 1.4946189325719306, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.376607894897461, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.710426390171051, + "num_tokens": 352158760.0, + "step": 13610 + }, + { + "epoch": 1.4947287502745443, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.8401668071746826, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7277977466583252, + "num_tokens": 352177506.0, + "step": 13611 + }, + { + "epoch": 1.4948385679771579, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6127357482910156, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7231496572494507, + "num_tokens": 352198833.0, + "step": 13612 + }, + { + "epoch": 1.4949483856797716, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3120782375335693, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7398815155029297, + "num_tokens": 352220545.0, + "step": 13613 + }, + { + "epoch": 1.4950582033823854, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0479929447174072, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7102431058883667, + "num_tokens": 352251267.0, + "step": 13614 + }, + { + "epoch": 1.495168021084999, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.276137351989746, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7014257907867432, + "num_tokens": 352277812.0, + "step": 13615 + }, + { + "epoch": 1.4952778387876124, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3851680755615234, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7119677662849426, + "num_tokens": 352302730.0, + "step": 13616 + }, + { + "epoch": 1.4953876564902262, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.554710865020752, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7086367011070251, + "num_tokens": 352323936.0, + "step": 13617 + }, + { + "epoch": 1.49549747419284, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1718387603759766, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7109146118164062, + "num_tokens": 352353122.0, + "step": 13618 + }, + { + "epoch": 1.4956072918954535, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3528811931610107, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7147685885429382, + "num_tokens": 352379397.0, + "step": 13619 + }, + { + "epoch": 1.4957171095980673, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2440993785858154, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7284325361251831, + "num_tokens": 352405394.0, + "step": 13620 + }, + { + "epoch": 1.4958269273006808, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3626708984375, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7354418039321899, + "num_tokens": 352427832.0, + "step": 13621 + }, + { + "epoch": 1.4959367450032945, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.315117835998535, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7333778738975525, + "num_tokens": 352451288.0, + "step": 13622 + }, + { + "epoch": 1.4960465627059083, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2219088077545166, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7206002473831177, + "num_tokens": 352480422.0, + "step": 13623 + }, + { + "epoch": 1.4961563804085218, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4970314502716064, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7175650000572205, + "num_tokens": 352502417.0, + "step": 13624 + }, + { + "epoch": 1.4962661981111356, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0232434272766113, + "learning_rate": 1e-06, + "loss": 1.002, + "mean_token_accuracy": 0.6916394829750061, + "num_tokens": 352534885.0, + "step": 13625 + }, + { + "epoch": 1.4963760158137491, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.018685817718506, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7280335426330566, + "num_tokens": 352564471.0, + "step": 13626 + }, + { + "epoch": 1.4964858335163629, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3910882472991943, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.6970796585083008, + "num_tokens": 352586736.0, + "step": 13627 + }, + { + "epoch": 1.4965956512189764, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0646586418151855, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7239418029785156, + "num_tokens": 352615469.0, + "step": 13628 + }, + { + "epoch": 1.4967054689215902, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.104301929473877, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7210265398025513, + "num_tokens": 352643181.0, + "step": 13629 + }, + { + "epoch": 1.4968152866242037, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.122922897338867, + "learning_rate": 1e-06, + "loss": 1.01, + "mean_token_accuracy": 0.6872944831848145, + "num_tokens": 352673504.0, + "step": 13630 + }, + { + "epoch": 1.4969251043268175, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0613555908203125, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7274754047393799, + "num_tokens": 352701778.0, + "step": 13631 + }, + { + "epoch": 1.4970349220294312, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2696540355682373, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7179207801818848, + "num_tokens": 352728066.0, + "step": 13632 + }, + { + "epoch": 1.4971447397320448, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2711422443389893, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7114973068237305, + "num_tokens": 352752048.0, + "step": 13633 + }, + { + "epoch": 1.4972545574346585, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9120670557022095, + "learning_rate": 1e-06, + "loss": 0.9753, + "mean_token_accuracy": 0.6959515810012817, + "num_tokens": 352786096.0, + "step": 13634 + }, + { + "epoch": 1.497364375137272, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.183652639389038, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7375794649124146, + "num_tokens": 352812176.0, + "step": 13635 + }, + { + "epoch": 1.4974741928398858, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3086135387420654, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7202838063240051, + "num_tokens": 352835227.0, + "step": 13636 + }, + { + "epoch": 1.4975840105424996, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2064590454101562, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7223688960075378, + "num_tokens": 352861196.0, + "step": 13637 + }, + { + "epoch": 1.497693828245113, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4251654148101807, + "learning_rate": 1e-06, + "loss": 0.7252, + "mean_token_accuracy": 0.766343355178833, + "num_tokens": 352882435.0, + "step": 13638 + }, + { + "epoch": 1.4978036459477266, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3030879497528076, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7004268169403076, + "num_tokens": 352907753.0, + "step": 13639 + }, + { + "epoch": 1.4979134636503404, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.476545810699463, + "learning_rate": 1e-06, + "loss": 0.7726, + "mean_token_accuracy": 0.7491898536682129, + "num_tokens": 352929689.0, + "step": 13640 + }, + { + "epoch": 1.4980232813529542, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.419496536254883, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7266087532043457, + "num_tokens": 352950689.0, + "step": 13641 + }, + { + "epoch": 1.4981330990555677, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.122187376022339, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7212533950805664, + "num_tokens": 352978013.0, + "step": 13642 + }, + { + "epoch": 1.4982429167581814, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1249618530273438, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.720954418182373, + "num_tokens": 353007336.0, + "step": 13643 + }, + { + "epoch": 1.498352734460795, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2639153003692627, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7279939651489258, + "num_tokens": 353032226.0, + "step": 13644 + }, + { + "epoch": 1.4984625521634087, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.406785488128662, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7364134788513184, + "num_tokens": 353054202.0, + "step": 13645 + }, + { + "epoch": 1.4985723698660225, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4132797718048096, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7256438732147217, + "num_tokens": 353077624.0, + "step": 13646 + }, + { + "epoch": 1.498682187568636, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1430411338806152, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7228087782859802, + "num_tokens": 353105436.0, + "step": 13647 + }, + { + "epoch": 1.4987920052712498, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.144211530685425, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7053150534629822, + "num_tokens": 353133795.0, + "step": 13648 + }, + { + "epoch": 1.4989018229738633, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5391757488250732, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7422813177108765, + "num_tokens": 353153980.0, + "step": 13649 + }, + { + "epoch": 1.499011640676477, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.230849266052246, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7262331247329712, + "num_tokens": 353180072.0, + "step": 13650 + }, + { + "epoch": 1.4991214583790908, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2332589626312256, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7059903144836426, + "num_tokens": 353209854.0, + "step": 13651 + }, + { + "epoch": 1.4992312760817044, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.8381640911102295, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7289305925369263, + "num_tokens": 353227772.0, + "step": 13652 + }, + { + "epoch": 1.499341093784318, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.474400281906128, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.7504176497459412, + "num_tokens": 353248130.0, + "step": 13653 + }, + { + "epoch": 1.4994509114869317, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7051377296447754, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7170780301094055, + "num_tokens": 353268005.0, + "step": 13654 + }, + { + "epoch": 1.4995607291895454, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2873191833496094, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7364253997802734, + "num_tokens": 353293531.0, + "step": 13655 + }, + { + "epoch": 1.499670546892159, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0964722633361816, + "learning_rate": 1e-06, + "loss": 0.964, + "mean_token_accuracy": 0.6993409395217896, + "num_tokens": 353323734.0, + "step": 13656 + }, + { + "epoch": 1.4997803645947727, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.464414358139038, + "learning_rate": 1e-06, + "loss": 0.9979, + "mean_token_accuracy": 0.7074954509735107, + "num_tokens": 353345463.0, + "step": 13657 + }, + { + "epoch": 1.4998901822973862, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3439841270446777, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7156873345375061, + "num_tokens": 353369808.0, + "step": 13658 + }, + { + "epoch": 1.5, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.283287525177002, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7476279735565186, + "num_tokens": 353393328.0, + "step": 13659 + }, + { + "epoch": 1.5001098177026138, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.438628911972046, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7235246896743774, + "num_tokens": 353417908.0, + "step": 13660 + }, + { + "epoch": 1.5002196354052273, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.265186071395874, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7191705703735352, + "num_tokens": 353443856.0, + "step": 13661 + }, + { + "epoch": 1.5003294531078408, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1828017234802246, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.722848653793335, + "num_tokens": 353469890.0, + "step": 13662 + }, + { + "epoch": 1.5004392708104546, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2330567836761475, + "learning_rate": 1e-06, + "loss": 0.9731, + "mean_token_accuracy": 0.703944742679596, + "num_tokens": 353496955.0, + "step": 13663 + }, + { + "epoch": 1.5005490885130683, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.045914888381958, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7081988453865051, + "num_tokens": 353527810.0, + "step": 13664 + }, + { + "epoch": 1.500658906215682, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.18437123298645, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7072492241859436, + "num_tokens": 353554731.0, + "step": 13665 + }, + { + "epoch": 1.5007687239182956, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.009880304336548, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7052562236785889, + "num_tokens": 353586332.0, + "step": 13666 + }, + { + "epoch": 1.5008785416209092, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1572718620300293, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7009108066558838, + "num_tokens": 353612790.0, + "step": 13667 + }, + { + "epoch": 1.500988359323523, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.09399151802063, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.717208743095398, + "num_tokens": 353640385.0, + "step": 13668 + }, + { + "epoch": 1.5010981770261367, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.314540386199951, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7029078006744385, + "num_tokens": 353665468.0, + "step": 13669 + }, + { + "epoch": 1.5012079947287504, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.230232000350952, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7342719435691833, + "num_tokens": 353690606.0, + "step": 13670 + }, + { + "epoch": 1.501317812431364, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3779983520507812, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7263497114181519, + "num_tokens": 353713806.0, + "step": 13671 + }, + { + "epoch": 1.5014276301339775, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2744300365448, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7272707223892212, + "num_tokens": 353738924.0, + "step": 13672 + }, + { + "epoch": 1.5015374478365913, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.243816375732422, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7154183387756348, + "num_tokens": 353763454.0, + "step": 13673 + }, + { + "epoch": 1.501647265539205, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.23372483253479, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7401409149169922, + "num_tokens": 353787635.0, + "step": 13674 + }, + { + "epoch": 1.5017570832418186, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.159278392791748, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7173044681549072, + "num_tokens": 353815796.0, + "step": 13675 + }, + { + "epoch": 1.501866900944432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2464864253997803, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7456983327865601, + "num_tokens": 353840383.0, + "step": 13676 + }, + { + "epoch": 1.5019767186470458, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.20251202583313, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7195518016815186, + "num_tokens": 353866233.0, + "step": 13677 + }, + { + "epoch": 1.5020865363496596, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0713019371032715, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7129614949226379, + "num_tokens": 353894026.0, + "step": 13678 + }, + { + "epoch": 1.5021963540522734, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0545666217803955, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6948821544647217, + "num_tokens": 353923991.0, + "step": 13679 + }, + { + "epoch": 1.502306171754887, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.074643135070801, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7294907569885254, + "num_tokens": 353954798.0, + "step": 13680 + }, + { + "epoch": 1.5024159894575004, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6757283210754395, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7259305715560913, + "num_tokens": 353975069.0, + "step": 13681 + }, + { + "epoch": 1.5025258071601142, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6083579063415527, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.744552731513977, + "num_tokens": 353994106.0, + "step": 13682 + }, + { + "epoch": 1.502635624862728, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3509740829467773, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.700032651424408, + "num_tokens": 354018331.0, + "step": 13683 + }, + { + "epoch": 1.5027454425653415, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.473578453063965, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7369714975357056, + "num_tokens": 354040243.0, + "step": 13684 + }, + { + "epoch": 1.5028552602679552, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.322174310684204, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7258929014205933, + "num_tokens": 354064205.0, + "step": 13685 + }, + { + "epoch": 1.5029650779705688, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3578147888183594, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7129167318344116, + "num_tokens": 354089638.0, + "step": 13686 + }, + { + "epoch": 1.5030748956731825, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.502530097961426, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.731849193572998, + "num_tokens": 354113131.0, + "step": 13687 + }, + { + "epoch": 1.5031847133757963, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3527543544769287, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7321597933769226, + "num_tokens": 354137117.0, + "step": 13688 + }, + { + "epoch": 1.5032945310784098, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3614654541015625, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7378480434417725, + "num_tokens": 354161436.0, + "step": 13689 + }, + { + "epoch": 1.5034043487810234, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2735607624053955, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7124587297439575, + "num_tokens": 354186678.0, + "step": 13690 + }, + { + "epoch": 1.5035141664836371, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0370776653289795, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7369281649589539, + "num_tokens": 354214855.0, + "step": 13691 + }, + { + "epoch": 1.5036239841862509, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.071106195449829, + "learning_rate": 1e-06, + "loss": 1.0123, + "mean_token_accuracy": 0.6809577345848083, + "num_tokens": 354247092.0, + "step": 13692 + }, + { + "epoch": 1.5037338018888646, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.479609251022339, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7387030720710754, + "num_tokens": 354268746.0, + "step": 13693 + }, + { + "epoch": 1.5038436195914782, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1437342166900635, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7227681279182434, + "num_tokens": 354296795.0, + "step": 13694 + }, + { + "epoch": 1.5039534372940917, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0743298530578613, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7273606061935425, + "num_tokens": 354325884.0, + "step": 13695 + }, + { + "epoch": 1.5040632549967055, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.353531837463379, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7305108904838562, + "num_tokens": 354348878.0, + "step": 13696 + }, + { + "epoch": 1.5041730726993192, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3050804138183594, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7486422061920166, + "num_tokens": 354373386.0, + "step": 13697 + }, + { + "epoch": 1.5042828904019327, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.145658493041992, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7213373184204102, + "num_tokens": 354399582.0, + "step": 13698 + }, + { + "epoch": 1.5043927081045465, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1730294227600098, + "learning_rate": 1e-06, + "loss": 1.0144, + "mean_token_accuracy": 0.6839028000831604, + "num_tokens": 354427759.0, + "step": 13699 + }, + { + "epoch": 1.50450252580716, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4739468097686768, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.7002120614051819, + "num_tokens": 354450286.0, + "step": 13700 + }, + { + "epoch": 1.5046123435097738, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2724485397338867, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7074970006942749, + "num_tokens": 354476107.0, + "step": 13701 + }, + { + "epoch": 1.5047221612123876, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1279613971710205, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6945703029632568, + "num_tokens": 354505040.0, + "step": 13702 + }, + { + "epoch": 1.504831978915001, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.155670404434204, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.715253472328186, + "num_tokens": 354534336.0, + "step": 13703 + }, + { + "epoch": 1.5049417966176146, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3551137447357178, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7190045118331909, + "num_tokens": 354559588.0, + "step": 13704 + }, + { + "epoch": 1.5050516143202284, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.287099838256836, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7155430316925049, + "num_tokens": 354583966.0, + "step": 13705 + }, + { + "epoch": 1.5051614320228421, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.284154176712036, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7216291427612305, + "num_tokens": 354608592.0, + "step": 13706 + }, + { + "epoch": 1.505271249725456, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4174115657806396, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7335917353630066, + "num_tokens": 354631381.0, + "step": 13707 + }, + { + "epoch": 1.5053810674280694, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3175559043884277, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7217593789100647, + "num_tokens": 354655687.0, + "step": 13708 + }, + { + "epoch": 1.505490885130683, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.282365322113037, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7102565765380859, + "num_tokens": 354680390.0, + "step": 13709 + }, + { + "epoch": 1.5056007028332967, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2262136936187744, + "learning_rate": 1e-06, + "loss": 0.9692, + "mean_token_accuracy": 0.700339138507843, + "num_tokens": 354706633.0, + "step": 13710 + }, + { + "epoch": 1.5057105205359105, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.393615961074829, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7215675711631775, + "num_tokens": 354730625.0, + "step": 13711 + }, + { + "epoch": 1.505820338238524, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7424473762512207, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7425416111946106, + "num_tokens": 354747197.0, + "step": 13712 + }, + { + "epoch": 1.5059301559411375, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.322376251220703, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7258836030960083, + "num_tokens": 354771713.0, + "step": 13713 + }, + { + "epoch": 1.5060399736437513, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.22796893119812, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7246280312538147, + "num_tokens": 354799653.0, + "step": 13714 + }, + { + "epoch": 1.506149791346365, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1905946731567383, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7076824903488159, + "num_tokens": 354827490.0, + "step": 13715 + }, + { + "epoch": 1.5062596090489788, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.35115385055542, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.710150957107544, + "num_tokens": 354854128.0, + "step": 13716 + }, + { + "epoch": 1.5063694267515924, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.285022258758545, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7148997783660889, + "num_tokens": 354879134.0, + "step": 13717 + }, + { + "epoch": 1.506479244454206, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5124151706695557, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7133697271347046, + "num_tokens": 354902971.0, + "step": 13718 + }, + { + "epoch": 1.5065890621568196, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3484816551208496, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7299931645393372, + "num_tokens": 354926622.0, + "step": 13719 + }, + { + "epoch": 1.5066988798594334, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.238205909729004, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7213742733001709, + "num_tokens": 354950837.0, + "step": 13720 + }, + { + "epoch": 1.5068086975620472, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2167303562164307, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7008087635040283, + "num_tokens": 354978412.0, + "step": 13721 + }, + { + "epoch": 1.5069185152646607, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3628628253936768, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.726477861404419, + "num_tokens": 355003622.0, + "step": 13722 + }, + { + "epoch": 1.5070283329672742, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.598754644393921, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7203325033187866, + "num_tokens": 355023549.0, + "step": 13723 + }, + { + "epoch": 1.507138150669888, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2633657455444336, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7095173597335815, + "num_tokens": 355048551.0, + "step": 13724 + }, + { + "epoch": 1.5072479683725017, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3020637035369873, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7171437740325928, + "num_tokens": 355075336.0, + "step": 13725 + }, + { + "epoch": 1.5073577860751153, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0812788009643555, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7122517824172974, + "num_tokens": 355105548.0, + "step": 13726 + }, + { + "epoch": 1.5074676037777288, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1337053775787354, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7301618456840515, + "num_tokens": 355133482.0, + "step": 13727 + }, + { + "epoch": 1.5075774214803426, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3018627166748047, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7120331525802612, + "num_tokens": 355158234.0, + "step": 13728 + }, + { + "epoch": 1.5076872391829563, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2160489559173584, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7163052558898926, + "num_tokens": 355184906.0, + "step": 13729 + }, + { + "epoch": 1.50779705688557, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4131698608398438, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7018183469772339, + "num_tokens": 355206627.0, + "step": 13730 + }, + { + "epoch": 1.5079068745881836, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.947365641593933, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7107419967651367, + "num_tokens": 355236100.0, + "step": 13731 + }, + { + "epoch": 1.5080166922907972, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9394848346710205, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6911398768424988, + "num_tokens": 355270428.0, + "step": 13732 + }, + { + "epoch": 1.508126509993411, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9824538230895996, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7122111320495605, + "num_tokens": 355303500.0, + "step": 13733 + }, + { + "epoch": 1.5082363276960247, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.120889902114868, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7095223665237427, + "num_tokens": 355333276.0, + "step": 13734 + }, + { + "epoch": 1.5083461453986382, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.362546682357788, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7352571487426758, + "num_tokens": 355356832.0, + "step": 13735 + }, + { + "epoch": 1.508455963101252, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.441563129425049, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7230976223945618, + "num_tokens": 355379912.0, + "step": 13736 + }, + { + "epoch": 1.5085657808038655, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2308053970336914, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7025994062423706, + "num_tokens": 355408113.0, + "step": 13737 + }, + { + "epoch": 1.5086755985064793, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.34816312789917, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7338215112686157, + "num_tokens": 355432286.0, + "step": 13738 + }, + { + "epoch": 1.508785416209093, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.364929676055908, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7006993293762207, + "num_tokens": 355457847.0, + "step": 13739 + }, + { + "epoch": 1.5088952339117065, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3009443283081055, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7283399105072021, + "num_tokens": 355483148.0, + "step": 13740 + }, + { + "epoch": 1.50900505161432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.141258716583252, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7300364971160889, + "num_tokens": 355510583.0, + "step": 13741 + }, + { + "epoch": 1.5091148693169338, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3219659328460693, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7448203563690186, + "num_tokens": 355535520.0, + "step": 13742 + }, + { + "epoch": 1.5092246870195476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1238389015197754, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7203831076622009, + "num_tokens": 355562645.0, + "step": 13743 + }, + { + "epoch": 1.5093345047221614, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1313974857330322, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.707756519317627, + "num_tokens": 355590352.0, + "step": 13744 + }, + { + "epoch": 1.5094443224247749, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3115735054016113, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7042921185493469, + "num_tokens": 355616447.0, + "step": 13745 + }, + { + "epoch": 1.5095541401273884, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2441937923431396, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7074554562568665, + "num_tokens": 355643921.0, + "step": 13746 + }, + { + "epoch": 1.5096639578300022, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.528198719024658, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7373129725456238, + "num_tokens": 355663688.0, + "step": 13747 + }, + { + "epoch": 1.509773775532616, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7132036685943604, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7187820076942444, + "num_tokens": 355684287.0, + "step": 13748 + }, + { + "epoch": 1.5098835932352295, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0917465686798096, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7061182260513306, + "num_tokens": 355714256.0, + "step": 13749 + }, + { + "epoch": 1.5099934109378432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1680071353912354, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7133424282073975, + "num_tokens": 355741462.0, + "step": 13750 + }, + { + "epoch": 1.5101032286404568, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.355130434036255, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7444665431976318, + "num_tokens": 355762296.0, + "step": 13751 + }, + { + "epoch": 1.5102130463430705, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0716400146484375, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6998714804649353, + "num_tokens": 355792887.0, + "step": 13752 + }, + { + "epoch": 1.5103228640456843, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1429443359375, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7135428190231323, + "num_tokens": 355821379.0, + "step": 13753 + }, + { + "epoch": 1.5104326817482978, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4376790523529053, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.726775586605072, + "num_tokens": 355843447.0, + "step": 13754 + }, + { + "epoch": 1.5105424994509113, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3798274993896484, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7133709788322449, + "num_tokens": 355867998.0, + "step": 13755 + }, + { + "epoch": 1.510652317153525, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3870654106140137, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7381787300109863, + "num_tokens": 355888455.0, + "step": 13756 + }, + { + "epoch": 1.5107621348561389, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1734085083007812, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7397838830947876, + "num_tokens": 355913597.0, + "step": 13757 + }, + { + "epoch": 1.5108719525587526, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.218315601348877, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7137405872344971, + "num_tokens": 355939107.0, + "step": 13758 + }, + { + "epoch": 1.5109817702613662, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9700920581817627, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7187908291816711, + "num_tokens": 355969652.0, + "step": 13759 + }, + { + "epoch": 1.5110915879639797, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2959723472595215, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7398297786712646, + "num_tokens": 355995923.0, + "step": 13760 + }, + { + "epoch": 1.5112014056665934, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.266437530517578, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7153184413909912, + "num_tokens": 356020220.0, + "step": 13761 + }, + { + "epoch": 1.5113112233692072, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.108206033706665, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7098656892776489, + "num_tokens": 356047911.0, + "step": 13762 + }, + { + "epoch": 1.5114210410718207, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1383702754974365, + "learning_rate": 1e-06, + "loss": 1.0356, + "mean_token_accuracy": 0.689599335193634, + "num_tokens": 356074462.0, + "step": 13763 + }, + { + "epoch": 1.5115308587744345, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.129960298538208, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7324596047401428, + "num_tokens": 356101105.0, + "step": 13764 + }, + { + "epoch": 1.511640676477048, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.237347364425659, + "learning_rate": 1e-06, + "loss": 0.9806, + "mean_token_accuracy": 0.7002265453338623, + "num_tokens": 356128051.0, + "step": 13765 + }, + { + "epoch": 1.5117504941796618, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4038214683532715, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7283679246902466, + "num_tokens": 356151505.0, + "step": 13766 + }, + { + "epoch": 1.5118603118822755, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3454031944274902, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7214834690093994, + "num_tokens": 356177332.0, + "step": 13767 + }, + { + "epoch": 1.511970129584889, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0228073596954346, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.747958779335022, + "num_tokens": 356205591.0, + "step": 13768 + }, + { + "epoch": 1.5120799472875026, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4031505584716797, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7459050416946411, + "num_tokens": 356226094.0, + "step": 13769 + }, + { + "epoch": 1.5121897649901164, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.253248929977417, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.6947380304336548, + "num_tokens": 356252556.0, + "step": 13770 + }, + { + "epoch": 1.5122995826927301, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2535135746002197, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7396551370620728, + "num_tokens": 356276592.0, + "step": 13771 + }, + { + "epoch": 1.5124094003953439, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4385805130004883, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7239370346069336, + "num_tokens": 356298457.0, + "step": 13772 + }, + { + "epoch": 1.5125192180979574, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.231300115585327, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7132484912872314, + "num_tokens": 356324337.0, + "step": 13773 + }, + { + "epoch": 1.512629035800571, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.488434076309204, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7276705503463745, + "num_tokens": 356345853.0, + "step": 13774 + }, + { + "epoch": 1.5127388535031847, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0973751544952393, + "learning_rate": 1e-06, + "loss": 1.0194, + "mean_token_accuracy": 0.6851277947425842, + "num_tokens": 356379117.0, + "step": 13775 + }, + { + "epoch": 1.5128486712057985, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3366754055023193, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7184035181999207, + "num_tokens": 356405071.0, + "step": 13776 + }, + { + "epoch": 1.512958488908412, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4878621101379395, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7245503664016724, + "num_tokens": 356425984.0, + "step": 13777 + }, + { + "epoch": 1.5130683066110255, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2907023429870605, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7265188694000244, + "num_tokens": 356448982.0, + "step": 13778 + }, + { + "epoch": 1.5131781243136393, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5296542644500732, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7324329018592834, + "num_tokens": 356468840.0, + "step": 13779 + }, + { + "epoch": 1.513287942016253, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1562955379486084, + "learning_rate": 1e-06, + "loss": 1.0137, + "mean_token_accuracy": 0.69645756483078, + "num_tokens": 356496916.0, + "step": 13780 + }, + { + "epoch": 1.5133977597188668, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3695127964019775, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7395610213279724, + "num_tokens": 356519639.0, + "step": 13781 + }, + { + "epoch": 1.5135075774214803, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.381181001663208, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7116272449493408, + "num_tokens": 356543945.0, + "step": 13782 + }, + { + "epoch": 1.5136173951240939, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.160782814025879, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.711862325668335, + "num_tokens": 356571028.0, + "step": 13783 + }, + { + "epoch": 1.5137272128267076, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1316518783569336, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7218524217605591, + "num_tokens": 356600256.0, + "step": 13784 + }, + { + "epoch": 1.5138370305293214, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.328422784805298, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.7061924934387207, + "num_tokens": 356625344.0, + "step": 13785 + }, + { + "epoch": 1.5139468482319351, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1734578609466553, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7266427278518677, + "num_tokens": 356650057.0, + "step": 13786 + }, + { + "epoch": 1.5140566659345487, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2594263553619385, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7202334403991699, + "num_tokens": 356675194.0, + "step": 13787 + }, + { + "epoch": 1.5141664836371622, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.432574510574341, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7329927086830139, + "num_tokens": 356698600.0, + "step": 13788 + }, + { + "epoch": 1.514276301339776, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.052466869354248, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7061364650726318, + "num_tokens": 356729849.0, + "step": 13789 + }, + { + "epoch": 1.5143861190423897, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3051488399505615, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.726523756980896, + "num_tokens": 356755212.0, + "step": 13790 + }, + { + "epoch": 1.5144959367450033, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9795845746994019, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.7024251222610474, + "num_tokens": 356786954.0, + "step": 13791 + }, + { + "epoch": 1.5146057544476168, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2830257415771484, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.703959047794342, + "num_tokens": 356811653.0, + "step": 13792 + }, + { + "epoch": 1.5147155721502306, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.000504493713379, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7154175043106079, + "num_tokens": 356841369.0, + "step": 13793 + }, + { + "epoch": 1.5148253898528443, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2167856693267822, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7087466716766357, + "num_tokens": 356870485.0, + "step": 13794 + }, + { + "epoch": 1.514935207555458, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9358441829681396, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7035864591598511, + "num_tokens": 356904064.0, + "step": 13795 + }, + { + "epoch": 1.5150450252580716, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.53739595413208, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7363486289978027, + "num_tokens": 356924167.0, + "step": 13796 + }, + { + "epoch": 1.5151548429606851, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5628514289855957, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7453693747520447, + "num_tokens": 356944442.0, + "step": 13797 + }, + { + "epoch": 1.515264660663299, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.724932909011841, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7153956890106201, + "num_tokens": 356963483.0, + "step": 13798 + }, + { + "epoch": 1.5153744783659127, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4132320880889893, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7058883309364319, + "num_tokens": 356986322.0, + "step": 13799 + }, + { + "epoch": 1.5154842960685262, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0723721981048584, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.6905444264411926, + "num_tokens": 357019143.0, + "step": 13800 + }, + { + "epoch": 1.51559411377114, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.41023588180542, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7255555987358093, + "num_tokens": 357041884.0, + "step": 13801 + }, + { + "epoch": 1.5157039314737535, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.060436487197876, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7048019170761108, + "num_tokens": 357072800.0, + "step": 13802 + }, + { + "epoch": 1.5158137491763672, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2228007316589355, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7223455309867859, + "num_tokens": 357096559.0, + "step": 13803 + }, + { + "epoch": 1.515923566878981, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4326727390289307, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7002713680267334, + "num_tokens": 357118220.0, + "step": 13804 + }, + { + "epoch": 1.5160333845815945, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0510239601135254, + "learning_rate": 1e-06, + "loss": 1.0444, + "mean_token_accuracy": 0.6833184957504272, + "num_tokens": 357150262.0, + "step": 13805 + }, + { + "epoch": 1.516143202284208, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2125823497772217, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7290088534355164, + "num_tokens": 357176008.0, + "step": 13806 + }, + { + "epoch": 1.5162530199868218, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.268167495727539, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7072118520736694, + "num_tokens": 357202813.0, + "step": 13807 + }, + { + "epoch": 1.5163628376894356, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.196131467819214, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7242642641067505, + "num_tokens": 357229265.0, + "step": 13808 + }, + { + "epoch": 1.5164726553920493, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1359622478485107, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.735652506351471, + "num_tokens": 357256720.0, + "step": 13809 + }, + { + "epoch": 1.5165824730946629, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2398393154144287, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7190604209899902, + "num_tokens": 357281438.0, + "step": 13810 + }, + { + "epoch": 1.5166922907972764, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.021665334701538, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7207746505737305, + "num_tokens": 357309993.0, + "step": 13811 + }, + { + "epoch": 1.5168021084998902, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1657989025115967, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7040002942085266, + "num_tokens": 357338261.0, + "step": 13812 + }, + { + "epoch": 1.516911926202504, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.255247116088867, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7223862409591675, + "num_tokens": 357364164.0, + "step": 13813 + }, + { + "epoch": 1.5170217439051175, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.167004346847534, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7333012819290161, + "num_tokens": 357390889.0, + "step": 13814 + }, + { + "epoch": 1.5171315616077312, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.211850643157959, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7110458016395569, + "num_tokens": 357417326.0, + "step": 13815 + }, + { + "epoch": 1.5172413793103448, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6250526905059814, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.730854332447052, + "num_tokens": 357437130.0, + "step": 13816 + }, + { + "epoch": 1.5173511970129585, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.315333366394043, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7221917510032654, + "num_tokens": 357462575.0, + "step": 13817 + }, + { + "epoch": 1.5174610147155723, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2585785388946533, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7359375953674316, + "num_tokens": 357485898.0, + "step": 13818 + }, + { + "epoch": 1.5175708324181858, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.467939615249634, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7198823690414429, + "num_tokens": 357508517.0, + "step": 13819 + }, + { + "epoch": 1.5176806501207993, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.192220449447632, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.730850338935852, + "num_tokens": 357535443.0, + "step": 13820 + }, + { + "epoch": 1.517790467823413, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0658066272735596, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7148246765136719, + "num_tokens": 357565210.0, + "step": 13821 + }, + { + "epoch": 1.5179002855260268, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3036913871765137, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7105408906936646, + "num_tokens": 357589676.0, + "step": 13822 + }, + { + "epoch": 1.5180101032286406, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.06123685836792, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7391851544380188, + "num_tokens": 357616481.0, + "step": 13823 + }, + { + "epoch": 1.5181199209312541, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3596699237823486, + "learning_rate": 1e-06, + "loss": 0.756, + "mean_token_accuracy": 0.7580174207687378, + "num_tokens": 357637844.0, + "step": 13824 + }, + { + "epoch": 1.5182297386338677, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.070965528488159, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7182561159133911, + "num_tokens": 357665617.0, + "step": 13825 + }, + { + "epoch": 1.5183395563364814, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.073687791824341, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7185084223747253, + "num_tokens": 357693346.0, + "step": 13826 + }, + { + "epoch": 1.5184493740390952, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.335463523864746, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7176963090896606, + "num_tokens": 357716822.0, + "step": 13827 + }, + { + "epoch": 1.5185591917417087, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1834449768066406, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7220691442489624, + "num_tokens": 357743016.0, + "step": 13828 + }, + { + "epoch": 1.5186690094443223, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.34877347946167, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7117645740509033, + "num_tokens": 357767385.0, + "step": 13829 + }, + { + "epoch": 1.518778827146936, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2429306507110596, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7086629867553711, + "num_tokens": 357794899.0, + "step": 13830 + }, + { + "epoch": 1.5188886448495498, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3576200008392334, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7356328964233398, + "num_tokens": 357817416.0, + "step": 13831 + }, + { + "epoch": 1.5189984625521635, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2532737255096436, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7078918218612671, + "num_tokens": 357843749.0, + "step": 13832 + }, + { + "epoch": 1.519108280254777, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.423765182495117, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7269241809844971, + "num_tokens": 357867032.0, + "step": 13833 + }, + { + "epoch": 1.5192180979573906, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4322855472564697, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7478494644165039, + "num_tokens": 357888379.0, + "step": 13834 + }, + { + "epoch": 1.5193279156600044, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3359529972076416, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7153019905090332, + "num_tokens": 357911911.0, + "step": 13835 + }, + { + "epoch": 1.5194377333626181, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6673567295074463, + "learning_rate": 1e-06, + "loss": 0.7451, + "mean_token_accuracy": 0.7591546773910522, + "num_tokens": 357929136.0, + "step": 13836 + }, + { + "epoch": 1.5195475510652319, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1668636798858643, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7218830585479736, + "num_tokens": 357955890.0, + "step": 13837 + }, + { + "epoch": 1.5196573687678454, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.41560697555542, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7266504168510437, + "num_tokens": 357978799.0, + "step": 13838 + }, + { + "epoch": 1.519767186470459, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7686169147491455, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7217340469360352, + "num_tokens": 357997097.0, + "step": 13839 + }, + { + "epoch": 1.5198770041730727, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1184372901916504, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7143288254737854, + "num_tokens": 358023895.0, + "step": 13840 + }, + { + "epoch": 1.5199868218756865, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0208094120025635, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7075743079185486, + "num_tokens": 358055221.0, + "step": 13841 + }, + { + "epoch": 1.5200966395783, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9578031301498413, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7263168692588806, + "num_tokens": 358086249.0, + "step": 13842 + }, + { + "epoch": 1.5202064572809135, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.366391181945801, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7327159643173218, + "num_tokens": 358108163.0, + "step": 13843 + }, + { + "epoch": 1.5203162749835273, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.175647497177124, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7100033164024353, + "num_tokens": 358136746.0, + "step": 13844 + }, + { + "epoch": 1.520426092686141, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2765069007873535, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7174415588378906, + "num_tokens": 358161247.0, + "step": 13845 + }, + { + "epoch": 1.5205359103887548, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.339163303375244, + "learning_rate": 1e-06, + "loss": 0.9535, + "mean_token_accuracy": 0.7082589864730835, + "num_tokens": 358185895.0, + "step": 13846 + }, + { + "epoch": 1.5206457280913683, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.376629114151001, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7366712093353271, + "num_tokens": 358207700.0, + "step": 13847 + }, + { + "epoch": 1.5207555457939819, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.479888916015625, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7298389077186584, + "num_tokens": 358229003.0, + "step": 13848 + }, + { + "epoch": 1.5208653634965956, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.15978741645813, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.725089430809021, + "num_tokens": 358254281.0, + "step": 13849 + }, + { + "epoch": 1.5209751811992094, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3325798511505127, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6994660496711731, + "num_tokens": 358280670.0, + "step": 13850 + }, + { + "epoch": 1.5210849989018231, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.007858991622925, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6993104219436646, + "num_tokens": 358310030.0, + "step": 13851 + }, + { + "epoch": 1.5211948166044367, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3698790073394775, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7061688899993896, + "num_tokens": 358335096.0, + "step": 13852 + }, + { + "epoch": 1.5213046343070502, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3500585556030273, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7285836338996887, + "num_tokens": 358356406.0, + "step": 13853 + }, + { + "epoch": 1.521414452009664, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.551196336746216, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7336723208427429, + "num_tokens": 358377855.0, + "step": 13854 + }, + { + "epoch": 1.5215242697122777, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.183886766433716, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7412074208259583, + "num_tokens": 358405613.0, + "step": 13855 + }, + { + "epoch": 1.5216340874148913, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.040877103805542, + "learning_rate": 1e-06, + "loss": 1.0401, + "mean_token_accuracy": 0.6871349811553955, + "num_tokens": 358434728.0, + "step": 13856 + }, + { + "epoch": 1.5217439051175048, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3317079544067383, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7259752750396729, + "num_tokens": 358458558.0, + "step": 13857 + }, + { + "epoch": 1.5218537228201185, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.689039468765259, + "learning_rate": 1e-06, + "loss": 0.7576, + "mean_token_accuracy": 0.7580468654632568, + "num_tokens": 358475837.0, + "step": 13858 + }, + { + "epoch": 1.5219635405227323, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3507111072540283, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7334097623825073, + "num_tokens": 358500531.0, + "step": 13859 + }, + { + "epoch": 1.522073358225346, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7236554622650146, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7376264333724976, + "num_tokens": 358519057.0, + "step": 13860 + }, + { + "epoch": 1.5221831759279596, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.255260944366455, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7200954556465149, + "num_tokens": 358543699.0, + "step": 13861 + }, + { + "epoch": 1.5222929936305731, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5510106086730957, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.728110671043396, + "num_tokens": 358564328.0, + "step": 13862 + }, + { + "epoch": 1.522402811333187, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.263674020767212, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.740390419960022, + "num_tokens": 358587141.0, + "step": 13863 + }, + { + "epoch": 1.5225126290358006, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.028423547744751, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7262157201766968, + "num_tokens": 358616127.0, + "step": 13864 + }, + { + "epoch": 1.5226224467384142, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4309422969818115, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7146461009979248, + "num_tokens": 358639200.0, + "step": 13865 + }, + { + "epoch": 1.522732264441028, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4088683128356934, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7140169739723206, + "num_tokens": 358666300.0, + "step": 13866 + }, + { + "epoch": 1.5228420821436415, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.047433853149414, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.704627513885498, + "num_tokens": 358696624.0, + "step": 13867 + }, + { + "epoch": 1.5229518998462552, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0728628635406494, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7098436951637268, + "num_tokens": 358725844.0, + "step": 13868 + }, + { + "epoch": 1.523061717548869, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.467876672744751, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7136920690536499, + "num_tokens": 358746996.0, + "step": 13869 + }, + { + "epoch": 1.5231715352514825, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2093756198883057, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7272944450378418, + "num_tokens": 358772315.0, + "step": 13870 + }, + { + "epoch": 1.523281352954096, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1803488731384277, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7203387022018433, + "num_tokens": 358798304.0, + "step": 13871 + }, + { + "epoch": 1.5233911706567098, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1987762451171875, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7213914394378662, + "num_tokens": 358823430.0, + "step": 13872 + }, + { + "epoch": 1.5235009883593236, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5340018272399902, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7076910734176636, + "num_tokens": 358845496.0, + "step": 13873 + }, + { + "epoch": 1.5236108060619373, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.063809633255005, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7270491123199463, + "num_tokens": 358874770.0, + "step": 13874 + }, + { + "epoch": 1.5237206237645509, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2636537551879883, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7260359525680542, + "num_tokens": 358898948.0, + "step": 13875 + }, + { + "epoch": 1.5238304414671644, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1579573154449463, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7398123741149902, + "num_tokens": 358925509.0, + "step": 13876 + }, + { + "epoch": 1.5239402591697782, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0152552127838135, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7158480882644653, + "num_tokens": 358956547.0, + "step": 13877 + }, + { + "epoch": 1.524050076872392, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0669538974761963, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7226482629776001, + "num_tokens": 358985992.0, + "step": 13878 + }, + { + "epoch": 1.5241598945750054, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2293310165405273, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7175358533859253, + "num_tokens": 359011171.0, + "step": 13879 + }, + { + "epoch": 1.5242697122776192, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7482571601867676, + "learning_rate": 1e-06, + "loss": 0.7695, + "mean_token_accuracy": 0.7503061294555664, + "num_tokens": 359029728.0, + "step": 13880 + }, + { + "epoch": 1.5243795299802327, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4797163009643555, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7101527452468872, + "num_tokens": 359051188.0, + "step": 13881 + }, + { + "epoch": 1.5244893476828465, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2326629161834717, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7227389812469482, + "num_tokens": 359075009.0, + "step": 13882 + }, + { + "epoch": 1.5245991653854603, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.516712188720703, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7435121536254883, + "num_tokens": 359096328.0, + "step": 13883 + }, + { + "epoch": 1.5247089830880738, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.439117193222046, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7327277064323425, + "num_tokens": 359120262.0, + "step": 13884 + }, + { + "epoch": 1.5248188007906873, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.450079917907715, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7261490225791931, + "num_tokens": 359140696.0, + "step": 13885 + }, + { + "epoch": 1.524928618493301, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.339266538619995, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7111620903015137, + "num_tokens": 359164468.0, + "step": 13886 + }, + { + "epoch": 1.5250384361959148, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0374042987823486, + "learning_rate": 1e-06, + "loss": 0.9589, + "mean_token_accuracy": 0.7034311890602112, + "num_tokens": 359195179.0, + "step": 13887 + }, + { + "epoch": 1.5251482538985286, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.256958484649658, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6974666714668274, + "num_tokens": 359222376.0, + "step": 13888 + }, + { + "epoch": 1.5252580716011421, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.184037685394287, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.6959589719772339, + "num_tokens": 359251589.0, + "step": 13889 + }, + { + "epoch": 1.5253678893037557, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0154130458831787, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7043054103851318, + "num_tokens": 359283194.0, + "step": 13890 + }, + { + "epoch": 1.5254777070063694, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2958710193634033, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7330785989761353, + "num_tokens": 359306658.0, + "step": 13891 + }, + { + "epoch": 1.5255875247089832, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0862345695495605, + "learning_rate": 1e-06, + "loss": 0.9801, + "mean_token_accuracy": 0.6974393129348755, + "num_tokens": 359338558.0, + "step": 13892 + }, + { + "epoch": 1.5256973424115967, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.112833261489868, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7218542098999023, + "num_tokens": 359365423.0, + "step": 13893 + }, + { + "epoch": 1.5258071601142102, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5820348262786865, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.704635500907898, + "num_tokens": 359387069.0, + "step": 13894 + }, + { + "epoch": 1.525916977816824, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.236342191696167, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7153868079185486, + "num_tokens": 359412170.0, + "step": 13895 + }, + { + "epoch": 1.5260267955194378, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.112992763519287, + "learning_rate": 1e-06, + "loss": 0.9694, + "mean_token_accuracy": 0.6975142359733582, + "num_tokens": 359441555.0, + "step": 13896 + }, + { + "epoch": 1.5261366132220515, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.461637258529663, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7479557394981384, + "num_tokens": 359463418.0, + "step": 13897 + }, + { + "epoch": 1.526246430924665, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.066931962966919, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7262015342712402, + "num_tokens": 359493384.0, + "step": 13898 + }, + { + "epoch": 1.5263562486272786, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9790911674499512, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7248034477233887, + "num_tokens": 359525207.0, + "step": 13899 + }, + { + "epoch": 1.5264660663298923, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1207385063171387, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7191802859306335, + "num_tokens": 359554399.0, + "step": 13900 + }, + { + "epoch": 1.526575884032506, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4975831508636475, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7279725074768066, + "num_tokens": 359574869.0, + "step": 13901 + }, + { + "epoch": 1.5266857017351199, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5826361179351807, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7171304225921631, + "num_tokens": 359598923.0, + "step": 13902 + }, + { + "epoch": 1.5267955194377334, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.039612293243408, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7058502435684204, + "num_tokens": 359633756.0, + "step": 13903 + }, + { + "epoch": 1.526905337140347, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7468972206115723, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7241188287734985, + "num_tokens": 359655317.0, + "step": 13904 + }, + { + "epoch": 1.5270151548429607, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4240009784698486, + "learning_rate": 1e-06, + "loss": 0.9899, + "mean_token_accuracy": 0.6923831701278687, + "num_tokens": 359679187.0, + "step": 13905 + }, + { + "epoch": 1.5271249725455744, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3790206909179688, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7242501378059387, + "num_tokens": 359701225.0, + "step": 13906 + }, + { + "epoch": 1.527234790248188, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1139323711395264, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7256814241409302, + "num_tokens": 359731034.0, + "step": 13907 + }, + { + "epoch": 1.5273446079508015, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1496262550354004, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.70501309633255, + "num_tokens": 359759267.0, + "step": 13908 + }, + { + "epoch": 1.5274544256534153, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3118531703948975, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7243276238441467, + "num_tokens": 359783366.0, + "step": 13909 + }, + { + "epoch": 1.527564243356029, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.084449529647827, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.6996798515319824, + "num_tokens": 359813623.0, + "step": 13910 + }, + { + "epoch": 1.5276740610586428, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.115272283554077, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7377448678016663, + "num_tokens": 359839622.0, + "step": 13911 + }, + { + "epoch": 1.5277838787612563, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5429844856262207, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7206823825836182, + "num_tokens": 359861511.0, + "step": 13912 + }, + { + "epoch": 1.5278936964638699, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.580838203430176, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7106307148933411, + "num_tokens": 359882319.0, + "step": 13913 + }, + { + "epoch": 1.5280035141664836, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.221818447113037, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7153949737548828, + "num_tokens": 359907346.0, + "step": 13914 + }, + { + "epoch": 1.5281133318690974, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.201462745666504, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.726978063583374, + "num_tokens": 359937308.0, + "step": 13915 + }, + { + "epoch": 1.5282231495717111, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9209660291671753, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.721308171749115, + "num_tokens": 359969915.0, + "step": 13916 + }, + { + "epoch": 1.5283329672743247, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.365530252456665, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7232750654220581, + "num_tokens": 359992266.0, + "step": 13917 + }, + { + "epoch": 1.5284427849769382, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.423959970474243, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7310401797294617, + "num_tokens": 360014096.0, + "step": 13918 + }, + { + "epoch": 1.528552602679552, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0388522148132324, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.6977106928825378, + "num_tokens": 360044076.0, + "step": 13919 + }, + { + "epoch": 1.5286624203821657, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9962652921676636, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7193666696548462, + "num_tokens": 360073970.0, + "step": 13920 + }, + { + "epoch": 1.5287722380847792, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1792116165161133, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7247000336647034, + "num_tokens": 360099803.0, + "step": 13921 + }, + { + "epoch": 1.5288820557873928, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.220895290374756, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7173575162887573, + "num_tokens": 360124665.0, + "step": 13922 + }, + { + "epoch": 1.5289918734900065, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0350897312164307, + "learning_rate": 1e-06, + "loss": 1.0177, + "mean_token_accuracy": 0.6982614398002625, + "num_tokens": 360154624.0, + "step": 13923 + }, + { + "epoch": 1.5291016911926203, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1518208980560303, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7263229489326477, + "num_tokens": 360181063.0, + "step": 13924 + }, + { + "epoch": 1.529211508895234, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0082178115844727, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7113838195800781, + "num_tokens": 360209406.0, + "step": 13925 + }, + { + "epoch": 1.5293213265978476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.423165798187256, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7367559671401978, + "num_tokens": 360231664.0, + "step": 13926 + }, + { + "epoch": 1.5294311443004611, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1380808353424072, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7243569493293762, + "num_tokens": 360259971.0, + "step": 13927 + }, + { + "epoch": 1.5295409620030749, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3393027782440186, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7330236434936523, + "num_tokens": 360283538.0, + "step": 13928 + }, + { + "epoch": 1.5296507797056886, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.09552264213562, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7021734714508057, + "num_tokens": 360311061.0, + "step": 13929 + }, + { + "epoch": 1.5297605974083022, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.8335556983947754, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7412686347961426, + "num_tokens": 360327539.0, + "step": 13930 + }, + { + "epoch": 1.529870415110916, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2766621112823486, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7166948318481445, + "num_tokens": 360352858.0, + "step": 13931 + }, + { + "epoch": 1.5299802328135295, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.100423574447632, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7117263078689575, + "num_tokens": 360383213.0, + "step": 13932 + }, + { + "epoch": 1.5300900505161432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2094106674194336, + "learning_rate": 1e-06, + "loss": 0.9659, + "mean_token_accuracy": 0.7116167545318604, + "num_tokens": 360411710.0, + "step": 13933 + }, + { + "epoch": 1.530199868218757, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3902883529663086, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7097100019454956, + "num_tokens": 360434777.0, + "step": 13934 + }, + { + "epoch": 1.5303096859213705, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9747718572616577, + "learning_rate": 1e-06, + "loss": 1.0075, + "mean_token_accuracy": 0.6901780366897583, + "num_tokens": 360467550.0, + "step": 13935 + }, + { + "epoch": 1.530419503623984, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2259178161621094, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7037107944488525, + "num_tokens": 360494011.0, + "step": 13936 + }, + { + "epoch": 1.5305293213265978, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.142448663711548, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7185683250427246, + "num_tokens": 360521569.0, + "step": 13937 + }, + { + "epoch": 1.5306391390292116, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2699618339538574, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7215512990951538, + "num_tokens": 360548911.0, + "step": 13938 + }, + { + "epoch": 1.5307489567318253, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2776939868927, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7096192836761475, + "num_tokens": 360574672.0, + "step": 13939 + }, + { + "epoch": 1.5308587744344389, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3030965328216553, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7232877016067505, + "num_tokens": 360599959.0, + "step": 13940 + }, + { + "epoch": 1.5309685921370524, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2005057334899902, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7193766832351685, + "num_tokens": 360626176.0, + "step": 13941 + }, + { + "epoch": 1.5310784098396661, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.18605637550354, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.727827787399292, + "num_tokens": 360650766.0, + "step": 13942 + }, + { + "epoch": 1.53118822754228, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3836441040039062, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7095168828964233, + "num_tokens": 360674516.0, + "step": 13943 + }, + { + "epoch": 1.5312980452448934, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0097882747650146, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.708882212638855, + "num_tokens": 360702725.0, + "step": 13944 + }, + { + "epoch": 1.5314078629475072, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3042185306549072, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7077468633651733, + "num_tokens": 360727098.0, + "step": 13945 + }, + { + "epoch": 1.5315176806501207, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.214289426803589, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7205734252929688, + "num_tokens": 360755180.0, + "step": 13946 + }, + { + "epoch": 1.5316274983527345, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3129756450653076, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7346576452255249, + "num_tokens": 360779517.0, + "step": 13947 + }, + { + "epoch": 1.5317373160553482, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.452723264694214, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.717828631401062, + "num_tokens": 360799708.0, + "step": 13948 + }, + { + "epoch": 1.5318471337579618, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.281508445739746, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7051928043365479, + "num_tokens": 360825254.0, + "step": 13949 + }, + { + "epoch": 1.5319569514605753, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5569522380828857, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7168651819229126, + "num_tokens": 360844500.0, + "step": 13950 + }, + { + "epoch": 1.532066769163189, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6117851734161377, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7146356105804443, + "num_tokens": 360864445.0, + "step": 13951 + }, + { + "epoch": 1.5321765868658028, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3536624908447266, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7428182363510132, + "num_tokens": 360887670.0, + "step": 13952 + }, + { + "epoch": 1.5322864045684166, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.297255754470825, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7393508553504944, + "num_tokens": 360911609.0, + "step": 13953 + }, + { + "epoch": 1.5323962222710301, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3551583290100098, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7191802263259888, + "num_tokens": 360935516.0, + "step": 13954 + }, + { + "epoch": 1.5325060399736437, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.54319167137146, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7142156362533569, + "num_tokens": 360957172.0, + "step": 13955 + }, + { + "epoch": 1.5326158576762574, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.13212251663208, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7424570322036743, + "num_tokens": 360984712.0, + "step": 13956 + }, + { + "epoch": 1.5327256753788712, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1959598064422607, + "learning_rate": 1e-06, + "loss": 0.9742, + "mean_token_accuracy": 0.7002779245376587, + "num_tokens": 361013040.0, + "step": 13957 + }, + { + "epoch": 1.5328354930814847, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.066288948059082, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.7015519738197327, + "num_tokens": 361045049.0, + "step": 13958 + }, + { + "epoch": 1.5329453107840982, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1739590167999268, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.6942801475524902, + "num_tokens": 361074053.0, + "step": 13959 + }, + { + "epoch": 1.533055128486712, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2036709785461426, + "learning_rate": 1e-06, + "loss": 1.0205, + "mean_token_accuracy": 0.6884422302246094, + "num_tokens": 361104096.0, + "step": 13960 + }, + { + "epoch": 1.5331649461893258, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3255622386932373, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7241446375846863, + "num_tokens": 361127413.0, + "step": 13961 + }, + { + "epoch": 1.5332747638919395, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.040621280670166, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7108727693557739, + "num_tokens": 361156264.0, + "step": 13962 + }, + { + "epoch": 1.533384581594553, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4180524349212646, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7178913354873657, + "num_tokens": 361179506.0, + "step": 13963 + }, + { + "epoch": 1.5334943992971666, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0138251781463623, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7105041742324829, + "num_tokens": 361208503.0, + "step": 13964 + }, + { + "epoch": 1.5336042169997803, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.052227258682251, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7088308334350586, + "num_tokens": 361236807.0, + "step": 13965 + }, + { + "epoch": 1.533714034702394, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.148409128189087, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7429083585739136, + "num_tokens": 361260891.0, + "step": 13966 + }, + { + "epoch": 1.5338238524050078, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5479514598846436, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7513651251792908, + "num_tokens": 361280470.0, + "step": 13967 + }, + { + "epoch": 1.5339336701076214, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.04435133934021, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7173643112182617, + "num_tokens": 361308667.0, + "step": 13968 + }, + { + "epoch": 1.534043487810235, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2469489574432373, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7276867628097534, + "num_tokens": 361333878.0, + "step": 13969 + }, + { + "epoch": 1.5341533055128487, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.221811294555664, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7015345096588135, + "num_tokens": 361360833.0, + "step": 13970 + }, + { + "epoch": 1.5342631232154624, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2460832595825195, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7114501595497131, + "num_tokens": 361388075.0, + "step": 13971 + }, + { + "epoch": 1.534372940918076, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.308760404586792, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7264178991317749, + "num_tokens": 361413148.0, + "step": 13972 + }, + { + "epoch": 1.5344827586206895, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.269451856613159, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7134600877761841, + "num_tokens": 361441051.0, + "step": 13973 + }, + { + "epoch": 1.5345925763233033, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3891701698303223, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7290055751800537, + "num_tokens": 361464033.0, + "step": 13974 + }, + { + "epoch": 1.534702394025917, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1923651695251465, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7232636213302612, + "num_tokens": 361490130.0, + "step": 13975 + }, + { + "epoch": 1.5348122117285308, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1017825603485107, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.733566403388977, + "num_tokens": 361516786.0, + "step": 13976 + }, + { + "epoch": 1.5349220294311443, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2907187938690186, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7356099486351013, + "num_tokens": 361541102.0, + "step": 13977 + }, + { + "epoch": 1.5350318471337578, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.533306837081909, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7060914039611816, + "num_tokens": 361564405.0, + "step": 13978 + }, + { + "epoch": 1.5351416648363716, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3450849056243896, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7381690144538879, + "num_tokens": 361587221.0, + "step": 13979 + }, + { + "epoch": 1.5352514825389854, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1584599018096924, + "learning_rate": 1e-06, + "loss": 1.0117, + "mean_token_accuracy": 0.6893631815910339, + "num_tokens": 361617470.0, + "step": 13980 + }, + { + "epoch": 1.535361300241599, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.412792444229126, + "learning_rate": 1e-06, + "loss": 0.9699, + "mean_token_accuracy": 0.7083097696304321, + "num_tokens": 361641579.0, + "step": 13981 + }, + { + "epoch": 1.5354711179442126, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.657052516937256, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.72039794921875, + "num_tokens": 361660463.0, + "step": 13982 + }, + { + "epoch": 1.5355809356468262, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.206861972808838, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7151846885681152, + "num_tokens": 361687467.0, + "step": 13983 + }, + { + "epoch": 1.53569075334944, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.300126314163208, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7324151992797852, + "num_tokens": 361711970.0, + "step": 13984 + }, + { + "epoch": 1.5358005710520537, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.483903646469116, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7232218980789185, + "num_tokens": 361734584.0, + "step": 13985 + }, + { + "epoch": 1.5359103887546672, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.05837082862854, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7137013673782349, + "num_tokens": 361764865.0, + "step": 13986 + }, + { + "epoch": 1.5360202064572808, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.060910940170288, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7128207683563232, + "num_tokens": 361794556.0, + "step": 13987 + }, + { + "epoch": 1.5361300241598945, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.341787338256836, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7235082983970642, + "num_tokens": 361820100.0, + "step": 13988 + }, + { + "epoch": 1.5362398418625083, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.191512107849121, + "learning_rate": 1e-06, + "loss": 0.9394, + "mean_token_accuracy": 0.7102879285812378, + "num_tokens": 361847555.0, + "step": 13989 + }, + { + "epoch": 1.536349659565122, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.260023832321167, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7256431579589844, + "num_tokens": 361873202.0, + "step": 13990 + }, + { + "epoch": 1.5364594772677356, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0644690990448, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7288479804992676, + "num_tokens": 361900641.0, + "step": 13991 + }, + { + "epoch": 1.536569294970349, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.533141613006592, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7212378978729248, + "num_tokens": 361920081.0, + "step": 13992 + }, + { + "epoch": 1.5366791126729629, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9866560697555542, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.711260974407196, + "num_tokens": 361951476.0, + "step": 13993 + }, + { + "epoch": 1.5367889303755766, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.179738759994507, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7156837582588196, + "num_tokens": 361978693.0, + "step": 13994 + }, + { + "epoch": 1.5368987480781902, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4822914600372314, + "learning_rate": 1e-06, + "loss": 1.0625, + "mean_token_accuracy": 0.6837517023086548, + "num_tokens": 362001740.0, + "step": 13995 + }, + { + "epoch": 1.537008565780804, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3525655269622803, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.718360960483551, + "num_tokens": 362024253.0, + "step": 13996 + }, + { + "epoch": 1.5371183834834174, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.092590093612671, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7125956416130066, + "num_tokens": 362052889.0, + "step": 13997 + }, + { + "epoch": 1.5372282011860312, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1527795791625977, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7430918216705322, + "num_tokens": 362081419.0, + "step": 13998 + }, + { + "epoch": 1.537338018888645, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7396912574768066, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7349244952201843, + "num_tokens": 362101127.0, + "step": 13999 + }, + { + "epoch": 1.5374478365912585, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.205458879470825, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.748159646987915, + "num_tokens": 362129145.0, + "step": 14000 + }, + { + "epoch": 1.537557654293872, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6134283542633057, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7342389822006226, + "num_tokens": 362148434.0, + "step": 14001 + }, + { + "epoch": 1.5376674719964858, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0764124393463135, + "learning_rate": 1e-06, + "loss": 0.9733, + "mean_token_accuracy": 0.7069076895713806, + "num_tokens": 362177949.0, + "step": 14002 + }, + { + "epoch": 1.5377772896990995, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.28743052482605, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7031693458557129, + "num_tokens": 362202940.0, + "step": 14003 + }, + { + "epoch": 1.5378871074017133, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2900123596191406, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7177385687828064, + "num_tokens": 362227009.0, + "step": 14004 + }, + { + "epoch": 1.5379969251043268, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.289855718612671, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7335245609283447, + "num_tokens": 362251623.0, + "step": 14005 + }, + { + "epoch": 1.5381067428069404, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0834126472473145, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6874111890792847, + "num_tokens": 362282130.0, + "step": 14006 + }, + { + "epoch": 1.5382165605095541, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2150321006774902, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.6979316473007202, + "num_tokens": 362309947.0, + "step": 14007 + }, + { + "epoch": 1.5383263782121679, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.521512508392334, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7240296602249146, + "num_tokens": 362331758.0, + "step": 14008 + }, + { + "epoch": 1.5384361959147814, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.284686803817749, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7160289883613586, + "num_tokens": 362355589.0, + "step": 14009 + }, + { + "epoch": 1.538546013617395, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.315657615661621, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7339188456535339, + "num_tokens": 362378441.0, + "step": 14010 + }, + { + "epoch": 1.5386558313200087, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1022720336914062, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7180887460708618, + "num_tokens": 362406264.0, + "step": 14011 + }, + { + "epoch": 1.5387656490226225, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3360366821289062, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7271716594696045, + "num_tokens": 362429698.0, + "step": 14012 + }, + { + "epoch": 1.5388754667252362, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.403042793273926, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7354285717010498, + "num_tokens": 362451815.0, + "step": 14013 + }, + { + "epoch": 1.5389852844278498, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.443495512008667, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7074317932128906, + "num_tokens": 362475655.0, + "step": 14014 + }, + { + "epoch": 1.5390951021304633, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9931639432907104, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6976132392883301, + "num_tokens": 362509816.0, + "step": 14015 + }, + { + "epoch": 1.539204919833077, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.220259666442871, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.728216290473938, + "num_tokens": 362536931.0, + "step": 14016 + }, + { + "epoch": 1.5393147375356908, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.413029193878174, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7304665446281433, + "num_tokens": 362558784.0, + "step": 14017 + }, + { + "epoch": 1.5394245552383046, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.454845428466797, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7532949447631836, + "num_tokens": 362579862.0, + "step": 14018 + }, + { + "epoch": 1.539534372940918, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.235112190246582, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7422876358032227, + "num_tokens": 362604486.0, + "step": 14019 + }, + { + "epoch": 1.5396441906435316, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.019360065460205, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7298007011413574, + "num_tokens": 362632912.0, + "step": 14020 + }, + { + "epoch": 1.5397540083461454, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1562154293060303, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7357675433158875, + "num_tokens": 362658011.0, + "step": 14021 + }, + { + "epoch": 1.5398638260487592, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5087897777557373, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7090864777565002, + "num_tokens": 362678574.0, + "step": 14022 + }, + { + "epoch": 1.5399736437513727, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.170698642730713, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7045853734016418, + "num_tokens": 362707936.0, + "step": 14023 + }, + { + "epoch": 1.5400834614539862, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2073445320129395, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7004464864730835, + "num_tokens": 362736487.0, + "step": 14024 + }, + { + "epoch": 1.5401932791566, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4303247928619385, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7125269174575806, + "num_tokens": 362759878.0, + "step": 14025 + }, + { + "epoch": 1.5403030968592137, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0503406524658203, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7017453908920288, + "num_tokens": 362788417.0, + "step": 14026 + }, + { + "epoch": 1.5404129145618275, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0302867889404297, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7222977876663208, + "num_tokens": 362817928.0, + "step": 14027 + }, + { + "epoch": 1.540522732264441, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1337478160858154, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.6887043118476868, + "num_tokens": 362846569.0, + "step": 14028 + }, + { + "epoch": 1.5406325499670546, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2390284538269043, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7366910576820374, + "num_tokens": 362872228.0, + "step": 14029 + }, + { + "epoch": 1.5407423676696683, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3177995681762695, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7339385747909546, + "num_tokens": 362894173.0, + "step": 14030 + }, + { + "epoch": 1.540852185372282, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.246088981628418, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.717115581035614, + "num_tokens": 362920259.0, + "step": 14031 + }, + { + "epoch": 1.5409620030748958, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.607520341873169, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7458964586257935, + "num_tokens": 362938237.0, + "step": 14032 + }, + { + "epoch": 1.5410718207775094, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3106369972229004, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7229444980621338, + "num_tokens": 362961984.0, + "step": 14033 + }, + { + "epoch": 1.541181638480123, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.481977939605713, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.6962561011314392, + "num_tokens": 362985064.0, + "step": 14034 + }, + { + "epoch": 1.5412914561827367, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.324636936187744, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7127435803413391, + "num_tokens": 363010611.0, + "step": 14035 + }, + { + "epoch": 1.5414012738853504, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4284591674804688, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7316427230834961, + "num_tokens": 363032997.0, + "step": 14036 + }, + { + "epoch": 1.541511091587964, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1532492637634277, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7017620205879211, + "num_tokens": 363060709.0, + "step": 14037 + }, + { + "epoch": 1.5416209092905775, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.380229949951172, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7142002582550049, + "num_tokens": 363087751.0, + "step": 14038 + }, + { + "epoch": 1.5417307269931912, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5029263496398926, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7171047925949097, + "num_tokens": 363109821.0, + "step": 14039 + }, + { + "epoch": 1.541840544695805, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1007609367370605, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7049190402030945, + "num_tokens": 363140215.0, + "step": 14040 + }, + { + "epoch": 1.5419503623984188, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2721176147460938, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7138091325759888, + "num_tokens": 363164361.0, + "step": 14041 + }, + { + "epoch": 1.5420601801010323, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9856681823730469, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7042162418365479, + "num_tokens": 363199268.0, + "step": 14042 + }, + { + "epoch": 1.5421699978036458, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.117588520050049, + "learning_rate": 1e-06, + "loss": 0.9546, + "mean_token_accuracy": 0.7013471722602844, + "num_tokens": 363229026.0, + "step": 14043 + }, + { + "epoch": 1.5422798155062596, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1467087268829346, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7291419506072998, + "num_tokens": 363257039.0, + "step": 14044 + }, + { + "epoch": 1.5423896332088733, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.216526508331299, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7059170603752136, + "num_tokens": 363282409.0, + "step": 14045 + }, + { + "epoch": 1.5424994509114869, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1898350715637207, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7107455730438232, + "num_tokens": 363311899.0, + "step": 14046 + }, + { + "epoch": 1.5426092686141006, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1908271312713623, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7245764136314392, + "num_tokens": 363337782.0, + "step": 14047 + }, + { + "epoch": 1.5427190863167142, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.324854850769043, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7313269376754761, + "num_tokens": 363361218.0, + "step": 14048 + }, + { + "epoch": 1.542828904019328, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9797353744506836, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7298270463943481, + "num_tokens": 363390987.0, + "step": 14049 + }, + { + "epoch": 1.5429387217219417, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.8251116275787354, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7393282651901245, + "num_tokens": 363423598.0, + "step": 14050 + }, + { + "epoch": 1.5430485394245552, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3534698486328125, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7305251359939575, + "num_tokens": 363445154.0, + "step": 14051 + }, + { + "epoch": 1.5431583571271688, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0248000621795654, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7034899592399597, + "num_tokens": 363475840.0, + "step": 14052 + }, + { + "epoch": 1.5432681748297825, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3340251445770264, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.738358199596405, + "num_tokens": 363497933.0, + "step": 14053 + }, + { + "epoch": 1.5433779925323963, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1016616821289062, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.717195451259613, + "num_tokens": 363524024.0, + "step": 14054 + }, + { + "epoch": 1.54348781023501, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2812163829803467, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7191632986068726, + "num_tokens": 363549265.0, + "step": 14055 + }, + { + "epoch": 1.5435976279376236, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5154953002929688, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7247589230537415, + "num_tokens": 363571016.0, + "step": 14056 + }, + { + "epoch": 1.543707445640237, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0737204551696777, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.6981773972511292, + "num_tokens": 363600326.0, + "step": 14057 + }, + { + "epoch": 1.5438172633428509, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1838266849517822, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7157511711120605, + "num_tokens": 363626848.0, + "step": 14058 + }, + { + "epoch": 1.5439270810454646, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0646615028381348, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7196657657623291, + "num_tokens": 363656209.0, + "step": 14059 + }, + { + "epoch": 1.5440368987480781, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1659300327301025, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7177915573120117, + "num_tokens": 363683720.0, + "step": 14060 + }, + { + "epoch": 1.544146716450692, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.374601125717163, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7171857357025146, + "num_tokens": 363705783.0, + "step": 14061 + }, + { + "epoch": 1.5442565341533054, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.338536500930786, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7204269170761108, + "num_tokens": 363729198.0, + "step": 14062 + }, + { + "epoch": 1.5443663518559192, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.027193784713745, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7386263608932495, + "num_tokens": 363757236.0, + "step": 14063 + }, + { + "epoch": 1.544476169558533, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2138493061065674, + "learning_rate": 1e-06, + "loss": 1.0105, + "mean_token_accuracy": 0.6895579695701599, + "num_tokens": 363784799.0, + "step": 14064 + }, + { + "epoch": 1.5445859872611465, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4133822917938232, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7268297672271729, + "num_tokens": 363806497.0, + "step": 14065 + }, + { + "epoch": 1.54469580496376, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2147178649902344, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7013120651245117, + "num_tokens": 363833430.0, + "step": 14066 + }, + { + "epoch": 1.5448056226663738, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.434295892715454, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7094442844390869, + "num_tokens": 363856361.0, + "step": 14067 + }, + { + "epoch": 1.5449154403689875, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2713921070098877, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7273555994033813, + "num_tokens": 363881557.0, + "step": 14068 + }, + { + "epoch": 1.5450252580716013, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.645484447479248, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7167109251022339, + "num_tokens": 363901078.0, + "step": 14069 + }, + { + "epoch": 1.5451350757742148, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4115493297576904, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7421197891235352, + "num_tokens": 363923541.0, + "step": 14070 + }, + { + "epoch": 1.5452448934768284, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.328073263168335, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7172285318374634, + "num_tokens": 363947757.0, + "step": 14071 + }, + { + "epoch": 1.5453547111794421, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.526951313018799, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7085447311401367, + "num_tokens": 363969680.0, + "step": 14072 + }, + { + "epoch": 1.5454645288820559, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.257838726043701, + "learning_rate": 1e-06, + "loss": 0.9914, + "mean_token_accuracy": 0.6982697248458862, + "num_tokens": 363995485.0, + "step": 14073 + }, + { + "epoch": 1.5455743465846694, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.469581365585327, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7104389667510986, + "num_tokens": 364017312.0, + "step": 14074 + }, + { + "epoch": 1.545684164287283, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.146517276763916, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7155390381813049, + "num_tokens": 364043938.0, + "step": 14075 + }, + { + "epoch": 1.5457939819898967, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3315908908843994, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7383451461791992, + "num_tokens": 364066453.0, + "step": 14076 + }, + { + "epoch": 1.5459037996925105, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3496010303497314, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7014524936676025, + "num_tokens": 364089828.0, + "step": 14077 + }, + { + "epoch": 1.5460136173951242, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.212496280670166, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.7207542657852173, + "num_tokens": 364117049.0, + "step": 14078 + }, + { + "epoch": 1.5461234350977378, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.324160575866699, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.72332364320755, + "num_tokens": 364141653.0, + "step": 14079 + }, + { + "epoch": 1.5462332528003513, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2078394889831543, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7467794418334961, + "num_tokens": 364167485.0, + "step": 14080 + }, + { + "epoch": 1.546343070502965, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.8162081241607666, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7514828443527222, + "num_tokens": 364183903.0, + "step": 14081 + }, + { + "epoch": 1.5464528882055788, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.210653066635132, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7063519954681396, + "num_tokens": 364210196.0, + "step": 14082 + }, + { + "epoch": 1.5465627059081926, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.156471014022827, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7308878898620605, + "num_tokens": 364235182.0, + "step": 14083 + }, + { + "epoch": 1.546672523610806, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1910979747772217, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7174142599105835, + "num_tokens": 364260900.0, + "step": 14084 + }, + { + "epoch": 1.5467823413134196, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.223073959350586, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.69926917552948, + "num_tokens": 364288023.0, + "step": 14085 + }, + { + "epoch": 1.5468921590160334, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9467765092849731, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7158448696136475, + "num_tokens": 364318840.0, + "step": 14086 + }, + { + "epoch": 1.5470019767186471, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1369683742523193, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7176514863967896, + "num_tokens": 364346546.0, + "step": 14087 + }, + { + "epoch": 1.5471117944212607, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.6928915977478027, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7312157154083252, + "num_tokens": 364367853.0, + "step": 14088 + }, + { + "epoch": 1.5472216121238742, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1033055782318115, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7118511199951172, + "num_tokens": 364395855.0, + "step": 14089 + }, + { + "epoch": 1.547331429826488, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.305403709411621, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.698340892791748, + "num_tokens": 364421719.0, + "step": 14090 + }, + { + "epoch": 1.5474412475291017, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5451886653900146, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7350790500640869, + "num_tokens": 364444366.0, + "step": 14091 + }, + { + "epoch": 1.5475510652317155, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4176371097564697, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.7106227874755859, + "num_tokens": 364467061.0, + "step": 14092 + }, + { + "epoch": 1.547660882934329, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.431438446044922, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7307953834533691, + "num_tokens": 364488964.0, + "step": 14093 + }, + { + "epoch": 1.5477707006369426, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4181764125823975, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.715148389339447, + "num_tokens": 364511490.0, + "step": 14094 + }, + { + "epoch": 1.5478805183395563, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3616445064544678, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7227746248245239, + "num_tokens": 364537702.0, + "step": 14095 + }, + { + "epoch": 1.54799033604217, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0360162258148193, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7233120203018188, + "num_tokens": 364566065.0, + "step": 14096 + }, + { + "epoch": 1.5481001537447838, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.262296438217163, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7243260145187378, + "num_tokens": 364591670.0, + "step": 14097 + }, + { + "epoch": 1.5482099714473974, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.960947036743164, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7304908037185669, + "num_tokens": 364622290.0, + "step": 14098 + }, + { + "epoch": 1.548319789150011, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.483342409133911, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7241672277450562, + "num_tokens": 364643694.0, + "step": 14099 + }, + { + "epoch": 1.5484296068526247, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2058379650115967, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7139095067977905, + "num_tokens": 364671334.0, + "step": 14100 + }, + { + "epoch": 1.5485394245552384, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3558897972106934, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7151607275009155, + "num_tokens": 364694191.0, + "step": 14101 + }, + { + "epoch": 1.548649242257852, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.388545036315918, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7115720510482788, + "num_tokens": 364718287.0, + "step": 14102 + }, + { + "epoch": 1.5487590599604655, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.282413959503174, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7464755773544312, + "num_tokens": 364742007.0, + "step": 14103 + }, + { + "epoch": 1.5488688776630792, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0500433444976807, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.713466465473175, + "num_tokens": 364772320.0, + "step": 14104 + }, + { + "epoch": 1.548978695365693, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.415290117263794, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7227997779846191, + "num_tokens": 364794693.0, + "step": 14105 + }, + { + "epoch": 1.5490885130683067, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2390644550323486, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7133775353431702, + "num_tokens": 364821681.0, + "step": 14106 + }, + { + "epoch": 1.5491983307709203, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 1.9592212438583374, + "learning_rate": 1e-06, + "loss": 0.9959, + "mean_token_accuracy": 0.6920822858810425, + "num_tokens": 364854392.0, + "step": 14107 + }, + { + "epoch": 1.5493081484735338, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1977155208587646, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7231658697128296, + "num_tokens": 364882341.0, + "step": 14108 + }, + { + "epoch": 1.5494179661761476, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3370683193206787, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7120419144630432, + "num_tokens": 364905926.0, + "step": 14109 + }, + { + "epoch": 1.5495277838787613, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.063840866088867, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7038199305534363, + "num_tokens": 364934990.0, + "step": 14110 + }, + { + "epoch": 1.5496376015813749, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3286333084106445, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7097700834274292, + "num_tokens": 364960789.0, + "step": 14111 + }, + { + "epoch": 1.5497474192839886, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.0928847789764404, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7094016075134277, + "num_tokens": 364986802.0, + "step": 14112 + }, + { + "epoch": 1.5498572369866022, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2449915409088135, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7147151827812195, + "num_tokens": 365012297.0, + "step": 14113 + }, + { + "epoch": 1.549967054689216, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1992907524108887, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7335835099220276, + "num_tokens": 365037997.0, + "step": 14114 + }, + { + "epoch": 1.5500768723918297, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.633357048034668, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7497759461402893, + "num_tokens": 365058239.0, + "step": 14115 + }, + { + "epoch": 1.5501866900944432, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1778926849365234, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7063993215560913, + "num_tokens": 365086407.0, + "step": 14116 + }, + { + "epoch": 1.5502965077970567, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.450317859649658, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7380479574203491, + "num_tokens": 365109176.0, + "step": 14117 + }, + { + "epoch": 1.5504063254996705, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4532618522644043, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7231281399726868, + "num_tokens": 365131615.0, + "step": 14118 + }, + { + "epoch": 1.5505161432022843, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3252711296081543, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.7036274671554565, + "num_tokens": 365158257.0, + "step": 14119 + }, + { + "epoch": 1.550625960904898, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.766261339187622, + "learning_rate": 1e-06, + "loss": 0.7565, + "mean_token_accuracy": 0.7548996210098267, + "num_tokens": 365175178.0, + "step": 14120 + }, + { + "epoch": 1.5507357786075116, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4326999187469482, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7259010076522827, + "num_tokens": 365197933.0, + "step": 14121 + }, + { + "epoch": 1.550845596310125, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.051175117492676, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7058596611022949, + "num_tokens": 365233179.0, + "step": 14122 + }, + { + "epoch": 1.5509554140127388, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.132807970046997, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7103801369667053, + "num_tokens": 365259249.0, + "step": 14123 + }, + { + "epoch": 1.5510652317153526, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3314568996429443, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7217716574668884, + "num_tokens": 365282260.0, + "step": 14124 + }, + { + "epoch": 1.5511750494179661, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1792657375335693, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7182172536849976, + "num_tokens": 365308282.0, + "step": 14125 + }, + { + "epoch": 1.55128486712058, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.152362585067749, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7244284152984619, + "num_tokens": 365335380.0, + "step": 14126 + }, + { + "epoch": 1.5513946848231934, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2464771270751953, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.7156168222427368, + "num_tokens": 365360887.0, + "step": 14127 + }, + { + "epoch": 1.5515045025258072, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.5369808673858643, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7421457767486572, + "num_tokens": 365380668.0, + "step": 14128 + }, + { + "epoch": 1.551614320228421, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.105356454849243, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7290219068527222, + "num_tokens": 365407323.0, + "step": 14129 + }, + { + "epoch": 1.5517241379310345, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.691053628921509, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7411830425262451, + "num_tokens": 365426765.0, + "step": 14130 + }, + { + "epoch": 1.551833955633648, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1419289112091064, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7230892777442932, + "num_tokens": 365452730.0, + "step": 14131 + }, + { + "epoch": 1.5519437733362618, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.210692882537842, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7256217002868652, + "num_tokens": 365478368.0, + "step": 14132 + }, + { + "epoch": 1.5520535910388755, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3317248821258545, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7158116698265076, + "num_tokens": 365502634.0, + "step": 14133 + }, + { + "epoch": 1.5521634087414893, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.540261745452881, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7298249006271362, + "num_tokens": 365523409.0, + "step": 14134 + }, + { + "epoch": 1.5522732264441028, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1747052669525146, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7229981422424316, + "num_tokens": 365550306.0, + "step": 14135 + }, + { + "epoch": 1.5523830441467164, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.178151845932007, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7463896870613098, + "num_tokens": 365575175.0, + "step": 14136 + }, + { + "epoch": 1.55249286184933, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.7163612842559814, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7489231824874878, + "num_tokens": 365592589.0, + "step": 14137 + }, + { + "epoch": 1.5526026795519439, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.4362683296203613, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7282848358154297, + "num_tokens": 365615450.0, + "step": 14138 + }, + { + "epoch": 1.5527124972545574, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.1958909034729004, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7216455936431885, + "num_tokens": 365643137.0, + "step": 14139 + }, + { + "epoch": 1.552822314957171, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.097666025161743, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7235748171806335, + "num_tokens": 365670653.0, + "step": 14140 + }, + { + "epoch": 1.5529321326597847, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.092542886734009, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7003894448280334, + "num_tokens": 365699304.0, + "step": 14141 + }, + { + "epoch": 1.5530419503623984, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.155734062194824, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7030649781227112, + "num_tokens": 365727050.0, + "step": 14142 + }, + { + "epoch": 1.5531517680650122, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.3865904808044434, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7193577885627747, + "num_tokens": 365752438.0, + "step": 14143 + }, + { + "epoch": 1.5532615857676257, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.2920007705688477, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7176330089569092, + "num_tokens": 365775100.0, + "step": 14144 + }, + { + "epoch": 1.5533714034702393, + "ewc_loss": 7.987022399902344e-06, + "grad_norm": 2.149824857711792, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6887719631195068, + "num_tokens": 365804574.0, + "step": 14145 + }, + { + "epoch": 1.553481221172853, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.4932217597961426, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7028964757919312, + "num_tokens": 365827418.0, + "step": 14146 + }, + { + "epoch": 1.5535910388754668, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.4466073513031006, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7544920444488525, + "num_tokens": 365847419.0, + "step": 14147 + }, + { + "epoch": 1.5537008565780805, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.161165475845337, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7244836091995239, + "num_tokens": 365874118.0, + "step": 14148 + }, + { + "epoch": 1.553810674280694, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.0972917079925537, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7122093439102173, + "num_tokens": 365902006.0, + "step": 14149 + }, + { + "epoch": 1.5539204919833076, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.4646034240722656, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7208119630813599, + "num_tokens": 365922732.0, + "step": 14150 + }, + { + "epoch": 1.5540303096859214, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.177828311920166, + "learning_rate": 1e-06, + "loss": 0.9864, + "mean_token_accuracy": 0.696097731590271, + "num_tokens": 365951237.0, + "step": 14151 + }, + { + "epoch": 1.5541401273885351, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2370598316192627, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7226475477218628, + "num_tokens": 365976298.0, + "step": 14152 + }, + { + "epoch": 1.5542499450911487, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.1949641704559326, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.701906681060791, + "num_tokens": 366003615.0, + "step": 14153 + }, + { + "epoch": 1.5543597627937622, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.3458938598632812, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7160754799842834, + "num_tokens": 366028463.0, + "step": 14154 + }, + { + "epoch": 1.554469580496376, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2526886463165283, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7138869762420654, + "num_tokens": 366054708.0, + "step": 14155 + }, + { + "epoch": 1.5545793981989897, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.070523738861084, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7210149765014648, + "num_tokens": 366084130.0, + "step": 14156 + }, + { + "epoch": 1.5546892159016035, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.728386640548706, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7428388595581055, + "num_tokens": 366105798.0, + "step": 14157 + }, + { + "epoch": 1.554799033604217, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 1.9035720825195312, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7061281204223633, + "num_tokens": 366138830.0, + "step": 14158 + }, + { + "epoch": 1.5549088513068305, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2882513999938965, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7310894727706909, + "num_tokens": 366164338.0, + "step": 14159 + }, + { + "epoch": 1.5550186690094443, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.406555652618408, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7186337113380432, + "num_tokens": 366188310.0, + "step": 14160 + }, + { + "epoch": 1.555128486712058, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.1088085174560547, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7294864058494568, + "num_tokens": 366218560.0, + "step": 14161 + }, + { + "epoch": 1.5552383044146716, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.4867501258850098, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7265084981918335, + "num_tokens": 366239808.0, + "step": 14162 + }, + { + "epoch": 1.5553481221172853, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.2771949768066406, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7327317595481873, + "num_tokens": 366265356.0, + "step": 14163 + }, + { + "epoch": 1.5554579398198989, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.129542112350464, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7186210751533508, + "num_tokens": 366294312.0, + "step": 14164 + }, + { + "epoch": 1.5555677575225126, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.514728546142578, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7126380205154419, + "num_tokens": 366315345.0, + "step": 14165 + }, + { + "epoch": 1.5556775752251264, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.291794538497925, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7088750004768372, + "num_tokens": 366338596.0, + "step": 14166 + }, + { + "epoch": 1.55578739292774, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.417210817337036, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7458499073982239, + "num_tokens": 366360721.0, + "step": 14167 + }, + { + "epoch": 1.5558972106303535, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.434746265411377, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7344779372215271, + "num_tokens": 366381748.0, + "step": 14168 + }, + { + "epoch": 1.5560070283329672, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.185819625854492, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7240148782730103, + "num_tokens": 366408008.0, + "step": 14169 + }, + { + "epoch": 1.556116846035581, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.337440252304077, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7226848602294922, + "num_tokens": 366431701.0, + "step": 14170 + }, + { + "epoch": 1.5562266637381947, + "ewc_loss": 8.106231689453125e-06, + "grad_norm": 2.133465051651001, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7288608551025391, + "num_tokens": 366457522.0, + "step": 14171 + }, + { + "epoch": 1.5563364814408083, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.197611093521118, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7379800081253052, + "num_tokens": 366481980.0, + "step": 14172 + }, + { + "epoch": 1.5564462991434218, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4618351459503174, + "learning_rate": 1e-06, + "loss": 0.976, + "mean_token_accuracy": 0.6988399028778076, + "num_tokens": 366504694.0, + "step": 14173 + }, + { + "epoch": 1.5565561168460356, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.5850346088409424, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7188053727149963, + "num_tokens": 366526182.0, + "step": 14174 + }, + { + "epoch": 1.5566659345486493, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1151046752929688, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7291333675384521, + "num_tokens": 366555008.0, + "step": 14175 + }, + { + "epoch": 1.5567757522512629, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2408177852630615, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7243633270263672, + "num_tokens": 366580796.0, + "step": 14176 + }, + { + "epoch": 1.5568855699538766, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.298034429550171, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7135111689567566, + "num_tokens": 366606322.0, + "step": 14177 + }, + { + "epoch": 1.5569953876564901, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.350769281387329, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7218532562255859, + "num_tokens": 366629526.0, + "step": 14178 + }, + { + "epoch": 1.557105205359104, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4382522106170654, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7354632616043091, + "num_tokens": 366650121.0, + "step": 14179 + }, + { + "epoch": 1.5572150230617177, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4602253437042236, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7184938192367554, + "num_tokens": 366673256.0, + "step": 14180 + }, + { + "epoch": 1.5573248407643312, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0927467346191406, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7158182263374329, + "num_tokens": 366703456.0, + "step": 14181 + }, + { + "epoch": 1.5574346584669447, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.066096782684326, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7046835422515869, + "num_tokens": 366734352.0, + "step": 14182 + }, + { + "epoch": 1.5575444761695585, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.205571174621582, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7243279218673706, + "num_tokens": 366761205.0, + "step": 14183 + }, + { + "epoch": 1.5576542938721722, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.258692741394043, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7251767516136169, + "num_tokens": 366785165.0, + "step": 14184 + }, + { + "epoch": 1.557764111574786, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.205260992050171, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.724777340888977, + "num_tokens": 366809717.0, + "step": 14185 + }, + { + "epoch": 1.5578739292773995, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.23323392868042, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7476731538772583, + "num_tokens": 366832449.0, + "step": 14186 + }, + { + "epoch": 1.557983746980013, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3925838470458984, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7248750925064087, + "num_tokens": 366857295.0, + "step": 14187 + }, + { + "epoch": 1.5580935646826268, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1798958778381348, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.6923414468765259, + "num_tokens": 366885833.0, + "step": 14188 + }, + { + "epoch": 1.5582033823852406, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1002583503723145, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7140353322029114, + "num_tokens": 366912767.0, + "step": 14189 + }, + { + "epoch": 1.5583132000878541, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4102933406829834, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.717481791973114, + "num_tokens": 366935518.0, + "step": 14190 + }, + { + "epoch": 1.5584230177904677, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2252016067504883, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7229923009872437, + "num_tokens": 366960821.0, + "step": 14191 + }, + { + "epoch": 1.5585328354930814, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9074722528457642, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.6961561441421509, + "num_tokens": 366996506.0, + "step": 14192 + }, + { + "epoch": 1.5586426531956952, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3806512355804443, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7115104794502258, + "num_tokens": 367021721.0, + "step": 14193 + }, + { + "epoch": 1.558752470898309, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.273284435272217, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7290065884590149, + "num_tokens": 367047808.0, + "step": 14194 + }, + { + "epoch": 1.5588622886009225, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2910566329956055, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.7036361694335938, + "num_tokens": 367073388.0, + "step": 14195 + }, + { + "epoch": 1.558972106303536, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.215785264968872, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7050768733024597, + "num_tokens": 367101125.0, + "step": 14196 + }, + { + "epoch": 1.5590819240061498, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.399677276611328, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.7048937678337097, + "num_tokens": 367126849.0, + "step": 14197 + }, + { + "epoch": 1.5591917417087635, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.354591131210327, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7136104106903076, + "num_tokens": 367150035.0, + "step": 14198 + }, + { + "epoch": 1.5593015594113773, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3619790077209473, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.741623044013977, + "num_tokens": 367171590.0, + "step": 14199 + }, + { + "epoch": 1.5594113771139908, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9865074157714844, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7316232919692993, + "num_tokens": 367199356.0, + "step": 14200 + }, + { + "epoch": 1.5595211948166043, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0418906211853027, + "learning_rate": 1e-06, + "loss": 0.9717, + "mean_token_accuracy": 0.7018535137176514, + "num_tokens": 367228959.0, + "step": 14201 + }, + { + "epoch": 1.559631012519218, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0915935039520264, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7198163866996765, + "num_tokens": 367256230.0, + "step": 14202 + }, + { + "epoch": 1.5597408302218319, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0298285484313965, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7087286710739136, + "num_tokens": 367287741.0, + "step": 14203 + }, + { + "epoch": 1.5598506479244454, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3984317779541016, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7359048128128052, + "num_tokens": 367310656.0, + "step": 14204 + }, + { + "epoch": 1.559960465627059, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.188232898712158, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7226622104644775, + "num_tokens": 367337462.0, + "step": 14205 + }, + { + "epoch": 1.5600702833296727, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2670748233795166, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7296959757804871, + "num_tokens": 367363685.0, + "step": 14206 + }, + { + "epoch": 1.5601801010322864, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.5341272354125977, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7203586101531982, + "num_tokens": 367386120.0, + "step": 14207 + }, + { + "epoch": 1.5602899187349002, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.107377767562866, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6895250678062439, + "num_tokens": 367417056.0, + "step": 14208 + }, + { + "epoch": 1.5603997364375137, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1720378398895264, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7232475280761719, + "num_tokens": 367442342.0, + "step": 14209 + }, + { + "epoch": 1.5605095541401273, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9636300802230835, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7209098935127258, + "num_tokens": 367472690.0, + "step": 14210 + }, + { + "epoch": 1.560619371842741, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.276939868927002, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7107498049736023, + "num_tokens": 367498412.0, + "step": 14211 + }, + { + "epoch": 1.5607291895453548, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1125495433807373, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7345954179763794, + "num_tokens": 367525955.0, + "step": 14212 + }, + { + "epoch": 1.5608390072479685, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0154225826263428, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7352198362350464, + "num_tokens": 367554794.0, + "step": 14213 + }, + { + "epoch": 1.560948824950582, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0330309867858887, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7094767689704895, + "num_tokens": 367586005.0, + "step": 14214 + }, + { + "epoch": 1.5610586426531956, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.155780553817749, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7284570932388306, + "num_tokens": 367612133.0, + "step": 14215 + }, + { + "epoch": 1.5611684603558094, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4339208602905273, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7260905504226685, + "num_tokens": 367633180.0, + "step": 14216 + }, + { + "epoch": 1.5612782780584231, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.135708808898926, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7081213593482971, + "num_tokens": 367663191.0, + "step": 14217 + }, + { + "epoch": 1.5613880957610367, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2668368816375732, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7104684114456177, + "num_tokens": 367689478.0, + "step": 14218 + }, + { + "epoch": 1.5614979134636502, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1860873699188232, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7397711277008057, + "num_tokens": 367713709.0, + "step": 14219 + }, + { + "epoch": 1.561607731166264, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0229339599609375, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7026640176773071, + "num_tokens": 367744479.0, + "step": 14220 + }, + { + "epoch": 1.5617175488688777, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.6151368618011475, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7382930517196655, + "num_tokens": 367765403.0, + "step": 14221 + }, + { + "epoch": 1.5618273665714915, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1588993072509766, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7270032167434692, + "num_tokens": 367792333.0, + "step": 14222 + }, + { + "epoch": 1.561937184274105, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.171983480453491, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7263832688331604, + "num_tokens": 367817925.0, + "step": 14223 + }, + { + "epoch": 1.5620470019767185, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1135942935943604, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7219910025596619, + "num_tokens": 367846064.0, + "step": 14224 + }, + { + "epoch": 1.5621568196793323, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.408313274383545, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7333546280860901, + "num_tokens": 367871089.0, + "step": 14225 + }, + { + "epoch": 1.562266637381946, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.758099317550659, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7282828092575073, + "num_tokens": 367891372.0, + "step": 14226 + }, + { + "epoch": 1.5623764550845596, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3518173694610596, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7285970449447632, + "num_tokens": 367914436.0, + "step": 14227 + }, + { + "epoch": 1.5624862727871733, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.7260611057281494, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7496060729026794, + "num_tokens": 367931973.0, + "step": 14228 + }, + { + "epoch": 1.5625960904897869, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1746256351470947, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7254096269607544, + "num_tokens": 367957829.0, + "step": 14229 + }, + { + "epoch": 1.5627059081924006, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3706867694854736, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7004905343055725, + "num_tokens": 367981594.0, + "step": 14230 + }, + { + "epoch": 1.5628157258950144, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.006570339202881, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7221529483795166, + "num_tokens": 368012875.0, + "step": 14231 + }, + { + "epoch": 1.562925543597628, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2312891483306885, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7253013253211975, + "num_tokens": 368038115.0, + "step": 14232 + }, + { + "epoch": 1.5630353613002415, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.265129327774048, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6980660557746887, + "num_tokens": 368064082.0, + "step": 14233 + }, + { + "epoch": 1.5631451790028552, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.49277925491333, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7240968942642212, + "num_tokens": 368085281.0, + "step": 14234 + }, + { + "epoch": 1.563254996705469, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3300065994262695, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.737470269203186, + "num_tokens": 368108808.0, + "step": 14235 + }, + { + "epoch": 1.5633648144080827, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.532287359237671, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7580217123031616, + "num_tokens": 368127261.0, + "step": 14236 + }, + { + "epoch": 1.5634746321106963, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.099839448928833, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7458328008651733, + "num_tokens": 368152865.0, + "step": 14237 + }, + { + "epoch": 1.5635844498133098, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9626346826553345, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7065247297286987, + "num_tokens": 368183795.0, + "step": 14238 + }, + { + "epoch": 1.5636942675159236, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2477939128875732, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.707217812538147, + "num_tokens": 368209576.0, + "step": 14239 + }, + { + "epoch": 1.5638040852185373, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9930505752563477, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7239323854446411, + "num_tokens": 368240376.0, + "step": 14240 + }, + { + "epoch": 1.5639139029211508, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9811381101608276, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7228958606719971, + "num_tokens": 368269971.0, + "step": 14241 + }, + { + "epoch": 1.5640237206237646, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.824592113494873, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7315794229507446, + "num_tokens": 368288264.0, + "step": 14242 + }, + { + "epoch": 1.5641335383263781, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.285917043685913, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.730617344379425, + "num_tokens": 368313174.0, + "step": 14243 + }, + { + "epoch": 1.564243356028992, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.290745973587036, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7229167819023132, + "num_tokens": 368338117.0, + "step": 14244 + }, + { + "epoch": 1.5643531737316057, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1803643703460693, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7294667959213257, + "num_tokens": 368363618.0, + "step": 14245 + }, + { + "epoch": 1.5644629914342192, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.5765926837921143, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7007153034210205, + "num_tokens": 368384049.0, + "step": 14246 + }, + { + "epoch": 1.5645728091368327, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.565160036087036, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7403333783149719, + "num_tokens": 368404667.0, + "step": 14247 + }, + { + "epoch": 1.5646826268394465, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3165833950042725, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7228528261184692, + "num_tokens": 368429066.0, + "step": 14248 + }, + { + "epoch": 1.5647924445420602, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.283212661743164, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7079545855522156, + "num_tokens": 368455380.0, + "step": 14249 + }, + { + "epoch": 1.564902262244674, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.675598382949829, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7315741777420044, + "num_tokens": 368474136.0, + "step": 14250 + }, + { + "epoch": 1.5650120799472875, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9428433179855347, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7077846527099609, + "num_tokens": 368505615.0, + "step": 14251 + }, + { + "epoch": 1.565121897649901, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2978785037994385, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7054920792579651, + "num_tokens": 368530575.0, + "step": 14252 + }, + { + "epoch": 1.5652317153525148, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.5701587200164795, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7296657562255859, + "num_tokens": 368550127.0, + "step": 14253 + }, + { + "epoch": 1.5653415330551286, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.533510208129883, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7343825101852417, + "num_tokens": 368570110.0, + "step": 14254 + }, + { + "epoch": 1.565451350757742, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2345783710479736, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7549228668212891, + "num_tokens": 368593773.0, + "step": 14255 + }, + { + "epoch": 1.5655611684603556, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.251278877258301, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7391201257705688, + "num_tokens": 368620602.0, + "step": 14256 + }, + { + "epoch": 1.5656709861629694, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.374140739440918, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7299870848655701, + "num_tokens": 368645461.0, + "step": 14257 + }, + { + "epoch": 1.5657808038655832, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.187335968017578, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7058587670326233, + "num_tokens": 368671850.0, + "step": 14258 + }, + { + "epoch": 1.565890621568197, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.2199463844299316, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7287807464599609, + "num_tokens": 368697958.0, + "step": 14259 + }, + { + "epoch": 1.5660004392708105, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1378445625305176, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7137513160705566, + "num_tokens": 368725058.0, + "step": 14260 + }, + { + "epoch": 1.566110256973424, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1954784393310547, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7062020301818848, + "num_tokens": 368750045.0, + "step": 14261 + }, + { + "epoch": 1.5662200746760377, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.198270559310913, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7075445055961609, + "num_tokens": 368773970.0, + "step": 14262 + }, + { + "epoch": 1.5663298923786515, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3216450214385986, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7283048629760742, + "num_tokens": 368797353.0, + "step": 14263 + }, + { + "epoch": 1.5664397100812653, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.1899771690368652, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.720341682434082, + "num_tokens": 368822956.0, + "step": 14264 + }, + { + "epoch": 1.5665495277838788, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.6284451484680176, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7014701962471008, + "num_tokens": 368844391.0, + "step": 14265 + }, + { + "epoch": 1.5666593454864923, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.130603075027466, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7294144034385681, + "num_tokens": 368871579.0, + "step": 14266 + }, + { + "epoch": 1.566769163189106, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4435529708862305, + "learning_rate": 1e-06, + "loss": 0.7143, + "mean_token_accuracy": 0.7677375078201294, + "num_tokens": 368892953.0, + "step": 14267 + }, + { + "epoch": 1.5668789808917198, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.334960699081421, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7350038290023804, + "num_tokens": 368915633.0, + "step": 14268 + }, + { + "epoch": 1.5669887985943334, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.417858600616455, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7508749961853027, + "num_tokens": 368937851.0, + "step": 14269 + }, + { + "epoch": 1.567098616296947, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3952274322509766, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7298542261123657, + "num_tokens": 368959935.0, + "step": 14270 + }, + { + "epoch": 1.5672084339995607, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.3437068462371826, + "learning_rate": 1e-06, + "loss": 1.0044, + "mean_token_accuracy": 0.6967736482620239, + "num_tokens": 368983887.0, + "step": 14271 + }, + { + "epoch": 1.5673182517021744, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.4156320095062256, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7094212770462036, + "num_tokens": 369007119.0, + "step": 14272 + }, + { + "epoch": 1.5674280694047882, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.0064563751220703, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7160817980766296, + "num_tokens": 369038244.0, + "step": 14273 + }, + { + "epoch": 1.5675378871074017, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.236037492752075, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7205401659011841, + "num_tokens": 369064040.0, + "step": 14274 + }, + { + "epoch": 1.5676477048100153, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.871259093284607, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7252504825592041, + "num_tokens": 369099983.0, + "step": 14275 + }, + { + "epoch": 1.567757522512629, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.350337266921997, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7194359302520752, + "num_tokens": 369124905.0, + "step": 14276 + }, + { + "epoch": 1.5678673402152428, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.17189359664917, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7077430486679077, + "num_tokens": 369152553.0, + "step": 14277 + }, + { + "epoch": 1.5679771579178565, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 2.52036190032959, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7381528615951538, + "num_tokens": 369173755.0, + "step": 14278 + }, + { + "epoch": 1.56808697562047, + "ewc_loss": 8.225440979003906e-06, + "grad_norm": 1.9310318231582642, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7285028100013733, + "num_tokens": 369204835.0, + "step": 14279 + }, + { + "epoch": 1.5681967933230836, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.3887276649475098, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.7032285928726196, + "num_tokens": 369229517.0, + "step": 14280 + }, + { + "epoch": 1.5683066110256974, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.3399617671966553, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7088147401809692, + "num_tokens": 369251994.0, + "step": 14281 + }, + { + "epoch": 1.568416428728311, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.216477155685425, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7082220315933228, + "num_tokens": 369278656.0, + "step": 14282 + }, + { + "epoch": 1.5685262464309246, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.2642219066619873, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7123965620994568, + "num_tokens": 369307060.0, + "step": 14283 + }, + { + "epoch": 1.5686360641335382, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.0661368370056152, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7092427015304565, + "num_tokens": 369334937.0, + "step": 14284 + }, + { + "epoch": 1.568745881836152, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.41225528717041, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7206019759178162, + "num_tokens": 369359214.0, + "step": 14285 + }, + { + "epoch": 1.5688556995387657, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.1369552612304688, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7293725609779358, + "num_tokens": 369387553.0, + "step": 14286 + }, + { + "epoch": 1.5689655172413794, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.2990221977233887, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7251625061035156, + "num_tokens": 369411021.0, + "step": 14287 + }, + { + "epoch": 1.569075334943993, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.421152114868164, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7524855732917786, + "num_tokens": 369433815.0, + "step": 14288 + }, + { + "epoch": 1.5691851526466065, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.3816614151000977, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.723168134689331, + "num_tokens": 369456333.0, + "step": 14289 + }, + { + "epoch": 1.5692949703492203, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.3182225227355957, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7210314273834229, + "num_tokens": 369479927.0, + "step": 14290 + }, + { + "epoch": 1.569404788051834, + "ewc_loss": 8.285045623779297e-06, + "grad_norm": 2.104142904281616, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7114569544792175, + "num_tokens": 369509725.0, + "step": 14291 + }, + { + "epoch": 1.5695146057544476, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2731781005859375, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.727627158164978, + "num_tokens": 369533129.0, + "step": 14292 + }, + { + "epoch": 1.5696244234570613, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.890356421470642, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7060105204582214, + "num_tokens": 369566005.0, + "step": 14293 + }, + { + "epoch": 1.5697342411596749, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1917836666107178, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7066913843154907, + "num_tokens": 369594485.0, + "step": 14294 + }, + { + "epoch": 1.5698440588622886, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.522664785385132, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7266029119491577, + "num_tokens": 369616813.0, + "step": 14295 + }, + { + "epoch": 1.5699538765649024, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.31758713722229, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7112553715705872, + "num_tokens": 369644035.0, + "step": 14296 + }, + { + "epoch": 1.570063694267516, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.238137722015381, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.716096043586731, + "num_tokens": 369671675.0, + "step": 14297 + }, + { + "epoch": 1.5701735119701294, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.224614381790161, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7371630072593689, + "num_tokens": 369697141.0, + "step": 14298 + }, + { + "epoch": 1.5702833296727432, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1972382068634033, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.721424400806427, + "num_tokens": 369723377.0, + "step": 14299 + }, + { + "epoch": 1.570393147375357, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4860141277313232, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7157745361328125, + "num_tokens": 369746629.0, + "step": 14300 + }, + { + "epoch": 1.5705029650779707, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.211122751235962, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7256988286972046, + "num_tokens": 369772301.0, + "step": 14301 + }, + { + "epoch": 1.5706127827805842, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4391555786132812, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7209672331809998, + "num_tokens": 369794383.0, + "step": 14302 + }, + { + "epoch": 1.5707226004831978, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.293208360671997, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7233339548110962, + "num_tokens": 369820624.0, + "step": 14303 + }, + { + "epoch": 1.5708324181858115, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1994123458862305, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7112002372741699, + "num_tokens": 369847660.0, + "step": 14304 + }, + { + "epoch": 1.5709422358884253, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4394330978393555, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7269021272659302, + "num_tokens": 369869638.0, + "step": 14305 + }, + { + "epoch": 1.5710520535910388, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1194872856140137, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7457301616668701, + "num_tokens": 369894309.0, + "step": 14306 + }, + { + "epoch": 1.5711618712936526, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.223106622695923, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7232934236526489, + "num_tokens": 369921955.0, + "step": 14307 + }, + { + "epoch": 1.5712716889962661, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2249362468719482, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7161561250686646, + "num_tokens": 369947219.0, + "step": 14308 + }, + { + "epoch": 1.5713815066988799, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3127214908599854, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7467008829116821, + "num_tokens": 369972667.0, + "step": 14309 + }, + { + "epoch": 1.5714913244014936, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2878406047821045, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7271760702133179, + "num_tokens": 369998788.0, + "step": 14310 + }, + { + "epoch": 1.5716011421041072, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9377087354660034, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7565143704414368, + "num_tokens": 370029169.0, + "step": 14311 + }, + { + "epoch": 1.5717109598067207, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2902913093566895, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7198951840400696, + "num_tokens": 370053808.0, + "step": 14312 + }, + { + "epoch": 1.5718207775093345, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.294442653656006, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7192035913467407, + "num_tokens": 370078890.0, + "step": 14313 + }, + { + "epoch": 1.5719305952119482, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3380093574523926, + "learning_rate": 1e-06, + "loss": 1.0403, + "mean_token_accuracy": 0.6848560571670532, + "num_tokens": 370104070.0, + "step": 14314 + }, + { + "epoch": 1.572040412914562, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2625725269317627, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7288126945495605, + "num_tokens": 370128170.0, + "step": 14315 + }, + { + "epoch": 1.5721502306171755, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.974272608757019, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7207857370376587, + "num_tokens": 370159853.0, + "step": 14316 + }, + { + "epoch": 1.572260048319789, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.049971580505371, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7266914248466492, + "num_tokens": 370188294.0, + "step": 14317 + }, + { + "epoch": 1.5723698660224028, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.27917218208313, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6985732913017273, + "num_tokens": 370213186.0, + "step": 14318 + }, + { + "epoch": 1.5724796837250166, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.035640001296997, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7252488136291504, + "num_tokens": 370242167.0, + "step": 14319 + }, + { + "epoch": 1.57258950142763, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9447020292282104, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.721980094909668, + "num_tokens": 370275462.0, + "step": 14320 + }, + { + "epoch": 1.5726993191302436, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.170419931411743, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7208162546157837, + "num_tokens": 370301238.0, + "step": 14321 + }, + { + "epoch": 1.5728091368328574, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.399794101715088, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.726409912109375, + "num_tokens": 370322420.0, + "step": 14322 + }, + { + "epoch": 1.5729189545354711, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2574803829193115, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.692106306552887, + "num_tokens": 370350201.0, + "step": 14323 + }, + { + "epoch": 1.573028772238085, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.21501088142395, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6985183358192444, + "num_tokens": 370379132.0, + "step": 14324 + }, + { + "epoch": 1.5731385899406984, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.31532621383667, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.6989258527755737, + "num_tokens": 370403156.0, + "step": 14325 + }, + { + "epoch": 1.573248407643312, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3797550201416016, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7257295846939087, + "num_tokens": 370425779.0, + "step": 14326 + }, + { + "epoch": 1.5733582253459257, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.096503973007202, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7209333777427673, + "num_tokens": 370456410.0, + "step": 14327 + }, + { + "epoch": 1.5734680430485395, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1738173961639404, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7353769540786743, + "num_tokens": 370482807.0, + "step": 14328 + }, + { + "epoch": 1.5735778607511532, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.31950306892395, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7229629755020142, + "num_tokens": 370504454.0, + "step": 14329 + }, + { + "epoch": 1.5736876784537668, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.124946117401123, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7071327567100525, + "num_tokens": 370531949.0, + "step": 14330 + }, + { + "epoch": 1.5737974961563803, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.16890811920166, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.733346700668335, + "num_tokens": 370557677.0, + "step": 14331 + }, + { + "epoch": 1.573907313858994, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3045480251312256, + "learning_rate": 1e-06, + "loss": 0.9996, + "mean_token_accuracy": 0.7002238035202026, + "num_tokens": 370584060.0, + "step": 14332 + }, + { + "epoch": 1.5740171315616078, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1939430236816406, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7280073761940002, + "num_tokens": 370608927.0, + "step": 14333 + }, + { + "epoch": 1.5741269492642214, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1959023475646973, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7239705920219421, + "num_tokens": 370634430.0, + "step": 14334 + }, + { + "epoch": 1.574236766966835, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.402675151824951, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7304354310035706, + "num_tokens": 370655568.0, + "step": 14335 + }, + { + "epoch": 1.5743465846694487, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.267385721206665, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7131920456886292, + "num_tokens": 370681602.0, + "step": 14336 + }, + { + "epoch": 1.5744564023720624, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.929315209388733, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7198096513748169, + "num_tokens": 370713455.0, + "step": 14337 + }, + { + "epoch": 1.5745662200746762, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2867343425750732, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7170993089675903, + "num_tokens": 370739938.0, + "step": 14338 + }, + { + "epoch": 1.5746760377772897, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3355348110198975, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.7171881198883057, + "num_tokens": 370763405.0, + "step": 14339 + }, + { + "epoch": 1.5747858554799032, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0252861976623535, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6882543563842773, + "num_tokens": 370795013.0, + "step": 14340 + }, + { + "epoch": 1.574895673182517, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.063668727874756, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7176840305328369, + "num_tokens": 370821450.0, + "step": 14341 + }, + { + "epoch": 1.5750054908851308, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.296191692352295, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7213311195373535, + "num_tokens": 370846368.0, + "step": 14342 + }, + { + "epoch": 1.5751153085877443, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.216820001602173, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7274659276008606, + "num_tokens": 370872383.0, + "step": 14343 + }, + { + "epoch": 1.575225126290358, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2203550338745117, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7031829953193665, + "num_tokens": 370901133.0, + "step": 14344 + }, + { + "epoch": 1.5753349439929716, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.364043712615967, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7007806301116943, + "num_tokens": 370926893.0, + "step": 14345 + }, + { + "epoch": 1.5754447616955853, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2456119060516357, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7263942956924438, + "num_tokens": 370953501.0, + "step": 14346 + }, + { + "epoch": 1.575554579398199, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0274465084075928, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7293559312820435, + "num_tokens": 370983747.0, + "step": 14347 + }, + { + "epoch": 1.5756643971008126, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.036644697189331, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7169323563575745, + "num_tokens": 371012900.0, + "step": 14348 + }, + { + "epoch": 1.5757742148034262, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3108763694763184, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7182345390319824, + "num_tokens": 371037220.0, + "step": 14349 + }, + { + "epoch": 1.57588403250604, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4416005611419678, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.713810384273529, + "num_tokens": 371059267.0, + "step": 14350 + }, + { + "epoch": 1.5759938502086537, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0560288429260254, + "learning_rate": 1e-06, + "loss": 1.0453, + "mean_token_accuracy": 0.6878465414047241, + "num_tokens": 371090398.0, + "step": 14351 + }, + { + "epoch": 1.5761036679112674, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.160562515258789, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.7019670009613037, + "num_tokens": 371119403.0, + "step": 14352 + }, + { + "epoch": 1.576213485613881, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9188741445541382, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7082422971725464, + "num_tokens": 371154075.0, + "step": 14353 + }, + { + "epoch": 1.5763233033164945, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.449507713317871, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.721847414970398, + "num_tokens": 371176912.0, + "step": 14354 + }, + { + "epoch": 1.5764331210191083, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1942086219787598, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7133743762969971, + "num_tokens": 371202272.0, + "step": 14355 + }, + { + "epoch": 1.576542938721722, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.274885892868042, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7199604511260986, + "num_tokens": 371227268.0, + "step": 14356 + }, + { + "epoch": 1.5766527564243356, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.386359691619873, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.73273766040802, + "num_tokens": 371252192.0, + "step": 14357 + }, + { + "epoch": 1.5767625741269493, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3973989486694336, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7242124676704407, + "num_tokens": 371274291.0, + "step": 14358 + }, + { + "epoch": 1.5768723918295628, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.044666290283203, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.6986629366874695, + "num_tokens": 371304575.0, + "step": 14359 + }, + { + "epoch": 1.5769822095321766, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.074451208114624, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7302521467208862, + "num_tokens": 371333681.0, + "step": 14360 + }, + { + "epoch": 1.5770920272347904, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0334436893463135, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7129881381988525, + "num_tokens": 371367450.0, + "step": 14361 + }, + { + "epoch": 1.577201844937404, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1762523651123047, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7146710753440857, + "num_tokens": 371393930.0, + "step": 14362 + }, + { + "epoch": 1.5773116626400174, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4333622455596924, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7194396257400513, + "num_tokens": 371415346.0, + "step": 14363 + }, + { + "epoch": 1.5774214803426312, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4457736015319824, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7062177658081055, + "num_tokens": 371436946.0, + "step": 14364 + }, + { + "epoch": 1.577531298045245, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1070873737335205, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7133466601371765, + "num_tokens": 371466547.0, + "step": 14365 + }, + { + "epoch": 1.5776411157478587, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.979921579360962, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7274901270866394, + "num_tokens": 371498664.0, + "step": 14366 + }, + { + "epoch": 1.5777509334504722, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.081821918487549, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.725853681564331, + "num_tokens": 371526832.0, + "step": 14367 + }, + { + "epoch": 1.5778607511530858, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0208582878112793, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.722896158695221, + "num_tokens": 371556697.0, + "step": 14368 + }, + { + "epoch": 1.5779705688556995, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9294425249099731, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7036124467849731, + "num_tokens": 371587747.0, + "step": 14369 + }, + { + "epoch": 1.5780803865583133, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3004772663116455, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7168252468109131, + "num_tokens": 371611011.0, + "step": 14370 + }, + { + "epoch": 1.5781902042609268, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2193844318389893, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7152137160301208, + "num_tokens": 371636310.0, + "step": 14371 + }, + { + "epoch": 1.5783000219635406, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0862889289855957, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7125835418701172, + "num_tokens": 371664235.0, + "step": 14372 + }, + { + "epoch": 1.5784098396661541, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.336081027984619, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7147209644317627, + "num_tokens": 371689834.0, + "step": 14373 + }, + { + "epoch": 1.5785196573687679, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.8902255296707153, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7068922519683838, + "num_tokens": 371725226.0, + "step": 14374 + }, + { + "epoch": 1.5786294750713816, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4166815280914307, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7412674427032471, + "num_tokens": 371748342.0, + "step": 14375 + }, + { + "epoch": 1.5787392927739952, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2418339252471924, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7211090922355652, + "num_tokens": 371771183.0, + "step": 14376 + }, + { + "epoch": 1.5788491104766087, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2942216396331787, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7156062126159668, + "num_tokens": 371797106.0, + "step": 14377 + }, + { + "epoch": 1.5789589281792225, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.229865074157715, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.7041852474212646, + "num_tokens": 371824608.0, + "step": 14378 + }, + { + "epoch": 1.5790687458818362, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3267483711242676, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7154261469841003, + "num_tokens": 371849493.0, + "step": 14379 + }, + { + "epoch": 1.57917856358445, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.057765007019043, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.726067304611206, + "num_tokens": 371877672.0, + "step": 14380 + }, + { + "epoch": 1.5792883812870635, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0995194911956787, + "learning_rate": 1e-06, + "loss": 1.0216, + "mean_token_accuracy": 0.6836323738098145, + "num_tokens": 371909281.0, + "step": 14381 + }, + { + "epoch": 1.579398198989677, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.410353183746338, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7133474946022034, + "num_tokens": 371932126.0, + "step": 14382 + }, + { + "epoch": 1.5795080166922908, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5687341690063477, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.74242103099823, + "num_tokens": 371952465.0, + "step": 14383 + }, + { + "epoch": 1.5796178343949046, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.787419319152832, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7384741306304932, + "num_tokens": 371971853.0, + "step": 14384 + }, + { + "epoch": 1.579727652097518, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.205644130706787, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6968587636947632, + "num_tokens": 371999889.0, + "step": 14385 + }, + { + "epoch": 1.5798374698001316, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.002453565597534, + "learning_rate": 1e-06, + "loss": 0.9655, + "mean_token_accuracy": 0.7028272747993469, + "num_tokens": 372031413.0, + "step": 14386 + }, + { + "epoch": 1.5799472875027454, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2322752475738525, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7297201156616211, + "num_tokens": 372055163.0, + "step": 14387 + }, + { + "epoch": 1.5800571052053591, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2187159061431885, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.708118200302124, + "num_tokens": 372081104.0, + "step": 14388 + }, + { + "epoch": 1.580166922907973, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.389545440673828, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7288186550140381, + "num_tokens": 372105152.0, + "step": 14389 + }, + { + "epoch": 1.5802767406105864, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1580448150634766, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7139636874198914, + "num_tokens": 372134673.0, + "step": 14390 + }, + { + "epoch": 1.5803865583132, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3988189697265625, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7205816507339478, + "num_tokens": 372157655.0, + "step": 14391 + }, + { + "epoch": 1.5804963760158137, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1412246227264404, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.732581615447998, + "num_tokens": 372181746.0, + "step": 14392 + }, + { + "epoch": 1.5806061937184275, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2096948623657227, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7047492265701294, + "num_tokens": 372209356.0, + "step": 14393 + }, + { + "epoch": 1.5807160114210412, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3910512924194336, + "learning_rate": 1e-06, + "loss": 1.0092, + "mean_token_accuracy": 0.6965005397796631, + "num_tokens": 372233574.0, + "step": 14394 + }, + { + "epoch": 1.5808258291236548, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2517952919006348, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7040038108825684, + "num_tokens": 372261343.0, + "step": 14395 + }, + { + "epoch": 1.5809356468262683, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3635809421539307, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7175987958908081, + "num_tokens": 372284609.0, + "step": 14396 + }, + { + "epoch": 1.581045464528882, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2371034622192383, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.725318193435669, + "num_tokens": 372309407.0, + "step": 14397 + }, + { + "epoch": 1.5811552822314958, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.367669105529785, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7313631772994995, + "num_tokens": 372330438.0, + "step": 14398 + }, + { + "epoch": 1.5812650999341094, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.286403179168701, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.72434401512146, + "num_tokens": 372358343.0, + "step": 14399 + }, + { + "epoch": 1.5813749176367229, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5505757331848145, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7456503510475159, + "num_tokens": 372377730.0, + "step": 14400 + }, + { + "epoch": 1.5814847353393366, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9868990182876587, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7064905762672424, + "num_tokens": 372410573.0, + "step": 14401 + }, + { + "epoch": 1.5815945530419504, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1241707801818848, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.7002976536750793, + "num_tokens": 372438940.0, + "step": 14402 + }, + { + "epoch": 1.5817043707445642, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3221776485443115, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7096235156059265, + "num_tokens": 372462985.0, + "step": 14403 + }, + { + "epoch": 1.5818141884471777, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3298568725585938, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7161931395530701, + "num_tokens": 372488502.0, + "step": 14404 + }, + { + "epoch": 1.5819240061497912, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2725956439971924, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7118447422981262, + "num_tokens": 372514893.0, + "step": 14405 + }, + { + "epoch": 1.582033823852405, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3833675384521484, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7161048054695129, + "num_tokens": 372538957.0, + "step": 14406 + }, + { + "epoch": 1.5821436415550187, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.049767255783081, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7051894664764404, + "num_tokens": 372568819.0, + "step": 14407 + }, + { + "epoch": 1.5822534592576323, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.175736904144287, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.6961371898651123, + "num_tokens": 372599091.0, + "step": 14408 + }, + { + "epoch": 1.582363276960246, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2250123023986816, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7407841682434082, + "num_tokens": 372627719.0, + "step": 14409 + }, + { + "epoch": 1.5824730946628596, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4958629608154297, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7051236629486084, + "num_tokens": 372650903.0, + "step": 14410 + }, + { + "epoch": 1.5825829123654733, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.267660617828369, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7151362299919128, + "num_tokens": 372678165.0, + "step": 14411 + }, + { + "epoch": 1.582692730068087, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9091013669967651, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7047759294509888, + "num_tokens": 372713150.0, + "step": 14412 + }, + { + "epoch": 1.5828025477707006, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1861484050750732, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7114571928977966, + "num_tokens": 372739846.0, + "step": 14413 + }, + { + "epoch": 1.5829123654733142, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.249650001525879, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7103927135467529, + "num_tokens": 372765815.0, + "step": 14414 + }, + { + "epoch": 1.583022183175928, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0301854610443115, + "learning_rate": 1e-06, + "loss": 0.9599, + "mean_token_accuracy": 0.6995340585708618, + "num_tokens": 372797191.0, + "step": 14415 + }, + { + "epoch": 1.5831320008785417, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.132394313812256, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.727292537689209, + "num_tokens": 372824195.0, + "step": 14416 + }, + { + "epoch": 1.5832418185811554, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.390787363052368, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.716143786907196, + "num_tokens": 372847841.0, + "step": 14417 + }, + { + "epoch": 1.583351636283769, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5301666259765625, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7270873188972473, + "num_tokens": 372867002.0, + "step": 14418 + }, + { + "epoch": 1.5834614539863825, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2687747478485107, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7144219875335693, + "num_tokens": 372892138.0, + "step": 14419 + }, + { + "epoch": 1.5835712716889963, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1240932941436768, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7153178453445435, + "num_tokens": 372917333.0, + "step": 14420 + }, + { + "epoch": 1.58368108939161, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2566280364990234, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7286642789840698, + "num_tokens": 372941779.0, + "step": 14421 + }, + { + "epoch": 1.5837909070942235, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4550681114196777, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7190235257148743, + "num_tokens": 372962844.0, + "step": 14422 + }, + { + "epoch": 1.5839007247968373, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.153475522994995, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7070845365524292, + "num_tokens": 372989413.0, + "step": 14423 + }, + { + "epoch": 1.5840105424994508, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.343756675720215, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7230193018913269, + "num_tokens": 373014865.0, + "step": 14424 + }, + { + "epoch": 1.5841203602020646, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1406407356262207, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7049449682235718, + "num_tokens": 373043360.0, + "step": 14425 + }, + { + "epoch": 1.5842301779046783, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0823171138763428, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7139931917190552, + "num_tokens": 373071402.0, + "step": 14426 + }, + { + "epoch": 1.5843399956072919, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5085549354553223, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7165847420692444, + "num_tokens": 373092339.0, + "step": 14427 + }, + { + "epoch": 1.5844498133099054, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5035362243652344, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7134880423545837, + "num_tokens": 373113931.0, + "step": 14428 + }, + { + "epoch": 1.5845596310125192, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.334822654724121, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7326890230178833, + "num_tokens": 373137454.0, + "step": 14429 + }, + { + "epoch": 1.584669448715133, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.175018548965454, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7020108103752136, + "num_tokens": 373163741.0, + "step": 14430 + }, + { + "epoch": 1.5847792664177467, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1488900184631348, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7136034965515137, + "num_tokens": 373191778.0, + "step": 14431 + }, + { + "epoch": 1.5848890841203602, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.340070962905884, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7109710574150085, + "num_tokens": 373215503.0, + "step": 14432 + }, + { + "epoch": 1.5849989018229738, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.372302293777466, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7207205295562744, + "num_tokens": 373239727.0, + "step": 14433 + }, + { + "epoch": 1.5851087195255875, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1826069355010986, + "learning_rate": 1e-06, + "loss": 1.005, + "mean_token_accuracy": 0.6906956434249878, + "num_tokens": 373268032.0, + "step": 14434 + }, + { + "epoch": 1.5852185372282013, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1517558097839355, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7149009108543396, + "num_tokens": 373295587.0, + "step": 14435 + }, + { + "epoch": 1.5853283549308148, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5449647903442383, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7132546901702881, + "num_tokens": 373316872.0, + "step": 14436 + }, + { + "epoch": 1.5854381726334283, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4162425994873047, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7383893728256226, + "num_tokens": 373337786.0, + "step": 14437 + }, + { + "epoch": 1.585547990336042, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4921021461486816, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7200631499290466, + "num_tokens": 373359915.0, + "step": 14438 + }, + { + "epoch": 1.5856578080386559, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3244917392730713, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7207369208335876, + "num_tokens": 373384787.0, + "step": 14439 + }, + { + "epoch": 1.5857676257412696, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.122386932373047, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7239951491355896, + "num_tokens": 373414612.0, + "step": 14440 + }, + { + "epoch": 1.5858774434438832, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3229832649230957, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7196758985519409, + "num_tokens": 373439049.0, + "step": 14441 + }, + { + "epoch": 1.5859872611464967, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 1.9824137687683105, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6903354525566101, + "num_tokens": 373474120.0, + "step": 14442 + }, + { + "epoch": 1.5860970788491104, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3370344638824463, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7415158152580261, + "num_tokens": 373497659.0, + "step": 14443 + }, + { + "epoch": 1.5862068965517242, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2157323360443115, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7369666695594788, + "num_tokens": 373522404.0, + "step": 14444 + }, + { + "epoch": 1.586316714254338, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.160081624984741, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7355179786682129, + "num_tokens": 373549575.0, + "step": 14445 + }, + { + "epoch": 1.5864265319569515, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4072275161743164, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7419518828392029, + "num_tokens": 373572238.0, + "step": 14446 + }, + { + "epoch": 1.586536349659565, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3613665103912354, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.718871533870697, + "num_tokens": 373597048.0, + "step": 14447 + }, + { + "epoch": 1.5866461673621788, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3263349533081055, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7199352383613586, + "num_tokens": 373620945.0, + "step": 14448 + }, + { + "epoch": 1.5867559850647925, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.41267991065979, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7143398523330688, + "num_tokens": 373644170.0, + "step": 14449 + }, + { + "epoch": 1.586865802767406, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.51196551322937, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7152727842330933, + "num_tokens": 373665553.0, + "step": 14450 + }, + { + "epoch": 1.5869756204700196, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.256439685821533, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.712514340877533, + "num_tokens": 373693133.0, + "step": 14451 + }, + { + "epoch": 1.5870854381726334, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3169143199920654, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7195930480957031, + "num_tokens": 373716781.0, + "step": 14452 + }, + { + "epoch": 1.5871952558752471, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.371056079864502, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7225559949874878, + "num_tokens": 373742232.0, + "step": 14453 + }, + { + "epoch": 1.5873050735778609, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.204514980316162, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7169118523597717, + "num_tokens": 373768072.0, + "step": 14454 + }, + { + "epoch": 1.5874148912804744, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1729202270507812, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7241774797439575, + "num_tokens": 373794979.0, + "step": 14455 + }, + { + "epoch": 1.587524708983088, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.374462366104126, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7265727519989014, + "num_tokens": 373817546.0, + "step": 14456 + }, + { + "epoch": 1.5876345266857017, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.122054100036621, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7202589511871338, + "num_tokens": 373845958.0, + "step": 14457 + }, + { + "epoch": 1.5877443443883155, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3988184928894043, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7317204475402832, + "num_tokens": 373869453.0, + "step": 14458 + }, + { + "epoch": 1.5878541620909292, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.240305185317993, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7166976928710938, + "num_tokens": 373894136.0, + "step": 14459 + }, + { + "epoch": 1.5879639797935428, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5261406898498535, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7148292064666748, + "num_tokens": 373914939.0, + "step": 14460 + }, + { + "epoch": 1.5880737974961563, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.13870906829834, + "learning_rate": 1e-06, + "loss": 0.991, + "mean_token_accuracy": 0.6930817365646362, + "num_tokens": 373942048.0, + "step": 14461 + }, + { + "epoch": 1.58818361519877, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1289188861846924, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7195433974266052, + "num_tokens": 373971014.0, + "step": 14462 + }, + { + "epoch": 1.5882934329013838, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2607834339141846, + "learning_rate": 1e-06, + "loss": 0.7582, + "mean_token_accuracy": 0.7605589628219604, + "num_tokens": 373994789.0, + "step": 14463 + }, + { + "epoch": 1.5884032506039973, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.344205856323242, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7282553911209106, + "num_tokens": 374016747.0, + "step": 14464 + }, + { + "epoch": 1.5885130683066109, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4871573448181152, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7302659749984741, + "num_tokens": 374039477.0, + "step": 14465 + }, + { + "epoch": 1.5886228860092246, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1039834022521973, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7100991606712341, + "num_tokens": 374068549.0, + "step": 14466 + }, + { + "epoch": 1.5887327037118384, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.5032737255096436, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7325572371482849, + "num_tokens": 374090115.0, + "step": 14467 + }, + { + "epoch": 1.5888425214144521, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2755961418151855, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7093269228935242, + "num_tokens": 374114433.0, + "step": 14468 + }, + { + "epoch": 1.5889523391170657, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1517269611358643, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7135813236236572, + "num_tokens": 374144386.0, + "step": 14469 + }, + { + "epoch": 1.5890621568196792, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.3926188945770264, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7339068651199341, + "num_tokens": 374166726.0, + "step": 14470 + }, + { + "epoch": 1.589171974522293, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.038252830505371, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7147972583770752, + "num_tokens": 374197369.0, + "step": 14471 + }, + { + "epoch": 1.5892817922249067, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.388747215270996, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7141230702400208, + "num_tokens": 374220021.0, + "step": 14472 + }, + { + "epoch": 1.5893916099275203, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.0221686363220215, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7054345011711121, + "num_tokens": 374252238.0, + "step": 14473 + }, + { + "epoch": 1.589501427630134, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1782729625701904, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7157005071640015, + "num_tokens": 374279155.0, + "step": 14474 + }, + { + "epoch": 1.5896112453327476, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.1229183673858643, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7194052934646606, + "num_tokens": 374308634.0, + "step": 14475 + }, + { + "epoch": 1.5897210630353613, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.407731533050537, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7207140326499939, + "num_tokens": 374331359.0, + "step": 14476 + }, + { + "epoch": 1.589830880737975, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.4973886013031006, + "learning_rate": 1e-06, + "loss": 0.9604, + "mean_token_accuracy": 0.697107195854187, + "num_tokens": 374355995.0, + "step": 14477 + }, + { + "epoch": 1.5899406984405886, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.430206775665283, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7225381731987, + "num_tokens": 374377534.0, + "step": 14478 + }, + { + "epoch": 1.5900505161432021, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.2225160598754883, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7235934138298035, + "num_tokens": 374403849.0, + "step": 14479 + }, + { + "epoch": 1.590160333845816, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.23506236076355, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7363202571868896, + "num_tokens": 374429684.0, + "step": 14480 + }, + { + "epoch": 1.5902701515484297, + "ewc_loss": 8.344650268554688e-06, + "grad_norm": 2.185678243637085, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7278646230697632, + "num_tokens": 374455969.0, + "step": 14481 + }, + { + "epoch": 1.5903799692510434, + "ewc_loss": 8.404254913330078e-06, + "grad_norm": 2.266536235809326, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7215040922164917, + "num_tokens": 374481316.0, + "step": 14482 + }, + { + "epoch": 1.590489786953657, + "ewc_loss": 8.404254913330078e-06, + "grad_norm": 2.207387685775757, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7049087285995483, + "num_tokens": 374505819.0, + "step": 14483 + }, + { + "epoch": 1.5905996046562705, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.150653600692749, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.755787193775177, + "num_tokens": 374529943.0, + "step": 14484 + }, + { + "epoch": 1.5907094223588842, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.6491966247558594, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.728155791759491, + "num_tokens": 374551720.0, + "step": 14485 + }, + { + "epoch": 1.590819240061498, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0740630626678467, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7123651504516602, + "num_tokens": 374580820.0, + "step": 14486 + }, + { + "epoch": 1.5909290577641115, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4575822353363037, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7270724773406982, + "num_tokens": 374603548.0, + "step": 14487 + }, + { + "epoch": 1.5910388754667253, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2434604167938232, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7106397151947021, + "num_tokens": 374629400.0, + "step": 14488 + }, + { + "epoch": 1.5911486931693388, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3626840114593506, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7185847163200378, + "num_tokens": 374651432.0, + "step": 14489 + }, + { + "epoch": 1.5912585108719526, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9327812194824219, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7132527828216553, + "num_tokens": 374684310.0, + "step": 14490 + }, + { + "epoch": 1.5913683285745663, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2537262439727783, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7069951295852661, + "num_tokens": 374709123.0, + "step": 14491 + }, + { + "epoch": 1.5914781462771799, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2292463779449463, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7186535596847534, + "num_tokens": 374737090.0, + "step": 14492 + }, + { + "epoch": 1.5915879639797934, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.240313768386841, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7192124724388123, + "num_tokens": 374762696.0, + "step": 14493 + }, + { + "epoch": 1.5916977816824072, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3582961559295654, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7259291410446167, + "num_tokens": 374787539.0, + "step": 14494 + }, + { + "epoch": 1.591807599385021, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1087734699249268, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7091073989868164, + "num_tokens": 374815548.0, + "step": 14495 + }, + { + "epoch": 1.5919174170876347, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1803362369537354, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7079140543937683, + "num_tokens": 374842855.0, + "step": 14496 + }, + { + "epoch": 1.5920272347902482, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9573794603347778, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7115573883056641, + "num_tokens": 374874927.0, + "step": 14497 + }, + { + "epoch": 1.5921370524928617, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.46233868598938, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.743047833442688, + "num_tokens": 374895034.0, + "step": 14498 + }, + { + "epoch": 1.5922468701954755, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.322019100189209, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7111541032791138, + "num_tokens": 374919496.0, + "step": 14499 + }, + { + "epoch": 1.5923566878980893, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.343902826309204, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.714573323726654, + "num_tokens": 374943244.0, + "step": 14500 + }, + { + "epoch": 1.5924665056007028, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1764938831329346, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7141885757446289, + "num_tokens": 374971146.0, + "step": 14501 + }, + { + "epoch": 1.5925763233033163, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1422994136810303, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7253382205963135, + "num_tokens": 374997752.0, + "step": 14502 + }, + { + "epoch": 1.59268614100593, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.196668863296509, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7044167518615723, + "num_tokens": 375025536.0, + "step": 14503 + }, + { + "epoch": 1.5927959587085438, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.375849962234497, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.723656952381134, + "num_tokens": 375049295.0, + "step": 14504 + }, + { + "epoch": 1.5929057764111576, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.244774580001831, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7347757816314697, + "num_tokens": 375073356.0, + "step": 14505 + }, + { + "epoch": 1.5930155941137711, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.348803997039795, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7341704368591309, + "num_tokens": 375097520.0, + "step": 14506 + }, + { + "epoch": 1.5931254118163847, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.145094394683838, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7176545858383179, + "num_tokens": 375124237.0, + "step": 14507 + }, + { + "epoch": 1.5932352295189984, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2624049186706543, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7482565641403198, + "num_tokens": 375148049.0, + "step": 14508 + }, + { + "epoch": 1.5933450472216122, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.118248224258423, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7184402942657471, + "num_tokens": 375177807.0, + "step": 14509 + }, + { + "epoch": 1.593454864924226, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2673401832580566, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7370495796203613, + "num_tokens": 375201645.0, + "step": 14510 + }, + { + "epoch": 1.5935646826268395, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.07401442527771, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7123850584030151, + "num_tokens": 375229559.0, + "step": 14511 + }, + { + "epoch": 1.593674500329453, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0835564136505127, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.7118342518806458, + "num_tokens": 375259333.0, + "step": 14512 + }, + { + "epoch": 1.5937843180320668, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0900685787200928, + "learning_rate": 1e-06, + "loss": 0.9932, + "mean_token_accuracy": 0.6966198682785034, + "num_tokens": 375288550.0, + "step": 14513 + }, + { + "epoch": 1.5938941357346805, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9328371286392212, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7338978052139282, + "num_tokens": 375320715.0, + "step": 14514 + }, + { + "epoch": 1.594003953437294, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1493256092071533, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.6935129761695862, + "num_tokens": 375350974.0, + "step": 14515 + }, + { + "epoch": 1.5941137711399076, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.080514669418335, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7080790400505066, + "num_tokens": 375380124.0, + "step": 14516 + }, + { + "epoch": 1.5942235888425214, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2959682941436768, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7390365600585938, + "num_tokens": 375403688.0, + "step": 14517 + }, + { + "epoch": 1.5943334065451351, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4701924324035645, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7304546236991882, + "num_tokens": 375424204.0, + "step": 14518 + }, + { + "epoch": 1.5944432242477489, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0412893295288086, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.6981443762779236, + "num_tokens": 375454822.0, + "step": 14519 + }, + { + "epoch": 1.5945530419503624, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3003056049346924, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7117632627487183, + "num_tokens": 375481295.0, + "step": 14520 + }, + { + "epoch": 1.594662859652976, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.234304428100586, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7283514738082886, + "num_tokens": 375505553.0, + "step": 14521 + }, + { + "epoch": 1.5947726773555897, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3394198417663574, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7222499251365662, + "num_tokens": 375529493.0, + "step": 14522 + }, + { + "epoch": 1.5948824950582035, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0310633182525635, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7175487875938416, + "num_tokens": 375560012.0, + "step": 14523 + }, + { + "epoch": 1.5949923127608172, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.314162492752075, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7218817472457886, + "num_tokens": 375584353.0, + "step": 14524 + }, + { + "epoch": 1.5951021304634307, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1808688640594482, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7278238534927368, + "num_tokens": 375610895.0, + "step": 14525 + }, + { + "epoch": 1.5952119481660443, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.419113874435425, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7284184098243713, + "num_tokens": 375632870.0, + "step": 14526 + }, + { + "epoch": 1.595321765868658, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2338805198669434, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7094308137893677, + "num_tokens": 375658316.0, + "step": 14527 + }, + { + "epoch": 1.5954315835712718, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.18176007270813, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7121734023094177, + "num_tokens": 375684916.0, + "step": 14528 + }, + { + "epoch": 1.5955414012738853, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1892404556274414, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.7167268991470337, + "num_tokens": 375711341.0, + "step": 14529 + }, + { + "epoch": 1.5956512189764989, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.22534441947937, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7001287937164307, + "num_tokens": 375738416.0, + "step": 14530 + }, + { + "epoch": 1.5957610366791126, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.391986608505249, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7425282597541809, + "num_tokens": 375760713.0, + "step": 14531 + }, + { + "epoch": 1.5958708543817264, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2857043743133545, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7439424991607666, + "num_tokens": 375786840.0, + "step": 14532 + }, + { + "epoch": 1.5959806720843401, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3430511951446533, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7188748121261597, + "num_tokens": 375811614.0, + "step": 14533 + }, + { + "epoch": 1.5960904897869537, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2260019779205322, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7197507619857788, + "num_tokens": 375838020.0, + "step": 14534 + }, + { + "epoch": 1.5962003074895672, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3486475944519043, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7177645564079285, + "num_tokens": 375861796.0, + "step": 14535 + }, + { + "epoch": 1.596310125192181, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3087410926818848, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7225637435913086, + "num_tokens": 375887156.0, + "step": 14536 + }, + { + "epoch": 1.5964199428947947, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2001500129699707, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.738394021987915, + "num_tokens": 375911343.0, + "step": 14537 + }, + { + "epoch": 1.5965297605974083, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3427536487579346, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6973787546157837, + "num_tokens": 375936167.0, + "step": 14538 + }, + { + "epoch": 1.596639578300022, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9535006284713745, + "learning_rate": 1e-06, + "loss": 0.9447, + "mean_token_accuracy": 0.7159910202026367, + "num_tokens": 375967462.0, + "step": 14539 + }, + { + "epoch": 1.5967493960026355, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2578108310699463, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.730859637260437, + "num_tokens": 375990868.0, + "step": 14540 + }, + { + "epoch": 1.5968592137052493, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3034956455230713, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7045142650604248, + "num_tokens": 376016493.0, + "step": 14541 + }, + { + "epoch": 1.596969031407863, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5253195762634277, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7211593389511108, + "num_tokens": 376039072.0, + "step": 14542 + }, + { + "epoch": 1.5970788491104766, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1243643760681152, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7223212122917175, + "num_tokens": 376067183.0, + "step": 14543 + }, + { + "epoch": 1.5971886668130901, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3326337337493896, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.730992317199707, + "num_tokens": 376091007.0, + "step": 14544 + }, + { + "epoch": 1.5972984845157039, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5197887420654297, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7432237267494202, + "num_tokens": 376110858.0, + "step": 14545 + }, + { + "epoch": 1.5974083022183176, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.455193519592285, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7454535961151123, + "num_tokens": 376131065.0, + "step": 14546 + }, + { + "epoch": 1.5975181199209314, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2441601753234863, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7146971821784973, + "num_tokens": 376156691.0, + "step": 14547 + }, + { + "epoch": 1.597627937623545, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3377537727355957, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.710576593875885, + "num_tokens": 376180701.0, + "step": 14548 + }, + { + "epoch": 1.5977377553261585, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.156386137008667, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7096215486526489, + "num_tokens": 376209886.0, + "step": 14549 + }, + { + "epoch": 1.5978475730287722, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.277988910675049, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.736112654209137, + "num_tokens": 376234719.0, + "step": 14550 + }, + { + "epoch": 1.597957390731386, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.489772081375122, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7301467061042786, + "num_tokens": 376255215.0, + "step": 14551 + }, + { + "epoch": 1.5980672084339995, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1999378204345703, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7189358472824097, + "num_tokens": 376280985.0, + "step": 14552 + }, + { + "epoch": 1.5981770261366133, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3951218128204346, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7218084335327148, + "num_tokens": 376305600.0, + "step": 14553 + }, + { + "epoch": 1.5982868438392268, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.487651824951172, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7160676717758179, + "num_tokens": 376329147.0, + "step": 14554 + }, + { + "epoch": 1.5983966615418406, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2418606281280518, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.718812108039856, + "num_tokens": 376353320.0, + "step": 14555 + }, + { + "epoch": 1.5985064792444543, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1519882678985596, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7038551568984985, + "num_tokens": 376380025.0, + "step": 14556 + }, + { + "epoch": 1.5986162969470679, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0226128101348877, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7125877737998962, + "num_tokens": 376411452.0, + "step": 14557 + }, + { + "epoch": 1.5987261146496814, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.979965329170227, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7041339874267578, + "num_tokens": 376444923.0, + "step": 14558 + }, + { + "epoch": 1.5988359323522952, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3988475799560547, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7161399126052856, + "num_tokens": 376467832.0, + "step": 14559 + }, + { + "epoch": 1.598945750054909, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2535319328308105, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7173342108726501, + "num_tokens": 376492742.0, + "step": 14560 + }, + { + "epoch": 1.5990555677575227, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.00004506111145, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7194284796714783, + "num_tokens": 376525201.0, + "step": 14561 + }, + { + "epoch": 1.5991653854601362, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3952136039733887, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7309591770172119, + "num_tokens": 376549642.0, + "step": 14562 + }, + { + "epoch": 1.5992752031627497, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.8141732215881348, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7137153148651123, + "num_tokens": 376568606.0, + "step": 14563 + }, + { + "epoch": 1.5993850208653635, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.321533441543579, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7031141519546509, + "num_tokens": 376593160.0, + "step": 14564 + }, + { + "epoch": 1.5994948385679773, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3265583515167236, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.740058422088623, + "num_tokens": 376615368.0, + "step": 14565 + }, + { + "epoch": 1.5996046562705908, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3578503131866455, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6961877942085266, + "num_tokens": 376639264.0, + "step": 14566 + }, + { + "epoch": 1.5997144739732043, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.051255464553833, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.7075662612915039, + "num_tokens": 376669428.0, + "step": 14567 + }, + { + "epoch": 1.599824291675818, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.126880168914795, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7385410070419312, + "num_tokens": 376694719.0, + "step": 14568 + }, + { + "epoch": 1.5999341093784318, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3062336444854736, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7167338728904724, + "num_tokens": 376718948.0, + "step": 14569 + }, + { + "epoch": 1.6000439270810456, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1498451232910156, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7196651697158813, + "num_tokens": 376747031.0, + "step": 14570 + }, + { + "epoch": 1.6001537447836591, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3717055320739746, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7341529726982117, + "num_tokens": 376770942.0, + "step": 14571 + }, + { + "epoch": 1.6002635624862727, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.117849349975586, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7229134440422058, + "num_tokens": 376800528.0, + "step": 14572 + }, + { + "epoch": 1.6003733801888864, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9496705532073975, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7137334942817688, + "num_tokens": 376833529.0, + "step": 14573 + }, + { + "epoch": 1.6004831978915002, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2422935962677, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7251598238945007, + "num_tokens": 376856645.0, + "step": 14574 + }, + { + "epoch": 1.600593015594114, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.033006191253662, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7225346565246582, + "num_tokens": 376888611.0, + "step": 14575 + }, + { + "epoch": 1.6007028332967275, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.271413564682007, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7192609906196594, + "num_tokens": 376912936.0, + "step": 14576 + }, + { + "epoch": 1.600812650999341, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.253251314163208, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7093961238861084, + "num_tokens": 376937269.0, + "step": 14577 + }, + { + "epoch": 1.6009224687019548, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.33136248588562, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7240354418754578, + "num_tokens": 376960702.0, + "step": 14578 + }, + { + "epoch": 1.6010322864045685, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.097188949584961, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7262811660766602, + "num_tokens": 376988619.0, + "step": 14579 + }, + { + "epoch": 1.601142104107182, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.974237084388733, + "learning_rate": 1e-06, + "loss": 1.02, + "mean_token_accuracy": 0.6877955198287964, + "num_tokens": 377020953.0, + "step": 14580 + }, + { + "epoch": 1.6012519218097956, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3171281814575195, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7271668314933777, + "num_tokens": 377044028.0, + "step": 14581 + }, + { + "epoch": 1.6013617395124093, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1861114501953125, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7134146094322205, + "num_tokens": 377076356.0, + "step": 14582 + }, + { + "epoch": 1.601471557215023, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2796809673309326, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7248458862304688, + "num_tokens": 377100286.0, + "step": 14583 + }, + { + "epoch": 1.6015813749176369, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0799989700317383, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.728685200214386, + "num_tokens": 377129006.0, + "step": 14584 + }, + { + "epoch": 1.6016911926202504, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.749948024749756, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7392233610153198, + "num_tokens": 377147317.0, + "step": 14585 + }, + { + "epoch": 1.601801010322864, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2024879455566406, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7092127203941345, + "num_tokens": 377174880.0, + "step": 14586 + }, + { + "epoch": 1.6019108280254777, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4154300689697266, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7205385565757751, + "num_tokens": 377197324.0, + "step": 14587 + }, + { + "epoch": 1.6020206457280914, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5326242446899414, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7137203216552734, + "num_tokens": 377218568.0, + "step": 14588 + }, + { + "epoch": 1.602130463430705, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.305616617202759, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7195636034011841, + "num_tokens": 377242710.0, + "step": 14589 + }, + { + "epoch": 1.6022402811333187, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.8810917139053345, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7175384759902954, + "num_tokens": 377282627.0, + "step": 14590 + }, + { + "epoch": 1.6023500988359323, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.465710163116455, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7325812578201294, + "num_tokens": 377304607.0, + "step": 14591 + }, + { + "epoch": 1.602459916538546, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.6739792823791504, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7403824329376221, + "num_tokens": 377324519.0, + "step": 14592 + }, + { + "epoch": 1.6025697342411598, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.8537861108779907, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6889358758926392, + "num_tokens": 377360265.0, + "step": 14593 + }, + { + "epoch": 1.6026795519437733, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.792907238006592, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7190166115760803, + "num_tokens": 377377975.0, + "step": 14594 + }, + { + "epoch": 1.6027893696463869, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0625412464141846, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.724259614944458, + "num_tokens": 377407738.0, + "step": 14595 + }, + { + "epoch": 1.6028991873490006, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3633220195770264, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7219994068145752, + "num_tokens": 377429190.0, + "step": 14596 + }, + { + "epoch": 1.6030090050516144, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3906514644622803, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7409436702728271, + "num_tokens": 377451639.0, + "step": 14597 + }, + { + "epoch": 1.6031188227542281, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3572371006011963, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7121191620826721, + "num_tokens": 377477193.0, + "step": 14598 + }, + { + "epoch": 1.6032286404568417, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4080810546875, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7223902940750122, + "num_tokens": 377503375.0, + "step": 14599 + }, + { + "epoch": 1.6033384581594552, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.501857042312622, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7292335033416748, + "num_tokens": 377523637.0, + "step": 14600 + }, + { + "epoch": 1.603448275862069, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2008495330810547, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.698870062828064, + "num_tokens": 377553443.0, + "step": 14601 + }, + { + "epoch": 1.6035580935646827, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9839568138122559, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.7551653385162354, + "num_tokens": 377582274.0, + "step": 14602 + }, + { + "epoch": 1.6036679112672962, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4466323852539062, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7082799077033997, + "num_tokens": 377604082.0, + "step": 14603 + }, + { + "epoch": 1.60377772896991, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1251606941223145, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7396037578582764, + "num_tokens": 377629865.0, + "step": 14604 + }, + { + "epoch": 1.6038875466725235, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4192259311676025, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7403096556663513, + "num_tokens": 377651745.0, + "step": 14605 + }, + { + "epoch": 1.6039973643751373, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.991228461265564, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7029930353164673, + "num_tokens": 377683487.0, + "step": 14606 + }, + { + "epoch": 1.604107182077751, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.314906597137451, + "learning_rate": 1e-06, + "loss": 0.9596, + "mean_token_accuracy": 0.699591875076294, + "num_tokens": 377708915.0, + "step": 14607 + }, + { + "epoch": 1.6042169997803646, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.162696123123169, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7197921276092529, + "num_tokens": 377733857.0, + "step": 14608 + }, + { + "epoch": 1.6043268174829781, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5566046237945557, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7470743656158447, + "num_tokens": 377752263.0, + "step": 14609 + }, + { + "epoch": 1.6044366351855919, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3103249073028564, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7192450761795044, + "num_tokens": 377775949.0, + "step": 14610 + }, + { + "epoch": 1.6045464528882056, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4608631134033203, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.737114429473877, + "num_tokens": 377797937.0, + "step": 14611 + }, + { + "epoch": 1.6046562705908194, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1485118865966797, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.72343909740448, + "num_tokens": 377828982.0, + "step": 14612 + }, + { + "epoch": 1.604766088293433, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9203490018844604, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7033212780952454, + "num_tokens": 377863360.0, + "step": 14613 + }, + { + "epoch": 1.6048759059960465, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.151719093322754, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7024895548820496, + "num_tokens": 377891909.0, + "step": 14614 + }, + { + "epoch": 1.6049857236986602, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.283968925476074, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7203266620635986, + "num_tokens": 377916869.0, + "step": 14615 + }, + { + "epoch": 1.605095541401274, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3558478355407715, + "learning_rate": 1e-06, + "loss": 1.0099, + "mean_token_accuracy": 0.6904432773590088, + "num_tokens": 377941425.0, + "step": 14616 + }, + { + "epoch": 1.6052053591038875, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.227891206741333, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7349655628204346, + "num_tokens": 377966567.0, + "step": 14617 + }, + { + "epoch": 1.605315176806501, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.235191822052002, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.6963897347450256, + "num_tokens": 377993994.0, + "step": 14618 + }, + { + "epoch": 1.6054249945091148, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.324209690093994, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7231194972991943, + "num_tokens": 378018392.0, + "step": 14619 + }, + { + "epoch": 1.6055348122117286, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5594518184661865, + "learning_rate": 1e-06, + "loss": 0.7433, + "mean_token_accuracy": 0.7562419176101685, + "num_tokens": 378038143.0, + "step": 14620 + }, + { + "epoch": 1.6056446299143423, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.271488666534424, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7107251882553101, + "num_tokens": 378063431.0, + "step": 14621 + }, + { + "epoch": 1.6057544476169558, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2776973247528076, + "learning_rate": 1e-06, + "loss": 0.8003, + "mean_token_accuracy": 0.746628999710083, + "num_tokens": 378086320.0, + "step": 14622 + }, + { + "epoch": 1.6058642653195694, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.630370616912842, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7197978496551514, + "num_tokens": 378107634.0, + "step": 14623 + }, + { + "epoch": 1.6059740830221831, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.324122190475464, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.722835898399353, + "num_tokens": 378134425.0, + "step": 14624 + }, + { + "epoch": 1.606083900724797, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.015942096710205, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7006153464317322, + "num_tokens": 378163654.0, + "step": 14625 + }, + { + "epoch": 1.6061937184274107, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1779117584228516, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7217952013015747, + "num_tokens": 378190317.0, + "step": 14626 + }, + { + "epoch": 1.6063035361300242, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1682357788085938, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6923617124557495, + "num_tokens": 378219657.0, + "step": 14627 + }, + { + "epoch": 1.6064133538326377, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5810153484344482, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7255040407180786, + "num_tokens": 378239882.0, + "step": 14628 + }, + { + "epoch": 1.6065231715352515, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1997485160827637, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6956747770309448, + "num_tokens": 378267995.0, + "step": 14629 + }, + { + "epoch": 1.6066329892378652, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3390064239501953, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7141424417495728, + "num_tokens": 378291931.0, + "step": 14630 + }, + { + "epoch": 1.6067428069404788, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3605735301971436, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.713502824306488, + "num_tokens": 378315418.0, + "step": 14631 + }, + { + "epoch": 1.6068526246430923, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.359684705734253, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7268548011779785, + "num_tokens": 378336854.0, + "step": 14632 + }, + { + "epoch": 1.606962442345706, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0681943893432617, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7163886427879333, + "num_tokens": 378365320.0, + "step": 14633 + }, + { + "epoch": 1.6070722600483198, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.643105983734131, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7276881337165833, + "num_tokens": 378387196.0, + "step": 14634 + }, + { + "epoch": 1.6071820777509336, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.617824077606201, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.738431990146637, + "num_tokens": 378406099.0, + "step": 14635 + }, + { + "epoch": 1.6072918954535471, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.046358108520508, + "learning_rate": 1e-06, + "loss": 1.0005, + "mean_token_accuracy": 0.6973012685775757, + "num_tokens": 378434516.0, + "step": 14636 + }, + { + "epoch": 1.6074017131561606, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0644659996032715, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7343464493751526, + "num_tokens": 378463211.0, + "step": 14637 + }, + { + "epoch": 1.6075115308587744, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.379685163497925, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7069879174232483, + "num_tokens": 378487206.0, + "step": 14638 + }, + { + "epoch": 1.6076213485613882, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3261778354644775, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7203393578529358, + "num_tokens": 378509810.0, + "step": 14639 + }, + { + "epoch": 1.607731166264002, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.424142837524414, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7119302749633789, + "num_tokens": 378535236.0, + "step": 14640 + }, + { + "epoch": 1.6078409839666155, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.118600606918335, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7165789008140564, + "num_tokens": 378564805.0, + "step": 14641 + }, + { + "epoch": 1.607950801669229, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2517240047454834, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7150395512580872, + "num_tokens": 378590294.0, + "step": 14642 + }, + { + "epoch": 1.6080606193718427, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.259232759475708, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.6972459554672241, + "num_tokens": 378616209.0, + "step": 14643 + }, + { + "epoch": 1.6081704370744565, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2382614612579346, + "learning_rate": 1e-06, + "loss": 0.9562, + "mean_token_accuracy": 0.7056179642677307, + "num_tokens": 378642574.0, + "step": 14644 + }, + { + "epoch": 1.60828025477707, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2263286113739014, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7166416645050049, + "num_tokens": 378667975.0, + "step": 14645 + }, + { + "epoch": 1.6083900724796836, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1620121002197266, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7318177223205566, + "num_tokens": 378697179.0, + "step": 14646 + }, + { + "epoch": 1.6084998901822973, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5955851078033447, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7261396050453186, + "num_tokens": 378718268.0, + "step": 14647 + }, + { + "epoch": 1.608609707884911, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2482831478118896, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7153621912002563, + "num_tokens": 378745182.0, + "step": 14648 + }, + { + "epoch": 1.6087195255875248, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.293192148208618, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7371674180030823, + "num_tokens": 378770063.0, + "step": 14649 + }, + { + "epoch": 1.6088293432901384, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2751495838165283, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7328193187713623, + "num_tokens": 378794104.0, + "step": 14650 + }, + { + "epoch": 1.608939160992752, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.18170428276062, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7180215120315552, + "num_tokens": 378820354.0, + "step": 14651 + }, + { + "epoch": 1.6090489786953657, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2660577297210693, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.716748058795929, + "num_tokens": 378847049.0, + "step": 14652 + }, + { + "epoch": 1.6091587963979794, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2611870765686035, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.737510621547699, + "num_tokens": 378871601.0, + "step": 14653 + }, + { + "epoch": 1.609268614100593, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2442800998687744, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7345239520072937, + "num_tokens": 378895536.0, + "step": 14654 + }, + { + "epoch": 1.6093784318032067, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.14931058883667, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7185242772102356, + "num_tokens": 378923499.0, + "step": 14655 + }, + { + "epoch": 1.6094882495058203, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0661208629608154, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7111334204673767, + "num_tokens": 378952595.0, + "step": 14656 + }, + { + "epoch": 1.609598067208434, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1034762859344482, + "learning_rate": 1e-06, + "loss": 0.9807, + "mean_token_accuracy": 0.7028731107711792, + "num_tokens": 378981185.0, + "step": 14657 + }, + { + "epoch": 1.6097078849110478, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1920111179351807, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.7109565734863281, + "num_tokens": 379009117.0, + "step": 14658 + }, + { + "epoch": 1.6098177026136613, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.601414680480957, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7360943555831909, + "num_tokens": 379028374.0, + "step": 14659 + }, + { + "epoch": 1.6099275203162748, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.025226593017578, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.6960123777389526, + "num_tokens": 379058265.0, + "step": 14660 + }, + { + "epoch": 1.6100373380188886, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4415981769561768, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7160435914993286, + "num_tokens": 379078961.0, + "step": 14661 + }, + { + "epoch": 1.6101471557215024, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.316272735595703, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.723511815071106, + "num_tokens": 379104371.0, + "step": 14662 + }, + { + "epoch": 1.6102569734241161, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5039243698120117, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7296921610832214, + "num_tokens": 379125368.0, + "step": 14663 + }, + { + "epoch": 1.6103667911267296, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.204702854156494, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.716580867767334, + "num_tokens": 379148661.0, + "step": 14664 + }, + { + "epoch": 1.6104766088293432, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.021595001220703, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7061820030212402, + "num_tokens": 379180087.0, + "step": 14665 + }, + { + "epoch": 1.610586426531957, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.181664228439331, + "learning_rate": 1e-06, + "loss": 0.989, + "mean_token_accuracy": 0.6954165697097778, + "num_tokens": 379208009.0, + "step": 14666 + }, + { + "epoch": 1.6106962442345707, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4047293663024902, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7184414267539978, + "num_tokens": 379229757.0, + "step": 14667 + }, + { + "epoch": 1.6108060619371842, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3391058444976807, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7340172529220581, + "num_tokens": 379252466.0, + "step": 14668 + }, + { + "epoch": 1.610915879639798, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.15152645111084, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7303094267845154, + "num_tokens": 379281949.0, + "step": 14669 + }, + { + "epoch": 1.6110256973424115, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2861757278442383, + "learning_rate": 1e-06, + "loss": 0.9849, + "mean_token_accuracy": 0.6960898637771606, + "num_tokens": 379314100.0, + "step": 14670 + }, + { + "epoch": 1.6111355150450253, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1863090991973877, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7223496437072754, + "num_tokens": 379341840.0, + "step": 14671 + }, + { + "epoch": 1.611245332747639, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.110708713531494, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.729026734828949, + "num_tokens": 379368847.0, + "step": 14672 + }, + { + "epoch": 1.6113551504502526, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4177048206329346, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.739107072353363, + "num_tokens": 379391704.0, + "step": 14673 + }, + { + "epoch": 1.611464968152866, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3958346843719482, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7318778038024902, + "num_tokens": 379415996.0, + "step": 14674 + }, + { + "epoch": 1.6115747858554799, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0682413578033447, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7267580628395081, + "num_tokens": 379445398.0, + "step": 14675 + }, + { + "epoch": 1.6116846035580936, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.125180721282959, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7245124578475952, + "num_tokens": 379473108.0, + "step": 14676 + }, + { + "epoch": 1.6117944212607074, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2815918922424316, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7297871112823486, + "num_tokens": 379497755.0, + "step": 14677 + }, + { + "epoch": 1.611904238963321, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5183119773864746, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7291114330291748, + "num_tokens": 379520189.0, + "step": 14678 + }, + { + "epoch": 1.6120140566659344, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.438495397567749, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7154477834701538, + "num_tokens": 379543320.0, + "step": 14679 + }, + { + "epoch": 1.6121238743685482, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.257636070251465, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7109131813049316, + "num_tokens": 379566938.0, + "step": 14680 + }, + { + "epoch": 1.612233692071162, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.275409698486328, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7214472889900208, + "num_tokens": 379591567.0, + "step": 14681 + }, + { + "epoch": 1.6123435097737755, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.170438766479492, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.731523334980011, + "num_tokens": 379617964.0, + "step": 14682 + }, + { + "epoch": 1.612453327476389, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.124917507171631, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7172209620475769, + "num_tokens": 379645750.0, + "step": 14683 + }, + { + "epoch": 1.6125631451790028, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.257617712020874, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7402153015136719, + "num_tokens": 379669620.0, + "step": 14684 + }, + { + "epoch": 1.6126729628816165, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2494447231292725, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7168485522270203, + "num_tokens": 379695559.0, + "step": 14685 + }, + { + "epoch": 1.6127827805842303, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.405897855758667, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7206560373306274, + "num_tokens": 379719060.0, + "step": 14686 + }, + { + "epoch": 1.6128925982868438, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2861251831054688, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7044486999511719, + "num_tokens": 379746301.0, + "step": 14687 + }, + { + "epoch": 1.6130024159894574, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1147830486297607, + "learning_rate": 1e-06, + "loss": 0.9634, + "mean_token_accuracy": 0.7066067457199097, + "num_tokens": 379774412.0, + "step": 14688 + }, + { + "epoch": 1.6131122336920711, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0805349349975586, + "learning_rate": 1e-06, + "loss": 0.9702, + "mean_token_accuracy": 0.6982161998748779, + "num_tokens": 379802433.0, + "step": 14689 + }, + { + "epoch": 1.6132220513946849, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.9456775188446045, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7137991189956665, + "num_tokens": 379834583.0, + "step": 14690 + }, + { + "epoch": 1.6133318690972986, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.694626808166504, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7233810424804688, + "num_tokens": 379854369.0, + "step": 14691 + }, + { + "epoch": 1.6134416867999122, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2443954944610596, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.720542311668396, + "num_tokens": 379880685.0, + "step": 14692 + }, + { + "epoch": 1.6135515045025257, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1260085105895996, + "learning_rate": 1e-06, + "loss": 1.0122, + "mean_token_accuracy": 0.6887853145599365, + "num_tokens": 379910731.0, + "step": 14693 + }, + { + "epoch": 1.6136613222051395, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.280564069747925, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7300075888633728, + "num_tokens": 379934116.0, + "step": 14694 + }, + { + "epoch": 1.6137711399077532, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2305257320404053, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.729399561882019, + "num_tokens": 379960684.0, + "step": 14695 + }, + { + "epoch": 1.6138809576103668, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2571797370910645, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.707202672958374, + "num_tokens": 379987816.0, + "step": 14696 + }, + { + "epoch": 1.6139907753129803, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0184171199798584, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7195776104927063, + "num_tokens": 380017144.0, + "step": 14697 + }, + { + "epoch": 1.614100593015594, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.399968147277832, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7276653051376343, + "num_tokens": 380039497.0, + "step": 14698 + }, + { + "epoch": 1.6142104107182078, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1238961219787598, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7164405584335327, + "num_tokens": 380068009.0, + "step": 14699 + }, + { + "epoch": 1.6143202284208216, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.046933650970459, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7060953378677368, + "num_tokens": 380101244.0, + "step": 14700 + }, + { + "epoch": 1.614430046123435, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.15427827835083, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7339769005775452, + "num_tokens": 380129242.0, + "step": 14701 + }, + { + "epoch": 1.6145398638260486, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4755806922912598, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7046488523483276, + "num_tokens": 380151774.0, + "step": 14702 + }, + { + "epoch": 1.6146496815286624, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0661396980285645, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7344919443130493, + "num_tokens": 380178855.0, + "step": 14703 + }, + { + "epoch": 1.6147594992312762, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4105799198150635, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7217332124710083, + "num_tokens": 380200391.0, + "step": 14704 + }, + { + "epoch": 1.61486931693389, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.055055618286133, + "learning_rate": 1e-06, + "loss": 0.9762, + "mean_token_accuracy": 0.7020370960235596, + "num_tokens": 380231561.0, + "step": 14705 + }, + { + "epoch": 1.6149791346365034, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.190732002258301, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7354335188865662, + "num_tokens": 380255188.0, + "step": 14706 + }, + { + "epoch": 1.615088952339117, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4275548458099365, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7262725830078125, + "num_tokens": 380278009.0, + "step": 14707 + }, + { + "epoch": 1.6151987700417307, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.941611886024475, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.69701087474823, + "num_tokens": 380311382.0, + "step": 14708 + }, + { + "epoch": 1.6153085877443445, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.8382573127746582, + "learning_rate": 1e-06, + "loss": 1.0217, + "mean_token_accuracy": 0.6862511038780212, + "num_tokens": 380348900.0, + "step": 14709 + }, + { + "epoch": 1.615418405446958, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2480099201202393, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7390387654304504, + "num_tokens": 380373273.0, + "step": 14710 + }, + { + "epoch": 1.6155282231495716, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0141196250915527, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7256428599357605, + "num_tokens": 380404588.0, + "step": 14711 + }, + { + "epoch": 1.6156380408521853, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.01899790763855, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7020989656448364, + "num_tokens": 380435705.0, + "step": 14712 + }, + { + "epoch": 1.615747858554799, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3377928733825684, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7093390226364136, + "num_tokens": 380459583.0, + "step": 14713 + }, + { + "epoch": 1.6158576762574128, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3315765857696533, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7473437786102295, + "num_tokens": 380483648.0, + "step": 14714 + }, + { + "epoch": 1.6159674939600264, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2226216793060303, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.7014756202697754, + "num_tokens": 380510835.0, + "step": 14715 + }, + { + "epoch": 1.61607731166264, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.426114559173584, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7059995532035828, + "num_tokens": 380534479.0, + "step": 14716 + }, + { + "epoch": 1.6161871293652537, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2760918140411377, + "learning_rate": 1e-06, + "loss": 0.9436, + "mean_token_accuracy": 0.71548992395401, + "num_tokens": 380558985.0, + "step": 14717 + }, + { + "epoch": 1.6162969470678674, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4329352378845215, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7128461003303528, + "num_tokens": 380582954.0, + "step": 14718 + }, + { + "epoch": 1.616406764770481, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1428399085998535, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.709580659866333, + "num_tokens": 380609904.0, + "step": 14719 + }, + { + "epoch": 1.6165165824730947, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4517321586608887, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7170435190200806, + "num_tokens": 380633853.0, + "step": 14720 + }, + { + "epoch": 1.6166264001757082, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3543272018432617, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7162646651268005, + "num_tokens": 380658581.0, + "step": 14721 + }, + { + "epoch": 1.616736217878322, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1871378421783447, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7342463731765747, + "num_tokens": 380684309.0, + "step": 14722 + }, + { + "epoch": 1.6168460355809358, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.124621629714966, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7094506025314331, + "num_tokens": 380714010.0, + "step": 14723 + }, + { + "epoch": 1.6169558532835493, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2206685543060303, + "learning_rate": 1e-06, + "loss": 0.7364, + "mean_token_accuracy": 0.7655867338180542, + "num_tokens": 380736775.0, + "step": 14724 + }, + { + "epoch": 1.6170656709861628, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.444927453994751, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7414791584014893, + "num_tokens": 380758733.0, + "step": 14725 + }, + { + "epoch": 1.6171754886887766, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.391308546066284, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7327346205711365, + "num_tokens": 380781061.0, + "step": 14726 + }, + { + "epoch": 1.6172853063913903, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5475709438323975, + "learning_rate": 1e-06, + "loss": 0.7741, + "mean_token_accuracy": 0.7492367625236511, + "num_tokens": 380800537.0, + "step": 14727 + }, + { + "epoch": 1.617395124094004, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.5006866455078125, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7119820713996887, + "num_tokens": 380822551.0, + "step": 14728 + }, + { + "epoch": 1.6175049417966176, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1424806118011475, + "learning_rate": 1e-06, + "loss": 1.0024, + "mean_token_accuracy": 0.6931896209716797, + "num_tokens": 380850508.0, + "step": 14729 + }, + { + "epoch": 1.6176147594992312, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2824158668518066, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7171554565429688, + "num_tokens": 380873596.0, + "step": 14730 + }, + { + "epoch": 1.617724577201845, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1275267601013184, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7172956466674805, + "num_tokens": 380900552.0, + "step": 14731 + }, + { + "epoch": 1.6178343949044587, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1449167728424072, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7244078516960144, + "num_tokens": 380927255.0, + "step": 14732 + }, + { + "epoch": 1.6179442126070722, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2166385650634766, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7028709650039673, + "num_tokens": 380952323.0, + "step": 14733 + }, + { + "epoch": 1.618054030309686, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4733622074127197, + "learning_rate": 1e-06, + "loss": 1.008, + "mean_token_accuracy": 0.693317711353302, + "num_tokens": 380978041.0, + "step": 14734 + }, + { + "epoch": 1.6181638480122995, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.383939743041992, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.7445448637008667, + "num_tokens": 380998558.0, + "step": 14735 + }, + { + "epoch": 1.6182736657149133, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.017244577407837, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.7046124339103699, + "num_tokens": 381029442.0, + "step": 14736 + }, + { + "epoch": 1.618383483417527, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3592112064361572, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7340040802955627, + "num_tokens": 381053022.0, + "step": 14737 + }, + { + "epoch": 1.6184933011201406, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4143106937408447, + "learning_rate": 1e-06, + "loss": 1.0115, + "mean_token_accuracy": 0.7018380165100098, + "num_tokens": 381078358.0, + "step": 14738 + }, + { + "epoch": 1.618603118822754, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1519041061401367, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7195925712585449, + "num_tokens": 381105573.0, + "step": 14739 + }, + { + "epoch": 1.6187129365253679, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.038142204284668, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7134344577789307, + "num_tokens": 381136566.0, + "step": 14740 + }, + { + "epoch": 1.6188227542279816, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.356732130050659, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7177304029464722, + "num_tokens": 381159490.0, + "step": 14741 + }, + { + "epoch": 1.6189325719305954, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.146331310272217, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7405639886856079, + "num_tokens": 381184986.0, + "step": 14742 + }, + { + "epoch": 1.619042389633209, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1393749713897705, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.72286057472229, + "num_tokens": 381210746.0, + "step": 14743 + }, + { + "epoch": 1.6191522073358224, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 1.975028157234192, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7092350721359253, + "num_tokens": 381241557.0, + "step": 14744 + }, + { + "epoch": 1.6192620250384362, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2529516220092773, + "learning_rate": 1e-06, + "loss": 0.7896, + "mean_token_accuracy": 0.7454143762588501, + "num_tokens": 381266486.0, + "step": 14745 + }, + { + "epoch": 1.61937184274105, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1814756393432617, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7327293157577515, + "num_tokens": 381292874.0, + "step": 14746 + }, + { + "epoch": 1.6194816604436635, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0631816387176514, + "learning_rate": 1e-06, + "loss": 1.0369, + "mean_token_accuracy": 0.6957062482833862, + "num_tokens": 381323286.0, + "step": 14747 + }, + { + "epoch": 1.619591478146277, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3471639156341553, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7319109439849854, + "num_tokens": 381344594.0, + "step": 14748 + }, + { + "epoch": 1.6197012958488908, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2286059856414795, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7211818099021912, + "num_tokens": 381371462.0, + "step": 14749 + }, + { + "epoch": 1.6198111135515045, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.293653726577759, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.6999638676643372, + "num_tokens": 381397383.0, + "step": 14750 + }, + { + "epoch": 1.6199209312541183, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0576038360595703, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.722980260848999, + "num_tokens": 381426112.0, + "step": 14751 + }, + { + "epoch": 1.6200307489567318, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.291912794113159, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7111421823501587, + "num_tokens": 381452303.0, + "step": 14752 + }, + { + "epoch": 1.6201405666593454, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3704025745391846, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7092828154563904, + "num_tokens": 381476583.0, + "step": 14753 + }, + { + "epoch": 1.6202503843619591, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.022491931915283, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.7080755233764648, + "num_tokens": 381509030.0, + "step": 14754 + }, + { + "epoch": 1.6203602020645729, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.4983129501342773, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7501263618469238, + "num_tokens": 381529557.0, + "step": 14755 + }, + { + "epoch": 1.6204700197671866, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.060854911804199, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7005102634429932, + "num_tokens": 381558586.0, + "step": 14756 + }, + { + "epoch": 1.6205798374698002, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.104109048843384, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6966699361801147, + "num_tokens": 381587177.0, + "step": 14757 + }, + { + "epoch": 1.6206896551724137, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.016508102416992, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7073978185653687, + "num_tokens": 381617923.0, + "step": 14758 + }, + { + "epoch": 1.6207994728750275, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.108308792114258, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.7059444785118103, + "num_tokens": 381647988.0, + "step": 14759 + }, + { + "epoch": 1.6209092905776412, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3372955322265625, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7150459885597229, + "num_tokens": 381671221.0, + "step": 14760 + }, + { + "epoch": 1.6210191082802548, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2340052127838135, + "learning_rate": 1e-06, + "loss": 1.018, + "mean_token_accuracy": 0.6981632709503174, + "num_tokens": 381698408.0, + "step": 14761 + }, + { + "epoch": 1.6211289259828683, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.643610954284668, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7272926568984985, + "num_tokens": 381717742.0, + "step": 14762 + }, + { + "epoch": 1.621238743685482, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0675387382507324, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.724392294883728, + "num_tokens": 381748107.0, + "step": 14763 + }, + { + "epoch": 1.6213485613880958, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3936939239501953, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.73052978515625, + "num_tokens": 381770856.0, + "step": 14764 + }, + { + "epoch": 1.6214583790907096, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0782253742218018, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7183057069778442, + "num_tokens": 381799450.0, + "step": 14765 + }, + { + "epoch": 1.621568196793323, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.217909097671509, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7125444412231445, + "num_tokens": 381825078.0, + "step": 14766 + }, + { + "epoch": 1.6216780144959366, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.1778316497802734, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.710159420967102, + "num_tokens": 381853370.0, + "step": 14767 + }, + { + "epoch": 1.6217878321985504, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.0994107723236084, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7252150774002075, + "num_tokens": 381880615.0, + "step": 14768 + }, + { + "epoch": 1.6218976499011641, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.361355781555176, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7185947895050049, + "num_tokens": 381904268.0, + "step": 14769 + }, + { + "epoch": 1.6220074676037777, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2741143703460693, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7136670351028442, + "num_tokens": 381928517.0, + "step": 14770 + }, + { + "epoch": 1.6221172853063914, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.424471139907837, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7186943292617798, + "num_tokens": 381951430.0, + "step": 14771 + }, + { + "epoch": 1.622227103009005, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.268918991088867, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7211828231811523, + "num_tokens": 381976894.0, + "step": 14772 + }, + { + "epoch": 1.6223369207116187, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.303907871246338, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7365801334381104, + "num_tokens": 382002349.0, + "step": 14773 + }, + { + "epoch": 1.6224467384142325, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4636013507843018, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7144690752029419, + "num_tokens": 382025309.0, + "step": 14774 + }, + { + "epoch": 1.622556556116846, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.409080743789673, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7121884226799011, + "num_tokens": 382047531.0, + "step": 14775 + }, + { + "epoch": 1.6226663738194596, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.466780185699463, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.698984682559967, + "num_tokens": 382070872.0, + "step": 14776 + }, + { + "epoch": 1.6227761915220733, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0378901958465576, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7213186025619507, + "num_tokens": 382100230.0, + "step": 14777 + }, + { + "epoch": 1.622886009224687, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9924339056015015, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.748336911201477, + "num_tokens": 382129822.0, + "step": 14778 + }, + { + "epoch": 1.6229958269273008, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.215055465698242, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7390475869178772, + "num_tokens": 382155243.0, + "step": 14779 + }, + { + "epoch": 1.6231056446299144, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2570109367370605, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7367843985557556, + "num_tokens": 382179428.0, + "step": 14780 + }, + { + "epoch": 1.623215462332528, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2122159004211426, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7153339385986328, + "num_tokens": 382205016.0, + "step": 14781 + }, + { + "epoch": 1.6233252800351416, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.007683515548706, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.6995176076889038, + "num_tokens": 382236986.0, + "step": 14782 + }, + { + "epoch": 1.6234350977377554, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.037365674972534, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7050774097442627, + "num_tokens": 382268483.0, + "step": 14783 + }, + { + "epoch": 1.623544915440369, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.294049024581909, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7145184278488159, + "num_tokens": 382294735.0, + "step": 14784 + }, + { + "epoch": 1.6236547331429827, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2006349563598633, + "learning_rate": 1e-06, + "loss": 0.7879, + "mean_token_accuracy": 0.749373197555542, + "num_tokens": 382319364.0, + "step": 14785 + }, + { + "epoch": 1.6237645508455962, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.211501121520996, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7085217833518982, + "num_tokens": 382343621.0, + "step": 14786 + }, + { + "epoch": 1.62387436854821, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.208859920501709, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7091994881629944, + "num_tokens": 382371121.0, + "step": 14787 + }, + { + "epoch": 1.6239841862508237, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2490248680114746, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7151330709457397, + "num_tokens": 382397503.0, + "step": 14788 + }, + { + "epoch": 1.6240940039534373, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1812827587127686, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.6999150514602661, + "num_tokens": 382424574.0, + "step": 14789 + }, + { + "epoch": 1.6242038216560508, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.118170738220215, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7121105790138245, + "num_tokens": 382450711.0, + "step": 14790 + }, + { + "epoch": 1.6243136393586646, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2452900409698486, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7221341133117676, + "num_tokens": 382476821.0, + "step": 14791 + }, + { + "epoch": 1.6244234570612783, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.20145845413208, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7169103622436523, + "num_tokens": 382503988.0, + "step": 14792 + }, + { + "epoch": 1.624533274763892, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.041285753250122, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.711125373840332, + "num_tokens": 382533558.0, + "step": 14793 + }, + { + "epoch": 1.6246430924665056, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.411747694015503, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7167713642120361, + "num_tokens": 382557276.0, + "step": 14794 + }, + { + "epoch": 1.6247529101691192, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.346789836883545, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7066135406494141, + "num_tokens": 382582157.0, + "step": 14795 + }, + { + "epoch": 1.624862727871733, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0643885135650635, + "learning_rate": 1e-06, + "loss": 1.0429, + "mean_token_accuracy": 0.698966383934021, + "num_tokens": 382612789.0, + "step": 14796 + }, + { + "epoch": 1.6249725455743467, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9657301902770996, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7011227607727051, + "num_tokens": 382646381.0, + "step": 14797 + }, + { + "epoch": 1.6250823632769602, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3317558765411377, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7177740335464478, + "num_tokens": 382670294.0, + "step": 14798 + }, + { + "epoch": 1.6251921809795737, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2546863555908203, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.728766918182373, + "num_tokens": 382694925.0, + "step": 14799 + }, + { + "epoch": 1.6253019986821875, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.082124948501587, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7054198980331421, + "num_tokens": 382724352.0, + "step": 14800 + }, + { + "epoch": 1.6254118163848013, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1662707328796387, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7228112816810608, + "num_tokens": 382751825.0, + "step": 14801 + }, + { + "epoch": 1.625521634087415, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.289628028869629, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7221511602401733, + "num_tokens": 382777700.0, + "step": 14802 + }, + { + "epoch": 1.6256314517900285, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.379650831222534, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7146905660629272, + "num_tokens": 382801147.0, + "step": 14803 + }, + { + "epoch": 1.625741269492642, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4117860794067383, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7087385654449463, + "num_tokens": 382823777.0, + "step": 14804 + }, + { + "epoch": 1.6258510871952558, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1260623931884766, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.715364933013916, + "num_tokens": 382851424.0, + "step": 14805 + }, + { + "epoch": 1.6259609048978696, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1669554710388184, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7327084541320801, + "num_tokens": 382879282.0, + "step": 14806 + }, + { + "epoch": 1.6260707226004834, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.392404317855835, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7105138897895813, + "num_tokens": 382905436.0, + "step": 14807 + }, + { + "epoch": 1.6261805403030969, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1632745265960693, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7151164412498474, + "num_tokens": 382933757.0, + "step": 14808 + }, + { + "epoch": 1.6262903580057104, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.456632137298584, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7127060890197754, + "num_tokens": 382956239.0, + "step": 14809 + }, + { + "epoch": 1.6264001757083242, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9919588565826416, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7213355302810669, + "num_tokens": 382988165.0, + "step": 14810 + }, + { + "epoch": 1.626509993410938, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3555290699005127, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7168110609054565, + "num_tokens": 383014012.0, + "step": 14811 + }, + { + "epoch": 1.6266198111135515, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.300229072570801, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7085878252983093, + "num_tokens": 383039558.0, + "step": 14812 + }, + { + "epoch": 1.626729628816165, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.064204216003418, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7082114219665527, + "num_tokens": 383068124.0, + "step": 14813 + }, + { + "epoch": 1.6268394465187788, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.3581771850585938, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7263973951339722, + "num_tokens": 383091127.0, + "step": 14814 + }, + { + "epoch": 1.6269492642213925, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2139790058135986, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7230255603790283, + "num_tokens": 383117348.0, + "step": 14815 + }, + { + "epoch": 1.6270590819240063, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.183586597442627, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7267182469367981, + "num_tokens": 383144594.0, + "step": 14816 + }, + { + "epoch": 1.6271688996266198, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.2421553134918213, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7071805596351624, + "num_tokens": 383173211.0, + "step": 14817 + }, + { + "epoch": 1.6272787173292333, + "ewc_loss": 8.463859558105469e-06, + "grad_norm": 2.30741548538208, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7092341780662537, + "num_tokens": 383196574.0, + "step": 14818 + }, + { + "epoch": 1.627388535031847, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.140972852706909, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.713931143283844, + "num_tokens": 383225233.0, + "step": 14819 + }, + { + "epoch": 1.6274983527344609, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.454132080078125, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7286205291748047, + "num_tokens": 383246301.0, + "step": 14820 + }, + { + "epoch": 1.6276081704370746, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.077549934387207, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7115117907524109, + "num_tokens": 383276684.0, + "step": 14821 + }, + { + "epoch": 1.6277179881396882, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4328160285949707, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7343268394470215, + "num_tokens": 383298194.0, + "step": 14822 + }, + { + "epoch": 1.6278278058423017, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4217121601104736, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7495111227035522, + "num_tokens": 383319361.0, + "step": 14823 + }, + { + "epoch": 1.6279376235449154, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.993195652961731, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7256513833999634, + "num_tokens": 383349452.0, + "step": 14824 + }, + { + "epoch": 1.6280474412475292, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.210137367248535, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7091312408447266, + "num_tokens": 383380037.0, + "step": 14825 + }, + { + "epoch": 1.6281572589501427, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.215334177017212, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7387716174125671, + "num_tokens": 383405249.0, + "step": 14826 + }, + { + "epoch": 1.6282670766527563, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2361133098602295, + "learning_rate": 1e-06, + "loss": 0.9684, + "mean_token_accuracy": 0.7033374905586243, + "num_tokens": 383430728.0, + "step": 14827 + }, + { + "epoch": 1.62837689435537, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.028486967086792, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.7158780694007874, + "num_tokens": 383460877.0, + "step": 14828 + }, + { + "epoch": 1.6284867120579838, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4549741744995117, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7259804010391235, + "num_tokens": 383484181.0, + "step": 14829 + }, + { + "epoch": 1.6285965297605975, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0685338973999023, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7181342244148254, + "num_tokens": 383513323.0, + "step": 14830 + }, + { + "epoch": 1.628706347463211, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.131303548812866, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7417430281639099, + "num_tokens": 383538472.0, + "step": 14831 + }, + { + "epoch": 1.6288161651658246, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2605857849121094, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7322547435760498, + "num_tokens": 383563054.0, + "step": 14832 + }, + { + "epoch": 1.6289259828684384, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4493064880371094, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7144806385040283, + "num_tokens": 383586167.0, + "step": 14833 + }, + { + "epoch": 1.6290358005710521, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.303438901901245, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7396572828292847, + "num_tokens": 383609904.0, + "step": 14834 + }, + { + "epoch": 1.6291456182736657, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3962502479553223, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7232617139816284, + "num_tokens": 383631601.0, + "step": 14835 + }, + { + "epoch": 1.6292554359762794, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.389021873474121, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7359312176704407, + "num_tokens": 383653670.0, + "step": 14836 + }, + { + "epoch": 1.629365253678893, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0954253673553467, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7044185996055603, + "num_tokens": 383683590.0, + "step": 14837 + }, + { + "epoch": 1.6294750713815067, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.944061040878296, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7338268756866455, + "num_tokens": 383716512.0, + "step": 14838 + }, + { + "epoch": 1.6295848890841205, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.087919235229492, + "learning_rate": 1e-06, + "loss": 0.9651, + "mean_token_accuracy": 0.6995397210121155, + "num_tokens": 383748673.0, + "step": 14839 + }, + { + "epoch": 1.629694706786734, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0566864013671875, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7369462847709656, + "num_tokens": 383775985.0, + "step": 14840 + }, + { + "epoch": 1.6298045244893475, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1906585693359375, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7092226147651672, + "num_tokens": 383803403.0, + "step": 14841 + }, + { + "epoch": 1.6299143421919613, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1757922172546387, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7414284944534302, + "num_tokens": 383829954.0, + "step": 14842 + }, + { + "epoch": 1.630024159894575, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3383419513702393, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7174729108810425, + "num_tokens": 383852254.0, + "step": 14843 + }, + { + "epoch": 1.6301339775971888, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.5279695987701416, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7121994495391846, + "num_tokens": 383873300.0, + "step": 14844 + }, + { + "epoch": 1.6302437952998023, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3363964557647705, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7290785312652588, + "num_tokens": 383898671.0, + "step": 14845 + }, + { + "epoch": 1.6303536130024159, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2617852687835693, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7215243577957153, + "num_tokens": 383924445.0, + "step": 14846 + }, + { + "epoch": 1.6304634307050296, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.144824266433716, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7222214341163635, + "num_tokens": 383952449.0, + "step": 14847 + }, + { + "epoch": 1.6305732484076434, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3813509941101074, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7227116823196411, + "num_tokens": 383976174.0, + "step": 14848 + }, + { + "epoch": 1.630683066110257, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1891114711761475, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7233739495277405, + "num_tokens": 384003191.0, + "step": 14849 + }, + { + "epoch": 1.6307928838128707, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1345183849334717, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7176229953765869, + "num_tokens": 384030783.0, + "step": 14850 + }, + { + "epoch": 1.6309027015154842, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.051060914993286, + "learning_rate": 1e-06, + "loss": 0.9862, + "mean_token_accuracy": 0.6960941553115845, + "num_tokens": 384061920.0, + "step": 14851 + }, + { + "epoch": 1.631012519218098, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3152811527252197, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7352406978607178, + "num_tokens": 384086460.0, + "step": 14852 + }, + { + "epoch": 1.6311223369207117, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.334120512008667, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7307407855987549, + "num_tokens": 384110573.0, + "step": 14853 + }, + { + "epoch": 1.6312321546233253, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3102691173553467, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7269615530967712, + "num_tokens": 384135820.0, + "step": 14854 + }, + { + "epoch": 1.6313419723259388, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1459391117095947, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7162050008773804, + "num_tokens": 384161870.0, + "step": 14855 + }, + { + "epoch": 1.6314517900285526, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3334672451019287, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7151054739952087, + "num_tokens": 384185167.0, + "step": 14856 + }, + { + "epoch": 1.6315616077311663, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2533328533172607, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.734373927116394, + "num_tokens": 384209478.0, + "step": 14857 + }, + { + "epoch": 1.63167142543378, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.095310688018799, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7140057682991028, + "num_tokens": 384238523.0, + "step": 14858 + }, + { + "epoch": 1.6317812431363936, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.126742362976074, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7338335514068604, + "num_tokens": 384264608.0, + "step": 14859 + }, + { + "epoch": 1.6318910608390071, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1936519145965576, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7175354957580566, + "num_tokens": 384292304.0, + "step": 14860 + }, + { + "epoch": 1.632000878541621, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.139660358428955, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.6887552738189697, + "num_tokens": 384322313.0, + "step": 14861 + }, + { + "epoch": 1.6321106962442347, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.068747043609619, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7066333889961243, + "num_tokens": 384351098.0, + "step": 14862 + }, + { + "epoch": 1.6322205139468482, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2814371585845947, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.719427764415741, + "num_tokens": 384376359.0, + "step": 14863 + }, + { + "epoch": 1.6323303316494617, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0698888301849365, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7199442982673645, + "num_tokens": 384404628.0, + "step": 14864 + }, + { + "epoch": 1.6324401493520755, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2395262718200684, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7086575031280518, + "num_tokens": 384431541.0, + "step": 14865 + }, + { + "epoch": 1.6325499670546892, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1710543632507324, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6996294856071472, + "num_tokens": 384460588.0, + "step": 14866 + }, + { + "epoch": 1.632659784757303, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.536963701248169, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7347441911697388, + "num_tokens": 384481212.0, + "step": 14867 + }, + { + "epoch": 1.6327696024599165, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.141056776046753, + "learning_rate": 1e-06, + "loss": 0.7816, + "mean_token_accuracy": 0.7448276281356812, + "num_tokens": 384507741.0, + "step": 14868 + }, + { + "epoch": 1.63287942016253, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.176442861557007, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7246891260147095, + "num_tokens": 384532661.0, + "step": 14869 + }, + { + "epoch": 1.6329892378651438, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1599137783050537, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.6905110478401184, + "num_tokens": 384560894.0, + "step": 14870 + }, + { + "epoch": 1.6330990555677576, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.105994939804077, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.7058442831039429, + "num_tokens": 384588914.0, + "step": 14871 + }, + { + "epoch": 1.6332088732703713, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1735403537750244, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7080341577529907, + "num_tokens": 384616629.0, + "step": 14872 + }, + { + "epoch": 1.6333186909729849, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.09088134765625, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7084894180297852, + "num_tokens": 384647781.0, + "step": 14873 + }, + { + "epoch": 1.6334285086755984, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4672274589538574, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7105927467346191, + "num_tokens": 384670700.0, + "step": 14874 + }, + { + "epoch": 1.6335383263782122, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.6050572395324707, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7303771376609802, + "num_tokens": 384689528.0, + "step": 14875 + }, + { + "epoch": 1.633648144080826, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9010990858078003, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.7008376121520996, + "num_tokens": 384723911.0, + "step": 14876 + }, + { + "epoch": 1.6337579617834395, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1564128398895264, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.717161238193512, + "num_tokens": 384750167.0, + "step": 14877 + }, + { + "epoch": 1.633867779486053, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.125682830810547, + "learning_rate": 1e-06, + "loss": 0.7401, + "mean_token_accuracy": 0.7630373239517212, + "num_tokens": 384775381.0, + "step": 14878 + }, + { + "epoch": 1.6339775971886668, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.507338523864746, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7256193161010742, + "num_tokens": 384795819.0, + "step": 14879 + }, + { + "epoch": 1.6340874148912805, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3411803245544434, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7166973352432251, + "num_tokens": 384818800.0, + "step": 14880 + }, + { + "epoch": 1.6341972325938943, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.23673415184021, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7076147794723511, + "num_tokens": 384844704.0, + "step": 14881 + }, + { + "epoch": 1.6343070502965078, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2454752922058105, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.71954345703125, + "num_tokens": 384871538.0, + "step": 14882 + }, + { + "epoch": 1.6344168679991213, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2291419506073, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7056900262832642, + "num_tokens": 384899619.0, + "step": 14883 + }, + { + "epoch": 1.634526685701735, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.236083745956421, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.7182513475418091, + "num_tokens": 384927967.0, + "step": 14884 + }, + { + "epoch": 1.6346365034043489, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.533205986022949, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7414857149124146, + "num_tokens": 384950049.0, + "step": 14885 + }, + { + "epoch": 1.6347463211069626, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3405468463897705, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7205901145935059, + "num_tokens": 384974721.0, + "step": 14886 + }, + { + "epoch": 1.6348561388095761, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.5164339542388916, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7258071899414062, + "num_tokens": 384997463.0, + "step": 14887 + }, + { + "epoch": 1.6349659565121897, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.8315377235412598, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7314482927322388, + "num_tokens": 385015660.0, + "step": 14888 + }, + { + "epoch": 1.6350757742148034, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3697011470794678, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7306684851646423, + "num_tokens": 385040549.0, + "step": 14889 + }, + { + "epoch": 1.6351855919174172, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.398982286453247, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7548844814300537, + "num_tokens": 385061746.0, + "step": 14890 + }, + { + "epoch": 1.6352954096200307, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.5613274574279785, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.713626503944397, + "num_tokens": 385082821.0, + "step": 14891 + }, + { + "epoch": 1.6354052273226443, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9495145082473755, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7058340311050415, + "num_tokens": 385115248.0, + "step": 14892 + }, + { + "epoch": 1.635515045025258, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3687424659729004, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7228624820709229, + "num_tokens": 385139636.0, + "step": 14893 + }, + { + "epoch": 1.6356248627278718, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3022334575653076, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7147101163864136, + "num_tokens": 385164473.0, + "step": 14894 + }, + { + "epoch": 1.6357346804304855, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1862616539001465, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6964675188064575, + "num_tokens": 385192452.0, + "step": 14895 + }, + { + "epoch": 1.635844498133099, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.213235378265381, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.726065993309021, + "num_tokens": 385215938.0, + "step": 14896 + }, + { + "epoch": 1.6359543158357126, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.302140712738037, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7296132445335388, + "num_tokens": 385239699.0, + "step": 14897 + }, + { + "epoch": 1.6360641335383264, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0321273803710938, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.697533369064331, + "num_tokens": 385271485.0, + "step": 14898 + }, + { + "epoch": 1.6361739512409401, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.5331523418426514, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7128845453262329, + "num_tokens": 385292962.0, + "step": 14899 + }, + { + "epoch": 1.6362837689435537, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.321608304977417, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7221609354019165, + "num_tokens": 385318331.0, + "step": 14900 + }, + { + "epoch": 1.6363935866461674, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3992919921875, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7168377637863159, + "num_tokens": 385342660.0, + "step": 14901 + }, + { + "epoch": 1.636503404348781, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1332695484161377, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.722043514251709, + "num_tokens": 385371735.0, + "step": 14902 + }, + { + "epoch": 1.6366132220513947, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.27864408493042, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7306662201881409, + "num_tokens": 385396048.0, + "step": 14903 + }, + { + "epoch": 1.6367230397540085, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2750954627990723, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7149031758308411, + "num_tokens": 385421690.0, + "step": 14904 + }, + { + "epoch": 1.636832857456622, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4339070320129395, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7193393707275391, + "num_tokens": 385444381.0, + "step": 14905 + }, + { + "epoch": 1.6369426751592355, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2491674423217773, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7263777256011963, + "num_tokens": 385470342.0, + "step": 14906 + }, + { + "epoch": 1.6370524928618493, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.296678304672241, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7398049235343933, + "num_tokens": 385493926.0, + "step": 14907 + }, + { + "epoch": 1.637162310564463, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.281254529953003, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7235255837440491, + "num_tokens": 385520209.0, + "step": 14908 + }, + { + "epoch": 1.6372721282670768, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.210496664047241, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7184900045394897, + "num_tokens": 385546552.0, + "step": 14909 + }, + { + "epoch": 1.6373819459696903, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.161454677581787, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7064797878265381, + "num_tokens": 385573161.0, + "step": 14910 + }, + { + "epoch": 1.6374917636723039, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2761194705963135, + "learning_rate": 1e-06, + "loss": 0.7533, + "mean_token_accuracy": 0.7596675157546997, + "num_tokens": 385597104.0, + "step": 14911 + }, + { + "epoch": 1.6376015813749176, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.307058811187744, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7346962690353394, + "num_tokens": 385619874.0, + "step": 14912 + }, + { + "epoch": 1.6377113990775314, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3367018699645996, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.6978722810745239, + "num_tokens": 385646407.0, + "step": 14913 + }, + { + "epoch": 1.637821216780145, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4019885063171387, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7179804444313049, + "num_tokens": 385671319.0, + "step": 14914 + }, + { + "epoch": 1.6379310344827587, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.155811309814453, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.698630154132843, + "num_tokens": 385701247.0, + "step": 14915 + }, + { + "epoch": 1.6380408521853722, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.062333822250366, + "learning_rate": 1e-06, + "loss": 0.9859, + "mean_token_accuracy": 0.6978272795677185, + "num_tokens": 385732322.0, + "step": 14916 + }, + { + "epoch": 1.638150669887986, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4939374923706055, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7273993492126465, + "num_tokens": 385754702.0, + "step": 14917 + }, + { + "epoch": 1.6382604875905997, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0565106868743896, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.705929160118103, + "num_tokens": 385785464.0, + "step": 14918 + }, + { + "epoch": 1.6383703052932133, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3197224140167236, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7234939336776733, + "num_tokens": 385810503.0, + "step": 14919 + }, + { + "epoch": 1.6384801229958268, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.6784653663635254, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7340778112411499, + "num_tokens": 385829758.0, + "step": 14920 + }, + { + "epoch": 1.6385899406984406, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.6336753368377686, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7272381782531738, + "num_tokens": 385849368.0, + "step": 14921 + }, + { + "epoch": 1.6386997584010543, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0911707878112793, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7136960029602051, + "num_tokens": 385876430.0, + "step": 14922 + }, + { + "epoch": 1.638809576103668, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3298556804656982, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7322158217430115, + "num_tokens": 385899573.0, + "step": 14923 + }, + { + "epoch": 1.6389193938062816, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3995542526245117, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7261055707931519, + "num_tokens": 385922626.0, + "step": 14924 + }, + { + "epoch": 1.6390292115088951, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2354538440704346, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7256898880004883, + "num_tokens": 385947999.0, + "step": 14925 + }, + { + "epoch": 1.639139029211509, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.612741470336914, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.724844217300415, + "num_tokens": 385968647.0, + "step": 14926 + }, + { + "epoch": 1.6392488469141226, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.60385799407959, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7371601462364197, + "num_tokens": 385988918.0, + "step": 14927 + }, + { + "epoch": 1.6393586646167362, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.230482339859009, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7232810258865356, + "num_tokens": 386014473.0, + "step": 14928 + }, + { + "epoch": 1.6394684823193497, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.969434142112732, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7183824777603149, + "num_tokens": 386044083.0, + "step": 14929 + }, + { + "epoch": 1.6395783000219635, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.442763328552246, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7205710411071777, + "num_tokens": 386066503.0, + "step": 14930 + }, + { + "epoch": 1.6396881177245772, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 1.9109673500061035, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7052313089370728, + "num_tokens": 386100803.0, + "step": 14931 + }, + { + "epoch": 1.639797935427191, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0007593631744385, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.708256721496582, + "num_tokens": 386133291.0, + "step": 14932 + }, + { + "epoch": 1.6399077531298045, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0217230319976807, + "learning_rate": 1e-06, + "loss": 0.9646, + "mean_token_accuracy": 0.7000017166137695, + "num_tokens": 386165398.0, + "step": 14933 + }, + { + "epoch": 1.640017570832418, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.003206729888916, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7066148519515991, + "num_tokens": 386196459.0, + "step": 14934 + }, + { + "epoch": 1.6401273885350318, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3609681129455566, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7071921825408936, + "num_tokens": 386221217.0, + "step": 14935 + }, + { + "epoch": 1.6402372062376456, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.375239133834839, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7231917381286621, + "num_tokens": 386244454.0, + "step": 14936 + }, + { + "epoch": 1.6403470239402593, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3346705436706543, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7139072418212891, + "num_tokens": 386271152.0, + "step": 14937 + }, + { + "epoch": 1.6404568416428729, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.2488327026367188, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.707832932472229, + "num_tokens": 386296627.0, + "step": 14938 + }, + { + "epoch": 1.6405666593454864, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.193694591522217, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7275708317756653, + "num_tokens": 386322958.0, + "step": 14939 + }, + { + "epoch": 1.6406764770481002, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.30865478515625, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7350790500640869, + "num_tokens": 386345027.0, + "step": 14940 + }, + { + "epoch": 1.640786294750714, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1052474975585938, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7254874110221863, + "num_tokens": 386375754.0, + "step": 14941 + }, + { + "epoch": 1.6408961124533274, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.243358850479126, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7042341232299805, + "num_tokens": 386405499.0, + "step": 14942 + }, + { + "epoch": 1.641005930155941, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.0854642391204834, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7273696064949036, + "num_tokens": 386435079.0, + "step": 14943 + }, + { + "epoch": 1.6411157478585547, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1729440689086914, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7141677141189575, + "num_tokens": 386461966.0, + "step": 14944 + }, + { + "epoch": 1.6412255655611685, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1917521953582764, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.72591233253479, + "num_tokens": 386487258.0, + "step": 14945 + }, + { + "epoch": 1.6413353832637823, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.4233505725860596, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.710264265537262, + "num_tokens": 386511094.0, + "step": 14946 + }, + { + "epoch": 1.6414452009663958, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1751301288604736, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7101482152938843, + "num_tokens": 386537644.0, + "step": 14947 + }, + { + "epoch": 1.6415550186690093, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.398621082305908, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.710050106048584, + "num_tokens": 386563414.0, + "step": 14948 + }, + { + "epoch": 1.641664836371623, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.3138294219970703, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7071419954299927, + "num_tokens": 386588261.0, + "step": 14949 + }, + { + "epoch": 1.6417746540742368, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1028740406036377, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7265647649765015, + "num_tokens": 386618288.0, + "step": 14950 + }, + { + "epoch": 1.6418844717768504, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.388817071914673, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7127149105072021, + "num_tokens": 386641683.0, + "step": 14951 + }, + { + "epoch": 1.6419942894794641, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.173915386199951, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7349805235862732, + "num_tokens": 386669054.0, + "step": 14952 + }, + { + "epoch": 1.6421041071820777, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.1008460521698, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7266476154327393, + "num_tokens": 386698654.0, + "step": 14953 + }, + { + "epoch": 1.6422139248846914, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.263751268386841, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7094244360923767, + "num_tokens": 386725386.0, + "step": 14954 + }, + { + "epoch": 1.6423237425873052, + "ewc_loss": 8.52346420288086e-06, + "grad_norm": 2.320502996444702, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7212550640106201, + "num_tokens": 386748385.0, + "step": 14955 + }, + { + "epoch": 1.6424335602899187, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1865673065185547, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7261857390403748, + "num_tokens": 386775453.0, + "step": 14956 + }, + { + "epoch": 1.6425433779925322, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1719231605529785, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7323773503303528, + "num_tokens": 386802118.0, + "step": 14957 + }, + { + "epoch": 1.642653195695146, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.133704662322998, + "learning_rate": 1e-06, + "loss": 1.0008, + "mean_token_accuracy": 0.6965920925140381, + "num_tokens": 386833173.0, + "step": 14958 + }, + { + "epoch": 1.6427630133977598, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.198930263519287, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7079461812973022, + "num_tokens": 386858346.0, + "step": 14959 + }, + { + "epoch": 1.6428728311003735, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2216849327087402, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.7207868099212646, + "num_tokens": 386886388.0, + "step": 14960 + }, + { + "epoch": 1.642982648802987, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5560333728790283, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7413840889930725, + "num_tokens": 386906007.0, + "step": 14961 + }, + { + "epoch": 1.6430924665056006, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0960450172424316, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7089051008224487, + "num_tokens": 386933428.0, + "step": 14962 + }, + { + "epoch": 1.6432022842082143, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.088573694229126, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7055995464324951, + "num_tokens": 386964853.0, + "step": 14963 + }, + { + "epoch": 1.643312101910828, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0874364376068115, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.706458568572998, + "num_tokens": 386993593.0, + "step": 14964 + }, + { + "epoch": 1.6434219196134416, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3396849632263184, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7091250419616699, + "num_tokens": 387017013.0, + "step": 14965 + }, + { + "epoch": 1.6435317373160554, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4379830360412598, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7409646511077881, + "num_tokens": 387037142.0, + "step": 14966 + }, + { + "epoch": 1.643641555018669, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4698328971862793, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7329801917076111, + "num_tokens": 387058297.0, + "step": 14967 + }, + { + "epoch": 1.6437513727212827, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.424562454223633, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7234981060028076, + "num_tokens": 387079291.0, + "step": 14968 + }, + { + "epoch": 1.6438611904238964, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5211234092712402, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7135192155838013, + "num_tokens": 387100506.0, + "step": 14969 + }, + { + "epoch": 1.64397100812651, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.177100658416748, + "learning_rate": 1e-06, + "loss": 0.9696, + "mean_token_accuracy": 0.7066518068313599, + "num_tokens": 387128650.0, + "step": 14970 + }, + { + "epoch": 1.6440808258291235, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.416259765625, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7246894836425781, + "num_tokens": 387151054.0, + "step": 14971 + }, + { + "epoch": 1.6441906435317373, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.415041923522949, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.724034309387207, + "num_tokens": 387173317.0, + "step": 14972 + }, + { + "epoch": 1.644300461234351, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4287047386169434, + "learning_rate": 1e-06, + "loss": 0.977, + "mean_token_accuracy": 0.7090578079223633, + "num_tokens": 387195398.0, + "step": 14973 + }, + { + "epoch": 1.6444102789369648, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5386080741882324, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7282552719116211, + "num_tokens": 387216806.0, + "step": 14974 + }, + { + "epoch": 1.6445200966395783, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.117435932159424, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7102484703063965, + "num_tokens": 387244637.0, + "step": 14975 + }, + { + "epoch": 1.6446299143421919, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.206441640853882, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7193890810012817, + "num_tokens": 387271443.0, + "step": 14976 + }, + { + "epoch": 1.6447397320448056, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1382784843444824, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7050357460975647, + "num_tokens": 387300901.0, + "step": 14977 + }, + { + "epoch": 1.6448495497474194, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.826420545578003, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7451955676078796, + "num_tokens": 387319865.0, + "step": 14978 + }, + { + "epoch": 1.644959367450033, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.306169271469116, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7157008647918701, + "num_tokens": 387343059.0, + "step": 14979 + }, + { + "epoch": 1.6450691851526464, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.12864089012146, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7225732803344727, + "num_tokens": 387372396.0, + "step": 14980 + }, + { + "epoch": 1.6451790028552602, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.8539321422576904, + "learning_rate": 1e-06, + "loss": 1.0238, + "mean_token_accuracy": 0.6803560256958008, + "num_tokens": 387405565.0, + "step": 14981 + }, + { + "epoch": 1.645288820557874, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5031254291534424, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7413855195045471, + "num_tokens": 387425517.0, + "step": 14982 + }, + { + "epoch": 1.6453986382604877, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.544907808303833, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7424367666244507, + "num_tokens": 387445500.0, + "step": 14983 + }, + { + "epoch": 1.6455084559631012, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.599179267883301, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7339521646499634, + "num_tokens": 387464702.0, + "step": 14984 + }, + { + "epoch": 1.6456182736657148, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2694685459136963, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7099306583404541, + "num_tokens": 387489147.0, + "step": 14985 + }, + { + "epoch": 1.6457280913683285, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1513495445251465, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7076398730278015, + "num_tokens": 387516359.0, + "step": 14986 + }, + { + "epoch": 1.6458379090709423, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3789141178131104, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7195214033126831, + "num_tokens": 387539169.0, + "step": 14987 + }, + { + "epoch": 1.645947726773556, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.473032236099243, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7442519068717957, + "num_tokens": 387561416.0, + "step": 14988 + }, + { + "epoch": 1.6460575444761696, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.163053274154663, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.6948255896568298, + "num_tokens": 387588704.0, + "step": 14989 + }, + { + "epoch": 1.6461673621787831, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5993499755859375, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7183212041854858, + "num_tokens": 387609671.0, + "step": 14990 + }, + { + "epoch": 1.6462771798813969, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.965391993522644, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.6995540857315063, + "num_tokens": 387644406.0, + "step": 14991 + }, + { + "epoch": 1.6463869975840106, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9705203771591187, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7196077704429626, + "num_tokens": 387675385.0, + "step": 14992 + }, + { + "epoch": 1.6464968152866242, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0600333213806152, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7301555871963501, + "num_tokens": 387702653.0, + "step": 14993 + }, + { + "epoch": 1.6466066329892377, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.324773073196411, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7404057383537292, + "num_tokens": 387726744.0, + "step": 14994 + }, + { + "epoch": 1.6467164506918515, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4195163249969482, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7285823225975037, + "num_tokens": 387749051.0, + "step": 14995 + }, + { + "epoch": 1.6468262683944652, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.274799346923828, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7124650478363037, + "num_tokens": 387773399.0, + "step": 14996 + }, + { + "epoch": 1.646936086097079, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.326364040374756, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7233636379241943, + "num_tokens": 387798562.0, + "step": 14997 + }, + { + "epoch": 1.6470459037996925, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3385374546051025, + "learning_rate": 1e-06, + "loss": 0.9739, + "mean_token_accuracy": 0.7146435976028442, + "num_tokens": 387825123.0, + "step": 14998 + }, + { + "epoch": 1.647155721502306, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0620036125183105, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7065612077713013, + "num_tokens": 387855785.0, + "step": 14999 + }, + { + "epoch": 1.6472655392049198, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1377644538879395, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7133070230484009, + "num_tokens": 387885037.0, + "step": 15000 + }, + { + "epoch": 1.6473753569075336, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.7177090644836426, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7374770045280457, + "num_tokens": 387903513.0, + "step": 15001 + }, + { + "epoch": 1.6474851746101473, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9801018238067627, + "learning_rate": 1e-06, + "loss": 0.9761, + "mean_token_accuracy": 0.7002348899841309, + "num_tokens": 387935721.0, + "step": 15002 + }, + { + "epoch": 1.6475949923127609, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1665241718292236, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7252715826034546, + "num_tokens": 387960680.0, + "step": 15003 + }, + { + "epoch": 1.6477048100153744, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.162703275680542, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7006350755691528, + "num_tokens": 387990376.0, + "step": 15004 + }, + { + "epoch": 1.6478146277179881, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.315401315689087, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7329437136650085, + "num_tokens": 388013280.0, + "step": 15005 + }, + { + "epoch": 1.647924445420602, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0980029106140137, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7041341662406921, + "num_tokens": 388044228.0, + "step": 15006 + }, + { + "epoch": 1.6480342631232154, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.154167652130127, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7199248671531677, + "num_tokens": 388069822.0, + "step": 15007 + }, + { + "epoch": 1.648144080825829, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9658366441726685, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.7058675289154053, + "num_tokens": 388102749.0, + "step": 15008 + }, + { + "epoch": 1.6482538985284427, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.613949775695801, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7372751235961914, + "num_tokens": 388122017.0, + "step": 15009 + }, + { + "epoch": 1.6483637162310565, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.435593605041504, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7090713977813721, + "num_tokens": 388146361.0, + "step": 15010 + }, + { + "epoch": 1.6484735339336702, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.370136260986328, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.698538601398468, + "num_tokens": 388170621.0, + "step": 15011 + }, + { + "epoch": 1.6485833516362838, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2065885066986084, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7228301763534546, + "num_tokens": 388198618.0, + "step": 15012 + }, + { + "epoch": 1.6486931693388973, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.255225896835327, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7108023166656494, + "num_tokens": 388224203.0, + "step": 15013 + }, + { + "epoch": 1.648802987041511, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1161084175109863, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.694858193397522, + "num_tokens": 388256049.0, + "step": 15014 + }, + { + "epoch": 1.6489128047441248, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2853574752807617, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7218790650367737, + "num_tokens": 388280606.0, + "step": 15015 + }, + { + "epoch": 1.6490226224467384, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.477263927459717, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7043514251708984, + "num_tokens": 388305205.0, + "step": 15016 + }, + { + "epoch": 1.6491324401493521, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1037023067474365, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7154989838600159, + "num_tokens": 388335252.0, + "step": 15017 + }, + { + "epoch": 1.6492422578519657, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5346314907073975, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.731168270111084, + "num_tokens": 388358098.0, + "step": 15018 + }, + { + "epoch": 1.6493520755545794, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.289497137069702, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7447240352630615, + "num_tokens": 388381800.0, + "step": 15019 + }, + { + "epoch": 1.6494618932571932, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2738943099975586, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7055442333221436, + "num_tokens": 388406651.0, + "step": 15020 + }, + { + "epoch": 1.6495717109598067, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.33651065826416, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7477272748947144, + "num_tokens": 388431031.0, + "step": 15021 + }, + { + "epoch": 1.6496815286624202, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1385855674743652, + "learning_rate": 1e-06, + "loss": 1.0006, + "mean_token_accuracy": 0.7039520144462585, + "num_tokens": 388459428.0, + "step": 15022 + }, + { + "epoch": 1.649791346365034, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.103671073913574, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7296514511108398, + "num_tokens": 388488077.0, + "step": 15023 + }, + { + "epoch": 1.6499011640676478, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.110398530960083, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7206634283065796, + "num_tokens": 388516556.0, + "step": 15024 + }, + { + "epoch": 1.6500109817702615, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3938510417938232, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.7084925174713135, + "num_tokens": 388542273.0, + "step": 15025 + }, + { + "epoch": 1.650120799472875, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1892266273498535, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7222973108291626, + "num_tokens": 388568517.0, + "step": 15026 + }, + { + "epoch": 1.6502306171754886, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3432350158691406, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7249423265457153, + "num_tokens": 388590410.0, + "step": 15027 + }, + { + "epoch": 1.6503404348781023, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.236326217651367, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7467833161354065, + "num_tokens": 388614421.0, + "step": 15028 + }, + { + "epoch": 1.650450252580716, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.007451295852661, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7339038848876953, + "num_tokens": 388643445.0, + "step": 15029 + }, + { + "epoch": 1.6505600702833296, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.259089231491089, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.6998838782310486, + "num_tokens": 388669637.0, + "step": 15030 + }, + { + "epoch": 1.6506698879859434, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.350069284439087, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7299631237983704, + "num_tokens": 388692316.0, + "step": 15031 + }, + { + "epoch": 1.650779705688557, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.480228900909424, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7389761805534363, + "num_tokens": 388714361.0, + "step": 15032 + }, + { + "epoch": 1.6508895233911707, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1285011768341064, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7141698598861694, + "num_tokens": 388742386.0, + "step": 15033 + }, + { + "epoch": 1.6509993410937844, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.44143009185791, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7115584015846252, + "num_tokens": 388765011.0, + "step": 15034 + }, + { + "epoch": 1.651109158796398, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2947237491607666, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7108877897262573, + "num_tokens": 388790939.0, + "step": 15035 + }, + { + "epoch": 1.6512189764990115, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3144729137420654, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7210662364959717, + "num_tokens": 388816139.0, + "step": 15036 + }, + { + "epoch": 1.6513287942016253, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3501274585723877, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7180991172790527, + "num_tokens": 388841709.0, + "step": 15037 + }, + { + "epoch": 1.651438611904239, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.337874174118042, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7120189070701599, + "num_tokens": 388868760.0, + "step": 15038 + }, + { + "epoch": 1.6515484296068528, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.396327495574951, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7283964157104492, + "num_tokens": 388889984.0, + "step": 15039 + }, + { + "epoch": 1.6516582473094663, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.140183210372925, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.725205659866333, + "num_tokens": 388915538.0, + "step": 15040 + }, + { + "epoch": 1.6517680650120798, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5156311988830566, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7408010959625244, + "num_tokens": 388935827.0, + "step": 15041 + }, + { + "epoch": 1.6518778827146936, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3835296630859375, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7162328958511353, + "num_tokens": 388959234.0, + "step": 15042 + }, + { + "epoch": 1.6519877004173074, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.268465518951416, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7096986174583435, + "num_tokens": 388985059.0, + "step": 15043 + }, + { + "epoch": 1.652097518119921, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4088387489318848, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6972827911376953, + "num_tokens": 389008956.0, + "step": 15044 + }, + { + "epoch": 1.6522073358225344, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2812325954437256, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7036209106445312, + "num_tokens": 389035428.0, + "step": 15045 + }, + { + "epoch": 1.6523171535251482, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.251133918762207, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7177526950836182, + "num_tokens": 389060170.0, + "step": 15046 + }, + { + "epoch": 1.652426971227762, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1629433631896973, + "learning_rate": 1e-06, + "loss": 0.968, + "mean_token_accuracy": 0.7061519622802734, + "num_tokens": 389087000.0, + "step": 15047 + }, + { + "epoch": 1.6525367889303757, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.357013463973999, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7034682035446167, + "num_tokens": 389110650.0, + "step": 15048 + }, + { + "epoch": 1.6526466066329892, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.495072364807129, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.728286862373352, + "num_tokens": 389131713.0, + "step": 15049 + }, + { + "epoch": 1.6527564243356028, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3284599781036377, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7010531425476074, + "num_tokens": 389155984.0, + "step": 15050 + }, + { + "epoch": 1.6528662420382165, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.6021218299865723, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.722567081451416, + "num_tokens": 389175024.0, + "step": 15051 + }, + { + "epoch": 1.6529760597408303, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3647069931030273, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7232393026351929, + "num_tokens": 389198396.0, + "step": 15052 + }, + { + "epoch": 1.653085877443444, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2191388607025146, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7166433930397034, + "num_tokens": 389224011.0, + "step": 15053 + }, + { + "epoch": 1.6531956951460576, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.097937822341919, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7266314029693604, + "num_tokens": 389250356.0, + "step": 15054 + }, + { + "epoch": 1.653305512848671, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2676026821136475, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7261568307876587, + "num_tokens": 389274008.0, + "step": 15055 + }, + { + "epoch": 1.6534153305512849, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4025368690490723, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6862704753875732, + "num_tokens": 389297790.0, + "step": 15056 + }, + { + "epoch": 1.6535251482538986, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.916914463043213, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.7047095894813538, + "num_tokens": 389329161.0, + "step": 15057 + }, + { + "epoch": 1.6536349659565122, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.17911696434021, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7284253835678101, + "num_tokens": 389355768.0, + "step": 15058 + }, + { + "epoch": 1.6537447836591257, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9722726345062256, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.7161529660224915, + "num_tokens": 389386013.0, + "step": 15059 + }, + { + "epoch": 1.6538546013617395, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2675399780273438, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7058413624763489, + "num_tokens": 389412209.0, + "step": 15060 + }, + { + "epoch": 1.6539644190643532, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.970335602760315, + "learning_rate": 1e-06, + "loss": 0.9607, + "mean_token_accuracy": 0.7012869119644165, + "num_tokens": 389446399.0, + "step": 15061 + }, + { + "epoch": 1.654074236766967, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1586246490478516, + "learning_rate": 1e-06, + "loss": 0.7542, + "mean_token_accuracy": 0.7579478621482849, + "num_tokens": 389471148.0, + "step": 15062 + }, + { + "epoch": 1.6541840544695805, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.8659148216247559, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6978697180747986, + "num_tokens": 389508397.0, + "step": 15063 + }, + { + "epoch": 1.654293872172194, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2455337047576904, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.7130434513092041, + "num_tokens": 389535562.0, + "step": 15064 + }, + { + "epoch": 1.6544036898748078, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2099616527557373, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7361915707588196, + "num_tokens": 389563004.0, + "step": 15065 + }, + { + "epoch": 1.6545135075774215, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.312901020050049, + "learning_rate": 1e-06, + "loss": 0.7551, + "mean_token_accuracy": 0.7576709389686584, + "num_tokens": 389586088.0, + "step": 15066 + }, + { + "epoch": 1.6546233252800353, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.190643548965454, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7344995737075806, + "num_tokens": 389612021.0, + "step": 15067 + }, + { + "epoch": 1.6547331429826488, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.943354845046997, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7242751121520996, + "num_tokens": 389642765.0, + "step": 15068 + }, + { + "epoch": 1.6548429606852624, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2815754413604736, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7407082319259644, + "num_tokens": 389665514.0, + "step": 15069 + }, + { + "epoch": 1.6549527783878761, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2038819789886475, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7100437879562378, + "num_tokens": 389692152.0, + "step": 15070 + }, + { + "epoch": 1.65506259609049, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3922808170318604, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7234084010124207, + "num_tokens": 389717493.0, + "step": 15071 + }, + { + "epoch": 1.6551724137931034, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3579564094543457, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7095131874084473, + "num_tokens": 389741451.0, + "step": 15072 + }, + { + "epoch": 1.655282231495717, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3483707904815674, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7195929884910583, + "num_tokens": 389767007.0, + "step": 15073 + }, + { + "epoch": 1.6553920491983307, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.6849889755249023, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7007104754447937, + "num_tokens": 389788434.0, + "step": 15074 + }, + { + "epoch": 1.6555018669009445, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1546316146850586, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7025718688964844, + "num_tokens": 389815287.0, + "step": 15075 + }, + { + "epoch": 1.6556116846035582, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.452667474746704, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7394126653671265, + "num_tokens": 389836107.0, + "step": 15076 + }, + { + "epoch": 1.6557215023061718, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.245913028717041, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7279625535011292, + "num_tokens": 389860566.0, + "step": 15077 + }, + { + "epoch": 1.6558313200087853, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9579027891159058, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7300773859024048, + "num_tokens": 389891930.0, + "step": 15078 + }, + { + "epoch": 1.655941137711399, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.320425271987915, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7117077708244324, + "num_tokens": 389916670.0, + "step": 15079 + }, + { + "epoch": 1.6560509554140128, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.089292526245117, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.6981504559516907, + "num_tokens": 389945835.0, + "step": 15080 + }, + { + "epoch": 1.6561607731166264, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.162235736846924, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7556339502334595, + "num_tokens": 389972301.0, + "step": 15081 + }, + { + "epoch": 1.65627059081924, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.618816614151001, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7498028874397278, + "num_tokens": 389992002.0, + "step": 15082 + }, + { + "epoch": 1.6563804085218536, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.313755750656128, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7018035650253296, + "num_tokens": 390018505.0, + "step": 15083 + }, + { + "epoch": 1.6564902262244674, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4211435317993164, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7315264940261841, + "num_tokens": 390040048.0, + "step": 15084 + }, + { + "epoch": 1.6566000439270812, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.069103240966797, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7096657156944275, + "num_tokens": 390068920.0, + "step": 15085 + }, + { + "epoch": 1.6567098616296947, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3515682220458984, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.697799563407898, + "num_tokens": 390095080.0, + "step": 15086 + }, + { + "epoch": 1.6568196793323082, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0927200317382812, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7261713147163391, + "num_tokens": 390123902.0, + "step": 15087 + }, + { + "epoch": 1.656929497034922, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.210179090499878, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.704905092716217, + "num_tokens": 390149633.0, + "step": 15088 + }, + { + "epoch": 1.6570393147375357, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.438173294067383, + "learning_rate": 1e-06, + "loss": 1.0252, + "mean_token_accuracy": 0.7025607824325562, + "num_tokens": 390173020.0, + "step": 15089 + }, + { + "epoch": 1.6571491324401495, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.321882724761963, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7125328779220581, + "num_tokens": 390198722.0, + "step": 15090 + }, + { + "epoch": 1.657258950142763, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.228541135787964, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7298159003257751, + "num_tokens": 390226112.0, + "step": 15091 + }, + { + "epoch": 1.6573687678453766, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2249326705932617, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7366626262664795, + "num_tokens": 390250826.0, + "step": 15092 + }, + { + "epoch": 1.6574785855479903, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.36753249168396, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7265392541885376, + "num_tokens": 390273521.0, + "step": 15093 + }, + { + "epoch": 1.657588403250604, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2833220958709717, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7251546382904053, + "num_tokens": 390300333.0, + "step": 15094 + }, + { + "epoch": 1.6576982209532176, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1648290157318115, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7498565912246704, + "num_tokens": 390327090.0, + "step": 15095 + }, + { + "epoch": 1.6578080386558314, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3731720447540283, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.7058842778205872, + "num_tokens": 390352112.0, + "step": 15096 + }, + { + "epoch": 1.657917856358445, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1871376037597656, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7385399341583252, + "num_tokens": 390379128.0, + "step": 15097 + }, + { + "epoch": 1.6580276740610587, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3629605770111084, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7194856405258179, + "num_tokens": 390402702.0, + "step": 15098 + }, + { + "epoch": 1.6581374917636724, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.6107373237609863, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7091325521469116, + "num_tokens": 390425482.0, + "step": 15099 + }, + { + "epoch": 1.658247309466286, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3692679405212402, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7236342430114746, + "num_tokens": 390448680.0, + "step": 15100 + }, + { + "epoch": 1.6583571271688995, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1456680297851562, + "learning_rate": 1e-06, + "loss": 0.9928, + "mean_token_accuracy": 0.6968072652816772, + "num_tokens": 390475776.0, + "step": 15101 + }, + { + "epoch": 1.6584669448715132, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.344299077987671, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7212660908699036, + "num_tokens": 390499965.0, + "step": 15102 + }, + { + "epoch": 1.658576762574127, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.245306968688965, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7297143936157227, + "num_tokens": 390525663.0, + "step": 15103 + }, + { + "epoch": 1.6586865802767408, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3535242080688477, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7074178457260132, + "num_tokens": 390549509.0, + "step": 15104 + }, + { + "epoch": 1.6587963979793543, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2216153144836426, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.6996471881866455, + "num_tokens": 390577154.0, + "step": 15105 + }, + { + "epoch": 1.6589062156819678, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0921664237976074, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7124036550521851, + "num_tokens": 390606950.0, + "step": 15106 + }, + { + "epoch": 1.6590160333845816, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.409445285797119, + "learning_rate": 1e-06, + "loss": 0.9913, + "mean_token_accuracy": 0.7064486145973206, + "num_tokens": 390630059.0, + "step": 15107 + }, + { + "epoch": 1.6591258510871953, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1581814289093018, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7183488607406616, + "num_tokens": 390656855.0, + "step": 15108 + }, + { + "epoch": 1.6592356687898089, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3857905864715576, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7464861273765564, + "num_tokens": 390681132.0, + "step": 15109 + }, + { + "epoch": 1.6593454864924224, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.118356227874756, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7182387709617615, + "num_tokens": 390710201.0, + "step": 15110 + }, + { + "epoch": 1.6594553041950362, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2046566009521484, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7389729022979736, + "num_tokens": 390735305.0, + "step": 15111 + }, + { + "epoch": 1.65956512189765, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4762110710144043, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7181333899497986, + "num_tokens": 390760313.0, + "step": 15112 + }, + { + "epoch": 1.6596749396002637, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3189291954040527, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7261891961097717, + "num_tokens": 390782847.0, + "step": 15113 + }, + { + "epoch": 1.6597847573028772, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.306466817855835, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7155650854110718, + "num_tokens": 390808668.0, + "step": 15114 + }, + { + "epoch": 1.6598945750054908, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2904205322265625, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7116312980651855, + "num_tokens": 390835133.0, + "step": 15115 + }, + { + "epoch": 1.6600043927081045, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1827495098114014, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.728210985660553, + "num_tokens": 390860844.0, + "step": 15116 + }, + { + "epoch": 1.6601142104107183, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.6354804039001465, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7192803621292114, + "num_tokens": 390881447.0, + "step": 15117 + }, + { + "epoch": 1.660224028113332, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9823118448257446, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7163949012756348, + "num_tokens": 390914311.0, + "step": 15118 + }, + { + "epoch": 1.6603338458159456, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4534857273101807, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7425004839897156, + "num_tokens": 390936130.0, + "step": 15119 + }, + { + "epoch": 1.660443663518559, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.029393434524536, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7331277132034302, + "num_tokens": 390963900.0, + "step": 15120 + }, + { + "epoch": 1.6605534812211729, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1499595642089844, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7417312264442444, + "num_tokens": 390988318.0, + "step": 15121 + }, + { + "epoch": 1.6606632989237866, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3148834705352783, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7179916501045227, + "num_tokens": 391013088.0, + "step": 15122 + }, + { + "epoch": 1.6607731166264001, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2710773944854736, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.71332848072052, + "num_tokens": 391038533.0, + "step": 15123 + }, + { + "epoch": 1.6608829343290137, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3118104934692383, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7167153358459473, + "num_tokens": 391063810.0, + "step": 15124 + }, + { + "epoch": 1.6609927520316274, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.275212287902832, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.7028778195381165, + "num_tokens": 391089542.0, + "step": 15125 + }, + { + "epoch": 1.6611025697342412, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2605180740356445, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7428262829780579, + "num_tokens": 391113833.0, + "step": 15126 + }, + { + "epoch": 1.661212387436855, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.344653606414795, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.728469729423523, + "num_tokens": 391138425.0, + "step": 15127 + }, + { + "epoch": 1.6613222051394685, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.360987901687622, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7434082627296448, + "num_tokens": 391161315.0, + "step": 15128 + }, + { + "epoch": 1.661432022842082, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.26745343208313, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7274836897850037, + "num_tokens": 391185436.0, + "step": 15129 + }, + { + "epoch": 1.6615418405446958, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3979220390319824, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7250452637672424, + "num_tokens": 391209066.0, + "step": 15130 + }, + { + "epoch": 1.6616516582473095, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.567845106124878, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7230846881866455, + "num_tokens": 391229470.0, + "step": 15131 + }, + { + "epoch": 1.661761475949923, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.036674976348877, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7185046672821045, + "num_tokens": 391259599.0, + "step": 15132 + }, + { + "epoch": 1.6618712936525368, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.289034366607666, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7290514707565308, + "num_tokens": 391284759.0, + "step": 15133 + }, + { + "epoch": 1.6619811113551504, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.280846118927002, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7235425710678101, + "num_tokens": 391310964.0, + "step": 15134 + }, + { + "epoch": 1.6620909290577641, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.264538288116455, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7267357110977173, + "num_tokens": 391336718.0, + "step": 15135 + }, + { + "epoch": 1.6622007467603779, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1965763568878174, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7074227333068848, + "num_tokens": 391363522.0, + "step": 15136 + }, + { + "epoch": 1.6623105644629914, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.128016710281372, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7233167886734009, + "num_tokens": 391390172.0, + "step": 15137 + }, + { + "epoch": 1.662420382165605, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.177145481109619, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7129189372062683, + "num_tokens": 391416941.0, + "step": 15138 + }, + { + "epoch": 1.6625301998682187, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0627708435058594, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.7050385475158691, + "num_tokens": 391444599.0, + "step": 15139 + }, + { + "epoch": 1.6626400175708325, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1840767860412598, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7052028179168701, + "num_tokens": 391471679.0, + "step": 15140 + }, + { + "epoch": 1.6627498352734462, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1002471446990967, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7157211899757385, + "num_tokens": 391500427.0, + "step": 15141 + }, + { + "epoch": 1.6628596529760598, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2006330490112305, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7512831091880798, + "num_tokens": 391526531.0, + "step": 15142 + }, + { + "epoch": 1.6629694706786733, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1946091651916504, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.747191309928894, + "num_tokens": 391549471.0, + "step": 15143 + }, + { + "epoch": 1.663079288381287, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.305396556854248, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7283076643943787, + "num_tokens": 391572180.0, + "step": 15144 + }, + { + "epoch": 1.6631891060839008, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.224632978439331, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7236137390136719, + "num_tokens": 391598234.0, + "step": 15145 + }, + { + "epoch": 1.6632989237865143, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0960824489593506, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7033675312995911, + "num_tokens": 391627804.0, + "step": 15146 + }, + { + "epoch": 1.663408741489128, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.602501392364502, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7306275367736816, + "num_tokens": 391650198.0, + "step": 15147 + }, + { + "epoch": 1.6635185591917416, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1303415298461914, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7207310199737549, + "num_tokens": 391679974.0, + "step": 15148 + }, + { + "epoch": 1.6636283768943554, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2498598098754883, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7155365943908691, + "num_tokens": 391706421.0, + "step": 15149 + }, + { + "epoch": 1.6637381945969691, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.038940906524658, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7143317461013794, + "num_tokens": 391737789.0, + "step": 15150 + }, + { + "epoch": 1.6638480122995827, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.005631685256958, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7322313785552979, + "num_tokens": 391767656.0, + "step": 15151 + }, + { + "epoch": 1.6639578300021962, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3828630447387695, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7280304431915283, + "num_tokens": 391788356.0, + "step": 15152 + }, + { + "epoch": 1.66406764770481, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.185142755508423, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7159529328346252, + "num_tokens": 391817753.0, + "step": 15153 + }, + { + "epoch": 1.6641774654074237, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3766143321990967, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.722011387348175, + "num_tokens": 391842091.0, + "step": 15154 + }, + { + "epoch": 1.6642872831100375, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.152536392211914, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7069658041000366, + "num_tokens": 391870270.0, + "step": 15155 + }, + { + "epoch": 1.664397100812651, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9120818376541138, + "learning_rate": 1e-06, + "loss": 1.0702, + "mean_token_accuracy": 0.6810184121131897, + "num_tokens": 391903803.0, + "step": 15156 + }, + { + "epoch": 1.6645069185152646, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3631160259246826, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7282512784004211, + "num_tokens": 391926968.0, + "step": 15157 + }, + { + "epoch": 1.6646167362178783, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2045044898986816, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7053423523902893, + "num_tokens": 391954503.0, + "step": 15158 + }, + { + "epoch": 1.664726553920492, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.5295462608337402, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7193312048912048, + "num_tokens": 391976220.0, + "step": 15159 + }, + { + "epoch": 1.6648363716231056, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1491293907165527, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7151461839675903, + "num_tokens": 392003237.0, + "step": 15160 + }, + { + "epoch": 1.6649461893257194, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9719542264938354, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7222815752029419, + "num_tokens": 392032988.0, + "step": 15161 + }, + { + "epoch": 1.665056007028333, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.224977731704712, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.738898754119873, + "num_tokens": 392057474.0, + "step": 15162 + }, + { + "epoch": 1.6651658247309467, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.192426919937134, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7197313904762268, + "num_tokens": 392084909.0, + "step": 15163 + }, + { + "epoch": 1.6652756424335604, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2045488357543945, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7074219584465027, + "num_tokens": 392111931.0, + "step": 15164 + }, + { + "epoch": 1.665385460136174, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1637442111968994, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7336296439170837, + "num_tokens": 392138413.0, + "step": 15165 + }, + { + "epoch": 1.6654952778387875, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2129554748535156, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7254409193992615, + "num_tokens": 392164030.0, + "step": 15166 + }, + { + "epoch": 1.6656050955414012, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2662353515625, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7198671698570251, + "num_tokens": 392189971.0, + "step": 15167 + }, + { + "epoch": 1.665714913244015, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1165285110473633, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7230240702629089, + "num_tokens": 392217168.0, + "step": 15168 + }, + { + "epoch": 1.6658247309466288, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.114287853240967, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.71891188621521, + "num_tokens": 392245303.0, + "step": 15169 + }, + { + "epoch": 1.6659345486492423, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.175518035888672, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7318899035453796, + "num_tokens": 392270791.0, + "step": 15170 + }, + { + "epoch": 1.6660443663518558, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3062350749969482, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7162781953811646, + "num_tokens": 392295622.0, + "step": 15171 + }, + { + "epoch": 1.6661541840544696, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2223713397979736, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.7112792134284973, + "num_tokens": 392322785.0, + "step": 15172 + }, + { + "epoch": 1.6662640017570833, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.586871385574341, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7274936437606812, + "num_tokens": 392343448.0, + "step": 15173 + }, + { + "epoch": 1.6663738194596969, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.6672608852386475, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7347958087921143, + "num_tokens": 392362200.0, + "step": 15174 + }, + { + "epoch": 1.6664836371623104, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.069669246673584, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7184485197067261, + "num_tokens": 392390515.0, + "step": 15175 + }, + { + "epoch": 1.6665934548649242, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1803929805755615, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7204142808914185, + "num_tokens": 392416679.0, + "step": 15176 + }, + { + "epoch": 1.666703272567538, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3993566036224365, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7165143489837646, + "num_tokens": 392441347.0, + "step": 15177 + }, + { + "epoch": 1.6668130902701517, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.07373046875, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7100783586502075, + "num_tokens": 392472060.0, + "step": 15178 + }, + { + "epoch": 1.6669229079727652, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.301793336868286, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.701728343963623, + "num_tokens": 392494595.0, + "step": 15179 + }, + { + "epoch": 1.6670327256753787, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2846760749816895, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7225220203399658, + "num_tokens": 392521890.0, + "step": 15180 + }, + { + "epoch": 1.6671425433779925, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4076945781707764, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7188674211502075, + "num_tokens": 392543064.0, + "step": 15181 + }, + { + "epoch": 1.6672523610806063, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.376432418823242, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.729885458946228, + "num_tokens": 392565624.0, + "step": 15182 + }, + { + "epoch": 1.66736217878322, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.4023821353912354, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7352205514907837, + "num_tokens": 392588081.0, + "step": 15183 + }, + { + "epoch": 1.6674719964858336, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.144726276397705, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7237183451652527, + "num_tokens": 392615830.0, + "step": 15184 + }, + { + "epoch": 1.667581814188447, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0557973384857178, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7355855107307434, + "num_tokens": 392643642.0, + "step": 15185 + }, + { + "epoch": 1.6676916318910608, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2367420196533203, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7146281003952026, + "num_tokens": 392668842.0, + "step": 15186 + }, + { + "epoch": 1.6678014495936746, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.375702142715454, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7340613603591919, + "num_tokens": 392691727.0, + "step": 15187 + }, + { + "epoch": 1.6679112672962881, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1993603706359863, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7170106172561646, + "num_tokens": 392718206.0, + "step": 15188 + }, + { + "epoch": 1.6680210849989017, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9949122667312622, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7190902829170227, + "num_tokens": 392749720.0, + "step": 15189 + }, + { + "epoch": 1.6681309027015154, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2623374462127686, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7038216590881348, + "num_tokens": 392775450.0, + "step": 15190 + }, + { + "epoch": 1.6682407204041292, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.149348258972168, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7113807201385498, + "num_tokens": 392805725.0, + "step": 15191 + }, + { + "epoch": 1.668350538106743, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.321031093597412, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.713572084903717, + "num_tokens": 392828935.0, + "step": 15192 + }, + { + "epoch": 1.6684603558093565, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.323324203491211, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7335754036903381, + "num_tokens": 392852371.0, + "step": 15193 + }, + { + "epoch": 1.66857017351197, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2567648887634277, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7223173975944519, + "num_tokens": 392876515.0, + "step": 15194 + }, + { + "epoch": 1.6686799912145838, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0164718627929688, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7123454809188843, + "num_tokens": 392909674.0, + "step": 15195 + }, + { + "epoch": 1.6687898089171975, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9886481761932373, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7044780850410461, + "num_tokens": 392942641.0, + "step": 15196 + }, + { + "epoch": 1.668899626619811, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.8839317560195923, + "learning_rate": 1e-06, + "loss": 0.9728, + "mean_token_accuracy": 0.705923855304718, + "num_tokens": 392976520.0, + "step": 15197 + }, + { + "epoch": 1.6690094443224248, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.449554204940796, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7106337547302246, + "num_tokens": 392999207.0, + "step": 15198 + }, + { + "epoch": 1.6691192620250384, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 1.9774112701416016, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7129380106925964, + "num_tokens": 393029558.0, + "step": 15199 + }, + { + "epoch": 1.669229079727652, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.254399061203003, + "learning_rate": 1e-06, + "loss": 0.7664, + "mean_token_accuracy": 0.7526655793190002, + "num_tokens": 393051392.0, + "step": 15200 + }, + { + "epoch": 1.6693388974302659, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2099854946136475, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7470178604125977, + "num_tokens": 393076234.0, + "step": 15201 + }, + { + "epoch": 1.6694487151328794, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2781009674072266, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7405966520309448, + "num_tokens": 393099892.0, + "step": 15202 + }, + { + "epoch": 1.669558532835493, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.126970052719116, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7146581411361694, + "num_tokens": 393128596.0, + "step": 15203 + }, + { + "epoch": 1.6696683505381067, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.0623865127563477, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7128169536590576, + "num_tokens": 393159301.0, + "step": 15204 + }, + { + "epoch": 1.6697781682407205, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.414128303527832, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7262431383132935, + "num_tokens": 393181250.0, + "step": 15205 + }, + { + "epoch": 1.6698879859433342, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1902172565460205, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.7041666507720947, + "num_tokens": 393208502.0, + "step": 15206 + }, + { + "epoch": 1.6699978036459477, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.113959789276123, + "learning_rate": 1e-06, + "loss": 0.9057, + "mean_token_accuracy": 0.720150351524353, + "num_tokens": 393234328.0, + "step": 15207 + }, + { + "epoch": 1.6701076213485613, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3587636947631836, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7220029234886169, + "num_tokens": 393257177.0, + "step": 15208 + }, + { + "epoch": 1.670217439051175, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.293553352355957, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6931356191635132, + "num_tokens": 393283471.0, + "step": 15209 + }, + { + "epoch": 1.6703272567537888, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.47017502784729, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7321597337722778, + "num_tokens": 393305955.0, + "step": 15210 + }, + { + "epoch": 1.6704370744564023, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1906161308288574, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7213932275772095, + "num_tokens": 393333716.0, + "step": 15211 + }, + { + "epoch": 1.670546892159016, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.1010773181915283, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7294292449951172, + "num_tokens": 393362875.0, + "step": 15212 + }, + { + "epoch": 1.6706567098616296, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.479609489440918, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7238800525665283, + "num_tokens": 393383648.0, + "step": 15213 + }, + { + "epoch": 1.6707665275642434, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.2314510345458984, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7157608270645142, + "num_tokens": 393410602.0, + "step": 15214 + }, + { + "epoch": 1.6708763452668571, + "ewc_loss": 8.58306884765625e-06, + "grad_norm": 2.3580543994903564, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7280070781707764, + "num_tokens": 393433729.0, + "step": 15215 + }, + { + "epoch": 1.6709861629694707, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.335418224334717, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.714970588684082, + "num_tokens": 393457823.0, + "step": 15216 + }, + { + "epoch": 1.6710959806720842, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0673277378082275, + "learning_rate": 1e-06, + "loss": 0.9682, + "mean_token_accuracy": 0.7039070129394531, + "num_tokens": 393487977.0, + "step": 15217 + }, + { + "epoch": 1.671205798374698, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1425797939300537, + "learning_rate": 1e-06, + "loss": 1.0108, + "mean_token_accuracy": 0.6907361745834351, + "num_tokens": 393520805.0, + "step": 15218 + }, + { + "epoch": 1.6713156160773117, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.050811290740967, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7072407007217407, + "num_tokens": 393550646.0, + "step": 15219 + }, + { + "epoch": 1.6714254337799255, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.323960542678833, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7360392808914185, + "num_tokens": 393572080.0, + "step": 15220 + }, + { + "epoch": 1.671535251482539, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2276735305786133, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7091384530067444, + "num_tokens": 393599908.0, + "step": 15221 + }, + { + "epoch": 1.6716450691851525, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3859446048736572, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7325066328048706, + "num_tokens": 393622182.0, + "step": 15222 + }, + { + "epoch": 1.6717548868877663, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2031185626983643, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7210109829902649, + "num_tokens": 393647899.0, + "step": 15223 + }, + { + "epoch": 1.67186470459038, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.359452724456787, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7349814176559448, + "num_tokens": 393671432.0, + "step": 15224 + }, + { + "epoch": 1.6719745222929936, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2670998573303223, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.732379674911499, + "num_tokens": 393695873.0, + "step": 15225 + }, + { + "epoch": 1.6720843399956071, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.31838321685791, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7248531579971313, + "num_tokens": 393721029.0, + "step": 15226 + }, + { + "epoch": 1.6721941576982209, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0411386489868164, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.716708779335022, + "num_tokens": 393751345.0, + "step": 15227 + }, + { + "epoch": 1.6723039754008346, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2120165824890137, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7124354839324951, + "num_tokens": 393777302.0, + "step": 15228 + }, + { + "epoch": 1.6724137931034484, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.116631507873535, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7002055644989014, + "num_tokens": 393805391.0, + "step": 15229 + }, + { + "epoch": 1.672523610806062, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.637079954147339, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7077656984329224, + "num_tokens": 393825653.0, + "step": 15230 + }, + { + "epoch": 1.6726334285086755, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.177887201309204, + "learning_rate": 1e-06, + "loss": 0.9836, + "mean_token_accuracy": 0.7040500640869141, + "num_tokens": 393852131.0, + "step": 15231 + }, + { + "epoch": 1.6727432462112892, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.159684419631958, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7217115759849548, + "num_tokens": 393879563.0, + "step": 15232 + }, + { + "epoch": 1.672853063913903, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0747857093811035, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7106878757476807, + "num_tokens": 393908384.0, + "step": 15233 + }, + { + "epoch": 1.6729628816165167, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.528261184692383, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7198247313499451, + "num_tokens": 393929500.0, + "step": 15234 + }, + { + "epoch": 1.6730726993191303, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2727959156036377, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7468743324279785, + "num_tokens": 393952337.0, + "step": 15235 + }, + { + "epoch": 1.6731825170217438, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.169938325881958, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7041589021682739, + "num_tokens": 393978801.0, + "step": 15236 + }, + { + "epoch": 1.6732923347243576, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.314701795578003, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7128131985664368, + "num_tokens": 394006803.0, + "step": 15237 + }, + { + "epoch": 1.6734021524269713, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2516064643859863, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7123200297355652, + "num_tokens": 394031596.0, + "step": 15238 + }, + { + "epoch": 1.6735119701295849, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.583854913711548, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7393330335617065, + "num_tokens": 394053285.0, + "step": 15239 + }, + { + "epoch": 1.6736217878321984, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0720770359039307, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7344211935997009, + "num_tokens": 394081476.0, + "step": 15240 + }, + { + "epoch": 1.6737316055348122, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0297985076904297, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7221596837043762, + "num_tokens": 394109698.0, + "step": 15241 + }, + { + "epoch": 1.673841423237426, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2346436977386475, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7543937563896179, + "num_tokens": 394133480.0, + "step": 15242 + }, + { + "epoch": 1.6739512409400397, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.286487340927124, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7309316396713257, + "num_tokens": 394159412.0, + "step": 15243 + }, + { + "epoch": 1.6740610586426532, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2415242195129395, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7118603587150574, + "num_tokens": 394185616.0, + "step": 15244 + }, + { + "epoch": 1.6741708763452667, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0654306411743164, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7308593392372131, + "num_tokens": 394214586.0, + "step": 15245 + }, + { + "epoch": 1.6742806940478805, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1581621170043945, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7091801762580872, + "num_tokens": 394241550.0, + "step": 15246 + }, + { + "epoch": 1.6743905117504942, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.314103364944458, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7330378293991089, + "num_tokens": 394266073.0, + "step": 15247 + }, + { + "epoch": 1.674500329453108, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3130576610565186, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7177153825759888, + "num_tokens": 394292595.0, + "step": 15248 + }, + { + "epoch": 1.6746101471557215, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0464723110198975, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.695465087890625, + "num_tokens": 394323940.0, + "step": 15249 + }, + { + "epoch": 1.674719964858335, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3144960403442383, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7086315751075745, + "num_tokens": 394349290.0, + "step": 15250 + }, + { + "epoch": 1.6748297825609488, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.5979835987091064, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.760230541229248, + "num_tokens": 394368900.0, + "step": 15251 + }, + { + "epoch": 1.6749396002635626, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.4991037845611572, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.726173996925354, + "num_tokens": 394390688.0, + "step": 15252 + }, + { + "epoch": 1.6750494179661761, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1100475788116455, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7070697546005249, + "num_tokens": 394419230.0, + "step": 15253 + }, + { + "epoch": 1.6751592356687897, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 1.892345905303955, + "learning_rate": 1e-06, + "loss": 0.9921, + "mean_token_accuracy": 0.6955399513244629, + "num_tokens": 394458588.0, + "step": 15254 + }, + { + "epoch": 1.6752690533714034, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2336599826812744, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.7036567330360413, + "num_tokens": 394486642.0, + "step": 15255 + }, + { + "epoch": 1.6753788710740172, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2451953887939453, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.714058518409729, + "num_tokens": 394513361.0, + "step": 15256 + }, + { + "epoch": 1.675488688776631, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0919861793518066, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7189491987228394, + "num_tokens": 394539867.0, + "step": 15257 + }, + { + "epoch": 1.6755985064792445, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.018540143966675, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.714202880859375, + "num_tokens": 394571484.0, + "step": 15258 + }, + { + "epoch": 1.675708324181858, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3843774795532227, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7524961233139038, + "num_tokens": 394592336.0, + "step": 15259 + }, + { + "epoch": 1.6758181418844718, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.295332670211792, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7140231132507324, + "num_tokens": 394617743.0, + "step": 15260 + }, + { + "epoch": 1.6759279595870855, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.300556182861328, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7294108867645264, + "num_tokens": 394642515.0, + "step": 15261 + }, + { + "epoch": 1.676037777289699, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.341036796569824, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7132812142372131, + "num_tokens": 394665608.0, + "step": 15262 + }, + { + "epoch": 1.6761475949923128, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.379136562347412, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7262719869613647, + "num_tokens": 394690613.0, + "step": 15263 + }, + { + "epoch": 1.6762574126949263, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.151038408279419, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.6987968683242798, + "num_tokens": 394720380.0, + "step": 15264 + }, + { + "epoch": 1.67636723039754, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.139267921447754, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7121368646621704, + "num_tokens": 394750186.0, + "step": 15265 + }, + { + "epoch": 1.6764770481001539, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1219449043273926, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7186676263809204, + "num_tokens": 394780990.0, + "step": 15266 + }, + { + "epoch": 1.6765868658027674, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.395189046859741, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7433270812034607, + "num_tokens": 394802607.0, + "step": 15267 + }, + { + "epoch": 1.676696683505381, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.489413261413574, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.726818323135376, + "num_tokens": 394825914.0, + "step": 15268 + }, + { + "epoch": 1.6768065012079947, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.022282361984253, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7272816896438599, + "num_tokens": 394855125.0, + "step": 15269 + }, + { + "epoch": 1.6769163189106084, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.224745750427246, + "learning_rate": 1e-06, + "loss": 0.9687, + "mean_token_accuracy": 0.7006979584693909, + "num_tokens": 394883159.0, + "step": 15270 + }, + { + "epoch": 1.6770261366132222, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.255424737930298, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7031978964805603, + "num_tokens": 394909074.0, + "step": 15271 + }, + { + "epoch": 1.6771359543158357, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.29878830909729, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7196170091629028, + "num_tokens": 394932896.0, + "step": 15272 + }, + { + "epoch": 1.6772457720184493, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.420785665512085, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7079427242279053, + "num_tokens": 394957311.0, + "step": 15273 + }, + { + "epoch": 1.677355589721063, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3270950317382812, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7050638198852539, + "num_tokens": 394981625.0, + "step": 15274 + }, + { + "epoch": 1.6774654074236768, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0930938720703125, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7157186269760132, + "num_tokens": 395011025.0, + "step": 15275 + }, + { + "epoch": 1.6775752251262903, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.583238124847412, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7178891897201538, + "num_tokens": 395031310.0, + "step": 15276 + }, + { + "epoch": 1.677685042828904, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 1.8922860622406006, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7067115306854248, + "num_tokens": 395064109.0, + "step": 15277 + }, + { + "epoch": 1.6777948605315176, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1101317405700684, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7091925144195557, + "num_tokens": 395092223.0, + "step": 15278 + }, + { + "epoch": 1.6779046782341314, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.095679759979248, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7187072038650513, + "num_tokens": 395121767.0, + "step": 15279 + }, + { + "epoch": 1.6780144959367451, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3007454872131348, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7390199899673462, + "num_tokens": 395145604.0, + "step": 15280 + }, + { + "epoch": 1.6781243136393587, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1282284259796143, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.718788743019104, + "num_tokens": 395172840.0, + "step": 15281 + }, + { + "epoch": 1.6782341313419722, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.538339853286743, + "learning_rate": 1e-06, + "loss": 0.7257, + "mean_token_accuracy": 0.7724708318710327, + "num_tokens": 395192119.0, + "step": 15282 + }, + { + "epoch": 1.678343949044586, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1500511169433594, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.704826831817627, + "num_tokens": 395220787.0, + "step": 15283 + }, + { + "epoch": 1.6784537667471997, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2316172122955322, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.6995068788528442, + "num_tokens": 395246050.0, + "step": 15284 + }, + { + "epoch": 1.6785635844498135, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.493302345275879, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7273871898651123, + "num_tokens": 395267040.0, + "step": 15285 + }, + { + "epoch": 1.678673402152427, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1917521953582764, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7125787734985352, + "num_tokens": 395293994.0, + "step": 15286 + }, + { + "epoch": 1.6787832198550405, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.353090286254883, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7233550548553467, + "num_tokens": 395319169.0, + "step": 15287 + }, + { + "epoch": 1.6788930375576543, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.5500833988189697, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7154622077941895, + "num_tokens": 395341689.0, + "step": 15288 + }, + { + "epoch": 1.679002855260268, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.169280767440796, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.7025656700134277, + "num_tokens": 395369341.0, + "step": 15289 + }, + { + "epoch": 1.6791126729628816, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.346970319747925, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7200385928153992, + "num_tokens": 395391825.0, + "step": 15290 + }, + { + "epoch": 1.6792224906654951, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.224966049194336, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7173916697502136, + "num_tokens": 395416109.0, + "step": 15291 + }, + { + "epoch": 1.6793323083681089, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3513665199279785, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.703079342842102, + "num_tokens": 395440664.0, + "step": 15292 + }, + { + "epoch": 1.6794421260707226, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.444831371307373, + "learning_rate": 1e-06, + "loss": 0.8133, + "mean_token_accuracy": 0.7437479496002197, + "num_tokens": 395460977.0, + "step": 15293 + }, + { + "epoch": 1.6795519437733364, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3905575275421143, + "learning_rate": 1e-06, + "loss": 1.0109, + "mean_token_accuracy": 0.6953319907188416, + "num_tokens": 395486277.0, + "step": 15294 + }, + { + "epoch": 1.67966176147595, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2319202423095703, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7231072783470154, + "num_tokens": 395511382.0, + "step": 15295 + }, + { + "epoch": 1.6797715791785635, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.224698543548584, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7398920059204102, + "num_tokens": 395535975.0, + "step": 15296 + }, + { + "epoch": 1.6798813968811772, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.213531732559204, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.709460437297821, + "num_tokens": 395561926.0, + "step": 15297 + }, + { + "epoch": 1.679991214583791, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.12467098236084, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6952574849128723, + "num_tokens": 395589193.0, + "step": 15298 + }, + { + "epoch": 1.6801010322864047, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 1.9480817317962646, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7109078168869019, + "num_tokens": 395622093.0, + "step": 15299 + }, + { + "epoch": 1.6802108499890183, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1023736000061035, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7388244867324829, + "num_tokens": 395649011.0, + "step": 15300 + }, + { + "epoch": 1.6803206676916318, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.290386199951172, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7293705940246582, + "num_tokens": 395673215.0, + "step": 15301 + }, + { + "epoch": 1.6804304853942456, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2260725498199463, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7060720324516296, + "num_tokens": 395699330.0, + "step": 15302 + }, + { + "epoch": 1.6805403030968593, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.3047478199005127, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7268763780593872, + "num_tokens": 395723199.0, + "step": 15303 + }, + { + "epoch": 1.6806501207994728, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.26127028465271, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7080153226852417, + "num_tokens": 395748689.0, + "step": 15304 + }, + { + "epoch": 1.6807599385020864, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.2081682682037354, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7417252063751221, + "num_tokens": 395775038.0, + "step": 15305 + }, + { + "epoch": 1.6808697562047001, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.0122029781341553, + "learning_rate": 1e-06, + "loss": 0.796, + "mean_token_accuracy": 0.7488757967948914, + "num_tokens": 395803491.0, + "step": 15306 + }, + { + "epoch": 1.680979573907314, + "ewc_loss": 8.64267349243164e-06, + "grad_norm": 2.1851344108581543, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7329932451248169, + "num_tokens": 395829483.0, + "step": 15307 + }, + { + "epoch": 1.6810893916099277, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.5231473445892334, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7184802293777466, + "num_tokens": 395849479.0, + "step": 15308 + }, + { + "epoch": 1.6811992093125412, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.2047529220581055, + "learning_rate": 1e-06, + "loss": 1.0127, + "mean_token_accuracy": 0.6924120187759399, + "num_tokens": 395877741.0, + "step": 15309 + }, + { + "epoch": 1.6813090270151547, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.412689685821533, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.73082435131073, + "num_tokens": 395900303.0, + "step": 15310 + }, + { + "epoch": 1.6814188447177685, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.259415864944458, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7168983817100525, + "num_tokens": 395928543.0, + "step": 15311 + }, + { + "epoch": 1.6815286624203822, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.170825958251953, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7303462028503418, + "num_tokens": 395954915.0, + "step": 15312 + }, + { + "epoch": 1.681638480122996, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.3504488468170166, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7480920553207397, + "num_tokens": 395977319.0, + "step": 15313 + }, + { + "epoch": 1.6817482978256095, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.3359878063201904, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7160782814025879, + "num_tokens": 396000471.0, + "step": 15314 + }, + { + "epoch": 1.681858115528223, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.006619691848755, + "learning_rate": 1e-06, + "loss": 0.9689, + "mean_token_accuracy": 0.712432861328125, + "num_tokens": 396033624.0, + "step": 15315 + }, + { + "epoch": 1.6819679332308368, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.258328437805176, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7223039269447327, + "num_tokens": 396060174.0, + "step": 15316 + }, + { + "epoch": 1.6820777509334506, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.8624558448791504, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7124090194702148, + "num_tokens": 396081819.0, + "step": 15317 + }, + { + "epoch": 1.6821875686360641, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.154494047164917, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6799257397651672, + "num_tokens": 396112089.0, + "step": 15318 + }, + { + "epoch": 1.6822973863386776, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.4340906143188477, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7285462617874146, + "num_tokens": 396134208.0, + "step": 15319 + }, + { + "epoch": 1.6824072040412914, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.16823148727417, + "learning_rate": 1e-06, + "loss": 1.0208, + "mean_token_accuracy": 0.6863290071487427, + "num_tokens": 396165284.0, + "step": 15320 + }, + { + "epoch": 1.6825170217439052, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.612764596939087, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7218214869499207, + "num_tokens": 396185911.0, + "step": 15321 + }, + { + "epoch": 1.682626839446519, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.365339517593384, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7207449674606323, + "num_tokens": 396208581.0, + "step": 15322 + }, + { + "epoch": 1.6827366571491325, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.0292930603027344, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7394418120384216, + "num_tokens": 396235996.0, + "step": 15323 + }, + { + "epoch": 1.682846474851746, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.332624673843384, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.708904504776001, + "num_tokens": 396259865.0, + "step": 15324 + }, + { + "epoch": 1.6829562925543597, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.1319003105163574, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.7120559215545654, + "num_tokens": 396288236.0, + "step": 15325 + }, + { + "epoch": 1.6830661102569735, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.263953685760498, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7281153202056885, + "num_tokens": 396310576.0, + "step": 15326 + }, + { + "epoch": 1.683175927959587, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.348529815673828, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7190017104148865, + "num_tokens": 396334790.0, + "step": 15327 + }, + { + "epoch": 1.6832857456622008, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.157193183898926, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7134124040603638, + "num_tokens": 396362040.0, + "step": 15328 + }, + { + "epoch": 1.6833955633648143, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.427117347717285, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.6946879625320435, + "num_tokens": 396385734.0, + "step": 15329 + }, + { + "epoch": 1.683505381067428, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.180344581604004, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7247463464736938, + "num_tokens": 396410754.0, + "step": 15330 + }, + { + "epoch": 1.6836151987700418, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.050463914871216, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.6987262964248657, + "num_tokens": 396440831.0, + "step": 15331 + }, + { + "epoch": 1.6837250164726554, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.2779154777526855, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.69941246509552, + "num_tokens": 396466952.0, + "step": 15332 + }, + { + "epoch": 1.683834834175269, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.1361279487609863, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7248876690864563, + "num_tokens": 396496122.0, + "step": 15333 + }, + { + "epoch": 1.6839446518778827, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.2921314239501953, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7194211483001709, + "num_tokens": 396518979.0, + "step": 15334 + }, + { + "epoch": 1.6840544695804964, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.0979557037353516, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7133023738861084, + "num_tokens": 396549076.0, + "step": 15335 + }, + { + "epoch": 1.6841642872831102, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.0395400524139404, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7101951241493225, + "num_tokens": 396579410.0, + "step": 15336 + }, + { + "epoch": 1.6842741049857237, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.1868207454681396, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7242323160171509, + "num_tokens": 396605417.0, + "step": 15337 + }, + { + "epoch": 1.6843839226883373, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.302206039428711, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7226552963256836, + "num_tokens": 396629173.0, + "step": 15338 + }, + { + "epoch": 1.684493740390951, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.153020143508911, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7141088247299194, + "num_tokens": 396657129.0, + "step": 15339 + }, + { + "epoch": 1.6846035580935648, + "ewc_loss": 8.702278137207031e-06, + "grad_norm": 2.4732279777526855, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7381106615066528, + "num_tokens": 396678250.0, + "step": 15340 + }, + { + "epoch": 1.6847133757961783, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2433130741119385, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7399372458457947, + "num_tokens": 396702668.0, + "step": 15341 + }, + { + "epoch": 1.684823193498792, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.316591501235962, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7260836362838745, + "num_tokens": 396727824.0, + "step": 15342 + }, + { + "epoch": 1.6849330112014056, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.172999858856201, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7183052897453308, + "num_tokens": 396756753.0, + "step": 15343 + }, + { + "epoch": 1.6850428289040194, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.419173002243042, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7204996347427368, + "num_tokens": 396780860.0, + "step": 15344 + }, + { + "epoch": 1.685152646606633, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4408628940582275, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7049874067306519, + "num_tokens": 396804809.0, + "step": 15345 + }, + { + "epoch": 1.6852624643092466, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4620893001556396, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7415602207183838, + "num_tokens": 396827140.0, + "step": 15346 + }, + { + "epoch": 1.6853722820118602, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3358240127563477, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7145100831985474, + "num_tokens": 396851950.0, + "step": 15347 + }, + { + "epoch": 1.685482099714474, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2167649269104004, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7326825857162476, + "num_tokens": 396876087.0, + "step": 15348 + }, + { + "epoch": 1.6855919174170877, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0492746829986572, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7138413190841675, + "num_tokens": 396905161.0, + "step": 15349 + }, + { + "epoch": 1.6857017351197014, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.385838270187378, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7219141125679016, + "num_tokens": 396928353.0, + "step": 15350 + }, + { + "epoch": 1.685811552822315, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2872984409332275, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7239259481430054, + "num_tokens": 396952381.0, + "step": 15351 + }, + { + "epoch": 1.6859213705249285, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.179550886154175, + "learning_rate": 1e-06, + "loss": 1.0195, + "mean_token_accuracy": 0.6883708834648132, + "num_tokens": 396981463.0, + "step": 15352 + }, + { + "epoch": 1.6860311882275423, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.17983078956604, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6910009980201721, + "num_tokens": 397008751.0, + "step": 15353 + }, + { + "epoch": 1.686141005930156, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.500669240951538, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7207434773445129, + "num_tokens": 397030367.0, + "step": 15354 + }, + { + "epoch": 1.6862508236327696, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.8739502429962158, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7158651351928711, + "num_tokens": 397063919.0, + "step": 15355 + }, + { + "epoch": 1.686360641335383, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.171847105026245, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.723371148109436, + "num_tokens": 397090232.0, + "step": 15356 + }, + { + "epoch": 1.6864704590379969, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.381030797958374, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7317049503326416, + "num_tokens": 397111520.0, + "step": 15357 + }, + { + "epoch": 1.6865802767406106, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.081697702407837, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.7018182277679443, + "num_tokens": 397144790.0, + "step": 15358 + }, + { + "epoch": 1.6866900944432244, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.145867109298706, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7188575267791748, + "num_tokens": 397171378.0, + "step": 15359 + }, + { + "epoch": 1.686799912145838, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4074649810791016, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7249935269355774, + "num_tokens": 397192336.0, + "step": 15360 + }, + { + "epoch": 1.6869097298484514, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.275571584701538, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7399044036865234, + "num_tokens": 397217011.0, + "step": 15361 + }, + { + "epoch": 1.6870195475510652, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.5533435344696045, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7079684138298035, + "num_tokens": 397236937.0, + "step": 15362 + }, + { + "epoch": 1.687129365253679, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3136603832244873, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.713592529296875, + "num_tokens": 397262375.0, + "step": 15363 + }, + { + "epoch": 1.6872391829562927, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2217509746551514, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7288532257080078, + "num_tokens": 397287748.0, + "step": 15364 + }, + { + "epoch": 1.6873490006589063, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.058997631072998, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7005513906478882, + "num_tokens": 397318339.0, + "step": 15365 + }, + { + "epoch": 1.6874588183615198, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.56030011177063, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7405322790145874, + "num_tokens": 397339436.0, + "step": 15366 + }, + { + "epoch": 1.6875686360641335, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2057764530181885, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7329572439193726, + "num_tokens": 397364289.0, + "step": 15367 + }, + { + "epoch": 1.6876784537667473, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.199817657470703, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6922202110290527, + "num_tokens": 397390891.0, + "step": 15368 + }, + { + "epoch": 1.6877882714693608, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.266686201095581, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7275462746620178, + "num_tokens": 397415729.0, + "step": 15369 + }, + { + "epoch": 1.6878980891719744, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0682547092437744, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7160013318061829, + "num_tokens": 397444613.0, + "step": 15370 + }, + { + "epoch": 1.6880079068745881, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.24043869972229, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7245376706123352, + "num_tokens": 397470442.0, + "step": 15371 + }, + { + "epoch": 1.6881177245772019, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3733203411102295, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7343064546585083, + "num_tokens": 397493162.0, + "step": 15372 + }, + { + "epoch": 1.6882275422798156, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.8974131345748901, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.701123058795929, + "num_tokens": 397527873.0, + "step": 15373 + }, + { + "epoch": 1.6883373599824292, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.303196430206299, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7362264394760132, + "num_tokens": 397551717.0, + "step": 15374 + }, + { + "epoch": 1.6884471776850427, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4270193576812744, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7239450812339783, + "num_tokens": 397576725.0, + "step": 15375 + }, + { + "epoch": 1.6885569953876565, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4186038970947266, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7234235405921936, + "num_tokens": 397597859.0, + "step": 15376 + }, + { + "epoch": 1.6886668130902702, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.33911395072937, + "learning_rate": 1e-06, + "loss": 0.9416, + "mean_token_accuracy": 0.720686674118042, + "num_tokens": 397620983.0, + "step": 15377 + }, + { + "epoch": 1.6887766307928838, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.9424036741256714, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.727001428604126, + "num_tokens": 397654202.0, + "step": 15378 + }, + { + "epoch": 1.6888864484954975, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0259017944335938, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.734799861907959, + "num_tokens": 397684205.0, + "step": 15379 + }, + { + "epoch": 1.688996266198111, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3270764350891113, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7220678925514221, + "num_tokens": 397709024.0, + "step": 15380 + }, + { + "epoch": 1.6891060839007248, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2937915325164795, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7339718341827393, + "num_tokens": 397734698.0, + "step": 15381 + }, + { + "epoch": 1.6892159016033386, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4002654552459717, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7228769063949585, + "num_tokens": 397757587.0, + "step": 15382 + }, + { + "epoch": 1.689325719305952, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.6322028636932373, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7364377975463867, + "num_tokens": 397776130.0, + "step": 15383 + }, + { + "epoch": 1.6894355370085656, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1781511306762695, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.72320556640625, + "num_tokens": 397801847.0, + "step": 15384 + }, + { + "epoch": 1.6895453547111794, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0246024131774902, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7307450771331787, + "num_tokens": 397830107.0, + "step": 15385 + }, + { + "epoch": 1.6896551724137931, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.183320999145508, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7076508402824402, + "num_tokens": 397856490.0, + "step": 15386 + }, + { + "epoch": 1.689764990116407, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.022257089614868, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7105715274810791, + "num_tokens": 397886156.0, + "step": 15387 + }, + { + "epoch": 1.6898748078190204, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.284344434738159, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7373419404029846, + "num_tokens": 397910565.0, + "step": 15388 + }, + { + "epoch": 1.689984625521634, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.6136319637298584, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7408947944641113, + "num_tokens": 397931254.0, + "step": 15389 + }, + { + "epoch": 1.6900944432242477, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.104097843170166, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7235379815101624, + "num_tokens": 397960044.0, + "step": 15390 + }, + { + "epoch": 1.6902042609268615, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.126347541809082, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7075510621070862, + "num_tokens": 397988078.0, + "step": 15391 + }, + { + "epoch": 1.690314078629475, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2957839965820312, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7030852437019348, + "num_tokens": 398014185.0, + "step": 15392 + }, + { + "epoch": 1.6904238963320888, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.092595338821411, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7200152277946472, + "num_tokens": 398041080.0, + "step": 15393 + }, + { + "epoch": 1.6905337140347023, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2786967754364014, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7261881828308105, + "num_tokens": 398066396.0, + "step": 15394 + }, + { + "epoch": 1.690643531737316, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3374722003936768, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.723387598991394, + "num_tokens": 398091231.0, + "step": 15395 + }, + { + "epoch": 1.6907533494399298, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.391605854034424, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7210103273391724, + "num_tokens": 398113988.0, + "step": 15396 + }, + { + "epoch": 1.6908631671425434, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2084622383117676, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7044618725776672, + "num_tokens": 398139097.0, + "step": 15397 + }, + { + "epoch": 1.690972984845157, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.235905170440674, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7331997156143188, + "num_tokens": 398166310.0, + "step": 15398 + }, + { + "epoch": 1.6910828025477707, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.355041980743408, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7322269082069397, + "num_tokens": 398188479.0, + "step": 15399 + }, + { + "epoch": 1.6911926202503844, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2148427963256836, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7228310108184814, + "num_tokens": 398214429.0, + "step": 15400 + }, + { + "epoch": 1.6913024379529982, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0040948390960693, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7118822336196899, + "num_tokens": 398246406.0, + "step": 15401 + }, + { + "epoch": 1.6914122556556117, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1400985717773438, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.716810941696167, + "num_tokens": 398273894.0, + "step": 15402 + }, + { + "epoch": 1.6915220733582252, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.508453369140625, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7261775135993958, + "num_tokens": 398295936.0, + "step": 15403 + }, + { + "epoch": 1.691631891060839, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1978392601013184, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7131954431533813, + "num_tokens": 398322470.0, + "step": 15404 + }, + { + "epoch": 1.6917417087634528, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3190956115722656, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7040387988090515, + "num_tokens": 398347501.0, + "step": 15405 + }, + { + "epoch": 1.6918515264660663, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0953898429870605, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.7177079916000366, + "num_tokens": 398377375.0, + "step": 15406 + }, + { + "epoch": 1.6919613441686798, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.538888931274414, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7288457155227661, + "num_tokens": 398397998.0, + "step": 15407 + }, + { + "epoch": 1.6920711618712936, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.880519986152649, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7129186391830444, + "num_tokens": 398433069.0, + "step": 15408 + }, + { + "epoch": 1.6921809795739073, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4464833736419678, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7312372922897339, + "num_tokens": 398455663.0, + "step": 15409 + }, + { + "epoch": 1.692290797276521, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3584682941436768, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7199774980545044, + "num_tokens": 398478737.0, + "step": 15410 + }, + { + "epoch": 1.6924006149791346, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.109539747238159, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7288420796394348, + "num_tokens": 398506106.0, + "step": 15411 + }, + { + "epoch": 1.6925104326817482, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4182097911834717, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7491543889045715, + "num_tokens": 398526706.0, + "step": 15412 + }, + { + "epoch": 1.692620250384362, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0996081829071045, + "learning_rate": 1e-06, + "loss": 1.0077, + "mean_token_accuracy": 0.6925375461578369, + "num_tokens": 398558623.0, + "step": 15413 + }, + { + "epoch": 1.6927300680869757, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.54765248298645, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7349198460578918, + "num_tokens": 398580334.0, + "step": 15414 + }, + { + "epoch": 1.6928398857895894, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4961657524108887, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.7517750859260559, + "num_tokens": 398601379.0, + "step": 15415 + }, + { + "epoch": 1.692949703492203, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.173543930053711, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7407509088516235, + "num_tokens": 398629504.0, + "step": 15416 + }, + { + "epoch": 1.6930595211948165, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.32525634765625, + "learning_rate": 1e-06, + "loss": 0.99, + "mean_token_accuracy": 0.6938886642456055, + "num_tokens": 398655967.0, + "step": 15417 + }, + { + "epoch": 1.6931693388974303, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.468956708908081, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7152036428451538, + "num_tokens": 398677510.0, + "step": 15418 + }, + { + "epoch": 1.693279156600044, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.065845012664795, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7141069173812866, + "num_tokens": 398705020.0, + "step": 15419 + }, + { + "epoch": 1.6933889743026576, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4414100646972656, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7205737829208374, + "num_tokens": 398730620.0, + "step": 15420 + }, + { + "epoch": 1.693498792005271, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3670661449432373, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7443861961364746, + "num_tokens": 398752590.0, + "step": 15421 + }, + { + "epoch": 1.6936086097078848, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2551474571228027, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7470807433128357, + "num_tokens": 398775873.0, + "step": 15422 + }, + { + "epoch": 1.6937184274104986, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.35366153717041, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7301023006439209, + "num_tokens": 398798566.0, + "step": 15423 + }, + { + "epoch": 1.6938282451131124, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3826684951782227, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7154722213745117, + "num_tokens": 398822146.0, + "step": 15424 + }, + { + "epoch": 1.693938062815726, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1164000034332275, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.6966562271118164, + "num_tokens": 398851007.0, + "step": 15425 + }, + { + "epoch": 1.6940478805183394, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3367559909820557, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.728625476360321, + "num_tokens": 398874032.0, + "step": 15426 + }, + { + "epoch": 1.6941576982209532, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.9574410915374756, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7140387892723083, + "num_tokens": 398907392.0, + "step": 15427 + }, + { + "epoch": 1.694267515923567, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.287736415863037, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.7030168175697327, + "num_tokens": 398932488.0, + "step": 15428 + }, + { + "epoch": 1.6943773336261807, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3051810264587402, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7400113940238953, + "num_tokens": 398955318.0, + "step": 15429 + }, + { + "epoch": 1.6944871513287942, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.758906841278076, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7422559261322021, + "num_tokens": 398973510.0, + "step": 15430 + }, + { + "epoch": 1.6945969690314078, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.279834270477295, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7291862964630127, + "num_tokens": 398998123.0, + "step": 15431 + }, + { + "epoch": 1.6947067867340215, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.357123374938965, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.729897141456604, + "num_tokens": 399021861.0, + "step": 15432 + }, + { + "epoch": 1.6948166044366353, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4856293201446533, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7373332977294922, + "num_tokens": 399042557.0, + "step": 15433 + }, + { + "epoch": 1.6949264221392488, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0789432525634766, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7299886345863342, + "num_tokens": 399072586.0, + "step": 15434 + }, + { + "epoch": 1.6950362398418624, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.5692458152770996, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6987522840499878, + "num_tokens": 399099061.0, + "step": 15435 + }, + { + "epoch": 1.6951460575444761, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.615152597427368, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.723186194896698, + "num_tokens": 399117783.0, + "step": 15436 + }, + { + "epoch": 1.6952558752470899, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3465068340301514, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7225595712661743, + "num_tokens": 399141358.0, + "step": 15437 + }, + { + "epoch": 1.6953656929497036, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1444132328033447, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.6977887153625488, + "num_tokens": 399169370.0, + "step": 15438 + }, + { + "epoch": 1.6954755106523172, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.062678337097168, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7268301248550415, + "num_tokens": 399197959.0, + "step": 15439 + }, + { + "epoch": 1.6955853283549307, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3619978427886963, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7229553461074829, + "num_tokens": 399220990.0, + "step": 15440 + }, + { + "epoch": 1.6956951460575445, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.269103765487671, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.709251880645752, + "num_tokens": 399246665.0, + "step": 15441 + }, + { + "epoch": 1.6958049637601582, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.018932342529297, + "learning_rate": 1e-06, + "loss": 0.9983, + "mean_token_accuracy": 0.6965120434761047, + "num_tokens": 399276337.0, + "step": 15442 + }, + { + "epoch": 1.6959147814627717, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3504507541656494, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7293318510055542, + "num_tokens": 399298346.0, + "step": 15443 + }, + { + "epoch": 1.6960245991653855, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.262603282928467, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7099946141242981, + "num_tokens": 399323963.0, + "step": 15444 + }, + { + "epoch": 1.696134416867999, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.5687389373779297, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7260671854019165, + "num_tokens": 399344041.0, + "step": 15445 + }, + { + "epoch": 1.6962442345706128, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.516915798187256, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7184998989105225, + "num_tokens": 399365154.0, + "step": 15446 + }, + { + "epoch": 1.6963540522732266, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.170818328857422, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7281869649887085, + "num_tokens": 399390134.0, + "step": 15447 + }, + { + "epoch": 1.69646386997584, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2546772956848145, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7244518995285034, + "num_tokens": 399414294.0, + "step": 15448 + }, + { + "epoch": 1.6965736876784536, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2898337841033936, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.693466067314148, + "num_tokens": 399442632.0, + "step": 15449 + }, + { + "epoch": 1.6966835053810674, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.175974130630493, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6866440773010254, + "num_tokens": 399471270.0, + "step": 15450 + }, + { + "epoch": 1.6967933230836811, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.258488893508911, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7230671644210815, + "num_tokens": 399495752.0, + "step": 15451 + }, + { + "epoch": 1.696903140786295, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.168332099914551, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7107720375061035, + "num_tokens": 399523581.0, + "step": 15452 + }, + { + "epoch": 1.6970129584889084, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.183772563934326, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7089985013008118, + "num_tokens": 399550092.0, + "step": 15453 + }, + { + "epoch": 1.697122776191522, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.9406859874725342, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7364447116851807, + "num_tokens": 399579278.0, + "step": 15454 + }, + { + "epoch": 1.6972325938941357, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.8646295070648193, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7231707572937012, + "num_tokens": 399610000.0, + "step": 15455 + }, + { + "epoch": 1.6973424115967495, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1874043941497803, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7362716794013977, + "num_tokens": 399635600.0, + "step": 15456 + }, + { + "epoch": 1.697452229299363, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.331178903579712, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7188355326652527, + "num_tokens": 399660664.0, + "step": 15457 + }, + { + "epoch": 1.6975620470019768, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3030636310577393, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7038329839706421, + "num_tokens": 399687216.0, + "step": 15458 + }, + { + "epoch": 1.6976718647045903, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.252223014831543, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6995166540145874, + "num_tokens": 399711844.0, + "step": 15459 + }, + { + "epoch": 1.697781682407204, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.9967286586761475, + "learning_rate": 1e-06, + "loss": 0.9772, + "mean_token_accuracy": 0.6941964626312256, + "num_tokens": 399743060.0, + "step": 15460 + }, + { + "epoch": 1.6978915001098178, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1718027591705322, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7069806456565857, + "num_tokens": 399769595.0, + "step": 15461 + }, + { + "epoch": 1.6980013178124314, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3167741298675537, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7168548107147217, + "num_tokens": 399794162.0, + "step": 15462 + }, + { + "epoch": 1.698111135515045, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.289268970489502, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.6979435086250305, + "num_tokens": 399820012.0, + "step": 15463 + }, + { + "epoch": 1.6982209532176586, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1569340229034424, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7117983102798462, + "num_tokens": 399846566.0, + "step": 15464 + }, + { + "epoch": 1.6983307709202724, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.146467685699463, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7452688217163086, + "num_tokens": 399872193.0, + "step": 15465 + }, + { + "epoch": 1.6984405886228862, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.188309907913208, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7156499028205872, + "num_tokens": 399897730.0, + "step": 15466 + }, + { + "epoch": 1.6985504063254997, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1611857414245605, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7079057693481445, + "num_tokens": 399926421.0, + "step": 15467 + }, + { + "epoch": 1.6986602240281132, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.4446778297424316, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7222338318824768, + "num_tokens": 399946819.0, + "step": 15468 + }, + { + "epoch": 1.698770041730727, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3793747425079346, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7301332950592041, + "num_tokens": 399970656.0, + "step": 15469 + }, + { + "epoch": 1.6988798594333407, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.089266061782837, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7206933498382568, + "num_tokens": 399998962.0, + "step": 15470 + }, + { + "epoch": 1.6989896771359543, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.030564785003662, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7212555408477783, + "num_tokens": 400029421.0, + "step": 15471 + }, + { + "epoch": 1.6990994948385678, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.332132339477539, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.707099437713623, + "num_tokens": 400053607.0, + "step": 15472 + }, + { + "epoch": 1.6992093125411816, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0643579959869385, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7336719632148743, + "num_tokens": 400080343.0, + "step": 15473 + }, + { + "epoch": 1.6993191302437953, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0760695934295654, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7060291171073914, + "num_tokens": 400109965.0, + "step": 15474 + }, + { + "epoch": 1.699428947946409, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3530216217041016, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.725940465927124, + "num_tokens": 400132559.0, + "step": 15475 + }, + { + "epoch": 1.6995387656490226, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1074135303497314, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7251632213592529, + "num_tokens": 400159520.0, + "step": 15476 + }, + { + "epoch": 1.6996485833516362, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.37304949760437, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7107884883880615, + "num_tokens": 400182574.0, + "step": 15477 + }, + { + "epoch": 1.69975840105425, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.131514310836792, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7115727066993713, + "num_tokens": 400212425.0, + "step": 15478 + }, + { + "epoch": 1.6998682187568637, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.341724395751953, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7349193692207336, + "num_tokens": 400234508.0, + "step": 15479 + }, + { + "epoch": 1.6999780364594774, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.111570358276367, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7281571626663208, + "num_tokens": 400261336.0, + "step": 15480 + }, + { + "epoch": 1.700087854162091, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.335753917694092, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7188630104064941, + "num_tokens": 400284616.0, + "step": 15481 + }, + { + "epoch": 1.7001976718647045, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2226686477661133, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7103018760681152, + "num_tokens": 400311134.0, + "step": 15482 + }, + { + "epoch": 1.7003074895673183, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3859379291534424, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7273358106613159, + "num_tokens": 400333231.0, + "step": 15483 + }, + { + "epoch": 1.700417307269932, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0052361488342285, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6925777196884155, + "num_tokens": 400366946.0, + "step": 15484 + }, + { + "epoch": 1.7005271249725455, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2328929901123047, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7257260084152222, + "num_tokens": 400391020.0, + "step": 15485 + }, + { + "epoch": 1.700636942675159, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0975701808929443, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.713429868221283, + "num_tokens": 400420854.0, + "step": 15486 + }, + { + "epoch": 1.7007467603777728, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2492470741271973, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7291638851165771, + "num_tokens": 400445400.0, + "step": 15487 + }, + { + "epoch": 1.7008565780803866, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2600624561309814, + "learning_rate": 1e-06, + "loss": 0.9758, + "mean_token_accuracy": 0.7043405771255493, + "num_tokens": 400472258.0, + "step": 15488 + }, + { + "epoch": 1.7009663957830004, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.291948080062866, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7270307540893555, + "num_tokens": 400495777.0, + "step": 15489 + }, + { + "epoch": 1.7010762134856139, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.317192792892456, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7087092399597168, + "num_tokens": 400523789.0, + "step": 15490 + }, + { + "epoch": 1.7011860311882274, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3747098445892334, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7292640805244446, + "num_tokens": 400548564.0, + "step": 15491 + }, + { + "epoch": 1.7012958488908412, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.281731367111206, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7092405557632446, + "num_tokens": 400572430.0, + "step": 15492 + }, + { + "epoch": 1.701405666593455, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0998833179473877, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7199815511703491, + "num_tokens": 400600268.0, + "step": 15493 + }, + { + "epoch": 1.7015154842960687, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3863861560821533, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7399307489395142, + "num_tokens": 400622347.0, + "step": 15494 + }, + { + "epoch": 1.7016253019986822, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.7467901706695557, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7345824241638184, + "num_tokens": 400640295.0, + "step": 15495 + }, + { + "epoch": 1.7017351197012958, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.72343111038208, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7189064025878906, + "num_tokens": 400657890.0, + "step": 15496 + }, + { + "epoch": 1.7018449374039095, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2532153129577637, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7220105528831482, + "num_tokens": 400682174.0, + "step": 15497 + }, + { + "epoch": 1.7019547551065233, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2503859996795654, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7260854244232178, + "num_tokens": 400707258.0, + "step": 15498 + }, + { + "epoch": 1.7020645728091368, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.126899242401123, + "learning_rate": 1e-06, + "loss": 0.9844, + "mean_token_accuracy": 0.6946662664413452, + "num_tokens": 400736267.0, + "step": 15499 + }, + { + "epoch": 1.7021743905117503, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.52095890045166, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7165528535842896, + "num_tokens": 400758113.0, + "step": 15500 + }, + { + "epoch": 1.702284208214364, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.5774459838867188, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7368091344833374, + "num_tokens": 400779398.0, + "step": 15501 + }, + { + "epoch": 1.7023940259169779, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.296654224395752, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.728800356388092, + "num_tokens": 400803902.0, + "step": 15502 + }, + { + "epoch": 1.7025038436195916, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2196619510650635, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7188395857810974, + "num_tokens": 400828366.0, + "step": 15503 + }, + { + "epoch": 1.7026136613222052, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0405683517456055, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7097875475883484, + "num_tokens": 400857298.0, + "step": 15504 + }, + { + "epoch": 1.7027234790248187, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.592052936553955, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7265374660491943, + "num_tokens": 400878272.0, + "step": 15505 + }, + { + "epoch": 1.7028332967274324, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2817280292510986, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7174569964408875, + "num_tokens": 400904791.0, + "step": 15506 + }, + { + "epoch": 1.7029431144300462, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0210800170898438, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7434935569763184, + "num_tokens": 400931359.0, + "step": 15507 + }, + { + "epoch": 1.7030529321326597, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.8783496618270874, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7058831453323364, + "num_tokens": 400968802.0, + "step": 15508 + }, + { + "epoch": 1.7031627498352735, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3607001304626465, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7132788896560669, + "num_tokens": 400993412.0, + "step": 15509 + }, + { + "epoch": 1.703272567537887, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.434927463531494, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7177034616470337, + "num_tokens": 401016039.0, + "step": 15510 + }, + { + "epoch": 1.7033823852405008, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.3776607513427734, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7310539484024048, + "num_tokens": 401037545.0, + "step": 15511 + }, + { + "epoch": 1.7034922029431145, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1941211223602295, + "learning_rate": 1e-06, + "loss": 1.0409, + "mean_token_accuracy": 0.682154655456543, + "num_tokens": 401064993.0, + "step": 15512 + }, + { + "epoch": 1.703602020645728, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.187366485595703, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7287256121635437, + "num_tokens": 401091084.0, + "step": 15513 + }, + { + "epoch": 1.7037118383483416, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 1.9789934158325195, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6957215070724487, + "num_tokens": 401122728.0, + "step": 15514 + }, + { + "epoch": 1.7038216560509554, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.5844411849975586, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7161766290664673, + "num_tokens": 401142675.0, + "step": 15515 + }, + { + "epoch": 1.7039314737535691, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1357779502868652, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7062879204750061, + "num_tokens": 401170929.0, + "step": 15516 + }, + { + "epoch": 1.7040412914561829, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.182281970977783, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.6980643272399902, + "num_tokens": 401198278.0, + "step": 15517 + }, + { + "epoch": 1.7041511091587964, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1355650424957275, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7198387384414673, + "num_tokens": 401225270.0, + "step": 15518 + }, + { + "epoch": 1.70426092686141, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.191030740737915, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.738223671913147, + "num_tokens": 401249881.0, + "step": 15519 + }, + { + "epoch": 1.7043707445640237, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.365647315979004, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7239043712615967, + "num_tokens": 401272146.0, + "step": 15520 + }, + { + "epoch": 1.7044805622666375, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.466583251953125, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7251710891723633, + "num_tokens": 401293937.0, + "step": 15521 + }, + { + "epoch": 1.704590379969251, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2575371265411377, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.735661506652832, + "num_tokens": 401317738.0, + "step": 15522 + }, + { + "epoch": 1.7047001976718648, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.1263504028320312, + "learning_rate": 1e-06, + "loss": 1.0161, + "mean_token_accuracy": 0.6841447949409485, + "num_tokens": 401348387.0, + "step": 15523 + }, + { + "epoch": 1.7048100153744783, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.0200958251953125, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7192485332489014, + "num_tokens": 401381340.0, + "step": 15524 + }, + { + "epoch": 1.704919833077092, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.461564302444458, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7239453196525574, + "num_tokens": 401402595.0, + "step": 15525 + }, + { + "epoch": 1.7050296507797058, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3402135372161865, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7304962873458862, + "num_tokens": 401425200.0, + "step": 15526 + }, + { + "epoch": 1.7051394684823193, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3906877040863037, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.7073209285736084, + "num_tokens": 401448836.0, + "step": 15527 + }, + { + "epoch": 1.7052492861849329, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.4395179748535156, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7251850366592407, + "num_tokens": 401470107.0, + "step": 15528 + }, + { + "epoch": 1.7053591038875466, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1725881099700928, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7055344581604004, + "num_tokens": 401497024.0, + "step": 15529 + }, + { + "epoch": 1.7054689215901604, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 1.9984053373336792, + "learning_rate": 1e-06, + "loss": 1.0287, + "mean_token_accuracy": 0.6811531186103821, + "num_tokens": 401529644.0, + "step": 15530 + }, + { + "epoch": 1.7055787392927741, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.245439291000366, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.716358482837677, + "num_tokens": 401553693.0, + "step": 15531 + }, + { + "epoch": 1.7056885569953877, + "ewc_loss": 8.821487426757812e-06, + "grad_norm": 2.2962088584899902, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7148751020431519, + "num_tokens": 401577197.0, + "step": 15532 + }, + { + "epoch": 1.7057983746980012, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.06526780128479, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6959517002105713, + "num_tokens": 401608064.0, + "step": 15533 + }, + { + "epoch": 1.705908192400615, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.582226037979126, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7431522011756897, + "num_tokens": 401628407.0, + "step": 15534 + }, + { + "epoch": 1.7060180101032287, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1440024375915527, + "learning_rate": 1e-06, + "loss": 1.0218, + "mean_token_accuracy": 0.6905065178871155, + "num_tokens": 401656515.0, + "step": 15535 + }, + { + "epoch": 1.7061278278058423, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.2837753295898438, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7224447727203369, + "num_tokens": 401681436.0, + "step": 15536 + }, + { + "epoch": 1.7062376455084558, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.2420601844787598, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7172467112541199, + "num_tokens": 401706904.0, + "step": 15537 + }, + { + "epoch": 1.7063474632110696, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.319401264190674, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7285916209220886, + "num_tokens": 401730977.0, + "step": 15538 + }, + { + "epoch": 1.7064572809136833, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.097818613052368, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7305707931518555, + "num_tokens": 401757877.0, + "step": 15539 + }, + { + "epoch": 1.706567098616297, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.190990924835205, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7265799641609192, + "num_tokens": 401784118.0, + "step": 15540 + }, + { + "epoch": 1.7066769163189106, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.222442626953125, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7532064914703369, + "num_tokens": 401809143.0, + "step": 15541 + }, + { + "epoch": 1.7067867340215241, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.5178282260894775, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7147507667541504, + "num_tokens": 401830948.0, + "step": 15542 + }, + { + "epoch": 1.706896551724138, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.4090306758880615, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7324713468551636, + "num_tokens": 401853041.0, + "step": 15543 + }, + { + "epoch": 1.7070063694267517, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1170456409454346, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7095126509666443, + "num_tokens": 401883021.0, + "step": 15544 + }, + { + "epoch": 1.7071161871293654, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.453970193862915, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7149994969367981, + "num_tokens": 401905419.0, + "step": 15545 + }, + { + "epoch": 1.707226004831979, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1018154621124268, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7045888900756836, + "num_tokens": 401934862.0, + "step": 15546 + }, + { + "epoch": 1.7073358225345925, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.0057761669158936, + "learning_rate": 1e-06, + "loss": 0.9887, + "mean_token_accuracy": 0.6927144527435303, + "num_tokens": 401965412.0, + "step": 15547 + }, + { + "epoch": 1.7074456402372062, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3091483116149902, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7042555212974548, + "num_tokens": 401990219.0, + "step": 15548 + }, + { + "epoch": 1.70755545793982, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.5154688358306885, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7265109419822693, + "num_tokens": 402010596.0, + "step": 15549 + }, + { + "epoch": 1.7076652756424335, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3376481533050537, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7244161367416382, + "num_tokens": 402035555.0, + "step": 15550 + }, + { + "epoch": 1.707775093345047, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.076024293899536, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7062435150146484, + "num_tokens": 402064140.0, + "step": 15551 + }, + { + "epoch": 1.7078849110476608, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.4488635063171387, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7208878993988037, + "num_tokens": 402085461.0, + "step": 15552 + }, + { + "epoch": 1.7079947287502746, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1090409755706787, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7251870036125183, + "num_tokens": 402113389.0, + "step": 15553 + }, + { + "epoch": 1.7081045464528883, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.455540895462036, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7207493782043457, + "num_tokens": 402134794.0, + "step": 15554 + }, + { + "epoch": 1.7082143641555019, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.20454740524292, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7023841142654419, + "num_tokens": 402162204.0, + "step": 15555 + }, + { + "epoch": 1.7083241818581154, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.651028633117676, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7323224544525146, + "num_tokens": 402180576.0, + "step": 15556 + }, + { + "epoch": 1.7084339995607292, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.300215721130371, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7429195642471313, + "num_tokens": 402202089.0, + "step": 15557 + }, + { + "epoch": 1.708543817263343, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.4920434951782227, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7411631345748901, + "num_tokens": 402221884.0, + "step": 15558 + }, + { + "epoch": 1.7086536349659565, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.568974494934082, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7180446982383728, + "num_tokens": 402243592.0, + "step": 15559 + }, + { + "epoch": 1.7087634526685702, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.264371633529663, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7037007808685303, + "num_tokens": 402268787.0, + "step": 15560 + }, + { + "epoch": 1.7088732703711838, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.0422422885894775, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7103750705718994, + "num_tokens": 402299378.0, + "step": 15561 + }, + { + "epoch": 1.7089830880737975, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.1377341747283936, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7282314300537109, + "num_tokens": 402324054.0, + "step": 15562 + }, + { + "epoch": 1.7090929057764113, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.29240083694458, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7262429594993591, + "num_tokens": 402346913.0, + "step": 15563 + }, + { + "epoch": 1.7092027234790248, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.2216105461120605, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.709515392780304, + "num_tokens": 402373163.0, + "step": 15564 + }, + { + "epoch": 1.7093125411816383, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.24851655960083, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7426889538764954, + "num_tokens": 402396775.0, + "step": 15565 + }, + { + "epoch": 1.709422358884252, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.110083818435669, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.721373975276947, + "num_tokens": 402426353.0, + "step": 15566 + }, + { + "epoch": 1.7095321765868658, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 1.9477144479751587, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.6972467303276062, + "num_tokens": 402457270.0, + "step": 15567 + }, + { + "epoch": 1.7096419942894796, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.206552028656006, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7152186632156372, + "num_tokens": 402483166.0, + "step": 15568 + }, + { + "epoch": 1.7097518119920931, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.13396954536438, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7318801283836365, + "num_tokens": 402510246.0, + "step": 15569 + }, + { + "epoch": 1.7098616296947067, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 1.9859755039215088, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7281247973442078, + "num_tokens": 402540081.0, + "step": 15570 + }, + { + "epoch": 1.7099714473973204, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.010397434234619, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7030857801437378, + "num_tokens": 402570877.0, + "step": 15571 + }, + { + "epoch": 1.7100812650999342, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.3424606323242188, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7295597195625305, + "num_tokens": 402594434.0, + "step": 15572 + }, + { + "epoch": 1.7101910828025477, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.27878999710083, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.726597785949707, + "num_tokens": 402618572.0, + "step": 15573 + }, + { + "epoch": 1.7103009005051615, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.6156320571899414, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7145841121673584, + "num_tokens": 402639406.0, + "step": 15574 + }, + { + "epoch": 1.710410718207775, + "ewc_loss": 8.940696716308594e-06, + "grad_norm": 2.706521511077881, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7129887938499451, + "num_tokens": 402658305.0, + "step": 15575 + }, + { + "epoch": 1.7105205359103888, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9965641498565674, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.6978040933609009, + "num_tokens": 402690186.0, + "step": 15576 + }, + { + "epoch": 1.7106303536130025, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0397098064422607, + "learning_rate": 1e-06, + "loss": 1.0376, + "mean_token_accuracy": 0.6874693036079407, + "num_tokens": 402722651.0, + "step": 15577 + }, + { + "epoch": 1.710740171315616, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3531582355499268, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7395563125610352, + "num_tokens": 402746767.0, + "step": 15578 + }, + { + "epoch": 1.7108499890182296, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9584165811538696, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7262080907821655, + "num_tokens": 402777349.0, + "step": 15579 + }, + { + "epoch": 1.7109598067208434, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.113762855529785, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7150738835334778, + "num_tokens": 402806854.0, + "step": 15580 + }, + { + "epoch": 1.7110696244234571, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0860626697540283, + "learning_rate": 1e-06, + "loss": 0.9841, + "mean_token_accuracy": 0.7048372030258179, + "num_tokens": 402836077.0, + "step": 15581 + }, + { + "epoch": 1.7111794421260709, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.20772123336792, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.708809494972229, + "num_tokens": 402862507.0, + "step": 15582 + }, + { + "epoch": 1.7112892598286844, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9181439876556396, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7207860946655273, + "num_tokens": 402895496.0, + "step": 15583 + }, + { + "epoch": 1.711399077531298, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.181671142578125, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7158923745155334, + "num_tokens": 402922000.0, + "step": 15584 + }, + { + "epoch": 1.7115088952339117, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.224494695663452, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7135215997695923, + "num_tokens": 402946708.0, + "step": 15585 + }, + { + "epoch": 1.7116187129365255, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.173128128051758, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7136595845222473, + "num_tokens": 402974797.0, + "step": 15586 + }, + { + "epoch": 1.711728530639139, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2660841941833496, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7249885201454163, + "num_tokens": 402999664.0, + "step": 15587 + }, + { + "epoch": 1.7118383483417525, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.1007039546966553, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6962406635284424, + "num_tokens": 403031461.0, + "step": 15588 + }, + { + "epoch": 1.7119481660443663, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.158078908920288, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7261099219322205, + "num_tokens": 403058704.0, + "step": 15589 + }, + { + "epoch": 1.71205798374698, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.01452898979187, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7349741458892822, + "num_tokens": 403089446.0, + "step": 15590 + }, + { + "epoch": 1.7121678014495938, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.304342031478882, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.6966004371643066, + "num_tokens": 403116757.0, + "step": 15591 + }, + { + "epoch": 1.7122776191522073, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3141045570373535, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7264593243598938, + "num_tokens": 403138993.0, + "step": 15592 + }, + { + "epoch": 1.7123874368548209, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.369009256362915, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7193474173545837, + "num_tokens": 403162916.0, + "step": 15593 + }, + { + "epoch": 1.7124972545574346, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0133063793182373, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7246036529541016, + "num_tokens": 403191582.0, + "step": 15594 + }, + { + "epoch": 1.7126070722600484, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.155322551727295, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7062922716140747, + "num_tokens": 403218747.0, + "step": 15595 + }, + { + "epoch": 1.7127168899626621, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.1093406677246094, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.7016777992248535, + "num_tokens": 403247716.0, + "step": 15596 + }, + { + "epoch": 1.7128267076652757, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9953129291534424, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7289614677429199, + "num_tokens": 403277059.0, + "step": 15597 + }, + { + "epoch": 1.7129365253678892, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.127708911895752, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7055134773254395, + "num_tokens": 403304477.0, + "step": 15598 + }, + { + "epoch": 1.713046343070503, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.4940550327301025, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7357031106948853, + "num_tokens": 403327045.0, + "step": 15599 + }, + { + "epoch": 1.7131561607731167, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.5954108238220215, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7332992553710938, + "num_tokens": 403346736.0, + "step": 15600 + }, + { + "epoch": 1.7132659784757303, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3178279399871826, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7212570905685425, + "num_tokens": 403372578.0, + "step": 15601 + }, + { + "epoch": 1.7133757961783438, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.5110585689544678, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7143895626068115, + "num_tokens": 403393365.0, + "step": 15602 + }, + { + "epoch": 1.7134856138809575, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.191030502319336, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7130376696586609, + "num_tokens": 403419727.0, + "step": 15603 + }, + { + "epoch": 1.7135954315835713, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.1076784133911133, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7193392515182495, + "num_tokens": 403447141.0, + "step": 15604 + }, + { + "epoch": 1.713705249286185, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3401424884796143, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7115820050239563, + "num_tokens": 403471657.0, + "step": 15605 + }, + { + "epoch": 1.7138150669887986, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2150323390960693, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7224926948547363, + "num_tokens": 403496854.0, + "step": 15606 + }, + { + "epoch": 1.7139248846914121, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0682191848754883, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7177679538726807, + "num_tokens": 403524503.0, + "step": 15607 + }, + { + "epoch": 1.7140347023940259, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.05795955657959, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.6984878778457642, + "num_tokens": 403554962.0, + "step": 15608 + }, + { + "epoch": 1.7141445200966396, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9196560382843018, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7113796472549438, + "num_tokens": 403587714.0, + "step": 15609 + }, + { + "epoch": 1.7142543377992534, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.232492208480835, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7009401321411133, + "num_tokens": 403616592.0, + "step": 15610 + }, + { + "epoch": 1.714364155501867, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.1815898418426514, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7236496806144714, + "num_tokens": 403641132.0, + "step": 15611 + }, + { + "epoch": 1.7144739732044805, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.90830659866333, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7181720733642578, + "num_tokens": 403673395.0, + "step": 15612 + }, + { + "epoch": 1.7145837909070942, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.4221880435943604, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7492899894714355, + "num_tokens": 403695074.0, + "step": 15613 + }, + { + "epoch": 1.714693608609708, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.28853178024292, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.713217556476593, + "num_tokens": 403722761.0, + "step": 15614 + }, + { + "epoch": 1.7148034263123215, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.5621540546417236, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7117340564727783, + "num_tokens": 403743307.0, + "step": 15615 + }, + { + "epoch": 1.714913244014935, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3579416275024414, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7151106595993042, + "num_tokens": 403767386.0, + "step": 15616 + }, + { + "epoch": 1.7150230617175488, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0995516777038574, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.726609468460083, + "num_tokens": 403795522.0, + "step": 15617 + }, + { + "epoch": 1.7151328794201626, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.289724111557007, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7154077291488647, + "num_tokens": 403821048.0, + "step": 15618 + }, + { + "epoch": 1.7152426971227763, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.878676414489746, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7220909595489502, + "num_tokens": 403854081.0, + "step": 15619 + }, + { + "epoch": 1.7153525148253899, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9144188165664673, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7255978584289551, + "num_tokens": 403887105.0, + "step": 15620 + }, + { + "epoch": 1.7154623325280034, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.322816848754883, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7118614912033081, + "num_tokens": 403911293.0, + "step": 15621 + }, + { + "epoch": 1.7155721502306172, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3865957260131836, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7206910848617554, + "num_tokens": 403932423.0, + "step": 15622 + }, + { + "epoch": 1.715681967933231, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.013310194015503, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7080747485160828, + "num_tokens": 403964248.0, + "step": 15623 + }, + { + "epoch": 1.7157917856358444, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.165696382522583, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7346206307411194, + "num_tokens": 403990120.0, + "step": 15624 + }, + { + "epoch": 1.7159016033384582, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.572835683822632, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7381769418716431, + "num_tokens": 404009131.0, + "step": 15625 + }, + { + "epoch": 1.7160114210410717, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0945863723754883, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.712098240852356, + "num_tokens": 404039592.0, + "step": 15626 + }, + { + "epoch": 1.7161212387436855, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3939621448516846, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7253974080085754, + "num_tokens": 404061611.0, + "step": 15627 + }, + { + "epoch": 1.7162310564462993, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.1545827388763428, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7049093246459961, + "num_tokens": 404089891.0, + "step": 15628 + }, + { + "epoch": 1.7163408741489128, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.098287582397461, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7091126441955566, + "num_tokens": 404118032.0, + "step": 15629 + }, + { + "epoch": 1.7164506918515263, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.664569616317749, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7203577756881714, + "num_tokens": 404139583.0, + "step": 15630 + }, + { + "epoch": 1.71656050955414, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.461582899093628, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7140135765075684, + "num_tokens": 404162484.0, + "step": 15631 + }, + { + "epoch": 1.7166703272567538, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.302881956100464, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7371925115585327, + "num_tokens": 404185439.0, + "step": 15632 + }, + { + "epoch": 1.7167801449593676, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.2604820728302, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7261616587638855, + "num_tokens": 404209797.0, + "step": 15633 + }, + { + "epoch": 1.7168899626619811, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.4118850231170654, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7435092926025391, + "num_tokens": 404230140.0, + "step": 15634 + }, + { + "epoch": 1.7169997803645947, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.423609495162964, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7099241018295288, + "num_tokens": 404254200.0, + "step": 15635 + }, + { + "epoch": 1.7171095980672084, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.288924217224121, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7225056290626526, + "num_tokens": 404278519.0, + "step": 15636 + }, + { + "epoch": 1.7172194157698222, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.298227548599243, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7058495879173279, + "num_tokens": 404303770.0, + "step": 15637 + }, + { + "epoch": 1.7173292334724357, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.507856607437134, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7304667830467224, + "num_tokens": 404325138.0, + "step": 15638 + }, + { + "epoch": 1.7174390511750495, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.3536808490753174, + "learning_rate": 1e-06, + "loss": 0.9591, + "mean_token_accuracy": 0.7051912546157837, + "num_tokens": 404349688.0, + "step": 15639 + }, + { + "epoch": 1.717548868877663, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9011201858520508, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.711322009563446, + "num_tokens": 404382347.0, + "step": 15640 + }, + { + "epoch": 1.7176586865802768, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.0506083965301514, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.716914713382721, + "num_tokens": 404409840.0, + "step": 15641 + }, + { + "epoch": 1.7177685042828905, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.032508373260498, + "learning_rate": 1e-06, + "loss": 1.026, + "mean_token_accuracy": 0.6915073394775391, + "num_tokens": 404441008.0, + "step": 15642 + }, + { + "epoch": 1.717878321985504, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.141676425933838, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7082780599594116, + "num_tokens": 404468403.0, + "step": 15643 + }, + { + "epoch": 1.7179881396881176, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 2.106144666671753, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7197823524475098, + "num_tokens": 404497085.0, + "step": 15644 + }, + { + "epoch": 1.7180979573907313, + "ewc_loss": 9.000301361083984e-06, + "grad_norm": 1.9608161449432373, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7035925388336182, + "num_tokens": 404529952.0, + "step": 15645 + }, + { + "epoch": 1.718207775093345, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.0100722312927246, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7021517753601074, + "num_tokens": 404559855.0, + "step": 15646 + }, + { + "epoch": 1.7183175927959589, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.1691088676452637, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7131403684616089, + "num_tokens": 404585447.0, + "step": 15647 + }, + { + "epoch": 1.7184274104985724, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.059209108352661, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7133416533470154, + "num_tokens": 404615258.0, + "step": 15648 + }, + { + "epoch": 1.718537228201186, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.245558738708496, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7112022042274475, + "num_tokens": 404640487.0, + "step": 15649 + }, + { + "epoch": 1.7186470459037997, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.133904218673706, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.7040928602218628, + "num_tokens": 404669984.0, + "step": 15650 + }, + { + "epoch": 1.7187568636064134, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.1755383014678955, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7123771905899048, + "num_tokens": 404695795.0, + "step": 15651 + }, + { + "epoch": 1.718866681309027, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.367213487625122, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7331804037094116, + "num_tokens": 404716863.0, + "step": 15652 + }, + { + "epoch": 1.7189764990116405, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.7462291717529297, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7344859838485718, + "num_tokens": 404733635.0, + "step": 15653 + }, + { + "epoch": 1.7190863167142543, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.5731406211853027, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7369651794433594, + "num_tokens": 404754534.0, + "step": 15654 + }, + { + "epoch": 1.719196134416868, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.037518262863159, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6993081569671631, + "num_tokens": 404786352.0, + "step": 15655 + }, + { + "epoch": 1.7193059521194818, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.2315287590026855, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7156392335891724, + "num_tokens": 404812670.0, + "step": 15656 + }, + { + "epoch": 1.7194157698220953, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.379204273223877, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7289240956306458, + "num_tokens": 404833971.0, + "step": 15657 + }, + { + "epoch": 1.7195255875247089, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 1.9602006673812866, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7147356271743774, + "num_tokens": 404865009.0, + "step": 15658 + }, + { + "epoch": 1.7196354052273226, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.2126002311706543, + "learning_rate": 1e-06, + "loss": 1.0193, + "mean_token_accuracy": 0.6980454921722412, + "num_tokens": 404891573.0, + "step": 15659 + }, + { + "epoch": 1.7197452229299364, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.219542980194092, + "learning_rate": 1e-06, + "loss": 1.0308, + "mean_token_accuracy": 0.711837649345398, + "num_tokens": 404918890.0, + "step": 15660 + }, + { + "epoch": 1.7198550406325501, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.1760969161987305, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7474437952041626, + "num_tokens": 404943132.0, + "step": 15661 + }, + { + "epoch": 1.7199648583351637, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.259758949279785, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.7102200388908386, + "num_tokens": 404970183.0, + "step": 15662 + }, + { + "epoch": 1.7200746760377772, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.217250108718872, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7144464254379272, + "num_tokens": 404996090.0, + "step": 15663 + }, + { + "epoch": 1.720184493740391, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.0528736114501953, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7216656804084778, + "num_tokens": 405024775.0, + "step": 15664 + }, + { + "epoch": 1.7202943114430047, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 1.9920631647109985, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7032912969589233, + "num_tokens": 405056691.0, + "step": 15665 + }, + { + "epoch": 1.7204041291456182, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.3193581104278564, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.708088219165802, + "num_tokens": 405082271.0, + "step": 15666 + }, + { + "epoch": 1.7205139468482318, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.6135404109954834, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7173556089401245, + "num_tokens": 405104542.0, + "step": 15667 + }, + { + "epoch": 1.7206237645508455, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.0124058723449707, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7240265607833862, + "num_tokens": 405135485.0, + "step": 15668 + }, + { + "epoch": 1.7207335822534593, + "ewc_loss": 9.059906005859375e-06, + "grad_norm": 2.294428586959839, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.726273775100708, + "num_tokens": 405159707.0, + "step": 15669 + }, + { + "epoch": 1.720843399956073, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3372278213500977, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7195876240730286, + "num_tokens": 405185350.0, + "step": 15670 + }, + { + "epoch": 1.7209532176586866, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.106769323348999, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7311049103736877, + "num_tokens": 405214522.0, + "step": 15671 + }, + { + "epoch": 1.7210630353613001, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.596381187438965, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7220768928527832, + "num_tokens": 405236811.0, + "step": 15672 + }, + { + "epoch": 1.7211728530639139, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.022632598876953, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7451045513153076, + "num_tokens": 405265540.0, + "step": 15673 + }, + { + "epoch": 1.7212826707665276, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1891589164733887, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7077306509017944, + "num_tokens": 405295474.0, + "step": 15674 + }, + { + "epoch": 1.7213924884691414, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.232043743133545, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7243728637695312, + "num_tokens": 405321196.0, + "step": 15675 + }, + { + "epoch": 1.721502306171755, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0641918182373047, + "learning_rate": 1e-06, + "loss": 0.9833, + "mean_token_accuracy": 0.7001415491104126, + "num_tokens": 405351944.0, + "step": 15676 + }, + { + "epoch": 1.7216121238743685, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.358783721923828, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7085702419281006, + "num_tokens": 405378336.0, + "step": 15677 + }, + { + "epoch": 1.7217219415769822, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1256556510925293, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7367161512374878, + "num_tokens": 405404115.0, + "step": 15678 + }, + { + "epoch": 1.721831759279596, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3790574073791504, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.713333785533905, + "num_tokens": 405427676.0, + "step": 15679 + }, + { + "epoch": 1.7219415769822095, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0726895332336426, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.6986511945724487, + "num_tokens": 405456316.0, + "step": 15680 + }, + { + "epoch": 1.722051394684823, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.513521194458008, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7149748206138611, + "num_tokens": 405478827.0, + "step": 15681 + }, + { + "epoch": 1.7221612123874368, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2115190029144287, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7396671175956726, + "num_tokens": 405503775.0, + "step": 15682 + }, + { + "epoch": 1.7222710300900506, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.271191358566284, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.7076925039291382, + "num_tokens": 405530284.0, + "step": 15683 + }, + { + "epoch": 1.7223808477926643, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1455023288726807, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7063942551612854, + "num_tokens": 405560487.0, + "step": 15684 + }, + { + "epoch": 1.7224906654952779, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2217650413513184, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7284681797027588, + "num_tokens": 405586029.0, + "step": 15685 + }, + { + "epoch": 1.7226004831978914, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.052199602127075, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.718256950378418, + "num_tokens": 405614303.0, + "step": 15686 + }, + { + "epoch": 1.7227103009005051, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.341681480407715, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.720173716545105, + "num_tokens": 405637574.0, + "step": 15687 + }, + { + "epoch": 1.722820118603119, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.224177122116089, + "learning_rate": 1e-06, + "loss": 0.9708, + "mean_token_accuracy": 0.6964590549468994, + "num_tokens": 405664647.0, + "step": 15688 + }, + { + "epoch": 1.7229299363057324, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.9473224878311157, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7302058935165405, + "num_tokens": 405695245.0, + "step": 15689 + }, + { + "epoch": 1.7230397540083462, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.134658098220825, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7397122383117676, + "num_tokens": 405721043.0, + "step": 15690 + }, + { + "epoch": 1.7231495717109597, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1664552688598633, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7063983678817749, + "num_tokens": 405748441.0, + "step": 15691 + }, + { + "epoch": 1.7232593894135735, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4442739486694336, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7145404815673828, + "num_tokens": 405770211.0, + "step": 15692 + }, + { + "epoch": 1.7233692071161872, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.167391538619995, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7355538606643677, + "num_tokens": 405799085.0, + "step": 15693 + }, + { + "epoch": 1.7234790248188008, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.19598126411438, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7308664321899414, + "num_tokens": 405825810.0, + "step": 15694 + }, + { + "epoch": 1.7235888425214143, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.206193447113037, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.7008423209190369, + "num_tokens": 405853812.0, + "step": 15695 + }, + { + "epoch": 1.723698660224028, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3456051349639893, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7304728627204895, + "num_tokens": 405879039.0, + "step": 15696 + }, + { + "epoch": 1.7238084779266418, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1584789752960205, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7130811810493469, + "num_tokens": 405908301.0, + "step": 15697 + }, + { + "epoch": 1.7239182956292556, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4958581924438477, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7374783754348755, + "num_tokens": 405928513.0, + "step": 15698 + }, + { + "epoch": 1.7240281133318691, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0549397468566895, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.708466112613678, + "num_tokens": 405958633.0, + "step": 15699 + }, + { + "epoch": 1.7241379310344827, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.259727716445923, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7299566864967346, + "num_tokens": 405983518.0, + "step": 15700 + }, + { + "epoch": 1.7242477487370964, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.165787696838379, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7261314392089844, + "num_tokens": 406009479.0, + "step": 15701 + }, + { + "epoch": 1.7243575664397102, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.667253017425537, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.720745325088501, + "num_tokens": 406030749.0, + "step": 15702 + }, + { + "epoch": 1.7244673841423237, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.320136070251465, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7185484766960144, + "num_tokens": 406056447.0, + "step": 15703 + }, + { + "epoch": 1.7245772018449375, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.766608476638794, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.753793478012085, + "num_tokens": 406074410.0, + "step": 15704 + }, + { + "epoch": 1.724687019547551, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.551628589630127, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7428269386291504, + "num_tokens": 406096397.0, + "step": 15705 + }, + { + "epoch": 1.7247968372501647, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2107250690460205, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7018411159515381, + "num_tokens": 406123869.0, + "step": 15706 + }, + { + "epoch": 1.7249066549527785, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.307913303375244, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.747689962387085, + "num_tokens": 406147337.0, + "step": 15707 + }, + { + "epoch": 1.725016472655392, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.367511749267578, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7132585048675537, + "num_tokens": 406171769.0, + "step": 15708 + }, + { + "epoch": 1.7251262903580056, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2712759971618652, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7123985290527344, + "num_tokens": 406196379.0, + "step": 15709 + }, + { + "epoch": 1.7252361080606193, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.108264207839966, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7167684435844421, + "num_tokens": 406226674.0, + "step": 15710 + }, + { + "epoch": 1.725345925763233, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.17240047454834, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7305017709732056, + "num_tokens": 406251785.0, + "step": 15711 + }, + { + "epoch": 1.7254557434658468, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4283270835876465, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7413430213928223, + "num_tokens": 406272446.0, + "step": 15712 + }, + { + "epoch": 1.7255655611684604, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3173413276672363, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7273392677307129, + "num_tokens": 406297309.0, + "step": 15713 + }, + { + "epoch": 1.725675378871074, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3126120567321777, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7167460322380066, + "num_tokens": 406322257.0, + "step": 15714 + }, + { + "epoch": 1.7257851965736877, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.302274703979492, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7174422144889832, + "num_tokens": 406346570.0, + "step": 15715 + }, + { + "epoch": 1.7258950142763014, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.12540340423584, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7228723764419556, + "num_tokens": 406372128.0, + "step": 15716 + }, + { + "epoch": 1.726004831978915, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.510965347290039, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7432070374488831, + "num_tokens": 406393200.0, + "step": 15717 + }, + { + "epoch": 1.7261146496815285, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2102138996124268, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6913002729415894, + "num_tokens": 406421258.0, + "step": 15718 + }, + { + "epoch": 1.7262244673841423, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.9010835886001587, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7355340123176575, + "num_tokens": 406454628.0, + "step": 15719 + }, + { + "epoch": 1.726334285086756, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.019270420074463, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7381336092948914, + "num_tokens": 406484194.0, + "step": 15720 + }, + { + "epoch": 1.7264441027893698, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4182536602020264, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7179111242294312, + "num_tokens": 406508641.0, + "step": 15721 + }, + { + "epoch": 1.7265539204919833, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1584858894348145, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.6972269415855408, + "num_tokens": 406535886.0, + "step": 15722 + }, + { + "epoch": 1.7266637381945968, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5489706993103027, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7207753658294678, + "num_tokens": 406557711.0, + "step": 15723 + }, + { + "epoch": 1.7267735558972106, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3731296062469482, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7261610627174377, + "num_tokens": 406579582.0, + "step": 15724 + }, + { + "epoch": 1.7268833735998244, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3127622604370117, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7392739057540894, + "num_tokens": 406603065.0, + "step": 15725 + }, + { + "epoch": 1.7269931913024381, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2614176273345947, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7135077714920044, + "num_tokens": 406627249.0, + "step": 15726 + }, + { + "epoch": 1.7271030090050516, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2184267044067383, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7413268685340881, + "num_tokens": 406651523.0, + "step": 15727 + }, + { + "epoch": 1.7272128267076652, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.189946174621582, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.6975712180137634, + "num_tokens": 406677929.0, + "step": 15728 + }, + { + "epoch": 1.727322644410279, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3640429973602295, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7180262804031372, + "num_tokens": 406703411.0, + "step": 15729 + }, + { + "epoch": 1.7274324621128927, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.270796537399292, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7141099572181702, + "num_tokens": 406727723.0, + "step": 15730 + }, + { + "epoch": 1.7275422798155062, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2485578060150146, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7046846151351929, + "num_tokens": 406755501.0, + "step": 15731 + }, + { + "epoch": 1.7276520975181198, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4842987060546875, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7343260049819946, + "num_tokens": 406775289.0, + "step": 15732 + }, + { + "epoch": 1.7277619152207335, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.150175094604492, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.729474663734436, + "num_tokens": 406804107.0, + "step": 15733 + }, + { + "epoch": 1.7278717329233473, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1771538257598877, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7176604270935059, + "num_tokens": 406832425.0, + "step": 15734 + }, + { + "epoch": 1.727981550625961, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2886252403259277, + "learning_rate": 1e-06, + "loss": 0.7393, + "mean_token_accuracy": 0.7613041400909424, + "num_tokens": 406855526.0, + "step": 15735 + }, + { + "epoch": 1.7280913683285746, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.526461362838745, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7227895855903625, + "num_tokens": 406876907.0, + "step": 15736 + }, + { + "epoch": 1.728201186031188, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2061100006103516, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6984601616859436, + "num_tokens": 406904760.0, + "step": 15737 + }, + { + "epoch": 1.7283110037338019, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4269142150878906, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.726601779460907, + "num_tokens": 406928549.0, + "step": 15738 + }, + { + "epoch": 1.7284208214364156, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.6349740028381348, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7144663333892822, + "num_tokens": 406949708.0, + "step": 15739 + }, + { + "epoch": 1.7285306391390292, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0753486156463623, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.7131308317184448, + "num_tokens": 406977852.0, + "step": 15740 + }, + { + "epoch": 1.728640456841643, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2190215587615967, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7106683254241943, + "num_tokens": 407004183.0, + "step": 15741 + }, + { + "epoch": 1.7287502745442564, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2062501907348633, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7250622510910034, + "num_tokens": 407029641.0, + "step": 15742 + }, + { + "epoch": 1.7288600922468702, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.9836498498916626, + "learning_rate": 1e-06, + "loss": 1.0118, + "mean_token_accuracy": 0.6955114603042603, + "num_tokens": 407061311.0, + "step": 15743 + }, + { + "epoch": 1.728969909949484, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4349281787872314, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.734437108039856, + "num_tokens": 407083842.0, + "step": 15744 + }, + { + "epoch": 1.7290797276520975, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4436309337615967, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7129186391830444, + "num_tokens": 407106403.0, + "step": 15745 + }, + { + "epoch": 1.729189545354711, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.538818359375, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7359058260917664, + "num_tokens": 407127611.0, + "step": 15746 + }, + { + "epoch": 1.7292993630573248, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.133023500442505, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7084549069404602, + "num_tokens": 407155998.0, + "step": 15747 + }, + { + "epoch": 1.7294091807599385, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1690196990966797, + "learning_rate": 1e-06, + "loss": 0.7784, + "mean_token_accuracy": 0.7547510862350464, + "num_tokens": 407180192.0, + "step": 15748 + }, + { + "epoch": 1.7295189984625523, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.196610689163208, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7086101770401001, + "num_tokens": 407207098.0, + "step": 15749 + }, + { + "epoch": 1.7296288161651658, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.365264654159546, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7416801452636719, + "num_tokens": 407229106.0, + "step": 15750 + }, + { + "epoch": 1.7297386338677794, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.41154408454895, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7336571216583252, + "num_tokens": 407251733.0, + "step": 15751 + }, + { + "epoch": 1.7298484515703931, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.241405963897705, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7103042006492615, + "num_tokens": 407276925.0, + "step": 15752 + }, + { + "epoch": 1.7299582692730069, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2475204467773438, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7141099572181702, + "num_tokens": 407301318.0, + "step": 15753 + }, + { + "epoch": 1.7300680869756204, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.486570119857788, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.73257976770401, + "num_tokens": 407323652.0, + "step": 15754 + }, + { + "epoch": 1.7301779046782342, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0270745754241943, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7188316583633423, + "num_tokens": 407353689.0, + "step": 15755 + }, + { + "epoch": 1.7302877223808477, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.302957057952881, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7077856063842773, + "num_tokens": 407378942.0, + "step": 15756 + }, + { + "epoch": 1.7303975400834615, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.263439655303955, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7238340377807617, + "num_tokens": 407404492.0, + "step": 15757 + }, + { + "epoch": 1.7305073577860752, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.325716733932495, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.722882866859436, + "num_tokens": 407429503.0, + "step": 15758 + }, + { + "epoch": 1.7306171754886888, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0672876834869385, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.721284031867981, + "num_tokens": 407459862.0, + "step": 15759 + }, + { + "epoch": 1.7307269931913023, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3316636085510254, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7309019565582275, + "num_tokens": 407485344.0, + "step": 15760 + }, + { + "epoch": 1.730836810893916, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.344362258911133, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.727653980255127, + "num_tokens": 407508639.0, + "step": 15761 + }, + { + "epoch": 1.7309466285965298, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2604732513427734, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7147665023803711, + "num_tokens": 407535977.0, + "step": 15762 + }, + { + "epoch": 1.7310564462991436, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.386180877685547, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7409329414367676, + "num_tokens": 407557839.0, + "step": 15763 + }, + { + "epoch": 1.731166264001757, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.595881462097168, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7302470207214355, + "num_tokens": 407577992.0, + "step": 15764 + }, + { + "epoch": 1.7312760817043706, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.238598585128784, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7397627830505371, + "num_tokens": 407602088.0, + "step": 15765 + }, + { + "epoch": 1.7313858994069844, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.571218490600586, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7368870973587036, + "num_tokens": 407624483.0, + "step": 15766 + }, + { + "epoch": 1.7314957171095982, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4569644927978516, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.698897123336792, + "num_tokens": 407648718.0, + "step": 15767 + }, + { + "epoch": 1.7316055348122117, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.419163227081299, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7220604419708252, + "num_tokens": 407671524.0, + "step": 15768 + }, + { + "epoch": 1.7317153525148252, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.361077308654785, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7172929048538208, + "num_tokens": 407696736.0, + "step": 15769 + }, + { + "epoch": 1.731825170217439, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.360447406768799, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7354569435119629, + "num_tokens": 407719242.0, + "step": 15770 + }, + { + "epoch": 1.7319349879200527, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.6256465911865234, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7314292192459106, + "num_tokens": 407741647.0, + "step": 15771 + }, + { + "epoch": 1.7320448056226665, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3168256282806396, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7298295497894287, + "num_tokens": 407766431.0, + "step": 15772 + }, + { + "epoch": 1.73215462332528, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1658668518066406, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.720619261264801, + "num_tokens": 407791725.0, + "step": 15773 + }, + { + "epoch": 1.7322644410278936, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.403083562850952, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7001944184303284, + "num_tokens": 407814167.0, + "step": 15774 + }, + { + "epoch": 1.7323742587305073, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.05656361579895, + "learning_rate": 1e-06, + "loss": 0.9712, + "mean_token_accuracy": 0.6965000629425049, + "num_tokens": 407843181.0, + "step": 15775 + }, + { + "epoch": 1.732484076433121, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2394351959228516, + "learning_rate": 1e-06, + "loss": 0.9965, + "mean_token_accuracy": 0.6974434852600098, + "num_tokens": 407868658.0, + "step": 15776 + }, + { + "epoch": 1.7325938941357348, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.125159978866577, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7357168197631836, + "num_tokens": 407895419.0, + "step": 15777 + }, + { + "epoch": 1.7327037118383484, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.212907552719116, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7275327444076538, + "num_tokens": 407920037.0, + "step": 15778 + }, + { + "epoch": 1.732813529540962, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.390385389328003, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.717932403087616, + "num_tokens": 407943819.0, + "step": 15779 + }, + { + "epoch": 1.7329233472435757, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0148251056671143, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7124297618865967, + "num_tokens": 407974168.0, + "step": 15780 + }, + { + "epoch": 1.7330331649461894, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.277761697769165, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7301428318023682, + "num_tokens": 408000107.0, + "step": 15781 + }, + { + "epoch": 1.733142982648803, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.186450481414795, + "learning_rate": 1e-06, + "loss": 1.0274, + "mean_token_accuracy": 0.6833674311637878, + "num_tokens": 408029346.0, + "step": 15782 + }, + { + "epoch": 1.7332528003514165, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.352964162826538, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7070835828781128, + "num_tokens": 408054076.0, + "step": 15783 + }, + { + "epoch": 1.7333626180540302, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.9506691694259644, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.705278754234314, + "num_tokens": 408087889.0, + "step": 15784 + }, + { + "epoch": 1.733472435756644, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1310362815856934, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7177091836929321, + "num_tokens": 408115327.0, + "step": 15785 + }, + { + "epoch": 1.7335822534592578, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4909558296203613, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.6988211870193481, + "num_tokens": 408139826.0, + "step": 15786 + }, + { + "epoch": 1.7336920711618713, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4517464637756348, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7156662344932556, + "num_tokens": 408163102.0, + "step": 15787 + }, + { + "epoch": 1.7338018888644848, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.545591354370117, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7158524394035339, + "num_tokens": 408186929.0, + "step": 15788 + }, + { + "epoch": 1.7339117065670986, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.259514093399048, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6864621639251709, + "num_tokens": 408213628.0, + "step": 15789 + }, + { + "epoch": 1.7340215242697123, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.124380111694336, + "learning_rate": 1e-06, + "loss": 0.9759, + "mean_token_accuracy": 0.6977325677871704, + "num_tokens": 408244207.0, + "step": 15790 + }, + { + "epoch": 1.734131341972326, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0333378314971924, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.730516254901886, + "num_tokens": 408273952.0, + "step": 15791 + }, + { + "epoch": 1.7342411596749396, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1536126136779785, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7230561971664429, + "num_tokens": 408300411.0, + "step": 15792 + }, + { + "epoch": 1.7343509773775532, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.489750623703003, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7366880178451538, + "num_tokens": 408321815.0, + "step": 15793 + }, + { + "epoch": 1.734460795080167, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4800021648406982, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.746833324432373, + "num_tokens": 408343248.0, + "step": 15794 + }, + { + "epoch": 1.7345706127827807, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3194739818573, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7030858993530273, + "num_tokens": 408368877.0, + "step": 15795 + }, + { + "epoch": 1.7346804304853942, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.30094838142395, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7034941911697388, + "num_tokens": 408394331.0, + "step": 15796 + }, + { + "epoch": 1.7347902481880078, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1899778842926025, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7019334435462952, + "num_tokens": 408419574.0, + "step": 15797 + }, + { + "epoch": 1.7349000658906215, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0836009979248047, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7215820550918579, + "num_tokens": 408446745.0, + "step": 15798 + }, + { + "epoch": 1.7350098835932353, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3355867862701416, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7127238512039185, + "num_tokens": 408470346.0, + "step": 15799 + }, + { + "epoch": 1.735119701295849, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.160555362701416, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7173100113868713, + "num_tokens": 408497927.0, + "step": 15800 + }, + { + "epoch": 1.7352295189984626, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.231027841567993, + "learning_rate": 1e-06, + "loss": 0.9608, + "mean_token_accuracy": 0.708827018737793, + "num_tokens": 408523528.0, + "step": 15801 + }, + { + "epoch": 1.735339336701076, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.8222496509552002, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7105463147163391, + "num_tokens": 408557516.0, + "step": 15802 + }, + { + "epoch": 1.7354491544036899, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2847132682800293, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.710088312625885, + "num_tokens": 408582193.0, + "step": 15803 + }, + { + "epoch": 1.7355589721063036, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5807244777679443, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7310322523117065, + "num_tokens": 408602622.0, + "step": 15804 + }, + { + "epoch": 1.7356687898089171, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.476731061935425, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7254252433776855, + "num_tokens": 408624316.0, + "step": 15805 + }, + { + "epoch": 1.735778607511531, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.111607074737549, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7258972525596619, + "num_tokens": 408652235.0, + "step": 15806 + }, + { + "epoch": 1.7358884252141444, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.281158924102783, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7138912081718445, + "num_tokens": 408677377.0, + "step": 15807 + }, + { + "epoch": 1.7359982429167582, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2266697883605957, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7139866352081299, + "num_tokens": 408703453.0, + "step": 15808 + }, + { + "epoch": 1.736108060619372, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.308894157409668, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7451152801513672, + "num_tokens": 408727630.0, + "step": 15809 + }, + { + "epoch": 1.7362178783219855, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.362877368927002, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.712846577167511, + "num_tokens": 408750923.0, + "step": 15810 + }, + { + "epoch": 1.736327696024599, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.190657377243042, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7120977640151978, + "num_tokens": 408778970.0, + "step": 15811 + }, + { + "epoch": 1.7364375137272128, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.282017946243286, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7450904250144958, + "num_tokens": 408803405.0, + "step": 15812 + }, + { + "epoch": 1.7365473314298265, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.233964681625366, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.6997108459472656, + "num_tokens": 408828951.0, + "step": 15813 + }, + { + "epoch": 1.7366571491324403, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.91309654712677, + "learning_rate": 1e-06, + "loss": 1.0016, + "mean_token_accuracy": 0.6806614995002747, + "num_tokens": 408864782.0, + "step": 15814 + }, + { + "epoch": 1.7367669668350538, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.610151767730713, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7382649779319763, + "num_tokens": 408883987.0, + "step": 15815 + }, + { + "epoch": 1.7368767845376674, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.185192823410034, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7067897319793701, + "num_tokens": 408911114.0, + "step": 15816 + }, + { + "epoch": 1.7369866022402811, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.490640878677368, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7369579076766968, + "num_tokens": 408932732.0, + "step": 15817 + }, + { + "epoch": 1.7370964199428949, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4235286712646484, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7415387630462646, + "num_tokens": 408956139.0, + "step": 15818 + }, + { + "epoch": 1.7372062376455084, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.474918842315674, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7290427684783936, + "num_tokens": 408979450.0, + "step": 15819 + }, + { + "epoch": 1.7373160553481222, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3935470581054688, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7375770807266235, + "num_tokens": 409002261.0, + "step": 15820 + }, + { + "epoch": 1.7374258730507357, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2184903621673584, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7113710641860962, + "num_tokens": 409031473.0, + "step": 15821 + }, + { + "epoch": 1.7375356907533495, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4114317893981934, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7217031717300415, + "num_tokens": 409053824.0, + "step": 15822 + }, + { + "epoch": 1.7376455084559632, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2976574897766113, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7078289985656738, + "num_tokens": 409081507.0, + "step": 15823 + }, + { + "epoch": 1.7377553261585768, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4482834339141846, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7439917325973511, + "num_tokens": 409104613.0, + "step": 15824 + }, + { + "epoch": 1.7378651438611903, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2024481296539307, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7147423028945923, + "num_tokens": 409131953.0, + "step": 15825 + }, + { + "epoch": 1.737974961563804, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3445513248443604, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7125258445739746, + "num_tokens": 409156621.0, + "step": 15826 + }, + { + "epoch": 1.7380847792664178, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.289323091506958, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7101982831954956, + "num_tokens": 409183001.0, + "step": 15827 + }, + { + "epoch": 1.7381945969690316, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4293060302734375, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7332613468170166, + "num_tokens": 409203539.0, + "step": 15828 + }, + { + "epoch": 1.738304414671645, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.202847719192505, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7252376079559326, + "num_tokens": 409230142.0, + "step": 15829 + }, + { + "epoch": 1.7384142323742586, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.24650239944458, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7004811763763428, + "num_tokens": 409255909.0, + "step": 15830 + }, + { + "epoch": 1.7385240500768724, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.18438982963562, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7351856231689453, + "num_tokens": 409282733.0, + "step": 15831 + }, + { + "epoch": 1.7386338677794861, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.146695852279663, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7315787076950073, + "num_tokens": 409309821.0, + "step": 15832 + }, + { + "epoch": 1.7387436854820997, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.240668773651123, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7447178959846497, + "num_tokens": 409334055.0, + "step": 15833 + }, + { + "epoch": 1.7388535031847132, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5672130584716797, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7167955040931702, + "num_tokens": 409354915.0, + "step": 15834 + }, + { + "epoch": 1.738963320887327, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.8256936073303223, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7499043941497803, + "num_tokens": 409372334.0, + "step": 15835 + }, + { + "epoch": 1.7390731385899407, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0738601684570312, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6944445371627808, + "num_tokens": 409401713.0, + "step": 15836 + }, + { + "epoch": 1.7391829562925545, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.080005407333374, + "learning_rate": 1e-06, + "loss": 0.9852, + "mean_token_accuracy": 0.6965913772583008, + "num_tokens": 409431433.0, + "step": 15837 + }, + { + "epoch": 1.739292773995168, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2776689529418945, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7209224700927734, + "num_tokens": 409454487.0, + "step": 15838 + }, + { + "epoch": 1.7394025916977816, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.192033529281616, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7261413335800171, + "num_tokens": 409479958.0, + "step": 15839 + }, + { + "epoch": 1.7395124094003953, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0537397861480713, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7236204743385315, + "num_tokens": 409509442.0, + "step": 15840 + }, + { + "epoch": 1.739622227103009, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5123050212860107, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7096425294876099, + "num_tokens": 409530593.0, + "step": 15841 + }, + { + "epoch": 1.7397320448056228, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.9481958150863647, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7155261039733887, + "num_tokens": 409563789.0, + "step": 15842 + }, + { + "epoch": 1.7398418625082364, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.060079336166382, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.709176778793335, + "num_tokens": 409592783.0, + "step": 15843 + }, + { + "epoch": 1.73995168021085, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2695534229278564, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7184605598449707, + "num_tokens": 409617309.0, + "step": 15844 + }, + { + "epoch": 1.7400614979134637, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2971134185791016, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7129596471786499, + "num_tokens": 409642941.0, + "step": 15845 + }, + { + "epoch": 1.7401713156160774, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.105830669403076, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7283791303634644, + "num_tokens": 409669861.0, + "step": 15846 + }, + { + "epoch": 1.740281133318691, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4773104190826416, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7098023891448975, + "num_tokens": 409691814.0, + "step": 15847 + }, + { + "epoch": 1.7403909510213045, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.279733657836914, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7116491198539734, + "num_tokens": 409717653.0, + "step": 15848 + }, + { + "epoch": 1.7405007687239182, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.4444377422332764, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7265610694885254, + "num_tokens": 409738514.0, + "step": 15849 + }, + { + "epoch": 1.740610586426532, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.359262704849243, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7035946846008301, + "num_tokens": 409762917.0, + "step": 15850 + }, + { + "epoch": 1.7407204041291457, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3695168495178223, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7282534837722778, + "num_tokens": 409786764.0, + "step": 15851 + }, + { + "epoch": 1.7408302218317593, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3062429428100586, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7114386558532715, + "num_tokens": 409814012.0, + "step": 15852 + }, + { + "epoch": 1.7409400395343728, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2516379356384277, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7145062685012817, + "num_tokens": 409841290.0, + "step": 15853 + }, + { + "epoch": 1.7410498572369866, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3167543411254883, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7143779397010803, + "num_tokens": 409865997.0, + "step": 15854 + }, + { + "epoch": 1.7411596749396003, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3840012550354004, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7067821025848389, + "num_tokens": 409888915.0, + "step": 15855 + }, + { + "epoch": 1.741269492642214, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.8486188650131226, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7005485892295837, + "num_tokens": 409922378.0, + "step": 15856 + }, + { + "epoch": 1.7413793103448276, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.239340305328369, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.754185676574707, + "num_tokens": 409945678.0, + "step": 15857 + }, + { + "epoch": 1.7414891280474412, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.301704168319702, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7135949730873108, + "num_tokens": 409968835.0, + "step": 15858 + }, + { + "epoch": 1.741598945750055, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.097297191619873, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7129051685333252, + "num_tokens": 409997469.0, + "step": 15859 + }, + { + "epoch": 1.7417087634526687, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2776260375976562, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7073643803596497, + "num_tokens": 410022375.0, + "step": 15860 + }, + { + "epoch": 1.7418185811552822, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0540480613708496, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6918197870254517, + "num_tokens": 410054020.0, + "step": 15861 + }, + { + "epoch": 1.7419283988578957, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.16385817527771, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7037820219993591, + "num_tokens": 410079572.0, + "step": 15862 + }, + { + "epoch": 1.7420382165605095, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.122687816619873, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7294825315475464, + "num_tokens": 410106514.0, + "step": 15863 + }, + { + "epoch": 1.7421480342631233, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.183426856994629, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7175994515419006, + "num_tokens": 410132343.0, + "step": 15864 + }, + { + "epoch": 1.742257851965737, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.11715030670166, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7183759212493896, + "num_tokens": 410161866.0, + "step": 15865 + }, + { + "epoch": 1.7423676696683505, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2996978759765625, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.6946538090705872, + "num_tokens": 410187236.0, + "step": 15866 + }, + { + "epoch": 1.742477487370964, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.124633550643921, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7185116410255432, + "num_tokens": 410216432.0, + "step": 15867 + }, + { + "epoch": 1.7425873050735778, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.094806432723999, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7256847620010376, + "num_tokens": 410244252.0, + "step": 15868 + }, + { + "epoch": 1.7426971227761916, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1865146160125732, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7015310525894165, + "num_tokens": 410271135.0, + "step": 15869 + }, + { + "epoch": 1.7428069404788051, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.23960018157959, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7207009196281433, + "num_tokens": 410296705.0, + "step": 15870 + }, + { + "epoch": 1.742916758181419, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2059857845306396, + "learning_rate": 1e-06, + "loss": 1.0064, + "mean_token_accuracy": 0.6876343488693237, + "num_tokens": 410324785.0, + "step": 15871 + }, + { + "epoch": 1.7430265758840324, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.18216609954834, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7278968095779419, + "num_tokens": 410350793.0, + "step": 15872 + }, + { + "epoch": 1.7431363935866462, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1815404891967773, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7223417162895203, + "num_tokens": 410380193.0, + "step": 15873 + }, + { + "epoch": 1.74324621128926, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3873729705810547, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7416893243789673, + "num_tokens": 410403374.0, + "step": 15874 + }, + { + "epoch": 1.7433560289918735, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.0623717308044434, + "learning_rate": 1e-06, + "loss": 0.9242, + "mean_token_accuracy": 0.707704484462738, + "num_tokens": 410431523.0, + "step": 15875 + }, + { + "epoch": 1.743465846694487, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.249866485595703, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7356185913085938, + "num_tokens": 410455780.0, + "step": 15876 + }, + { + "epoch": 1.7435756643971008, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.149250030517578, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6989302635192871, + "num_tokens": 410485812.0, + "step": 15877 + }, + { + "epoch": 1.7436854820997145, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.614806652069092, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7123144865036011, + "num_tokens": 410507164.0, + "step": 15878 + }, + { + "epoch": 1.7437952998023283, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1034815311431885, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7299294471740723, + "num_tokens": 410534324.0, + "step": 15879 + }, + { + "epoch": 1.7439051175049418, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.163221597671509, + "learning_rate": 1e-06, + "loss": 1.0245, + "mean_token_accuracy": 0.6863775253295898, + "num_tokens": 410564035.0, + "step": 15880 + }, + { + "epoch": 1.7440149352075554, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.7204980850219727, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7295330166816711, + "num_tokens": 410582722.0, + "step": 15881 + }, + { + "epoch": 1.744124752910169, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3528192043304443, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7351869344711304, + "num_tokens": 410606232.0, + "step": 15882 + }, + { + "epoch": 1.7442345706127829, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.145970344543457, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7081953287124634, + "num_tokens": 410631951.0, + "step": 15883 + }, + { + "epoch": 1.7443443883153964, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.016916513442993, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.748749315738678, + "num_tokens": 410660421.0, + "step": 15884 + }, + { + "epoch": 1.7444542060180102, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.268592119216919, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7029675245285034, + "num_tokens": 410688148.0, + "step": 15885 + }, + { + "epoch": 1.7445640237206237, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 1.983898639678955, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7040531635284424, + "num_tokens": 410721790.0, + "step": 15886 + }, + { + "epoch": 1.7446738414232374, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.7759556770324707, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.731856107711792, + "num_tokens": 410740257.0, + "step": 15887 + }, + { + "epoch": 1.7447836591258512, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2291531562805176, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.73372483253479, + "num_tokens": 410763734.0, + "step": 15888 + }, + { + "epoch": 1.7448934768284647, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.486433267593384, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7307838201522827, + "num_tokens": 410785032.0, + "step": 15889 + }, + { + "epoch": 1.7450032945310783, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2265331745147705, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.6936942338943481, + "num_tokens": 410810970.0, + "step": 15890 + }, + { + "epoch": 1.745113112233692, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.1336419582366943, + "learning_rate": 1e-06, + "loss": 1.0054, + "mean_token_accuracy": 0.6912117004394531, + "num_tokens": 410841451.0, + "step": 15891 + }, + { + "epoch": 1.7452229299363058, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3345420360565186, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7216734886169434, + "num_tokens": 410865315.0, + "step": 15892 + }, + { + "epoch": 1.7453327476389195, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3774378299713135, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7114971280097961, + "num_tokens": 410889008.0, + "step": 15893 + }, + { + "epoch": 1.745442565341533, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.474210500717163, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7026839852333069, + "num_tokens": 410912808.0, + "step": 15894 + }, + { + "epoch": 1.7455523830441466, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.2259457111358643, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7176109552383423, + "num_tokens": 410940907.0, + "step": 15895 + }, + { + "epoch": 1.7456622007467604, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3658056259155273, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7184364795684814, + "num_tokens": 410964706.0, + "step": 15896 + }, + { + "epoch": 1.7457720184493741, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.190002679824829, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7083253264427185, + "num_tokens": 410992747.0, + "step": 15897 + }, + { + "epoch": 1.7458818361519877, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.307429790496826, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.756895899772644, + "num_tokens": 411016194.0, + "step": 15898 + }, + { + "epoch": 1.7459916538546012, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.5145275592803955, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7059997320175171, + "num_tokens": 411037537.0, + "step": 15899 + }, + { + "epoch": 1.746101471557215, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.3267180919647217, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7293831706047058, + "num_tokens": 411061701.0, + "step": 15900 + }, + { + "epoch": 1.7462112892598287, + "ewc_loss": 9.119510650634766e-06, + "grad_norm": 2.294240951538086, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7255472540855408, + "num_tokens": 411086094.0, + "step": 15901 + }, + { + "epoch": 1.7463211069624425, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.46941876411438, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7167123556137085, + "num_tokens": 411107452.0, + "step": 15902 + }, + { + "epoch": 1.746430924665056, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.424853563308716, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7399662733078003, + "num_tokens": 411129754.0, + "step": 15903 + }, + { + "epoch": 1.7465407423676695, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.338444709777832, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7114980220794678, + "num_tokens": 411155350.0, + "step": 15904 + }, + { + "epoch": 1.7466505600702833, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.0568246841430664, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7205791473388672, + "num_tokens": 411183490.0, + "step": 15905 + }, + { + "epoch": 1.746760377772897, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.307246208190918, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.710152804851532, + "num_tokens": 411210915.0, + "step": 15906 + }, + { + "epoch": 1.7468701954755108, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1986873149871826, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7245813608169556, + "num_tokens": 411237205.0, + "step": 15907 + }, + { + "epoch": 1.7469800131781243, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.275150775909424, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7067787051200867, + "num_tokens": 411264664.0, + "step": 15908 + }, + { + "epoch": 1.7470898308807379, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.307192087173462, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7208132147789001, + "num_tokens": 411291043.0, + "step": 15909 + }, + { + "epoch": 1.7471996485833516, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.428274631500244, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7085015773773193, + "num_tokens": 411313227.0, + "step": 15910 + }, + { + "epoch": 1.7473094662859654, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.305959939956665, + "learning_rate": 1e-06, + "loss": 0.9797, + "mean_token_accuracy": 0.6962064504623413, + "num_tokens": 411339882.0, + "step": 15911 + }, + { + "epoch": 1.747419283988579, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2278220653533936, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7178797125816345, + "num_tokens": 411366020.0, + "step": 15912 + }, + { + "epoch": 1.7475291016911925, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.232935905456543, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.727986216545105, + "num_tokens": 411388987.0, + "step": 15913 + }, + { + "epoch": 1.7476389193938062, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1015422344207764, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7349433898925781, + "num_tokens": 411417028.0, + "step": 15914 + }, + { + "epoch": 1.74774873709642, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.4640958309173584, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7095941305160522, + "num_tokens": 411439589.0, + "step": 15915 + }, + { + "epoch": 1.7478585547990337, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.5374717712402344, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.7009999752044678, + "num_tokens": 411461548.0, + "step": 15916 + }, + { + "epoch": 1.7479683725016473, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.0332794189453125, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7302826642990112, + "num_tokens": 411489361.0, + "step": 15917 + }, + { + "epoch": 1.7480781902042608, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.030123710632324, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7114426493644714, + "num_tokens": 411519407.0, + "step": 15918 + }, + { + "epoch": 1.7481880079068746, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.066840410232544, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7242254614830017, + "num_tokens": 411547927.0, + "step": 15919 + }, + { + "epoch": 1.7482978256094883, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1162683963775635, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.7117122411727905, + "num_tokens": 411575453.0, + "step": 15920 + }, + { + "epoch": 1.7484076433121019, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1477506160736084, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7292875051498413, + "num_tokens": 411603169.0, + "step": 15921 + }, + { + "epoch": 1.7485174610147156, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9671496152877808, + "learning_rate": 1e-06, + "loss": 0.9543, + "mean_token_accuracy": 0.7042065858840942, + "num_tokens": 411635310.0, + "step": 15922 + }, + { + "epoch": 1.7486272787173291, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9091686010360718, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7363414168357849, + "num_tokens": 411666693.0, + "step": 15923 + }, + { + "epoch": 1.748737096419943, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1629436016082764, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7101314663887024, + "num_tokens": 411694091.0, + "step": 15924 + }, + { + "epoch": 1.7488469141225567, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.6359002590179443, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7213605046272278, + "num_tokens": 411715521.0, + "step": 15925 + }, + { + "epoch": 1.7489567318251702, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.0579206943511963, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.723124623298645, + "num_tokens": 411744107.0, + "step": 15926 + }, + { + "epoch": 1.7490665495277837, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9835233688354492, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.712315022945404, + "num_tokens": 411774228.0, + "step": 15927 + }, + { + "epoch": 1.7491763672303975, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.7473371028900146, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7284311056137085, + "num_tokens": 411791279.0, + "step": 15928 + }, + { + "epoch": 1.7492861849330112, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.350689649581909, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7319908142089844, + "num_tokens": 411813971.0, + "step": 15929 + }, + { + "epoch": 1.749396002635625, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.3768398761749268, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7239121794700623, + "num_tokens": 411836624.0, + "step": 15930 + }, + { + "epoch": 1.7495058203382385, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2564945220947266, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7167127728462219, + "num_tokens": 411862879.0, + "step": 15931 + }, + { + "epoch": 1.749615638040852, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1205267906188965, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7209359407424927, + "num_tokens": 411890777.0, + "step": 15932 + }, + { + "epoch": 1.7497254557434658, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.8856362104415894, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7062411308288574, + "num_tokens": 411929648.0, + "step": 15933 + }, + { + "epoch": 1.7498352734460796, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9569072723388672, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.740896463394165, + "num_tokens": 411959716.0, + "step": 15934 + }, + { + "epoch": 1.7499450911486931, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2456283569335938, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7229592204093933, + "num_tokens": 411984294.0, + "step": 15935 + }, + { + "epoch": 1.7500549088513069, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2440185546875, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.71707683801651, + "num_tokens": 412008278.0, + "step": 15936 + }, + { + "epoch": 1.7501647265539204, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.207233428955078, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.730980396270752, + "num_tokens": 412035117.0, + "step": 15937 + }, + { + "epoch": 1.7502745442565342, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2424237728118896, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7231668829917908, + "num_tokens": 412059205.0, + "step": 15938 + }, + { + "epoch": 1.750384361959148, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.4200706481933594, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7082823514938354, + "num_tokens": 412082773.0, + "step": 15939 + }, + { + "epoch": 1.7504941796617615, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1129379272460938, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7167391180992126, + "num_tokens": 412111838.0, + "step": 15940 + }, + { + "epoch": 1.750603997364375, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.6963682174682617, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7350496053695679, + "num_tokens": 412131623.0, + "step": 15941 + }, + { + "epoch": 1.7507138150669888, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.066066265106201, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.7179434299468994, + "num_tokens": 412161242.0, + "step": 15942 + }, + { + "epoch": 1.7508236327696025, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.476980686187744, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7385271191596985, + "num_tokens": 412182544.0, + "step": 15943 + }, + { + "epoch": 1.7509334504722163, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2894856929779053, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7121180891990662, + "num_tokens": 412207490.0, + "step": 15944 + }, + { + "epoch": 1.7510432681748298, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1290085315704346, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7082377076148987, + "num_tokens": 412235966.0, + "step": 15945 + }, + { + "epoch": 1.7511530858774433, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.4508917331695557, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7293172478675842, + "num_tokens": 412259956.0, + "step": 15946 + }, + { + "epoch": 1.751262903580057, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.148183822631836, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.6981397867202759, + "num_tokens": 412289410.0, + "step": 15947 + }, + { + "epoch": 1.7513727212826709, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1818621158599854, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7172093391418457, + "num_tokens": 412316136.0, + "step": 15948 + }, + { + "epoch": 1.7514825389852844, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9996896982192993, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7284660935401917, + "num_tokens": 412344879.0, + "step": 15949 + }, + { + "epoch": 1.7515923566878981, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1181962490081787, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7221574783325195, + "num_tokens": 412371632.0, + "step": 15950 + }, + { + "epoch": 1.7517021743905117, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.9775519371032715, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7284629940986633, + "num_tokens": 412400581.0, + "step": 15951 + }, + { + "epoch": 1.7518119920931254, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.287869691848755, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.6986553072929382, + "num_tokens": 412425480.0, + "step": 15952 + }, + { + "epoch": 1.7519218097957392, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.4094433784484863, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7228718996047974, + "num_tokens": 412448255.0, + "step": 15953 + }, + { + "epoch": 1.7520316274983527, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.349613666534424, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7202479839324951, + "num_tokens": 412471255.0, + "step": 15954 + }, + { + "epoch": 1.7521414452009663, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.7419395446777344, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7419605851173401, + "num_tokens": 412491773.0, + "step": 15955 + }, + { + "epoch": 1.75225126290358, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.123851776123047, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7199283838272095, + "num_tokens": 412521308.0, + "step": 15956 + }, + { + "epoch": 1.7523610806061938, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2239320278167725, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7118771076202393, + "num_tokens": 412547588.0, + "step": 15957 + }, + { + "epoch": 1.7524708983088075, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.1486966609954834, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7046473026275635, + "num_tokens": 412576696.0, + "step": 15958 + }, + { + "epoch": 1.752580716011421, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.464780330657959, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7113218307495117, + "num_tokens": 412598512.0, + "step": 15959 + }, + { + "epoch": 1.7526905337140346, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.3200736045837402, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7186904549598694, + "num_tokens": 412624152.0, + "step": 15960 + }, + { + "epoch": 1.7528003514166484, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.164860486984253, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7398189306259155, + "num_tokens": 412649925.0, + "step": 15961 + }, + { + "epoch": 1.7529101691192621, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 1.788887619972229, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7042748928070068, + "num_tokens": 412689788.0, + "step": 15962 + }, + { + "epoch": 1.7530199868218757, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.45420241355896, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7411553859710693, + "num_tokens": 412712857.0, + "step": 15963 + }, + { + "epoch": 1.7531298045244892, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.301464557647705, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7073080539703369, + "num_tokens": 412739070.0, + "step": 15964 + }, + { + "epoch": 1.753239622227103, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.084552764892578, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7238420844078064, + "num_tokens": 412768299.0, + "step": 15965 + }, + { + "epoch": 1.7533494399297167, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.46370005607605, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7148304581642151, + "num_tokens": 412790910.0, + "step": 15966 + }, + { + "epoch": 1.7534592576323305, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.335338830947876, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7259851098060608, + "num_tokens": 412816326.0, + "step": 15967 + }, + { + "epoch": 1.753569075334944, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2360339164733887, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7365140914916992, + "num_tokens": 412842083.0, + "step": 15968 + }, + { + "epoch": 1.7536788930375575, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.5298585891723633, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7145398855209351, + "num_tokens": 412862592.0, + "step": 15969 + }, + { + "epoch": 1.7537887107401713, + "ewc_loss": 9.238719940185547e-06, + "grad_norm": 2.2657291889190674, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7195677161216736, + "num_tokens": 412886549.0, + "step": 15970 + }, + { + "epoch": 1.753898528442785, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1760904788970947, + "learning_rate": 1e-06, + "loss": 0.9778, + "mean_token_accuracy": 0.6972264051437378, + "num_tokens": 412912929.0, + "step": 15971 + }, + { + "epoch": 1.7540083461453988, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.4462931156158447, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7253671884536743, + "num_tokens": 412934306.0, + "step": 15972 + }, + { + "epoch": 1.7541181638480123, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.151559591293335, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7189322113990784, + "num_tokens": 412963595.0, + "step": 15973 + }, + { + "epoch": 1.7542279815506259, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 1.9354534149169922, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7130810022354126, + "num_tokens": 412999685.0, + "step": 15974 + }, + { + "epoch": 1.7543377992532396, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.0991878509521484, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.71236252784729, + "num_tokens": 413030652.0, + "step": 15975 + }, + { + "epoch": 1.7544476169558534, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1963019371032715, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.7062276601791382, + "num_tokens": 413058995.0, + "step": 15976 + }, + { + "epoch": 1.754557434658467, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.4349870681762695, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7103005647659302, + "num_tokens": 413081066.0, + "step": 15977 + }, + { + "epoch": 1.7546672523610805, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1591267585754395, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7221234440803528, + "num_tokens": 413109710.0, + "step": 15978 + }, + { + "epoch": 1.7547770700636942, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.2932283878326416, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7290852069854736, + "num_tokens": 413133397.0, + "step": 15979 + }, + { + "epoch": 1.754886887766308, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 1.9931883811950684, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7106912136077881, + "num_tokens": 413165579.0, + "step": 15980 + }, + { + "epoch": 1.7549967054689217, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.360499382019043, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7345077991485596, + "num_tokens": 413190182.0, + "step": 15981 + }, + { + "epoch": 1.7551065231715353, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.2986390590667725, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7180934548377991, + "num_tokens": 413214874.0, + "step": 15982 + }, + { + "epoch": 1.7552163408741488, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.0441277027130127, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7157718539237976, + "num_tokens": 413244392.0, + "step": 15983 + }, + { + "epoch": 1.7553261585767626, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 1.9710193872451782, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.717327356338501, + "num_tokens": 413276386.0, + "step": 15984 + }, + { + "epoch": 1.7554359762793763, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.166781187057495, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7235662937164307, + "num_tokens": 413304406.0, + "step": 15985 + }, + { + "epoch": 1.7555457939819898, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1970486640930176, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7272112369537354, + "num_tokens": 413329957.0, + "step": 15986 + }, + { + "epoch": 1.7556556116846036, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.1750504970550537, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7186926603317261, + "num_tokens": 413355676.0, + "step": 15987 + }, + { + "epoch": 1.7557654293872171, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.359607219696045, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7249131798744202, + "num_tokens": 413377392.0, + "step": 15988 + }, + { + "epoch": 1.755875247089831, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.350686550140381, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7413504123687744, + "num_tokens": 413398830.0, + "step": 15989 + }, + { + "epoch": 1.7559850647924446, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.0167341232299805, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7050584554672241, + "num_tokens": 413429855.0, + "step": 15990 + }, + { + "epoch": 1.7560948824950582, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.206787347793579, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7244423627853394, + "num_tokens": 413456455.0, + "step": 15991 + }, + { + "epoch": 1.7562047001976717, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.3914358615875244, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.718877911567688, + "num_tokens": 413479740.0, + "step": 15992 + }, + { + "epoch": 1.7563145179002855, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.283700942993164, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7029756903648376, + "num_tokens": 413505148.0, + "step": 15993 + }, + { + "epoch": 1.7564243356028992, + "ewc_loss": 9.298324584960938e-06, + "grad_norm": 2.239814281463623, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7202694416046143, + "num_tokens": 413530087.0, + "step": 15994 + }, + { + "epoch": 1.756534153305513, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2808971405029297, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7361859083175659, + "num_tokens": 413556001.0, + "step": 15995 + }, + { + "epoch": 1.7566439710081265, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.376909017562866, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7484935522079468, + "num_tokens": 413578697.0, + "step": 15996 + }, + { + "epoch": 1.75675378871074, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.173924207687378, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.701755940914154, + "num_tokens": 413607179.0, + "step": 15997 + }, + { + "epoch": 1.7568636064133538, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.043888807296753, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7129549980163574, + "num_tokens": 413635705.0, + "step": 15998 + }, + { + "epoch": 1.7569734241159676, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.335052013397217, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7454346418380737, + "num_tokens": 413658062.0, + "step": 15999 + }, + { + "epoch": 1.757083241818581, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.325517177581787, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7368059158325195, + "num_tokens": 413679936.0, + "step": 16000 + }, + { + "epoch": 1.7571930595211949, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.483323335647583, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7169265747070312, + "num_tokens": 413700924.0, + "step": 16001 + }, + { + "epoch": 1.7573028772238084, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.506624221801758, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7314061522483826, + "num_tokens": 413720728.0, + "step": 16002 + }, + { + "epoch": 1.7574126949264222, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.19846773147583, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7362140417098999, + "num_tokens": 413747592.0, + "step": 16003 + }, + { + "epoch": 1.757522512629036, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 1.9610751867294312, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7192090749740601, + "num_tokens": 413779379.0, + "step": 16004 + }, + { + "epoch": 1.7576323303316495, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.474945306777954, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7129768133163452, + "num_tokens": 413801675.0, + "step": 16005 + }, + { + "epoch": 1.757742148034263, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1789097785949707, + "learning_rate": 1e-06, + "loss": 0.9826, + "mean_token_accuracy": 0.7064229846000671, + "num_tokens": 413828275.0, + "step": 16006 + }, + { + "epoch": 1.7578519657368767, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0699262619018555, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7240948677062988, + "num_tokens": 413859114.0, + "step": 16007 + }, + { + "epoch": 1.7579617834394905, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0317904949188232, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7345744967460632, + "num_tokens": 413886825.0, + "step": 16008 + }, + { + "epoch": 1.7580716011421043, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 1.8714351654052734, + "learning_rate": 1e-06, + "loss": 1.0576, + "mean_token_accuracy": 0.6853079795837402, + "num_tokens": 413925084.0, + "step": 16009 + }, + { + "epoch": 1.7581814188447178, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2969534397125244, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6959307193756104, + "num_tokens": 413952206.0, + "step": 16010 + }, + { + "epoch": 1.7582912365473313, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.3491251468658447, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7038072943687439, + "num_tokens": 413975722.0, + "step": 16011 + }, + { + "epoch": 1.758401054249945, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.169421672821045, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7383663654327393, + "num_tokens": 414000120.0, + "step": 16012 + }, + { + "epoch": 1.7585108719525588, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.4303712844848633, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7238161563873291, + "num_tokens": 414022521.0, + "step": 16013 + }, + { + "epoch": 1.7586206896551724, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2223494052886963, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7068185806274414, + "num_tokens": 414047699.0, + "step": 16014 + }, + { + "epoch": 1.758730507357786, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.4136650562286377, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7147730588912964, + "num_tokens": 414071095.0, + "step": 16015 + }, + { + "epoch": 1.7588403250603997, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1729564666748047, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7342678308486938, + "num_tokens": 414096434.0, + "step": 16016 + }, + { + "epoch": 1.7589501427630134, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.253013849258423, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7221150994300842, + "num_tokens": 414120608.0, + "step": 16017 + }, + { + "epoch": 1.7590599604656272, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.315835475921631, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.7599951028823853, + "num_tokens": 414142862.0, + "step": 16018 + }, + { + "epoch": 1.7591697781682407, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0394322872161865, + "learning_rate": 1e-06, + "loss": 0.9721, + "mean_token_accuracy": 0.6989549994468689, + "num_tokens": 414173434.0, + "step": 16019 + }, + { + "epoch": 1.7592795958708543, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.27834415435791, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7339994311332703, + "num_tokens": 414199609.0, + "step": 16020 + }, + { + "epoch": 1.759389413573468, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.3296701908111572, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7145425081253052, + "num_tokens": 414225107.0, + "step": 16021 + }, + { + "epoch": 1.7594992312760818, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1023333072662354, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7368120551109314, + "num_tokens": 414251708.0, + "step": 16022 + }, + { + "epoch": 1.7596090489786955, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.110459566116333, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.750727653503418, + "num_tokens": 414279664.0, + "step": 16023 + }, + { + "epoch": 1.759718866681309, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0594377517700195, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.7013665437698364, + "num_tokens": 414309844.0, + "step": 16024 + }, + { + "epoch": 1.7598286843839226, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.198573589324951, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7216460108757019, + "num_tokens": 414333773.0, + "step": 16025 + }, + { + "epoch": 1.7599385020865363, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.194829225540161, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7137243747711182, + "num_tokens": 414360364.0, + "step": 16026 + }, + { + "epoch": 1.76004831978915, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1715633869171143, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7430528402328491, + "num_tokens": 414387365.0, + "step": 16027 + }, + { + "epoch": 1.7601581374917636, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.184419870376587, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7288401126861572, + "num_tokens": 414413067.0, + "step": 16028 + }, + { + "epoch": 1.7602679551943772, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1101067066192627, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.7092785835266113, + "num_tokens": 414440864.0, + "step": 16029 + }, + { + "epoch": 1.760377772896991, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1748359203338623, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7120823860168457, + "num_tokens": 414467877.0, + "step": 16030 + }, + { + "epoch": 1.7604875905996047, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.225405693054199, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7366973757743835, + "num_tokens": 414492795.0, + "step": 16031 + }, + { + "epoch": 1.7605974083022184, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.184596538543701, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7308738231658936, + "num_tokens": 414519885.0, + "step": 16032 + }, + { + "epoch": 1.760707226004832, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.008622646331787, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7246543169021606, + "num_tokens": 414549805.0, + "step": 16033 + }, + { + "epoch": 1.7608170437074455, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1760528087615967, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7295187711715698, + "num_tokens": 414576865.0, + "step": 16034 + }, + { + "epoch": 1.7609268614100593, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.114529609680176, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7097938060760498, + "num_tokens": 414605167.0, + "step": 16035 + }, + { + "epoch": 1.761036679112673, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.4753451347351074, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7247511148452759, + "num_tokens": 414626986.0, + "step": 16036 + }, + { + "epoch": 1.7611464968152868, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.21409010887146, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7094986438751221, + "num_tokens": 414655905.0, + "step": 16037 + }, + { + "epoch": 1.7612563145179003, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.210287094116211, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7116208672523499, + "num_tokens": 414682696.0, + "step": 16038 + }, + { + "epoch": 1.7613661322205139, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.189124345779419, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7249417901039124, + "num_tokens": 414709072.0, + "step": 16039 + }, + { + "epoch": 1.7614759499231276, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1098721027374268, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7399258613586426, + "num_tokens": 414734163.0, + "step": 16040 + }, + { + "epoch": 1.7615857676257414, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2802019119262695, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7433737516403198, + "num_tokens": 414757528.0, + "step": 16041 + }, + { + "epoch": 1.761695585328355, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.3292739391326904, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7153571248054504, + "num_tokens": 414781026.0, + "step": 16042 + }, + { + "epoch": 1.7618054030309684, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 1.9850564002990723, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7203772068023682, + "num_tokens": 414811374.0, + "step": 16043 + }, + { + "epoch": 1.7619152207335822, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.355034589767456, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7240914702415466, + "num_tokens": 414834105.0, + "step": 16044 + }, + { + "epoch": 1.762025038436196, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1189169883728027, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.706937849521637, + "num_tokens": 414863277.0, + "step": 16045 + }, + { + "epoch": 1.7621348561388097, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.3002076148986816, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7303056716918945, + "num_tokens": 414888365.0, + "step": 16046 + }, + { + "epoch": 1.7622446738414232, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.140105962753296, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7306253910064697, + "num_tokens": 414914796.0, + "step": 16047 + }, + { + "epoch": 1.7623544915440368, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 1.9809514284133911, + "learning_rate": 1e-06, + "loss": 0.9593, + "mean_token_accuracy": 0.7090787887573242, + "num_tokens": 414946541.0, + "step": 16048 + }, + { + "epoch": 1.7624643092466505, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.3884034156799316, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7212374210357666, + "num_tokens": 414970385.0, + "step": 16049 + }, + { + "epoch": 1.7625741269492643, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1482226848602295, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7174351215362549, + "num_tokens": 414997241.0, + "step": 16050 + }, + { + "epoch": 1.7626839446518778, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.268031358718872, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7198275923728943, + "num_tokens": 415025244.0, + "step": 16051 + }, + { + "epoch": 1.7627937623544916, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.090015411376953, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7460688948631287, + "num_tokens": 415051255.0, + "step": 16052 + }, + { + "epoch": 1.7629035800571051, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0497944355010986, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7345531582832336, + "num_tokens": 415080247.0, + "step": 16053 + }, + { + "epoch": 1.7630133977597189, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.0360159873962402, + "learning_rate": 1e-06, + "loss": 0.9189, + "mean_token_accuracy": 0.711687445640564, + "num_tokens": 415109155.0, + "step": 16054 + }, + { + "epoch": 1.7631232154623326, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.300097942352295, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.72599196434021, + "num_tokens": 415133469.0, + "step": 16055 + }, + { + "epoch": 1.7632330331649462, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.492295265197754, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.708983302116394, + "num_tokens": 415155715.0, + "step": 16056 + }, + { + "epoch": 1.7633428508675597, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 1.7729209661483765, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7104811668395996, + "num_tokens": 415192246.0, + "step": 16057 + }, + { + "epoch": 1.7634526685701735, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.332181930541992, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.720568060874939, + "num_tokens": 415219216.0, + "step": 16058 + }, + { + "epoch": 1.7635624862727872, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.276188611984253, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7258932590484619, + "num_tokens": 415245103.0, + "step": 16059 + }, + { + "epoch": 1.763672303975401, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.204558849334717, + "learning_rate": 1e-06, + "loss": 0.9973, + "mean_token_accuracy": 0.7008472681045532, + "num_tokens": 415276220.0, + "step": 16060 + }, + { + "epoch": 1.7637821216780145, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2193007469177246, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7310029864311218, + "num_tokens": 415303037.0, + "step": 16061 + }, + { + "epoch": 1.763891939380628, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.469625473022461, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7183181643486023, + "num_tokens": 415324617.0, + "step": 16062 + }, + { + "epoch": 1.7640017570832418, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.283134937286377, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7196159362792969, + "num_tokens": 415349364.0, + "step": 16063 + }, + { + "epoch": 1.7641115747858556, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.091536521911621, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.7079371213912964, + "num_tokens": 415376114.0, + "step": 16064 + }, + { + "epoch": 1.764221392488469, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.1153197288513184, + "learning_rate": 1e-06, + "loss": 1.0493, + "mean_token_accuracy": 0.6841706037521362, + "num_tokens": 415406244.0, + "step": 16065 + }, + { + "epoch": 1.7643312101910829, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.510411024093628, + "learning_rate": 1e-06, + "loss": 0.9736, + "mean_token_accuracy": 0.7034385204315186, + "num_tokens": 415428018.0, + "step": 16066 + }, + { + "epoch": 1.7644410278936964, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2207493782043457, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.7030776739120483, + "num_tokens": 415455075.0, + "step": 16067 + }, + { + "epoch": 1.7645508455963101, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.03908371925354, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.702672004699707, + "num_tokens": 415485792.0, + "step": 16068 + }, + { + "epoch": 1.764660663298924, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.115356922149658, + "learning_rate": 1e-06, + "loss": 0.9493, + "mean_token_accuracy": 0.7074429988861084, + "num_tokens": 415514560.0, + "step": 16069 + }, + { + "epoch": 1.7647704810015374, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.31943678855896, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.723700225353241, + "num_tokens": 415537026.0, + "step": 16070 + }, + { + "epoch": 1.764880298704151, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1055872440338135, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6991556882858276, + "num_tokens": 415565677.0, + "step": 16071 + }, + { + "epoch": 1.7649901164067647, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1580796241760254, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7116137742996216, + "num_tokens": 415594587.0, + "step": 16072 + }, + { + "epoch": 1.7650999341093785, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1090142726898193, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7147992253303528, + "num_tokens": 415622755.0, + "step": 16073 + }, + { + "epoch": 1.7652097518119922, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9839050769805908, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7214069366455078, + "num_tokens": 415652436.0, + "step": 16074 + }, + { + "epoch": 1.7653195695146058, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.483595132827759, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7151851058006287, + "num_tokens": 415676475.0, + "step": 16075 + }, + { + "epoch": 1.7654293872172193, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2156729698181152, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7109686136245728, + "num_tokens": 415703020.0, + "step": 16076 + }, + { + "epoch": 1.765539204919833, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1676597595214844, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.709273099899292, + "num_tokens": 415729866.0, + "step": 16077 + }, + { + "epoch": 1.7656490226224468, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.475433826446533, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7254908084869385, + "num_tokens": 415752142.0, + "step": 16078 + }, + { + "epoch": 1.7657588403250604, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1199259757995605, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7118222713470459, + "num_tokens": 415780469.0, + "step": 16079 + }, + { + "epoch": 1.765868658027674, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0972087383270264, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7066299915313721, + "num_tokens": 415809635.0, + "step": 16080 + }, + { + "epoch": 1.7659784757302877, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.414475440979004, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7140318751335144, + "num_tokens": 415832723.0, + "step": 16081 + }, + { + "epoch": 1.7660882934329014, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2864954471588135, + "learning_rate": 1e-06, + "loss": 0.9353, + "mean_token_accuracy": 0.7028253674507141, + "num_tokens": 415857939.0, + "step": 16082 + }, + { + "epoch": 1.7661981111355152, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.343425989151001, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7189890146255493, + "num_tokens": 415881968.0, + "step": 16083 + }, + { + "epoch": 1.7663079288381287, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.103060722351074, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7138268947601318, + "num_tokens": 415911357.0, + "step": 16084 + }, + { + "epoch": 1.7664177465407422, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1594438552856445, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7327271699905396, + "num_tokens": 415936779.0, + "step": 16085 + }, + { + "epoch": 1.766527564243356, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1246628761291504, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7036628127098083, + "num_tokens": 415964168.0, + "step": 16086 + }, + { + "epoch": 1.7666373819459698, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2806806564331055, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7297248840332031, + "num_tokens": 415988411.0, + "step": 16087 + }, + { + "epoch": 1.7667471996485835, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9615422487258911, + "learning_rate": 1e-06, + "loss": 0.9585, + "mean_token_accuracy": 0.6988252997398376, + "num_tokens": 416018363.0, + "step": 16088 + }, + { + "epoch": 1.766857017351197, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.162191867828369, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7324731349945068, + "num_tokens": 416046232.0, + "step": 16089 + }, + { + "epoch": 1.7669668350538106, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1557207107543945, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7379114627838135, + "num_tokens": 416073759.0, + "step": 16090 + }, + { + "epoch": 1.7670766527564243, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2490487098693848, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7172655463218689, + "num_tokens": 416098163.0, + "step": 16091 + }, + { + "epoch": 1.767186470459038, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1011595726013184, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7162456512451172, + "num_tokens": 416126423.0, + "step": 16092 + }, + { + "epoch": 1.7672962881616516, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.5605268478393555, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7396941184997559, + "num_tokens": 416147882.0, + "step": 16093 + }, + { + "epoch": 1.7674061058642652, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4079909324645996, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7205672264099121, + "num_tokens": 416169273.0, + "step": 16094 + }, + { + "epoch": 1.767515923566879, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.35628604888916, + "learning_rate": 1e-06, + "loss": 0.9579, + "mean_token_accuracy": 0.7008470892906189, + "num_tokens": 416192765.0, + "step": 16095 + }, + { + "epoch": 1.7676257412694927, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2860069274902344, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7274360656738281, + "num_tokens": 416216952.0, + "step": 16096 + }, + { + "epoch": 1.7677355589721064, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.136622667312622, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7279297113418579, + "num_tokens": 416243681.0, + "step": 16097 + }, + { + "epoch": 1.76784537667472, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1816585063934326, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7296054363250732, + "num_tokens": 416270792.0, + "step": 16098 + }, + { + "epoch": 1.7679551943773335, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1952600479125977, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7115576267242432, + "num_tokens": 416298995.0, + "step": 16099 + }, + { + "epoch": 1.7680650120799473, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1561851501464844, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.727544903755188, + "num_tokens": 416324419.0, + "step": 16100 + }, + { + "epoch": 1.768174829782561, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4720752239227295, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7305202484130859, + "num_tokens": 416347089.0, + "step": 16101 + }, + { + "epoch": 1.7682846474851748, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2593798637390137, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6931672096252441, + "num_tokens": 416372595.0, + "step": 16102 + }, + { + "epoch": 1.7683944651877883, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.4238007068634033, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7213022708892822, + "num_tokens": 416393401.0, + "step": 16103 + }, + { + "epoch": 1.7685042828904018, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.451334238052368, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7272217273712158, + "num_tokens": 416416802.0, + "step": 16104 + }, + { + "epoch": 1.7686141005930156, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.2586240768432617, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7051559686660767, + "num_tokens": 416446841.0, + "step": 16105 + }, + { + "epoch": 1.7687239182956294, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.474024772644043, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7386230230331421, + "num_tokens": 416471615.0, + "step": 16106 + }, + { + "epoch": 1.768833735998243, + "ewc_loss": 9.357929229736328e-06, + "grad_norm": 2.140713691711426, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7085840702056885, + "num_tokens": 416500878.0, + "step": 16107 + }, + { + "epoch": 1.7689435537008564, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4515228271484375, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7317248582839966, + "num_tokens": 416524480.0, + "step": 16108 + }, + { + "epoch": 1.7690533714034702, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.222724199295044, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7322203516960144, + "num_tokens": 416549422.0, + "step": 16109 + }, + { + "epoch": 1.769163189106084, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.310438394546509, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7341116666793823, + "num_tokens": 416574267.0, + "step": 16110 + }, + { + "epoch": 1.7692730068086977, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1383001804351807, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7065899968147278, + "num_tokens": 416602495.0, + "step": 16111 + }, + { + "epoch": 1.7693828245113112, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.111456871032715, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7067165374755859, + "num_tokens": 416631999.0, + "step": 16112 + }, + { + "epoch": 1.7694926422139248, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.042839527130127, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7087202072143555, + "num_tokens": 416664289.0, + "step": 16113 + }, + { + "epoch": 1.7696024599165385, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4511168003082275, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7199565172195435, + "num_tokens": 416686554.0, + "step": 16114 + }, + { + "epoch": 1.7697122776191523, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.12204647064209, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7119884490966797, + "num_tokens": 416715570.0, + "step": 16115 + }, + { + "epoch": 1.7698220953217658, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.162923812866211, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7093110084533691, + "num_tokens": 416743489.0, + "step": 16116 + }, + { + "epoch": 1.7699319130243796, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0443789958953857, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7186727523803711, + "num_tokens": 416771487.0, + "step": 16117 + }, + { + "epoch": 1.7700417307269931, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0817785263061523, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7177281379699707, + "num_tokens": 416801416.0, + "step": 16118 + }, + { + "epoch": 1.7701515484296069, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4557790756225586, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7308432459831238, + "num_tokens": 416823188.0, + "step": 16119 + }, + { + "epoch": 1.7702613661322206, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.246033191680908, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7186380624771118, + "num_tokens": 416849104.0, + "step": 16120 + }, + { + "epoch": 1.7703711838348342, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2216453552246094, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7053633332252502, + "num_tokens": 416876435.0, + "step": 16121 + }, + { + "epoch": 1.7704810015374477, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9251757860183716, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7168500423431396, + "num_tokens": 416910832.0, + "step": 16122 + }, + { + "epoch": 1.7705908192400615, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2317001819610596, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7064225077629089, + "num_tokens": 416939844.0, + "step": 16123 + }, + { + "epoch": 1.7707006369426752, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1990842819213867, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7334719896316528, + "num_tokens": 416966003.0, + "step": 16124 + }, + { + "epoch": 1.770810454645289, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.538417100906372, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7248028516769409, + "num_tokens": 416985736.0, + "step": 16125 + }, + { + "epoch": 1.7709202723479025, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.45987606048584, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.720760703086853, + "num_tokens": 417007408.0, + "step": 16126 + }, + { + "epoch": 1.771030090050516, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9129323959350586, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7336163520812988, + "num_tokens": 417039428.0, + "step": 16127 + }, + { + "epoch": 1.7711399077531298, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.269125461578369, + "learning_rate": 1e-06, + "loss": 0.7794, + "mean_token_accuracy": 0.7435957193374634, + "num_tokens": 417062279.0, + "step": 16128 + }, + { + "epoch": 1.7712497254557436, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.279402017593384, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7312303185462952, + "num_tokens": 417086262.0, + "step": 16129 + }, + { + "epoch": 1.771359543158357, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1742825508117676, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7061545252799988, + "num_tokens": 417113676.0, + "step": 16130 + }, + { + "epoch": 1.7714693608609708, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.337214708328247, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7242547869682312, + "num_tokens": 417137542.0, + "step": 16131 + }, + { + "epoch": 1.7715791785635844, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3852336406707764, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7195343375205994, + "num_tokens": 417160678.0, + "step": 16132 + }, + { + "epoch": 1.7716889962661981, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4745500087738037, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7124421000480652, + "num_tokens": 417183766.0, + "step": 16133 + }, + { + "epoch": 1.771798813968812, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.371387243270874, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.6992496252059937, + "num_tokens": 417207715.0, + "step": 16134 + }, + { + "epoch": 1.7719086316714254, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3445396423339844, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7271325588226318, + "num_tokens": 417232832.0, + "step": 16135 + }, + { + "epoch": 1.772018449374039, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.033120632171631, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7195289134979248, + "num_tokens": 417262091.0, + "step": 16136 + }, + { + "epoch": 1.7721282670766527, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0488474369049072, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7116910219192505, + "num_tokens": 417293057.0, + "step": 16137 + }, + { + "epoch": 1.7722380847792665, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2748963832855225, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7254273891448975, + "num_tokens": 417317893.0, + "step": 16138 + }, + { + "epoch": 1.7723479024818802, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.110980749130249, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7320101857185364, + "num_tokens": 417345068.0, + "step": 16139 + }, + { + "epoch": 1.7724577201844938, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.117481231689453, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7306218147277832, + "num_tokens": 417372931.0, + "step": 16140 + }, + { + "epoch": 1.7725675378871073, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3215394020080566, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7156191468238831, + "num_tokens": 417395646.0, + "step": 16141 + }, + { + "epoch": 1.772677355589721, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.125157594680786, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7095351219177246, + "num_tokens": 417426697.0, + "step": 16142 + }, + { + "epoch": 1.7727871732923348, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.34444522857666, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7480614185333252, + "num_tokens": 417448738.0, + "step": 16143 + }, + { + "epoch": 1.7728969909949484, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2863454818725586, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7309032678604126, + "num_tokens": 417472669.0, + "step": 16144 + }, + { + "epoch": 1.7730068086975619, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.142655849456787, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7279583811759949, + "num_tokens": 417500389.0, + "step": 16145 + }, + { + "epoch": 1.7731166264001756, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1078171730041504, + "learning_rate": 1e-06, + "loss": 0.9863, + "mean_token_accuracy": 0.7043838500976562, + "num_tokens": 417529568.0, + "step": 16146 + }, + { + "epoch": 1.7732264441027894, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2254908084869385, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7202191352844238, + "num_tokens": 417554658.0, + "step": 16147 + }, + { + "epoch": 1.7733362618054032, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9938368797302246, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7139596343040466, + "num_tokens": 417587294.0, + "step": 16148 + }, + { + "epoch": 1.7734460795080167, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.283364772796631, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7224626541137695, + "num_tokens": 417612958.0, + "step": 16149 + }, + { + "epoch": 1.7735558972106302, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4836134910583496, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7294773459434509, + "num_tokens": 417634148.0, + "step": 16150 + }, + { + "epoch": 1.773665714913244, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1694436073303223, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7128678560256958, + "num_tokens": 417659456.0, + "step": 16151 + }, + { + "epoch": 1.7737755326158577, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.156381130218506, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7034464478492737, + "num_tokens": 417687557.0, + "step": 16152 + }, + { + "epoch": 1.7738853503184715, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2666633129119873, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7140523195266724, + "num_tokens": 417712667.0, + "step": 16153 + }, + { + "epoch": 1.773995168021085, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3097496032714844, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7409185767173767, + "num_tokens": 417735849.0, + "step": 16154 + }, + { + "epoch": 1.7741049857236986, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.083526849746704, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7020609378814697, + "num_tokens": 417765359.0, + "step": 16155 + }, + { + "epoch": 1.7742148034263123, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9514458179473877, + "learning_rate": 1e-06, + "loss": 0.9775, + "mean_token_accuracy": 0.7037228345870972, + "num_tokens": 417797089.0, + "step": 16156 + }, + { + "epoch": 1.774324621128926, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3246090412139893, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7240309715270996, + "num_tokens": 417820062.0, + "step": 16157 + }, + { + "epoch": 1.7744344388315396, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.14780330657959, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.6988174319267273, + "num_tokens": 417848752.0, + "step": 16158 + }, + { + "epoch": 1.7745442565341532, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9630903005599976, + "learning_rate": 1e-06, + "loss": 1.0311, + "mean_token_accuracy": 0.6876332759857178, + "num_tokens": 417881739.0, + "step": 16159 + }, + { + "epoch": 1.774654074236767, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0605180263519287, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7188706994056702, + "num_tokens": 417913173.0, + "step": 16160 + }, + { + "epoch": 1.7747638919393807, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3050127029418945, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7194724678993225, + "num_tokens": 417939617.0, + "step": 16161 + }, + { + "epoch": 1.7748737096419944, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.225224733352661, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7255685329437256, + "num_tokens": 417963686.0, + "step": 16162 + }, + { + "epoch": 1.774983527344608, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1237406730651855, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7077046632766724, + "num_tokens": 417992361.0, + "step": 16163 + }, + { + "epoch": 1.7750933450472215, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1938493251800537, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7439261674880981, + "num_tokens": 418017981.0, + "step": 16164 + }, + { + "epoch": 1.7752031627498353, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.1750950813293457, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7370812892913818, + "num_tokens": 418043341.0, + "step": 16165 + }, + { + "epoch": 1.775312980452449, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.235473394393921, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.6967920064926147, + "num_tokens": 418069128.0, + "step": 16166 + }, + { + "epoch": 1.7754227981550625, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.463275194168091, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7266054749488831, + "num_tokens": 418090271.0, + "step": 16167 + }, + { + "epoch": 1.7755326158576763, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3788561820983887, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7345589399337769, + "num_tokens": 418113119.0, + "step": 16168 + }, + { + "epoch": 1.7756424335602898, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2558341026306152, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7235884070396423, + "num_tokens": 418138183.0, + "step": 16169 + }, + { + "epoch": 1.7757522512629036, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.0189123153686523, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6971946358680725, + "num_tokens": 418171349.0, + "step": 16170 + }, + { + "epoch": 1.7758620689655173, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9169577360153198, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7223364114761353, + "num_tokens": 418201351.0, + "step": 16171 + }, + { + "epoch": 1.7759718866681309, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.293473482131958, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7382817268371582, + "num_tokens": 418226757.0, + "step": 16172 + }, + { + "epoch": 1.7760817043707444, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.347649335861206, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7255614995956421, + "num_tokens": 418250369.0, + "step": 16173 + }, + { + "epoch": 1.7761915220733582, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.483572006225586, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.7520997524261475, + "num_tokens": 418271077.0, + "step": 16174 + }, + { + "epoch": 1.776301339775972, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3953664302825928, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7344071865081787, + "num_tokens": 418292164.0, + "step": 16175 + }, + { + "epoch": 1.7764111574785857, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.42244291305542, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.728361964225769, + "num_tokens": 418315319.0, + "step": 16176 + }, + { + "epoch": 1.7765209751811992, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.382660150527954, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7030577659606934, + "num_tokens": 418339650.0, + "step": 16177 + }, + { + "epoch": 1.7766307928838128, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 1.9838520288467407, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7201454043388367, + "num_tokens": 418370376.0, + "step": 16178 + }, + { + "epoch": 1.7767406105864265, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3253798484802246, + "learning_rate": 1e-06, + "loss": 0.9685, + "mean_token_accuracy": 0.7000527381896973, + "num_tokens": 418395633.0, + "step": 16179 + }, + { + "epoch": 1.7768504282890403, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.4047465324401855, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7249875664710999, + "num_tokens": 418417308.0, + "step": 16180 + }, + { + "epoch": 1.7769602459916538, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3150365352630615, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7252507209777832, + "num_tokens": 418441138.0, + "step": 16181 + }, + { + "epoch": 1.7770700636942676, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.049956798553467, + "learning_rate": 1e-06, + "loss": 0.9971, + "mean_token_accuracy": 0.7019298076629639, + "num_tokens": 418470919.0, + "step": 16182 + }, + { + "epoch": 1.777179881396881, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.537301778793335, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7168211936950684, + "num_tokens": 418493758.0, + "step": 16183 + }, + { + "epoch": 1.7772896990994949, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.3110435009002686, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7045916318893433, + "num_tokens": 418520258.0, + "step": 16184 + }, + { + "epoch": 1.7773995168021086, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.2908644676208496, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7385258078575134, + "num_tokens": 418544191.0, + "step": 16185 + }, + { + "epoch": 1.7775093345047221, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.5127363204956055, + "learning_rate": 1e-06, + "loss": 0.7391, + "mean_token_accuracy": 0.7617853283882141, + "num_tokens": 418564807.0, + "step": 16186 + }, + { + "epoch": 1.7776191522073357, + "ewc_loss": 9.417533874511719e-06, + "grad_norm": 2.168558359146118, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7405804991722107, + "num_tokens": 418590708.0, + "step": 16187 + }, + { + "epoch": 1.7777289699099494, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2471203804016113, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7114464044570923, + "num_tokens": 418617068.0, + "step": 16188 + }, + { + "epoch": 1.7778387876125632, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2578821182250977, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7222914099693298, + "num_tokens": 418640811.0, + "step": 16189 + }, + { + "epoch": 1.777948605315177, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.090980052947998, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7209398150444031, + "num_tokens": 418667841.0, + "step": 16190 + }, + { + "epoch": 1.7780584230177905, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0684754848480225, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7179197072982788, + "num_tokens": 418696645.0, + "step": 16191 + }, + { + "epoch": 1.778168240720404, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.02659010887146, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.72897869348526, + "num_tokens": 418726463.0, + "step": 16192 + }, + { + "epoch": 1.7782780584230178, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.087312936782837, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7119214534759521, + "num_tokens": 418755149.0, + "step": 16193 + }, + { + "epoch": 1.7783878761256315, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3571393489837646, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.727989912033081, + "num_tokens": 418778806.0, + "step": 16194 + }, + { + "epoch": 1.778497693828245, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.43133807182312, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.7380011677742004, + "num_tokens": 418799077.0, + "step": 16195 + }, + { + "epoch": 1.7786075115308586, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.331275701522827, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7352310419082642, + "num_tokens": 418823886.0, + "step": 16196 + }, + { + "epoch": 1.7787173292334724, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.182168960571289, + "learning_rate": 1e-06, + "loss": 0.9489, + "mean_token_accuracy": 0.7140066623687744, + "num_tokens": 418850668.0, + "step": 16197 + }, + { + "epoch": 1.7788271469360861, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1534409523010254, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.723937451839447, + "num_tokens": 418876540.0, + "step": 16198 + }, + { + "epoch": 1.7789369646386999, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.70702862739563, + "learning_rate": 1e-06, + "loss": 0.7454, + "mean_token_accuracy": 0.7576488852500916, + "num_tokens": 418894376.0, + "step": 16199 + }, + { + "epoch": 1.7790467823413134, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2653238773345947, + "learning_rate": 1e-06, + "loss": 0.9664, + "mean_token_accuracy": 0.7045383453369141, + "num_tokens": 418919237.0, + "step": 16200 + }, + { + "epoch": 1.779156600043927, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9266525506973267, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.6878795623779297, + "num_tokens": 418954808.0, + "step": 16201 + }, + { + "epoch": 1.7792664177465407, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.404824733734131, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7140090465545654, + "num_tokens": 418979222.0, + "step": 16202 + }, + { + "epoch": 1.7793762354491545, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2498226165771484, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7333751916885376, + "num_tokens": 419005298.0, + "step": 16203 + }, + { + "epoch": 1.7794860531517682, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.343723773956299, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7128006219863892, + "num_tokens": 419031203.0, + "step": 16204 + }, + { + "epoch": 1.7795958708543818, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2342300415039062, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7375063896179199, + "num_tokens": 419056050.0, + "step": 16205 + }, + { + "epoch": 1.7797056885569953, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2077693939208984, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7118836641311646, + "num_tokens": 419085664.0, + "step": 16206 + }, + { + "epoch": 1.779815506259609, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.5246424674987793, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7205597758293152, + "num_tokens": 419107035.0, + "step": 16207 + }, + { + "epoch": 1.7799253239622228, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0843665599823, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.720307469367981, + "num_tokens": 419134962.0, + "step": 16208 + }, + { + "epoch": 1.7800351416648363, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4902021884918213, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7222064137458801, + "num_tokens": 419156469.0, + "step": 16209 + }, + { + "epoch": 1.7801449593674499, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.417942523956299, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7116533517837524, + "num_tokens": 419180412.0, + "step": 16210 + }, + { + "epoch": 1.7802547770700636, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.161677122116089, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7173915505409241, + "num_tokens": 419209102.0, + "step": 16211 + }, + { + "epoch": 1.7803645947726774, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.522200107574463, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7327272295951843, + "num_tokens": 419230360.0, + "step": 16212 + }, + { + "epoch": 1.7804744124752911, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.984895944595337, + "learning_rate": 1e-06, + "loss": 1.0551, + "mean_token_accuracy": 0.6781232357025146, + "num_tokens": 419263585.0, + "step": 16213 + }, + { + "epoch": 1.7805842301779047, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3108086585998535, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7118903994560242, + "num_tokens": 419290289.0, + "step": 16214 + }, + { + "epoch": 1.7806940478805182, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.173316717147827, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7068915963172913, + "num_tokens": 419315191.0, + "step": 16215 + }, + { + "epoch": 1.780803865583132, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.012544631958008, + "learning_rate": 1e-06, + "loss": 1.0297, + "mean_token_accuracy": 0.6831860542297363, + "num_tokens": 419345979.0, + "step": 16216 + }, + { + "epoch": 1.7809136832857457, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3177504539489746, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7460581660270691, + "num_tokens": 419369204.0, + "step": 16217 + }, + { + "epoch": 1.7810235009883595, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.239483594894409, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7402768135070801, + "num_tokens": 419393122.0, + "step": 16218 + }, + { + "epoch": 1.781133318690973, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.161215305328369, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7123408317565918, + "num_tokens": 419420255.0, + "step": 16219 + }, + { + "epoch": 1.7812431363935866, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.265395164489746, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7189849019050598, + "num_tokens": 419445421.0, + "step": 16220 + }, + { + "epoch": 1.7813529540962003, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1647145748138428, + "learning_rate": 1e-06, + "loss": 0.9853, + "mean_token_accuracy": 0.6990989446640015, + "num_tokens": 419473130.0, + "step": 16221 + }, + { + "epoch": 1.781462771798814, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.392575979232788, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7132909297943115, + "num_tokens": 419494795.0, + "step": 16222 + }, + { + "epoch": 1.7815725895014276, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4206008911132812, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7099814414978027, + "num_tokens": 419519442.0, + "step": 16223 + }, + { + "epoch": 1.7816824072040411, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3034889698028564, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7114953398704529, + "num_tokens": 419543835.0, + "step": 16224 + }, + { + "epoch": 1.781792224906655, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.5593621730804443, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7201259136199951, + "num_tokens": 419565539.0, + "step": 16225 + }, + { + "epoch": 1.7819020426092687, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.267652988433838, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7232115268707275, + "num_tokens": 419589785.0, + "step": 16226 + }, + { + "epoch": 1.7820118603118824, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1193370819091797, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.7546126842498779, + "num_tokens": 419614859.0, + "step": 16227 + }, + { + "epoch": 1.782121678014496, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0984294414520264, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7325571775436401, + "num_tokens": 419645316.0, + "step": 16228 + }, + { + "epoch": 1.7822314957171095, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.321485996246338, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7347569465637207, + "num_tokens": 419669390.0, + "step": 16229 + }, + { + "epoch": 1.7823413134197232, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.455343723297119, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.703968346118927, + "num_tokens": 419692156.0, + "step": 16230 + }, + { + "epoch": 1.782451131122337, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.215731382369995, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7235856056213379, + "num_tokens": 419718398.0, + "step": 16231 + }, + { + "epoch": 1.7825609488249505, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1823201179504395, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7249984741210938, + "num_tokens": 419743950.0, + "step": 16232 + }, + { + "epoch": 1.7826707665275643, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.477937698364258, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7433642148971558, + "num_tokens": 419764640.0, + "step": 16233 + }, + { + "epoch": 1.7827805842301778, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1817784309387207, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7222634553909302, + "num_tokens": 419792119.0, + "step": 16234 + }, + { + "epoch": 1.7828904019327916, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.081718921661377, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7251677513122559, + "num_tokens": 419820532.0, + "step": 16235 + }, + { + "epoch": 1.7830002196354053, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2010018825531006, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7445074915885925, + "num_tokens": 419845548.0, + "step": 16236 + }, + { + "epoch": 1.7831100373380189, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4042041301727295, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7234104871749878, + "num_tokens": 419867079.0, + "step": 16237 + }, + { + "epoch": 1.7832198550406324, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.417428731918335, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7270192503929138, + "num_tokens": 419891368.0, + "step": 16238 + }, + { + "epoch": 1.7833296727432462, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.370321273803711, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.713593602180481, + "num_tokens": 419915401.0, + "step": 16239 + }, + { + "epoch": 1.78343949044586, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1846373081207275, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7213670015335083, + "num_tokens": 419941102.0, + "step": 16240 + }, + { + "epoch": 1.7835493081484737, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.6879286766052246, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.728082537651062, + "num_tokens": 419960710.0, + "step": 16241 + }, + { + "epoch": 1.7836591258510872, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.570647954940796, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7333574295043945, + "num_tokens": 419979796.0, + "step": 16242 + }, + { + "epoch": 1.7837689435537007, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.6161885261535645, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7268073558807373, + "num_tokens": 419999905.0, + "step": 16243 + }, + { + "epoch": 1.7838787612563145, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0426783561706543, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7134643793106079, + "num_tokens": 420029631.0, + "step": 16244 + }, + { + "epoch": 1.7839885789589283, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.429637908935547, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7304874658584595, + "num_tokens": 420051474.0, + "step": 16245 + }, + { + "epoch": 1.7840983966615418, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.5383963584899902, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.726660966873169, + "num_tokens": 420071650.0, + "step": 16246 + }, + { + "epoch": 1.7842082143641556, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.513958215713501, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7196907997131348, + "num_tokens": 420094751.0, + "step": 16247 + }, + { + "epoch": 1.784318032066769, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.273747444152832, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7240220308303833, + "num_tokens": 420121656.0, + "step": 16248 + }, + { + "epoch": 1.7844278497693828, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.454765796661377, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7349721193313599, + "num_tokens": 420146275.0, + "step": 16249 + }, + { + "epoch": 1.7845376674719966, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2374165058135986, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7055222988128662, + "num_tokens": 420173068.0, + "step": 16250 + }, + { + "epoch": 1.7846474851746101, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0829243659973145, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7153944969177246, + "num_tokens": 420202313.0, + "step": 16251 + }, + { + "epoch": 1.7847573028772237, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.373896360397339, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7510303854942322, + "num_tokens": 420225258.0, + "step": 16252 + }, + { + "epoch": 1.7848671205798374, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0796701908111572, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7194262742996216, + "num_tokens": 420254837.0, + "step": 16253 + }, + { + "epoch": 1.7849769382824512, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.066234588623047, + "learning_rate": 1e-06, + "loss": 0.9553, + "mean_token_accuracy": 0.7031477689743042, + "num_tokens": 420284882.0, + "step": 16254 + }, + { + "epoch": 1.785086755985065, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2575459480285645, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7166414856910706, + "num_tokens": 420312526.0, + "step": 16255 + }, + { + "epoch": 1.7851965736876785, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1909427642822266, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7118717432022095, + "num_tokens": 420341823.0, + "step": 16256 + }, + { + "epoch": 1.785306391390292, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9295231103897095, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.723879873752594, + "num_tokens": 420372663.0, + "step": 16257 + }, + { + "epoch": 1.7854162090929058, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.274423360824585, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7124985456466675, + "num_tokens": 420396983.0, + "step": 16258 + }, + { + "epoch": 1.7855260267955195, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2232179641723633, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7220983505249023, + "num_tokens": 420421503.0, + "step": 16259 + }, + { + "epoch": 1.785635844498133, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1629080772399902, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7015066146850586, + "num_tokens": 420451332.0, + "step": 16260 + }, + { + "epoch": 1.7857456622007466, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0980801582336426, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7156491279602051, + "num_tokens": 420481896.0, + "step": 16261 + }, + { + "epoch": 1.7858554799033604, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2355380058288574, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7294319868087769, + "num_tokens": 420507325.0, + "step": 16262 + }, + { + "epoch": 1.7859652976059741, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.236945390701294, + "learning_rate": 1e-06, + "loss": 0.7649, + "mean_token_accuracy": 0.7550545930862427, + "num_tokens": 420532423.0, + "step": 16263 + }, + { + "epoch": 1.7860751153085879, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1533000469207764, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7209088206291199, + "num_tokens": 420558188.0, + "step": 16264 + }, + { + "epoch": 1.7861849330112014, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.175006628036499, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7301334142684937, + "num_tokens": 420583440.0, + "step": 16265 + }, + { + "epoch": 1.786294750713815, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.160613536834717, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7302491664886475, + "num_tokens": 420608649.0, + "step": 16266 + }, + { + "epoch": 1.7864045684164287, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.7340245246887207, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.727169394493103, + "num_tokens": 420627120.0, + "step": 16267 + }, + { + "epoch": 1.7865143861190425, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.081037998199463, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7101731300354004, + "num_tokens": 420656474.0, + "step": 16268 + }, + { + "epoch": 1.7866242038216562, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1536977291107178, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7319664359092712, + "num_tokens": 420683699.0, + "step": 16269 + }, + { + "epoch": 1.7867340215242697, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1866629123687744, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7122684717178345, + "num_tokens": 420708766.0, + "step": 16270 + }, + { + "epoch": 1.7868438392268833, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9460461139678955, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.6982454061508179, + "num_tokens": 420741260.0, + "step": 16271 + }, + { + "epoch": 1.786953656929497, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1457414627075195, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7180102467536926, + "num_tokens": 420768544.0, + "step": 16272 + }, + { + "epoch": 1.7870634746321108, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.354509115219116, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7442372441291809, + "num_tokens": 420790272.0, + "step": 16273 + }, + { + "epoch": 1.7871732923347243, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.048124313354492, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7101797461509705, + "num_tokens": 420818546.0, + "step": 16274 + }, + { + "epoch": 1.7872831100373379, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0641515254974365, + "learning_rate": 1e-06, + "loss": 0.9835, + "mean_token_accuracy": 0.7002263069152832, + "num_tokens": 420848155.0, + "step": 16275 + }, + { + "epoch": 1.7873929277399516, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.177069664001465, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7276438474655151, + "num_tokens": 420873991.0, + "step": 16276 + }, + { + "epoch": 1.7875027454425654, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.28669810295105, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7220373153686523, + "num_tokens": 420899750.0, + "step": 16277 + }, + { + "epoch": 1.7876125631451791, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2917399406433105, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.716755211353302, + "num_tokens": 420926433.0, + "step": 16278 + }, + { + "epoch": 1.7877223808477927, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0944716930389404, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7315007448196411, + "num_tokens": 420951268.0, + "step": 16279 + }, + { + "epoch": 1.7878321985504062, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0518877506256104, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7154980301856995, + "num_tokens": 420981931.0, + "step": 16280 + }, + { + "epoch": 1.78794201625302, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4309182167053223, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.691017746925354, + "num_tokens": 421007442.0, + "step": 16281 + }, + { + "epoch": 1.7880518339556337, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.187299966812134, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7103040218353271, + "num_tokens": 421034920.0, + "step": 16282 + }, + { + "epoch": 1.7881616516582475, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.984126329421997, + "learning_rate": 1e-06, + "loss": 0.9861, + "mean_token_accuracy": 0.7058238983154297, + "num_tokens": 421066890.0, + "step": 16283 + }, + { + "epoch": 1.788271469360861, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.346440315246582, + "learning_rate": 1e-06, + "loss": 0.9644, + "mean_token_accuracy": 0.6999456882476807, + "num_tokens": 421090461.0, + "step": 16284 + }, + { + "epoch": 1.7883812870634745, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4424285888671875, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.722937822341919, + "num_tokens": 421112929.0, + "step": 16285 + }, + { + "epoch": 1.7884911047660883, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2462520599365234, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7275499701499939, + "num_tokens": 421137933.0, + "step": 16286 + }, + { + "epoch": 1.788600922468702, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.218823194503784, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7218114137649536, + "num_tokens": 421165253.0, + "step": 16287 + }, + { + "epoch": 1.7887107401713156, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.274918556213379, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7493465542793274, + "num_tokens": 421187552.0, + "step": 16288 + }, + { + "epoch": 1.7888205578739291, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2024710178375244, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.729520857334137, + "num_tokens": 421213194.0, + "step": 16289 + }, + { + "epoch": 1.7889303755765429, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.298043727874756, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7180176973342896, + "num_tokens": 421236247.0, + "step": 16290 + }, + { + "epoch": 1.7890401932791566, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.55741024017334, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7255522608757019, + "num_tokens": 421257100.0, + "step": 16291 + }, + { + "epoch": 1.7891500109817704, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2068848609924316, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7325609922409058, + "num_tokens": 421280175.0, + "step": 16292 + }, + { + "epoch": 1.789259828684384, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.446382999420166, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7490061521530151, + "num_tokens": 421300730.0, + "step": 16293 + }, + { + "epoch": 1.7893696463869975, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3280444145202637, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7416477203369141, + "num_tokens": 421323715.0, + "step": 16294 + }, + { + "epoch": 1.7894794640896112, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.472153902053833, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.711079478263855, + "num_tokens": 421346426.0, + "step": 16295 + }, + { + "epoch": 1.789589281792225, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2993505001068115, + "learning_rate": 1e-06, + "loss": 1.0116, + "mean_token_accuracy": 0.6932623386383057, + "num_tokens": 421371371.0, + "step": 16296 + }, + { + "epoch": 1.7896990994948385, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.247464656829834, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.7044158577919006, + "num_tokens": 421396099.0, + "step": 16297 + }, + { + "epoch": 1.7898089171974523, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.181885242462158, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7235429286956787, + "num_tokens": 421424550.0, + "step": 16298 + }, + { + "epoch": 1.7899187349000658, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3759169578552246, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.74034184217453, + "num_tokens": 421446253.0, + "step": 16299 + }, + { + "epoch": 1.7900285526026796, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9250327348709106, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7126563787460327, + "num_tokens": 421480936.0, + "step": 16300 + }, + { + "epoch": 1.7901383703052933, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4157512187957764, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7218210697174072, + "num_tokens": 421501546.0, + "step": 16301 + }, + { + "epoch": 1.7902481880079069, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4414548873901367, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7167279720306396, + "num_tokens": 421523198.0, + "step": 16302 + }, + { + "epoch": 1.7903580057105204, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.174128293991089, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.720760703086853, + "num_tokens": 421549623.0, + "step": 16303 + }, + { + "epoch": 1.7904678234131342, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2399675846099854, + "learning_rate": 1e-06, + "loss": 0.9581, + "mean_token_accuracy": 0.6979895830154419, + "num_tokens": 421574848.0, + "step": 16304 + }, + { + "epoch": 1.790577641115748, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3870933055877686, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7227354645729065, + "num_tokens": 421596570.0, + "step": 16305 + }, + { + "epoch": 1.7906874588183617, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0413551330566406, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7129957675933838, + "num_tokens": 421626890.0, + "step": 16306 + }, + { + "epoch": 1.7907972765209752, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2759897708892822, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7235336303710938, + "num_tokens": 421652410.0, + "step": 16307 + }, + { + "epoch": 1.7909070942235887, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2462401390075684, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7117975950241089, + "num_tokens": 421677817.0, + "step": 16308 + }, + { + "epoch": 1.7910169119262025, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.248086929321289, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7320393323898315, + "num_tokens": 421701895.0, + "step": 16309 + }, + { + "epoch": 1.7911267296288162, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.163865566253662, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7231314182281494, + "num_tokens": 421727887.0, + "step": 16310 + }, + { + "epoch": 1.7912365473314298, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1930124759674072, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7213121652603149, + "num_tokens": 421754232.0, + "step": 16311 + }, + { + "epoch": 1.7913463650340435, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.6212494373321533, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.714055061340332, + "num_tokens": 421774594.0, + "step": 16312 + }, + { + "epoch": 1.791456182736657, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3615002632141113, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.7057059407234192, + "num_tokens": 421799392.0, + "step": 16313 + }, + { + "epoch": 1.7915660004392708, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.496088981628418, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7052539587020874, + "num_tokens": 421827295.0, + "step": 16314 + }, + { + "epoch": 1.7916758181418846, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.269705057144165, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7183822393417358, + "num_tokens": 421853525.0, + "step": 16315 + }, + { + "epoch": 1.7917856358444981, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1874239444732666, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.7335482835769653, + "num_tokens": 421880922.0, + "step": 16316 + }, + { + "epoch": 1.7918954535471117, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.304445266723633, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7327213287353516, + "num_tokens": 421905972.0, + "step": 16317 + }, + { + "epoch": 1.7920052712497254, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.367543935775757, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7162078619003296, + "num_tokens": 421928724.0, + "step": 16318 + }, + { + "epoch": 1.7921150889523392, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.102954864501953, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7139232754707336, + "num_tokens": 421956659.0, + "step": 16319 + }, + { + "epoch": 1.792224906654953, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3056623935699463, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.71175616979599, + "num_tokens": 421979564.0, + "step": 16320 + }, + { + "epoch": 1.7923347243575665, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3105223178863525, + "learning_rate": 1e-06, + "loss": 0.7648, + "mean_token_accuracy": 0.7568677663803101, + "num_tokens": 422000886.0, + "step": 16321 + }, + { + "epoch": 1.79244454206018, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.18806791305542, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7073237299919128, + "num_tokens": 422027894.0, + "step": 16322 + }, + { + "epoch": 1.7925543597627938, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3425047397613525, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7179523706436157, + "num_tokens": 422050366.0, + "step": 16323 + }, + { + "epoch": 1.7926641774654075, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1540701389312744, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7231537103652954, + "num_tokens": 422078367.0, + "step": 16324 + }, + { + "epoch": 1.792773995168021, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2994470596313477, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7325652837753296, + "num_tokens": 422102979.0, + "step": 16325 + }, + { + "epoch": 1.7928838128706346, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3169026374816895, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7247582674026489, + "num_tokens": 422127190.0, + "step": 16326 + }, + { + "epoch": 1.7929936305732483, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.324770927429199, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7064853310585022, + "num_tokens": 422151054.0, + "step": 16327 + }, + { + "epoch": 1.793103448275862, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1565747261047363, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7398557662963867, + "num_tokens": 422176163.0, + "step": 16328 + }, + { + "epoch": 1.7932132659784759, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1904208660125732, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7177467346191406, + "num_tokens": 422200960.0, + "step": 16329 + }, + { + "epoch": 1.7933230836810894, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.988741159439087, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7300662994384766, + "num_tokens": 422229630.0, + "step": 16330 + }, + { + "epoch": 1.793432901383703, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.0326082706451416, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7103304862976074, + "num_tokens": 422259552.0, + "step": 16331 + }, + { + "epoch": 1.7935427190863167, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1214888095855713, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7144618034362793, + "num_tokens": 422289751.0, + "step": 16332 + }, + { + "epoch": 1.7936525367889304, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.306046485900879, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7293250560760498, + "num_tokens": 422314633.0, + "step": 16333 + }, + { + "epoch": 1.7937623544915442, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.073247194290161, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7249082326889038, + "num_tokens": 422341077.0, + "step": 16334 + }, + { + "epoch": 1.7938721721941577, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.241196393966675, + "learning_rate": 1e-06, + "loss": 0.9881, + "mean_token_accuracy": 0.6956178545951843, + "num_tokens": 422369625.0, + "step": 16335 + }, + { + "epoch": 1.7939819898967713, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3114404678344727, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7240384817123413, + "num_tokens": 422394613.0, + "step": 16336 + }, + { + "epoch": 1.794091807599385, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4100301265716553, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7247941493988037, + "num_tokens": 422417982.0, + "step": 16337 + }, + { + "epoch": 1.7942016253019988, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1735947132110596, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7362491488456726, + "num_tokens": 422443533.0, + "step": 16338 + }, + { + "epoch": 1.7943114430046123, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3481898307800293, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7087150812149048, + "num_tokens": 422468024.0, + "step": 16339 + }, + { + "epoch": 1.7944212607072259, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.455202579498291, + "learning_rate": 1e-06, + "loss": 0.9408, + "mean_token_accuracy": 0.7143024206161499, + "num_tokens": 422490614.0, + "step": 16340 + }, + { + "epoch": 1.7945310784098396, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.46768856048584, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7293472290039062, + "num_tokens": 422512313.0, + "step": 16341 + }, + { + "epoch": 1.7946408961124534, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.123875856399536, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7285523414611816, + "num_tokens": 422537722.0, + "step": 16342 + }, + { + "epoch": 1.7947507138150671, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9296175241470337, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7245107889175415, + "num_tokens": 422567382.0, + "step": 16343 + }, + { + "epoch": 1.7948605315176807, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.135666608810425, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7040106058120728, + "num_tokens": 422595383.0, + "step": 16344 + }, + { + "epoch": 1.7949703492202942, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.205686569213867, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7184463143348694, + "num_tokens": 422623755.0, + "step": 16345 + }, + { + "epoch": 1.795080166922908, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2313551902770996, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7115610837936401, + "num_tokens": 422649779.0, + "step": 16346 + }, + { + "epoch": 1.7951899846255217, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4081971645355225, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7304103374481201, + "num_tokens": 422671435.0, + "step": 16347 + }, + { + "epoch": 1.7952998023281352, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.164832830429077, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.7409629225730896, + "num_tokens": 422696878.0, + "step": 16348 + }, + { + "epoch": 1.795409620030749, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.294457197189331, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7425304651260376, + "num_tokens": 422720433.0, + "step": 16349 + }, + { + "epoch": 1.7955194377333625, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4912948608398438, + "learning_rate": 1e-06, + "loss": 0.9713, + "mean_token_accuracy": 0.7025418281555176, + "num_tokens": 422743653.0, + "step": 16350 + }, + { + "epoch": 1.7956292554359763, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.1825907230377197, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7140578031539917, + "num_tokens": 422771067.0, + "step": 16351 + }, + { + "epoch": 1.79573907313859, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.153125286102295, + "learning_rate": 1e-06, + "loss": 0.9821, + "mean_token_accuracy": 0.6995011568069458, + "num_tokens": 422798121.0, + "step": 16352 + }, + { + "epoch": 1.7958488908412036, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.4096500873565674, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7142478227615356, + "num_tokens": 422821499.0, + "step": 16353 + }, + { + "epoch": 1.7959587085438171, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.3176305294036865, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7345898747444153, + "num_tokens": 422843980.0, + "step": 16354 + }, + { + "epoch": 1.7960685262464309, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.352800130844116, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7260070443153381, + "num_tokens": 422866031.0, + "step": 16355 + }, + { + "epoch": 1.7961783439490446, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9319484233856201, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7203589677810669, + "num_tokens": 422900563.0, + "step": 16356 + }, + { + "epoch": 1.7962881616516584, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2025253772735596, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7143392562866211, + "num_tokens": 422926656.0, + "step": 16357 + }, + { + "epoch": 1.796397979354272, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.225353240966797, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7223372459411621, + "num_tokens": 422952485.0, + "step": 16358 + }, + { + "epoch": 1.7965077970568855, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 1.9965320825576782, + "learning_rate": 1e-06, + "loss": 0.9947, + "mean_token_accuracy": 0.6978905200958252, + "num_tokens": 422984735.0, + "step": 16359 + }, + { + "epoch": 1.7966176147594992, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.2535996437072754, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7319261431694031, + "num_tokens": 423008382.0, + "step": 16360 + }, + { + "epoch": 1.796727432462113, + "ewc_loss": 9.47713851928711e-06, + "grad_norm": 2.591797113418579, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7384912967681885, + "num_tokens": 423028419.0, + "step": 16361 + }, + { + "epoch": 1.7968372501647265, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.2607107162475586, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7229135036468506, + "num_tokens": 423053152.0, + "step": 16362 + }, + { + "epoch": 1.7969470678673403, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.152212381362915, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.7355508804321289, + "num_tokens": 423079331.0, + "step": 16363 + }, + { + "epoch": 1.7970568855699538, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.426903486251831, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7158715128898621, + "num_tokens": 423102215.0, + "step": 16364 + }, + { + "epoch": 1.7971667032725676, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9969805479049683, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7496747374534607, + "num_tokens": 423131041.0, + "step": 16365 + }, + { + "epoch": 1.7972765209751813, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.146578788757324, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7289823293685913, + "num_tokens": 423158313.0, + "step": 16366 + }, + { + "epoch": 1.7973863386777948, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.620563507080078, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7245925664901733, + "num_tokens": 423177086.0, + "step": 16367 + }, + { + "epoch": 1.7974961563804084, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.227525472640991, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7224098443984985, + "num_tokens": 423202341.0, + "step": 16368 + }, + { + "epoch": 1.7976059740830221, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0012996196746826, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7033097147941589, + "num_tokens": 423232975.0, + "step": 16369 + }, + { + "epoch": 1.797715791785636, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4840009212493896, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.714266836643219, + "num_tokens": 423255271.0, + "step": 16370 + }, + { + "epoch": 1.7978256094882497, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.277294635772705, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7097230553627014, + "num_tokens": 423279840.0, + "step": 16371 + }, + { + "epoch": 1.7979354271908632, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.385061264038086, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7333523035049438, + "num_tokens": 423301839.0, + "step": 16372 + }, + { + "epoch": 1.7980452448934767, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.2814505100250244, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7095476388931274, + "num_tokens": 423325332.0, + "step": 16373 + }, + { + "epoch": 1.7981550625960905, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.06347918510437, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7308932542800903, + "num_tokens": 423354366.0, + "step": 16374 + }, + { + "epoch": 1.7982648802987042, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3468449115753174, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.7069666385650635, + "num_tokens": 423378460.0, + "step": 16375 + }, + { + "epoch": 1.7983746980013178, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0459694862365723, + "learning_rate": 1e-06, + "loss": 1.0067, + "mean_token_accuracy": 0.7073627710342407, + "num_tokens": 423409450.0, + "step": 16376 + }, + { + "epoch": 1.7984845157039313, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3308942317962646, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7212410569190979, + "num_tokens": 423432078.0, + "step": 16377 + }, + { + "epoch": 1.798594333406545, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9926565885543823, + "learning_rate": 1e-06, + "loss": 0.9808, + "mean_token_accuracy": 0.6977895498275757, + "num_tokens": 423463737.0, + "step": 16378 + }, + { + "epoch": 1.7987041511091588, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.233558416366577, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7366924285888672, + "num_tokens": 423490202.0, + "step": 16379 + }, + { + "epoch": 1.7988139688117726, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0638339519500732, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7285749316215515, + "num_tokens": 423518171.0, + "step": 16380 + }, + { + "epoch": 1.7989237865143861, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.996801733970642, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7048412561416626, + "num_tokens": 423549214.0, + "step": 16381 + }, + { + "epoch": 1.7990336042169996, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.968483328819275, + "learning_rate": 1e-06, + "loss": 0.9976, + "mean_token_accuracy": 0.6916885375976562, + "num_tokens": 423583822.0, + "step": 16382 + }, + { + "epoch": 1.7991434219196134, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.050976276397705, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7184025049209595, + "num_tokens": 423611450.0, + "step": 16383 + }, + { + "epoch": 1.7992532396222272, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.128077983856201, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7273148894309998, + "num_tokens": 423639246.0, + "step": 16384 + }, + { + "epoch": 1.799363057324841, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.342576503753662, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7182348966598511, + "num_tokens": 423663237.0, + "step": 16385 + }, + { + "epoch": 1.7994728750274545, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.8481669425964355, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7197643518447876, + "num_tokens": 423682575.0, + "step": 16386 + }, + { + "epoch": 1.799582692730068, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.19159197807312, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7135413885116577, + "num_tokens": 423709095.0, + "step": 16387 + }, + { + "epoch": 1.7996925104326817, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.525052070617676, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.744606077671051, + "num_tokens": 423729133.0, + "step": 16388 + }, + { + "epoch": 1.7998023281352955, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.050766944885254, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7371299266815186, + "num_tokens": 423756251.0, + "step": 16389 + }, + { + "epoch": 1.799912145837909, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.2952499389648438, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7058365941047668, + "num_tokens": 423780483.0, + "step": 16390 + }, + { + "epoch": 1.8000219635405226, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1183228492736816, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7204383611679077, + "num_tokens": 423807886.0, + "step": 16391 + }, + { + "epoch": 1.8001317812431363, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3928239345550537, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7354260087013245, + "num_tokens": 423830403.0, + "step": 16392 + }, + { + "epoch": 1.80024159894575, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.008793592453003, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7045026421546936, + "num_tokens": 423862554.0, + "step": 16393 + }, + { + "epoch": 1.8003514166483638, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.232672691345215, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7316477298736572, + "num_tokens": 423889404.0, + "step": 16394 + }, + { + "epoch": 1.8004612343509774, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9833133220672607, + "learning_rate": 1e-06, + "loss": 0.9234, + "mean_token_accuracy": 0.7153286933898926, + "num_tokens": 423920984.0, + "step": 16395 + }, + { + "epoch": 1.800571052053591, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4483349323272705, + "learning_rate": 1e-06, + "loss": 0.966, + "mean_token_accuracy": 0.7040829658508301, + "num_tokens": 423944673.0, + "step": 16396 + }, + { + "epoch": 1.8006808697562047, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0759849548339844, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7112240195274353, + "num_tokens": 423971785.0, + "step": 16397 + }, + { + "epoch": 1.8007906874588184, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.553744077682495, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7372763156890869, + "num_tokens": 423993324.0, + "step": 16398 + }, + { + "epoch": 1.8009005051614322, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.079228162765503, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7206563949584961, + "num_tokens": 424022594.0, + "step": 16399 + }, + { + "epoch": 1.8010103228640457, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.142404079437256, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.713895857334137, + "num_tokens": 424050715.0, + "step": 16400 + }, + { + "epoch": 1.8011201405666593, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.378786563873291, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7104849815368652, + "num_tokens": 424075604.0, + "step": 16401 + }, + { + "epoch": 1.801229958269273, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.729004383087158, + "learning_rate": 1e-06, + "loss": 0.7583, + "mean_token_accuracy": 0.7540390491485596, + "num_tokens": 424093160.0, + "step": 16402 + }, + { + "epoch": 1.8013397759718868, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.097459077835083, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.6994489431381226, + "num_tokens": 424121916.0, + "step": 16403 + }, + { + "epoch": 1.8014495936745003, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1462790966033936, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7273682355880737, + "num_tokens": 424146915.0, + "step": 16404 + }, + { + "epoch": 1.8015594113771138, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.217468738555908, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7016173601150513, + "num_tokens": 424174791.0, + "step": 16405 + }, + { + "epoch": 1.8016692290797276, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9235526323318481, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.714203417301178, + "num_tokens": 424206433.0, + "step": 16406 + }, + { + "epoch": 1.8017790467823414, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.327461004257202, + "learning_rate": 1e-06, + "loss": 0.7359, + "mean_token_accuracy": 0.7634795904159546, + "num_tokens": 424229767.0, + "step": 16407 + }, + { + "epoch": 1.801888864484955, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.228344202041626, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7204253077507019, + "num_tokens": 424258047.0, + "step": 16408 + }, + { + "epoch": 1.8019986821875686, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3264694213867188, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7053520679473877, + "num_tokens": 424281964.0, + "step": 16409 + }, + { + "epoch": 1.8021084998901822, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.126589298248291, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7182372212409973, + "num_tokens": 424309618.0, + "step": 16410 + }, + { + "epoch": 1.802218317592796, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.2143259048461914, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7263095378875732, + "num_tokens": 424336400.0, + "step": 16411 + }, + { + "epoch": 1.8023281352954097, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4666662216186523, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7245294451713562, + "num_tokens": 424358957.0, + "step": 16412 + }, + { + "epoch": 1.8024379529980232, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.9389222860336304, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7109397053718567, + "num_tokens": 424390698.0, + "step": 16413 + }, + { + "epoch": 1.802547770700637, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.7210710048675537, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.732659637928009, + "num_tokens": 424408802.0, + "step": 16414 + }, + { + "epoch": 1.8026575884032505, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4232590198516846, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7207996845245361, + "num_tokens": 424432093.0, + "step": 16415 + }, + { + "epoch": 1.8027674061058643, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.240769624710083, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7206048965454102, + "num_tokens": 424457333.0, + "step": 16416 + }, + { + "epoch": 1.802877223808478, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4387590885162354, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7272752523422241, + "num_tokens": 424479372.0, + "step": 16417 + }, + { + "epoch": 1.8029870415110916, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.20583176612854, + "learning_rate": 1e-06, + "loss": 0.9738, + "mean_token_accuracy": 0.7002160549163818, + "num_tokens": 424506414.0, + "step": 16418 + }, + { + "epoch": 1.803096859213705, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4723401069641113, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7002911567687988, + "num_tokens": 424530747.0, + "step": 16419 + }, + { + "epoch": 1.8032066769163189, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3242809772491455, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7192781567573547, + "num_tokens": 424554944.0, + "step": 16420 + }, + { + "epoch": 1.8033164946189326, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.2902743816375732, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.749761164188385, + "num_tokens": 424577934.0, + "step": 16421 + }, + { + "epoch": 1.8034263123215464, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.309765100479126, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7105323672294617, + "num_tokens": 424600956.0, + "step": 16422 + }, + { + "epoch": 1.80353613002416, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.177485227584839, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7273057699203491, + "num_tokens": 424625842.0, + "step": 16423 + }, + { + "epoch": 1.8036459477267734, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.221895694732666, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7251203060150146, + "num_tokens": 424652210.0, + "step": 16424 + }, + { + "epoch": 1.8037557654293872, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.35491943359375, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7297135591506958, + "num_tokens": 424676026.0, + "step": 16425 + }, + { + "epoch": 1.803865583132001, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.513580799102783, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7128402590751648, + "num_tokens": 424697535.0, + "step": 16426 + }, + { + "epoch": 1.8039754008346145, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.258958578109741, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.731133759021759, + "num_tokens": 424725156.0, + "step": 16427 + }, + { + "epoch": 1.8040852185372283, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.5510172843933105, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7141063809394836, + "num_tokens": 424746333.0, + "step": 16428 + }, + { + "epoch": 1.8041950362398418, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.277095317840576, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7352516055107117, + "num_tokens": 424771557.0, + "step": 16429 + }, + { + "epoch": 1.8043048539424555, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.468986749649048, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7137649059295654, + "num_tokens": 424792970.0, + "step": 16430 + }, + { + "epoch": 1.8044146716450693, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.5185060501098633, + "learning_rate": 1e-06, + "loss": 0.7453, + "mean_token_accuracy": 0.7591334581375122, + "num_tokens": 424812348.0, + "step": 16431 + }, + { + "epoch": 1.8045244893476828, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.120467185974121, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7075411081314087, + "num_tokens": 424840825.0, + "step": 16432 + }, + { + "epoch": 1.8046343070502964, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3372433185577393, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7405779361724854, + "num_tokens": 424864673.0, + "step": 16433 + }, + { + "epoch": 1.8047441247529101, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.448308229446411, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7293775081634521, + "num_tokens": 424885996.0, + "step": 16434 + }, + { + "epoch": 1.8048539424555239, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.246063232421875, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7148419618606567, + "num_tokens": 424913553.0, + "step": 16435 + }, + { + "epoch": 1.8049637601581376, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.4518136978149414, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7106278538703918, + "num_tokens": 424936463.0, + "step": 16436 + }, + { + "epoch": 1.8050735778607512, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1288177967071533, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.7032367587089539, + "num_tokens": 424967622.0, + "step": 16437 + }, + { + "epoch": 1.8051833955633647, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1650824546813965, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7096949815750122, + "num_tokens": 424996346.0, + "step": 16438 + }, + { + "epoch": 1.8052932132659785, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.5247817039489746, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7297062873840332, + "num_tokens": 425016697.0, + "step": 16439 + }, + { + "epoch": 1.8054030309685922, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3535561561584473, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7218188047409058, + "num_tokens": 425040663.0, + "step": 16440 + }, + { + "epoch": 1.8055128486712058, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1998848915100098, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7359808683395386, + "num_tokens": 425066619.0, + "step": 16441 + }, + { + "epoch": 1.8056226663738193, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.0781335830688477, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7145934104919434, + "num_tokens": 425095651.0, + "step": 16442 + }, + { + "epoch": 1.805732484076433, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.927396535873413, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7030363082885742, + "num_tokens": 425129103.0, + "step": 16443 + }, + { + "epoch": 1.8058423017790468, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1200788021087646, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7191592454910278, + "num_tokens": 425158032.0, + "step": 16444 + }, + { + "epoch": 1.8059521194816606, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3768820762634277, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7174474000930786, + "num_tokens": 425179784.0, + "step": 16445 + }, + { + "epoch": 1.806061937184274, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.214012384414673, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7356123328208923, + "num_tokens": 425204646.0, + "step": 16446 + }, + { + "epoch": 1.8061717548868876, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.3867716789245605, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7002394795417786, + "num_tokens": 425230154.0, + "step": 16447 + }, + { + "epoch": 1.8062815725895014, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.279273271560669, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7305012345314026, + "num_tokens": 425252949.0, + "step": 16448 + }, + { + "epoch": 1.8063913902921152, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.192919969558716, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7204219698905945, + "num_tokens": 425279039.0, + "step": 16449 + }, + { + "epoch": 1.806501207994729, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 1.7971115112304688, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7133635878562927, + "num_tokens": 425314120.0, + "step": 16450 + }, + { + "epoch": 1.8066110256973424, + "ewc_loss": 9.5367431640625e-06, + "grad_norm": 2.1547505855560303, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7262306213378906, + "num_tokens": 425341235.0, + "step": 16451 + }, + { + "epoch": 1.806720843399956, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.366863965988159, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7471317052841187, + "num_tokens": 425362936.0, + "step": 16452 + }, + { + "epoch": 1.8068306611025697, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.282562732696533, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7145531177520752, + "num_tokens": 425387980.0, + "step": 16453 + }, + { + "epoch": 1.8069404788051835, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.199946165084839, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.729582667350769, + "num_tokens": 425415000.0, + "step": 16454 + }, + { + "epoch": 1.807050296507797, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.395648956298828, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7222709655761719, + "num_tokens": 425438014.0, + "step": 16455 + }, + { + "epoch": 1.8071601142104106, + "ewc_loss": 9.59634780883789e-06, + "grad_norm": 2.0957765579223633, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7092470526695251, + "num_tokens": 425465809.0, + "step": 16456 + }, + { + "epoch": 1.8072699319130243, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.1121532917022705, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7075458765029907, + "num_tokens": 425494739.0, + "step": 16457 + }, + { + "epoch": 1.807379749615638, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2030279636383057, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7172226905822754, + "num_tokens": 425519222.0, + "step": 16458 + }, + { + "epoch": 1.8074895673182518, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.189906120300293, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7013972997665405, + "num_tokens": 425546860.0, + "step": 16459 + }, + { + "epoch": 1.8075993850208654, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.0790207386016846, + "learning_rate": 1e-06, + "loss": 1.0189, + "mean_token_accuracy": 0.6911906599998474, + "num_tokens": 425578098.0, + "step": 16460 + }, + { + "epoch": 1.807709202723479, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2395215034484863, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7128701210021973, + "num_tokens": 425603109.0, + "step": 16461 + }, + { + "epoch": 1.8078190204260927, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.4826231002807617, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7011653184890747, + "num_tokens": 425624711.0, + "step": 16462 + }, + { + "epoch": 1.8079288381287064, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.158895254135132, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7131644487380981, + "num_tokens": 425651333.0, + "step": 16463 + }, + { + "epoch": 1.8080386558313202, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.169867753982544, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7187533378601074, + "num_tokens": 425681369.0, + "step": 16464 + }, + { + "epoch": 1.8081484735339337, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2552833557128906, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7514856457710266, + "num_tokens": 425705748.0, + "step": 16465 + }, + { + "epoch": 1.8082582912365472, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.193134307861328, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7500251531600952, + "num_tokens": 425730636.0, + "step": 16466 + }, + { + "epoch": 1.808368108939161, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.149728298187256, + "learning_rate": 1e-06, + "loss": 0.9845, + "mean_token_accuracy": 0.7107500433921814, + "num_tokens": 425758932.0, + "step": 16467 + }, + { + "epoch": 1.8084779266417748, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.5528554916381836, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7250176668167114, + "num_tokens": 425778446.0, + "step": 16468 + }, + { + "epoch": 1.8085877443443883, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.0514419078826904, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7293142080307007, + "num_tokens": 425807780.0, + "step": 16469 + }, + { + "epoch": 1.8086975620470018, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2270243167877197, + "learning_rate": 1e-06, + "loss": 0.9668, + "mean_token_accuracy": 0.6983616352081299, + "num_tokens": 425834352.0, + "step": 16470 + }, + { + "epoch": 1.8088073797496156, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.009162664413452, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7088269591331482, + "num_tokens": 425864174.0, + "step": 16471 + }, + { + "epoch": 1.8089171974522293, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.0043258666992188, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.726561963558197, + "num_tokens": 425894335.0, + "step": 16472 + }, + { + "epoch": 1.809027015154843, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.1193103790283203, + "learning_rate": 1e-06, + "loss": 0.9533, + "mean_token_accuracy": 0.7075589895248413, + "num_tokens": 425925002.0, + "step": 16473 + }, + { + "epoch": 1.8091368328574566, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.2879343032836914, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7115409970283508, + "num_tokens": 425950776.0, + "step": 16474 + }, + { + "epoch": 1.8092466505600702, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.0933785438537598, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7111654281616211, + "num_tokens": 425977702.0, + "step": 16475 + }, + { + "epoch": 1.809356468262684, + "ewc_loss": 9.655952453613281e-06, + "grad_norm": 2.398341178894043, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7427636384963989, + "num_tokens": 425999777.0, + "step": 16476 + }, + { + "epoch": 1.8094662859652977, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4001340866088867, + "learning_rate": 1e-06, + "loss": 0.9462, + "mean_token_accuracy": 0.7056370973587036, + "num_tokens": 426023601.0, + "step": 16477 + }, + { + "epoch": 1.8095761036679112, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.040132999420166, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7225774526596069, + "num_tokens": 426052195.0, + "step": 16478 + }, + { + "epoch": 1.809685921370525, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.236227512359619, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7201774716377258, + "num_tokens": 426076543.0, + "step": 16479 + }, + { + "epoch": 1.8097957390731385, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.245281457901001, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7082476615905762, + "num_tokens": 426102646.0, + "step": 16480 + }, + { + "epoch": 1.8099055567757523, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1972784996032715, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7278625965118408, + "num_tokens": 426128498.0, + "step": 16481 + }, + { + "epoch": 1.810015374478366, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.272073268890381, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7227143049240112, + "num_tokens": 426152242.0, + "step": 16482 + }, + { + "epoch": 1.8101251921809796, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1615450382232666, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7178889513015747, + "num_tokens": 426179491.0, + "step": 16483 + }, + { + "epoch": 1.810235009883593, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0713140964508057, + "learning_rate": 1e-06, + "loss": 1.0365, + "mean_token_accuracy": 0.6870380640029907, + "num_tokens": 426210144.0, + "step": 16484 + }, + { + "epoch": 1.8103448275862069, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.044644832611084, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7083034515380859, + "num_tokens": 426240115.0, + "step": 16485 + }, + { + "epoch": 1.8104546452888206, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.29396915435791, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7205638885498047, + "num_tokens": 426264571.0, + "step": 16486 + }, + { + "epoch": 1.8105644629914344, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.9032233953475952, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.6972944140434265, + "num_tokens": 426299083.0, + "step": 16487 + }, + { + "epoch": 1.810674280694048, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.175884485244751, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7166093587875366, + "num_tokens": 426324702.0, + "step": 16488 + }, + { + "epoch": 1.8107840983966614, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.374467611312866, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7337075471878052, + "num_tokens": 426346607.0, + "step": 16489 + }, + { + "epoch": 1.8108939160992752, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2281057834625244, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7073596715927124, + "num_tokens": 426370961.0, + "step": 16490 + }, + { + "epoch": 1.811003733801889, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.039361000061035, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7221522331237793, + "num_tokens": 426401716.0, + "step": 16491 + }, + { + "epoch": 1.8111135515045025, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5673129558563232, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.714516818523407, + "num_tokens": 426423232.0, + "step": 16492 + }, + { + "epoch": 1.8112233692071162, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.581773042678833, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7198794484138489, + "num_tokens": 426444908.0, + "step": 16493 + }, + { + "epoch": 1.8113331869097298, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.571336269378662, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7475125193595886, + "num_tokens": 426464873.0, + "step": 16494 + }, + { + "epoch": 1.8114430046123435, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.008945941925049, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7134177684783936, + "num_tokens": 426493448.0, + "step": 16495 + }, + { + "epoch": 1.8115528223149573, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5213840007781982, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7207973003387451, + "num_tokens": 426514810.0, + "step": 16496 + }, + { + "epoch": 1.8116626400175708, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.497331142425537, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7387828826904297, + "num_tokens": 426535324.0, + "step": 16497 + }, + { + "epoch": 1.8117724577201844, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.124652862548828, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7212759256362915, + "num_tokens": 426563419.0, + "step": 16498 + }, + { + "epoch": 1.8118822754227981, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2061545848846436, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7310831546783447, + "num_tokens": 426589380.0, + "step": 16499 + }, + { + "epoch": 1.8119920931254119, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4176552295684814, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.726077139377594, + "num_tokens": 426613698.0, + "step": 16500 + }, + { + "epoch": 1.8121019108280256, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.8124107122421265, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6924145221710205, + "num_tokens": 426652593.0, + "step": 16501 + }, + { + "epoch": 1.8122117285306392, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0672607421875, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7050448656082153, + "num_tokens": 426682565.0, + "step": 16502 + }, + { + "epoch": 1.8123215462332527, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.8056280612945557, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7143739461898804, + "num_tokens": 426718171.0, + "step": 16503 + }, + { + "epoch": 1.8124313639358665, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.8859436511993408, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7286351919174194, + "num_tokens": 426751561.0, + "step": 16504 + }, + { + "epoch": 1.8125411816384802, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.272853136062622, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7039118409156799, + "num_tokens": 426776114.0, + "step": 16505 + }, + { + "epoch": 1.8126509993410937, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1055409908294678, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7102614641189575, + "num_tokens": 426805410.0, + "step": 16506 + }, + { + "epoch": 1.8127608170437073, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2273426055908203, + "learning_rate": 1e-06, + "loss": 0.9827, + "mean_token_accuracy": 0.702211320400238, + "num_tokens": 426831192.0, + "step": 16507 + }, + { + "epoch": 1.812870634746321, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.174011707305908, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.7045447826385498, + "num_tokens": 426858715.0, + "step": 16508 + }, + { + "epoch": 1.8129804524489348, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1268746852874756, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.70057612657547, + "num_tokens": 426888925.0, + "step": 16509 + }, + { + "epoch": 1.8130902701515486, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.358123779296875, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7257487177848816, + "num_tokens": 426911423.0, + "step": 16510 + }, + { + "epoch": 1.813200087854162, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.449376344680786, + "learning_rate": 1e-06, + "loss": 0.9706, + "mean_token_accuracy": 0.7045217752456665, + "num_tokens": 426935761.0, + "step": 16511 + }, + { + "epoch": 1.8133099055567756, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1868820190429688, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7280688881874084, + "num_tokens": 426963565.0, + "step": 16512 + }, + { + "epoch": 1.8134197232593894, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.21687912940979, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7367867231369019, + "num_tokens": 426988193.0, + "step": 16513 + }, + { + "epoch": 1.8135295409620031, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.867148756980896, + "learning_rate": 1e-06, + "loss": 1.0465, + "mean_token_accuracy": 0.6874978542327881, + "num_tokens": 427022948.0, + "step": 16514 + }, + { + "epoch": 1.813639358664617, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.078340530395508, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7198824882507324, + "num_tokens": 427052540.0, + "step": 16515 + }, + { + "epoch": 1.8137491763672304, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.333921194076538, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7201365232467651, + "num_tokens": 427075944.0, + "step": 16516 + }, + { + "epoch": 1.813858994069844, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.8906564712524414, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7251696586608887, + "num_tokens": 427108854.0, + "step": 16517 + }, + { + "epoch": 1.8139688117724577, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.404406785964966, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7186536192893982, + "num_tokens": 427131274.0, + "step": 16518 + }, + { + "epoch": 1.8140786294750715, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.401700735092163, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7156575918197632, + "num_tokens": 427155856.0, + "step": 16519 + }, + { + "epoch": 1.814188447177685, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.124612808227539, + "learning_rate": 1e-06, + "loss": 1.0438, + "mean_token_accuracy": 0.6846872568130493, + "num_tokens": 427185417.0, + "step": 16520 + }, + { + "epoch": 1.8142982648802986, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1590726375579834, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6951425671577454, + "num_tokens": 427212866.0, + "step": 16521 + }, + { + "epoch": 1.8144080825829123, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1687676906585693, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7103082537651062, + "num_tokens": 427238942.0, + "step": 16522 + }, + { + "epoch": 1.814517900285526, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.188786745071411, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7226020097732544, + "num_tokens": 427264887.0, + "step": 16523 + }, + { + "epoch": 1.8146277179881398, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0537219047546387, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7341653108596802, + "num_tokens": 427292572.0, + "step": 16524 + }, + { + "epoch": 1.8147375356907534, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.3539257049560547, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7100629806518555, + "num_tokens": 427316787.0, + "step": 16525 + }, + { + "epoch": 1.814847353393367, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1247949600219727, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6905332803726196, + "num_tokens": 427348035.0, + "step": 16526 + }, + { + "epoch": 1.8149571710959806, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.294235944747925, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.72086101770401, + "num_tokens": 427373501.0, + "step": 16527 + }, + { + "epoch": 1.8150669887985944, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.6029446125030518, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7333551645278931, + "num_tokens": 427392578.0, + "step": 16528 + }, + { + "epoch": 1.815176806501208, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1188337802886963, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7213443517684937, + "num_tokens": 427420076.0, + "step": 16529 + }, + { + "epoch": 1.8152866242038217, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.262188196182251, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.705511748790741, + "num_tokens": 427447000.0, + "step": 16530 + }, + { + "epoch": 1.8153964419064352, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5378036499023438, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7146098613739014, + "num_tokens": 427467724.0, + "step": 16531 + }, + { + "epoch": 1.815506259609049, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0016677379608154, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7298635244369507, + "num_tokens": 427498695.0, + "step": 16532 + }, + { + "epoch": 1.8156160773116627, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2648391723632812, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7220882177352905, + "num_tokens": 427524312.0, + "step": 16533 + }, + { + "epoch": 1.8157258950142763, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.271860122680664, + "learning_rate": 1e-06, + "loss": 0.9916, + "mean_token_accuracy": 0.6936013698577881, + "num_tokens": 427552420.0, + "step": 16534 + }, + { + "epoch": 1.8158357127168898, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.297498941421509, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7241056561470032, + "num_tokens": 427577265.0, + "step": 16535 + }, + { + "epoch": 1.8159455304195036, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0225629806518555, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7300025820732117, + "num_tokens": 427607241.0, + "step": 16536 + }, + { + "epoch": 1.8160553481221173, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.6258316040039062, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7280448079109192, + "num_tokens": 427626819.0, + "step": 16537 + }, + { + "epoch": 1.816165165824731, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0384774208068848, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7292413711547852, + "num_tokens": 427656171.0, + "step": 16538 + }, + { + "epoch": 1.8162749835273446, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1214523315429688, + "learning_rate": 1e-06, + "loss": 0.9511, + "mean_token_accuracy": 0.6989362835884094, + "num_tokens": 427683797.0, + "step": 16539 + }, + { + "epoch": 1.8163848012299582, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0978872776031494, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7409751415252686, + "num_tokens": 427710480.0, + "step": 16540 + }, + { + "epoch": 1.816494618932572, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1517841815948486, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7236080169677734, + "num_tokens": 427736371.0, + "step": 16541 + }, + { + "epoch": 1.8166044366351857, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4119322299957275, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7083215713500977, + "num_tokens": 427758803.0, + "step": 16542 + }, + { + "epoch": 1.8167142543377992, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.520637273788452, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7331050634384155, + "num_tokens": 427779095.0, + "step": 16543 + }, + { + "epoch": 1.816824072040413, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5361244678497314, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7084338068962097, + "num_tokens": 427799114.0, + "step": 16544 + }, + { + "epoch": 1.8169338897430265, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1960854530334473, + "learning_rate": 1e-06, + "loss": 0.9866, + "mean_token_accuracy": 0.6959834694862366, + "num_tokens": 427826069.0, + "step": 16545 + }, + { + "epoch": 1.8170437074456403, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.3269307613372803, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.72163325548172, + "num_tokens": 427849250.0, + "step": 16546 + }, + { + "epoch": 1.817153525148254, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.099574089050293, + "learning_rate": 1e-06, + "loss": 0.9392, + "mean_token_accuracy": 0.7065647840499878, + "num_tokens": 427880158.0, + "step": 16547 + }, + { + "epoch": 1.8172633428508675, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 1.9398871660232544, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7231022119522095, + "num_tokens": 427911055.0, + "step": 16548 + }, + { + "epoch": 1.817373160553481, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.003265857696533, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.6999465227127075, + "num_tokens": 427941798.0, + "step": 16549 + }, + { + "epoch": 1.8174829782560948, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2986183166503906, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7291870713233948, + "num_tokens": 427965018.0, + "step": 16550 + }, + { + "epoch": 1.8175927959587086, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4389476776123047, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7207036018371582, + "num_tokens": 427986831.0, + "step": 16551 + }, + { + "epoch": 1.8177026136613224, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.286468505859375, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7418196201324463, + "num_tokens": 428011835.0, + "step": 16552 + }, + { + "epoch": 1.8178124313639359, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.7373437881469727, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7285876870155334, + "num_tokens": 428030219.0, + "step": 16553 + }, + { + "epoch": 1.8179222490665494, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5621745586395264, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7250418663024902, + "num_tokens": 428050693.0, + "step": 16554 + }, + { + "epoch": 1.8180320667691632, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2264413833618164, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7273541688919067, + "num_tokens": 428075890.0, + "step": 16555 + }, + { + "epoch": 1.818141884471777, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.13653564453125, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7097278237342834, + "num_tokens": 428104492.0, + "step": 16556 + }, + { + "epoch": 1.8182517021743905, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4966375827789307, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7250620126724243, + "num_tokens": 428125142.0, + "step": 16557 + }, + { + "epoch": 1.818361519877004, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2959179878234863, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.7170653343200684, + "num_tokens": 428149052.0, + "step": 16558 + }, + { + "epoch": 1.8184713375796178, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.075836420059204, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.747179388999939, + "num_tokens": 428175138.0, + "step": 16559 + }, + { + "epoch": 1.8185811552822315, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1942086219787598, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7208311557769775, + "num_tokens": 428201616.0, + "step": 16560 + }, + { + "epoch": 1.8186909729848453, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1064229011535645, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7277928590774536, + "num_tokens": 428229204.0, + "step": 16561 + }, + { + "epoch": 1.8188007906874588, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.310331344604492, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.719450831413269, + "num_tokens": 428252398.0, + "step": 16562 + }, + { + "epoch": 1.8189106083900723, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1334590911865234, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7353004813194275, + "num_tokens": 428280091.0, + "step": 16563 + }, + { + "epoch": 1.819020426092686, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.5616567134857178, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7437721490859985, + "num_tokens": 428303463.0, + "step": 16564 + }, + { + "epoch": 1.8191302437952999, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.347559690475464, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7223471999168396, + "num_tokens": 428326495.0, + "step": 16565 + }, + { + "epoch": 1.8192400614979136, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.0472252368927, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7000232934951782, + "num_tokens": 428356987.0, + "step": 16566 + }, + { + "epoch": 1.8193498792005272, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.3706154823303223, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7123778462409973, + "num_tokens": 428382039.0, + "step": 16567 + }, + { + "epoch": 1.8194596969031407, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.1466147899627686, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7230274677276611, + "num_tokens": 428407816.0, + "step": 16568 + }, + { + "epoch": 1.8195695146057544, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.153435468673706, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7188889980316162, + "num_tokens": 428434878.0, + "step": 16569 + }, + { + "epoch": 1.8196793323083682, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.135524272918701, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7139765024185181, + "num_tokens": 428461640.0, + "step": 16570 + }, + { + "epoch": 1.8197891500109817, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2853195667266846, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7156751155853271, + "num_tokens": 428486922.0, + "step": 16571 + }, + { + "epoch": 1.8198989677135953, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.4411978721618652, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7509337067604065, + "num_tokens": 428509178.0, + "step": 16572 + }, + { + "epoch": 1.820008785416209, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.7990219593048096, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7316175699234009, + "num_tokens": 428526688.0, + "step": 16573 + }, + { + "epoch": 1.8201186031188228, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.2759385108947754, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.707760214805603, + "num_tokens": 428552595.0, + "step": 16574 + }, + { + "epoch": 1.8202284208214365, + "ewc_loss": 9.715557098388672e-06, + "grad_norm": 2.176098585128784, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7315313220024109, + "num_tokens": 428580590.0, + "step": 16575 + }, + { + "epoch": 1.82033823852405, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.003396987915039, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7292916178703308, + "num_tokens": 428610668.0, + "step": 16576 + }, + { + "epoch": 1.8204480562266636, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.4795358180999756, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7327761650085449, + "num_tokens": 428630137.0, + "step": 16577 + }, + { + "epoch": 1.8205578739292774, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2045254707336426, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.734552264213562, + "num_tokens": 428656268.0, + "step": 16578 + }, + { + "epoch": 1.8206676916318911, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.123429536819458, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7268434762954712, + "num_tokens": 428684183.0, + "step": 16579 + }, + { + "epoch": 1.8207775093345049, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.240302562713623, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7386115193367004, + "num_tokens": 428708305.0, + "step": 16580 + }, + { + "epoch": 1.8208873270371184, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.267773389816284, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7295396327972412, + "num_tokens": 428732776.0, + "step": 16581 + }, + { + "epoch": 1.820997144739732, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 1.9648258686065674, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7119227051734924, + "num_tokens": 428764808.0, + "step": 16582 + }, + { + "epoch": 1.8211069624423457, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.5582427978515625, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7157509326934814, + "num_tokens": 428784878.0, + "step": 16583 + }, + { + "epoch": 1.8212167801449595, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 1.970876693725586, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7380266189575195, + "num_tokens": 428815897.0, + "step": 16584 + }, + { + "epoch": 1.821326597847573, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.463765859603882, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7339068651199341, + "num_tokens": 428837386.0, + "step": 16585 + }, + { + "epoch": 1.8214364155501865, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.071964979171753, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7294372320175171, + "num_tokens": 428863800.0, + "step": 16586 + }, + { + "epoch": 1.8215462332528003, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.434544563293457, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7509146332740784, + "num_tokens": 428884082.0, + "step": 16587 + }, + { + "epoch": 1.821656050955414, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.534857988357544, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7437904477119446, + "num_tokens": 428904607.0, + "step": 16588 + }, + { + "epoch": 1.8217658686580278, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.22591233253479, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7175171375274658, + "num_tokens": 428930390.0, + "step": 16589 + }, + { + "epoch": 1.8218756863606413, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.343677282333374, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7206513285636902, + "num_tokens": 428955621.0, + "step": 16590 + }, + { + "epoch": 1.8219855040632549, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0564799308776855, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7384340763092041, + "num_tokens": 428986531.0, + "step": 16591 + }, + { + "epoch": 1.8220953217658686, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1520938873291016, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7361700534820557, + "num_tokens": 429011547.0, + "step": 16592 + }, + { + "epoch": 1.8222051394684824, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.217353105545044, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7193871140480042, + "num_tokens": 429037503.0, + "step": 16593 + }, + { + "epoch": 1.822314957171096, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.3099863529205322, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7124826312065125, + "num_tokens": 429063438.0, + "step": 16594 + }, + { + "epoch": 1.8224247748737097, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2314844131469727, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7267253994941711, + "num_tokens": 429089960.0, + "step": 16595 + }, + { + "epoch": 1.8225345925763232, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.3879575729370117, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.72132807970047, + "num_tokens": 429112940.0, + "step": 16596 + }, + { + "epoch": 1.822644410278937, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1418297290802, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7407887578010559, + "num_tokens": 429139904.0, + "step": 16597 + }, + { + "epoch": 1.8227542279815507, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 1.9398374557495117, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7112201452255249, + "num_tokens": 429172666.0, + "step": 16598 + }, + { + "epoch": 1.8228640456841643, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.402926445007324, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7267789840698242, + "num_tokens": 429195114.0, + "step": 16599 + }, + { + "epoch": 1.8229738633867778, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.117396593093872, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7028913497924805, + "num_tokens": 429222275.0, + "step": 16600 + }, + { + "epoch": 1.8230836810893916, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.200504779815674, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7040113210678101, + "num_tokens": 429248468.0, + "step": 16601 + }, + { + "epoch": 1.8231934987920053, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1029367446899414, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7123150825500488, + "num_tokens": 429276517.0, + "step": 16602 + }, + { + "epoch": 1.823303316494619, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1206603050231934, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7098355889320374, + "num_tokens": 429305965.0, + "step": 16603 + }, + { + "epoch": 1.8234131341972326, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.3025290966033936, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7126578092575073, + "num_tokens": 429329134.0, + "step": 16604 + }, + { + "epoch": 1.8235229518998461, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.172226667404175, + "learning_rate": 1e-06, + "loss": 0.9483, + "mean_token_accuracy": 0.7102588415145874, + "num_tokens": 429358836.0, + "step": 16605 + }, + { + "epoch": 1.82363276960246, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.558366060256958, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7384135723114014, + "num_tokens": 429380611.0, + "step": 16606 + }, + { + "epoch": 1.8237425873050737, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.199208974838257, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.703536331653595, + "num_tokens": 429407180.0, + "step": 16607 + }, + { + "epoch": 1.8238524050076872, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0929906368255615, + "learning_rate": 1e-06, + "loss": 0.9943, + "mean_token_accuracy": 0.6957335472106934, + "num_tokens": 429435290.0, + "step": 16608 + }, + { + "epoch": 1.823962222710301, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.3515660762786865, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7211763858795166, + "num_tokens": 429460877.0, + "step": 16609 + }, + { + "epoch": 1.8240720404129145, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0931813716888428, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7255687713623047, + "num_tokens": 429490257.0, + "step": 16610 + }, + { + "epoch": 1.8241818581155282, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.126495599746704, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.7003440856933594, + "num_tokens": 429518656.0, + "step": 16611 + }, + { + "epoch": 1.824291675818142, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 1.9573229551315308, + "learning_rate": 1e-06, + "loss": 0.9911, + "mean_token_accuracy": 0.6932625770568848, + "num_tokens": 429551292.0, + "step": 16612 + }, + { + "epoch": 1.8244014935207555, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 1.9845808744430542, + "learning_rate": 1e-06, + "loss": 1.0113, + "mean_token_accuracy": 0.7004798650741577, + "num_tokens": 429581395.0, + "step": 16613 + }, + { + "epoch": 1.824511311223369, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0745651721954346, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7244953513145447, + "num_tokens": 429610545.0, + "step": 16614 + }, + { + "epoch": 1.8246211289259828, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0556631088256836, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7052174210548401, + "num_tokens": 429640094.0, + "step": 16615 + }, + { + "epoch": 1.8247309466285966, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2473978996276855, + "learning_rate": 1e-06, + "loss": 0.9621, + "mean_token_accuracy": 0.6947294473648071, + "num_tokens": 429664718.0, + "step": 16616 + }, + { + "epoch": 1.8248407643312103, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0165164470672607, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7247219085693359, + "num_tokens": 429693462.0, + "step": 16617 + }, + { + "epoch": 1.8249505820338239, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1267282962799072, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6937135457992554, + "num_tokens": 429722076.0, + "step": 16618 + }, + { + "epoch": 1.8250603997364374, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0586514472961426, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7048065066337585, + "num_tokens": 429751375.0, + "step": 16619 + }, + { + "epoch": 1.8251702174390512, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.431814193725586, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7222875356674194, + "num_tokens": 429773201.0, + "step": 16620 + }, + { + "epoch": 1.825280035141665, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0966897010803223, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7436426877975464, + "num_tokens": 429801196.0, + "step": 16621 + }, + { + "epoch": 1.8253898528442785, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.34342885017395, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7250642776489258, + "num_tokens": 429823930.0, + "step": 16622 + }, + { + "epoch": 1.825499670546892, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1462666988372803, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7152602672576904, + "num_tokens": 429852022.0, + "step": 16623 + }, + { + "epoch": 1.8256094882495058, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0722873210906982, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.698694109916687, + "num_tokens": 429881626.0, + "step": 16624 + }, + { + "epoch": 1.8257193059521195, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.066187858581543, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7187575101852417, + "num_tokens": 429910784.0, + "step": 16625 + }, + { + "epoch": 1.8258291236547333, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.0236124992370605, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7219892740249634, + "num_tokens": 429940358.0, + "step": 16626 + }, + { + "epoch": 1.8259389413573468, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.089641809463501, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.6988598108291626, + "num_tokens": 429969579.0, + "step": 16627 + }, + { + "epoch": 1.8260487590599603, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.1863913536071777, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7068319320678711, + "num_tokens": 429996881.0, + "step": 16628 + }, + { + "epoch": 1.826158576762574, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2421693801879883, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7265332937240601, + "num_tokens": 430022276.0, + "step": 16629 + }, + { + "epoch": 1.8262683944651878, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.2261531352996826, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7255931496620178, + "num_tokens": 430047514.0, + "step": 16630 + }, + { + "epoch": 1.8263782121678016, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.3663413524627686, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7201979756355286, + "num_tokens": 430068455.0, + "step": 16631 + }, + { + "epoch": 1.8264880298704151, + "ewc_loss": 9.775161743164062e-06, + "grad_norm": 2.382622480392456, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7191435694694519, + "num_tokens": 430091890.0, + "step": 16632 + }, + { + "epoch": 1.8265978475730287, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.012589931488037, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6911661028862, + "num_tokens": 430122910.0, + "step": 16633 + }, + { + "epoch": 1.8267076652756424, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.286487579345703, + "learning_rate": 1e-06, + "loss": 0.9573, + "mean_token_accuracy": 0.7103433012962341, + "num_tokens": 430148544.0, + "step": 16634 + }, + { + "epoch": 1.8268174829782562, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1181275844573975, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.696627140045166, + "num_tokens": 430177519.0, + "step": 16635 + }, + { + "epoch": 1.8269273006808697, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.334106683731079, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7179295420646667, + "num_tokens": 430202491.0, + "step": 16636 + }, + { + "epoch": 1.8270371183834833, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2426624298095703, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.725658655166626, + "num_tokens": 430228359.0, + "step": 16637 + }, + { + "epoch": 1.827146936086097, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.6619839668273926, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7243184447288513, + "num_tokens": 430246665.0, + "step": 16638 + }, + { + "epoch": 1.8272567537887108, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.263219118118286, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7257867455482483, + "num_tokens": 430271257.0, + "step": 16639 + }, + { + "epoch": 1.8273665714913245, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1709165573120117, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7454439401626587, + "num_tokens": 430295702.0, + "step": 16640 + }, + { + "epoch": 1.827476389193938, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.303445816040039, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7267951965332031, + "num_tokens": 430319395.0, + "step": 16641 + }, + { + "epoch": 1.8275862068965516, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2937190532684326, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7153608798980713, + "num_tokens": 430344876.0, + "step": 16642 + }, + { + "epoch": 1.8276960245991654, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.14835524559021, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.737027645111084, + "num_tokens": 430369863.0, + "step": 16643 + }, + { + "epoch": 1.8278058423017791, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3801896572113037, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7121307849884033, + "num_tokens": 430394954.0, + "step": 16644 + }, + { + "epoch": 1.8279156600043929, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9641048908233643, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.6980627775192261, + "num_tokens": 430426202.0, + "step": 16645 + }, + { + "epoch": 1.8280254777070064, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.250195026397705, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7088484764099121, + "num_tokens": 430452728.0, + "step": 16646 + }, + { + "epoch": 1.82813529540962, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.32340407371521, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7213354706764221, + "num_tokens": 430476884.0, + "step": 16647 + }, + { + "epoch": 1.8282451131122337, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2255375385284424, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.721045732498169, + "num_tokens": 430501811.0, + "step": 16648 + }, + { + "epoch": 1.8283549308148475, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1969282627105713, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7348718643188477, + "num_tokens": 430525781.0, + "step": 16649 + }, + { + "epoch": 1.828464748517461, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9874571561813354, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7273873090744019, + "num_tokens": 430556236.0, + "step": 16650 + }, + { + "epoch": 1.8285745662200745, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.6865451335906982, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7328842878341675, + "num_tokens": 430575557.0, + "step": 16651 + }, + { + "epoch": 1.8286843839226883, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.046123743057251, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7381799817085266, + "num_tokens": 430602998.0, + "step": 16652 + }, + { + "epoch": 1.828794201625302, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3009588718414307, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7228690385818481, + "num_tokens": 430626485.0, + "step": 16653 + }, + { + "epoch": 1.8289040193279158, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.146679162979126, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7229340076446533, + "num_tokens": 430651993.0, + "step": 16654 + }, + { + "epoch": 1.8290138370305293, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.136951446533203, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7094331979751587, + "num_tokens": 430677563.0, + "step": 16655 + }, + { + "epoch": 1.8291236547331429, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.078915596008301, + "learning_rate": 1e-06, + "loss": 0.9834, + "mean_token_accuracy": 0.697758674621582, + "num_tokens": 430705592.0, + "step": 16656 + }, + { + "epoch": 1.8292334724357566, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.7941551208496094, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.7201496958732605, + "num_tokens": 430725170.0, + "step": 16657 + }, + { + "epoch": 1.8293432901383704, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.179398536682129, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7065149545669556, + "num_tokens": 430751860.0, + "step": 16658 + }, + { + "epoch": 1.829453107840984, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1424338817596436, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7074145674705505, + "num_tokens": 430781725.0, + "step": 16659 + }, + { + "epoch": 1.8295629255435977, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.150904893875122, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7394891977310181, + "num_tokens": 430807699.0, + "step": 16660 + }, + { + "epoch": 1.8296727432462112, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.283644914627075, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7281550168991089, + "num_tokens": 430832221.0, + "step": 16661 + }, + { + "epoch": 1.829782560948825, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.284553050994873, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7225426435470581, + "num_tokens": 430856288.0, + "step": 16662 + }, + { + "epoch": 1.8298923786514387, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.370173454284668, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7177208662033081, + "num_tokens": 430878692.0, + "step": 16663 + }, + { + "epoch": 1.8300021963540523, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2074267864227295, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7336499691009521, + "num_tokens": 430902390.0, + "step": 16664 + }, + { + "epoch": 1.8301120140566658, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.216843366622925, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7161386609077454, + "num_tokens": 430927656.0, + "step": 16665 + }, + { + "epoch": 1.8302218317592795, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.204802989959717, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7216371297836304, + "num_tokens": 430955307.0, + "step": 16666 + }, + { + "epoch": 1.8303316494618933, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.090583324432373, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7185088396072388, + "num_tokens": 430984328.0, + "step": 16667 + }, + { + "epoch": 1.830441467164507, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.876081943511963, + "learning_rate": 1e-06, + "loss": 0.9517, + "mean_token_accuracy": 0.7068631052970886, + "num_tokens": 431019994.0, + "step": 16668 + }, + { + "epoch": 1.8305512848671206, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.340548038482666, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7369304895401001, + "num_tokens": 431042177.0, + "step": 16669 + }, + { + "epoch": 1.8306611025697341, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1182565689086914, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.717025876045227, + "num_tokens": 431068939.0, + "step": 16670 + }, + { + "epoch": 1.830770920272348, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4469265937805176, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7075245380401611, + "num_tokens": 431089876.0, + "step": 16671 + }, + { + "epoch": 1.8308807379749616, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2829654216766357, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.6979166269302368, + "num_tokens": 431114712.0, + "step": 16672 + }, + { + "epoch": 1.8309905556775752, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.353513240814209, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7328435182571411, + "num_tokens": 431136945.0, + "step": 16673 + }, + { + "epoch": 1.831100373380189, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.473487377166748, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.719186007976532, + "num_tokens": 431159197.0, + "step": 16674 + }, + { + "epoch": 1.8312101910828025, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3254737854003906, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7114089131355286, + "num_tokens": 431183436.0, + "step": 16675 + }, + { + "epoch": 1.8313200087854162, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.229856491088867, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.716923713684082, + "num_tokens": 431208000.0, + "step": 16676 + }, + { + "epoch": 1.83142982648803, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4602010250091553, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7166913747787476, + "num_tokens": 431229066.0, + "step": 16677 + }, + { + "epoch": 1.8315396441906435, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.442547559738159, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7322710752487183, + "num_tokens": 431253228.0, + "step": 16678 + }, + { + "epoch": 1.831649461893257, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.280978202819824, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.715480625629425, + "num_tokens": 431277606.0, + "step": 16679 + }, + { + "epoch": 1.8317592795958708, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.0401792526245117, + "learning_rate": 1e-06, + "loss": 0.9701, + "mean_token_accuracy": 0.6968618035316467, + "num_tokens": 431306796.0, + "step": 16680 + }, + { + "epoch": 1.8318690972984846, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1894304752349854, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7128353118896484, + "num_tokens": 431332376.0, + "step": 16681 + }, + { + "epoch": 1.8319789150010983, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9497877359390259, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7233563661575317, + "num_tokens": 431361858.0, + "step": 16682 + }, + { + "epoch": 1.8320887327037119, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.358640193939209, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7221790552139282, + "num_tokens": 431385909.0, + "step": 16683 + }, + { + "epoch": 1.8321985504063254, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2031922340393066, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.6999934911727905, + "num_tokens": 431411587.0, + "step": 16684 + }, + { + "epoch": 1.8323083681089392, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.258659839630127, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7171642184257507, + "num_tokens": 431438484.0, + "step": 16685 + }, + { + "epoch": 1.832418185811553, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2913312911987305, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7029078006744385, + "num_tokens": 431463481.0, + "step": 16686 + }, + { + "epoch": 1.8325280035141664, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1031343936920166, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7213327884674072, + "num_tokens": 431491588.0, + "step": 16687 + }, + { + "epoch": 1.83263782121678, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2698209285736084, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7464487552642822, + "num_tokens": 431517339.0, + "step": 16688 + }, + { + "epoch": 1.8327476389193937, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.169933557510376, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7493636608123779, + "num_tokens": 431543952.0, + "step": 16689 + }, + { + "epoch": 1.8328574566220075, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4350624084472656, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7148293852806091, + "num_tokens": 431566628.0, + "step": 16690 + }, + { + "epoch": 1.8329672743246213, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2029943466186523, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7098667025566101, + "num_tokens": 431595692.0, + "step": 16691 + }, + { + "epoch": 1.8330770920272348, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4637303352355957, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7140002250671387, + "num_tokens": 431617405.0, + "step": 16692 + }, + { + "epoch": 1.8331869097298483, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.5457143783569336, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7241904735565186, + "num_tokens": 431639808.0, + "step": 16693 + }, + { + "epoch": 1.833296727432462, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2381508350372314, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7260413765907288, + "num_tokens": 431665541.0, + "step": 16694 + }, + { + "epoch": 1.8334065451350758, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.439408302307129, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7454420328140259, + "num_tokens": 431686833.0, + "step": 16695 + }, + { + "epoch": 1.8335163628376896, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.509463310241699, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.715910792350769, + "num_tokens": 431707512.0, + "step": 16696 + }, + { + "epoch": 1.8336261805403031, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.192514181137085, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7232508659362793, + "num_tokens": 431734053.0, + "step": 16697 + }, + { + "epoch": 1.8337359982429167, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1874070167541504, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7107864618301392, + "num_tokens": 431762374.0, + "step": 16698 + }, + { + "epoch": 1.8338458159455304, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.181544542312622, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7100726962089539, + "num_tokens": 431791432.0, + "step": 16699 + }, + { + "epoch": 1.8339556336481442, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.199880838394165, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7201900482177734, + "num_tokens": 431815212.0, + "step": 16700 + }, + { + "epoch": 1.8340654513507577, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.161853551864624, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.740045428276062, + "num_tokens": 431840246.0, + "step": 16701 + }, + { + "epoch": 1.8341752690533712, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.234121561050415, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7128134369850159, + "num_tokens": 431865239.0, + "step": 16702 + }, + { + "epoch": 1.834285086755985, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.28698468208313, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7148927450180054, + "num_tokens": 431889686.0, + "step": 16703 + }, + { + "epoch": 1.8343949044585988, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.0505948066711426, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7129005193710327, + "num_tokens": 431916534.0, + "step": 16704 + }, + { + "epoch": 1.8345047221612125, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.40702223777771, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7167508602142334, + "num_tokens": 431939591.0, + "step": 16705 + }, + { + "epoch": 1.834614539863826, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2539408206939697, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7365750074386597, + "num_tokens": 431962665.0, + "step": 16706 + }, + { + "epoch": 1.8347243575664396, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2941513061523438, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7255752086639404, + "num_tokens": 431986591.0, + "step": 16707 + }, + { + "epoch": 1.8348341752690533, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.200650930404663, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7328232526779175, + "num_tokens": 432012179.0, + "step": 16708 + }, + { + "epoch": 1.834943992971667, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.35649037361145, + "learning_rate": 1e-06, + "loss": 0.8923, + "mean_token_accuracy": 0.7231547832489014, + "num_tokens": 432035559.0, + "step": 16709 + }, + { + "epoch": 1.8350538106742806, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.359376907348633, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7340998649597168, + "num_tokens": 432058218.0, + "step": 16710 + }, + { + "epoch": 1.8351636283768944, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1972923278808594, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7288472652435303, + "num_tokens": 432087561.0, + "step": 16711 + }, + { + "epoch": 1.835273446079508, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.062832832336426, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7336766719818115, + "num_tokens": 432115675.0, + "step": 16712 + }, + { + "epoch": 1.8353832637821217, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1490769386291504, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.7057701945304871, + "num_tokens": 432143933.0, + "step": 16713 + }, + { + "epoch": 1.8354930814847354, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9450476169586182, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7055714130401611, + "num_tokens": 432174272.0, + "step": 16714 + }, + { + "epoch": 1.835602899187349, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.467897891998291, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7045189738273621, + "num_tokens": 432196508.0, + "step": 16715 + }, + { + "epoch": 1.8357127168899625, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.450108766555786, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.716267466545105, + "num_tokens": 432219663.0, + "step": 16716 + }, + { + "epoch": 1.8358225345925763, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.042231798171997, + "learning_rate": 1e-06, + "loss": 1.0702, + "mean_token_accuracy": 0.6815162897109985, + "num_tokens": 432251645.0, + "step": 16717 + }, + { + "epoch": 1.83593235229519, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.005164384841919, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7175175547599792, + "num_tokens": 432279998.0, + "step": 16718 + }, + { + "epoch": 1.8360421699978038, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1436963081359863, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.7098382711410522, + "num_tokens": 432309369.0, + "step": 16719 + }, + { + "epoch": 1.8361519877004173, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.141387462615967, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7395106554031372, + "num_tokens": 432335263.0, + "step": 16720 + }, + { + "epoch": 1.8362618054030309, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3133015632629395, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7151710987091064, + "num_tokens": 432362386.0, + "step": 16721 + }, + { + "epoch": 1.8363716231056446, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.306962251663208, + "learning_rate": 1e-06, + "loss": 1.0274, + "mean_token_accuracy": 0.6946272850036621, + "num_tokens": 432388428.0, + "step": 16722 + }, + { + "epoch": 1.8364814408082584, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1448118686676025, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7059774398803711, + "num_tokens": 432415030.0, + "step": 16723 + }, + { + "epoch": 1.836591258510872, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2342591285705566, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7212097644805908, + "num_tokens": 432438623.0, + "step": 16724 + }, + { + "epoch": 1.8367010762134857, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.5475826263427734, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7378650903701782, + "num_tokens": 432458752.0, + "step": 16725 + }, + { + "epoch": 1.8368108939160992, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2902894020080566, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7134599685668945, + "num_tokens": 432482944.0, + "step": 16726 + }, + { + "epoch": 1.836920711618713, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1001429557800293, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7167286276817322, + "num_tokens": 432510600.0, + "step": 16727 + }, + { + "epoch": 1.8370305293213267, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.424604654312134, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7438050508499146, + "num_tokens": 432532220.0, + "step": 16728 + }, + { + "epoch": 1.8371403470239402, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3555808067321777, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7289605736732483, + "num_tokens": 432556467.0, + "step": 16729 + }, + { + "epoch": 1.8372501647265538, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9961600303649902, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7079538106918335, + "num_tokens": 432588323.0, + "step": 16730 + }, + { + "epoch": 1.8373599824291675, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.361490488052368, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7070590257644653, + "num_tokens": 432610998.0, + "step": 16731 + }, + { + "epoch": 1.8374698001317813, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1776905059814453, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7097699642181396, + "num_tokens": 432639635.0, + "step": 16732 + }, + { + "epoch": 1.837579617834395, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2497687339782715, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7397726774215698, + "num_tokens": 432665412.0, + "step": 16733 + }, + { + "epoch": 1.8376894355370086, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.637887477874756, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7319204807281494, + "num_tokens": 432684610.0, + "step": 16734 + }, + { + "epoch": 1.8377992532396221, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.435499906539917, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.716720461845398, + "num_tokens": 432707798.0, + "step": 16735 + }, + { + "epoch": 1.8379090709422359, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3955695629119873, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7296028137207031, + "num_tokens": 432731428.0, + "step": 16736 + }, + { + "epoch": 1.8380188886448496, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4191455841064453, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7139561176300049, + "num_tokens": 432755040.0, + "step": 16737 + }, + { + "epoch": 1.8381287063474632, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.524425506591797, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7458865642547607, + "num_tokens": 432775161.0, + "step": 16738 + }, + { + "epoch": 1.838238524050077, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.437246322631836, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7267687320709229, + "num_tokens": 432797686.0, + "step": 16739 + }, + { + "epoch": 1.8383483417526905, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.9427881240844727, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7027766704559326, + "num_tokens": 432831722.0, + "step": 16740 + }, + { + "epoch": 1.8384581594553042, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.190918445587158, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7228624820709229, + "num_tokens": 432857199.0, + "step": 16741 + }, + { + "epoch": 1.838567977157918, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2836596965789795, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7078608274459839, + "num_tokens": 432884640.0, + "step": 16742 + }, + { + "epoch": 1.8386777948605315, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.5550220012664795, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7184373140335083, + "num_tokens": 432906955.0, + "step": 16743 + }, + { + "epoch": 1.838787612563145, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.0301930904388428, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7236002683639526, + "num_tokens": 432936136.0, + "step": 16744 + }, + { + "epoch": 1.8388974302657588, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.13312029838562, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7164065837860107, + "num_tokens": 432962710.0, + "step": 16745 + }, + { + "epoch": 1.8390072479683726, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.365497350692749, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7245466709136963, + "num_tokens": 432985909.0, + "step": 16746 + }, + { + "epoch": 1.8391170656709863, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4313018321990967, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.732559323310852, + "num_tokens": 433005395.0, + "step": 16747 + }, + { + "epoch": 1.8392268833735999, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.181270122528076, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7379306554794312, + "num_tokens": 433031676.0, + "step": 16748 + }, + { + "epoch": 1.8393367010762134, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2140538692474365, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7211127281188965, + "num_tokens": 433058830.0, + "step": 16749 + }, + { + "epoch": 1.8394465187788271, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.3287484645843506, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7295656800270081, + "num_tokens": 433082843.0, + "step": 16750 + }, + { + "epoch": 1.839556336481441, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.5263473987579346, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7284484505653381, + "num_tokens": 433102919.0, + "step": 16751 + }, + { + "epoch": 1.8396661541840544, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.7732093334198, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7401218414306641, + "num_tokens": 433120615.0, + "step": 16752 + }, + { + "epoch": 1.839775971886668, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 3.0075676441192627, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7169897556304932, + "num_tokens": 433144967.0, + "step": 16753 + }, + { + "epoch": 1.8398857895892817, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.4740352630615234, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7165195941925049, + "num_tokens": 433167433.0, + "step": 16754 + }, + { + "epoch": 1.8399956072918955, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1628854274749756, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7151680588722229, + "num_tokens": 433194566.0, + "step": 16755 + }, + { + "epoch": 1.8401054249945092, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.049299478530884, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7280402779579163, + "num_tokens": 433222826.0, + "step": 16756 + }, + { + "epoch": 1.8402152426971228, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2765953540802, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7213762998580933, + "num_tokens": 433246856.0, + "step": 16757 + }, + { + "epoch": 1.8403250603997363, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.224073648452759, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7326567769050598, + "num_tokens": 433270364.0, + "step": 16758 + }, + { + "epoch": 1.84043487810235, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.2452359199523926, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7045271992683411, + "num_tokens": 433298091.0, + "step": 16759 + }, + { + "epoch": 1.8405446958049638, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.0677947998046875, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7439039349555969, + "num_tokens": 433326190.0, + "step": 16760 + }, + { + "epoch": 1.8406545135075776, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 1.925741195678711, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.722133994102478, + "num_tokens": 433356604.0, + "step": 16761 + }, + { + "epoch": 1.8407643312101911, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.179657220840454, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.711969792842865, + "num_tokens": 433383447.0, + "step": 16762 + }, + { + "epoch": 1.8408741489128047, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1748058795928955, + "learning_rate": 1e-06, + "loss": 0.9964, + "mean_token_accuracy": 0.6954438090324402, + "num_tokens": 433411016.0, + "step": 16763 + }, + { + "epoch": 1.8409839666154184, + "ewc_loss": 9.834766387939453e-06, + "grad_norm": 2.1972577571868896, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7109798192977905, + "num_tokens": 433437278.0, + "step": 16764 + }, + { + "epoch": 1.8410937843180322, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.426384449005127, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7339308261871338, + "num_tokens": 433458780.0, + "step": 16765 + }, + { + "epoch": 1.8412036020206457, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.1632933616638184, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7057748436927795, + "num_tokens": 433483918.0, + "step": 16766 + }, + { + "epoch": 1.8413134197232592, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.4408857822418213, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7245240807533264, + "num_tokens": 433505207.0, + "step": 16767 + }, + { + "epoch": 1.841423237425873, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.4018101692199707, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7458155155181885, + "num_tokens": 433527139.0, + "step": 16768 + }, + { + "epoch": 1.8415330551284868, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.2659034729003906, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7174229025840759, + "num_tokens": 433552161.0, + "step": 16769 + }, + { + "epoch": 1.8416428728311005, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.302025079727173, + "learning_rate": 1e-06, + "loss": 1.001, + "mean_token_accuracy": 0.692933976650238, + "num_tokens": 433578670.0, + "step": 16770 + }, + { + "epoch": 1.841752690533714, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.4030463695526123, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7159741520881653, + "num_tokens": 433599905.0, + "step": 16771 + }, + { + "epoch": 1.8418625082363276, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.168959617614746, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7043979167938232, + "num_tokens": 433628044.0, + "step": 16772 + }, + { + "epoch": 1.8419723259389413, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.2039437294006348, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7186925411224365, + "num_tokens": 433654532.0, + "step": 16773 + }, + { + "epoch": 1.842082143641555, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 1.9681010246276855, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7255599498748779, + "num_tokens": 433685377.0, + "step": 16774 + }, + { + "epoch": 1.8421919613441686, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.292731761932373, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7177273035049438, + "num_tokens": 433708832.0, + "step": 16775 + }, + { + "epoch": 1.8423017790467824, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.156642436981201, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7479989528656006, + "num_tokens": 433733598.0, + "step": 16776 + }, + { + "epoch": 1.842411596749396, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.2566661834716797, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.740818440914154, + "num_tokens": 433757291.0, + "step": 16777 + }, + { + "epoch": 1.8425214144520097, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.2098584175109863, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7311086654663086, + "num_tokens": 433781578.0, + "step": 16778 + }, + { + "epoch": 1.8426312321546234, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 2.1987531185150146, + "learning_rate": 1e-06, + "loss": 0.9547, + "mean_token_accuracy": 0.7063862085342407, + "num_tokens": 433807409.0, + "step": 16779 + }, + { + "epoch": 1.842741049857237, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.981921911239624, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7212074398994446, + "num_tokens": 433836391.0, + "step": 16780 + }, + { + "epoch": 1.8428508675598505, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 1.9557913541793823, + "learning_rate": 1e-06, + "loss": 1.0102, + "mean_token_accuracy": 0.6950017213821411, + "num_tokens": 433869430.0, + "step": 16781 + }, + { + "epoch": 1.8429606852624643, + "ewc_loss": 9.894371032714844e-06, + "grad_norm": 1.9942437410354614, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7211078405380249, + "num_tokens": 433899256.0, + "step": 16782 + }, + { + "epoch": 1.843070502965078, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3282124996185303, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7152305841445923, + "num_tokens": 433922331.0, + "step": 16783 + }, + { + "epoch": 1.8431803206676918, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.9821919202804565, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7190003395080566, + "num_tokens": 433952389.0, + "step": 16784 + }, + { + "epoch": 1.8432901383703053, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5809218883514404, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.732428789138794, + "num_tokens": 433972878.0, + "step": 16785 + }, + { + "epoch": 1.8433999560729188, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.378551721572876, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7291619181632996, + "num_tokens": 433996169.0, + "step": 16786 + }, + { + "epoch": 1.8435097737755326, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1077823638916016, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7430539131164551, + "num_tokens": 434023402.0, + "step": 16787 + }, + { + "epoch": 1.8436195914781464, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2427823543548584, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7252611517906189, + "num_tokens": 434048443.0, + "step": 16788 + }, + { + "epoch": 1.84372940918076, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2242512702941895, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7331507205963135, + "num_tokens": 434074861.0, + "step": 16789 + }, + { + "epoch": 1.8438392268833736, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.136486768722534, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7085561752319336, + "num_tokens": 434105980.0, + "step": 16790 + }, + { + "epoch": 1.8439490445859872, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.321143865585327, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7201731204986572, + "num_tokens": 434129951.0, + "step": 16791 + }, + { + "epoch": 1.844058862288601, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4768271446228027, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7398480176925659, + "num_tokens": 434151178.0, + "step": 16792 + }, + { + "epoch": 1.8441686799912147, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1311142444610596, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.732086181640625, + "num_tokens": 434179016.0, + "step": 16793 + }, + { + "epoch": 1.8442784976938282, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.6605639457702637, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7140575647354126, + "num_tokens": 434198588.0, + "step": 16794 + }, + { + "epoch": 1.8443883153964418, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.324472427368164, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7398065328598022, + "num_tokens": 434220242.0, + "step": 16795 + }, + { + "epoch": 1.8444981330990555, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1772732734680176, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7154759168624878, + "num_tokens": 434246939.0, + "step": 16796 + }, + { + "epoch": 1.8446079508016693, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0608692169189453, + "learning_rate": 1e-06, + "loss": 0.9939, + "mean_token_accuracy": 0.6968296766281128, + "num_tokens": 434275946.0, + "step": 16797 + }, + { + "epoch": 1.844717768504283, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4032819271087646, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7119415998458862, + "num_tokens": 434299740.0, + "step": 16798 + }, + { + "epoch": 1.8448275862068966, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.9959886074066162, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7005681991577148, + "num_tokens": 434330720.0, + "step": 16799 + }, + { + "epoch": 1.84493740390951, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.098545789718628, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7185897827148438, + "num_tokens": 434359474.0, + "step": 16800 + }, + { + "epoch": 1.8450472216121239, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2401580810546875, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7312142252922058, + "num_tokens": 434383772.0, + "step": 16801 + }, + { + "epoch": 1.8451570393147376, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0305678844451904, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7202944159507751, + "num_tokens": 434412512.0, + "step": 16802 + }, + { + "epoch": 1.8452668570173512, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.317513942718506, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7201263904571533, + "num_tokens": 434435345.0, + "step": 16803 + }, + { + "epoch": 1.8453766747199647, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.319474697113037, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7146577835083008, + "num_tokens": 434459575.0, + "step": 16804 + }, + { + "epoch": 1.8454864924225785, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.18955397605896, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7109016180038452, + "num_tokens": 434487834.0, + "step": 16805 + }, + { + "epoch": 1.8455963101251922, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5857884883880615, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7235522270202637, + "num_tokens": 434508043.0, + "step": 16806 + }, + { + "epoch": 1.845706127827806, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.244213104248047, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7363495826721191, + "num_tokens": 434531909.0, + "step": 16807 + }, + { + "epoch": 1.8458159455304195, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.255225658416748, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7198203802108765, + "num_tokens": 434556812.0, + "step": 16808 + }, + { + "epoch": 1.845925763233033, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.096801280975342, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7289409041404724, + "num_tokens": 434585807.0, + "step": 16809 + }, + { + "epoch": 1.8460355809356468, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.9613757133483887, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7131552696228027, + "num_tokens": 434619430.0, + "step": 16810 + }, + { + "epoch": 1.8461453986382605, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.385942220687866, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7136216759681702, + "num_tokens": 434644649.0, + "step": 16811 + }, + { + "epoch": 1.8462552163408743, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0367705821990967, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.726718544960022, + "num_tokens": 434675536.0, + "step": 16812 + }, + { + "epoch": 1.8463650340434878, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1924405097961426, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7156776785850525, + "num_tokens": 434702111.0, + "step": 16813 + }, + { + "epoch": 1.8464748517461014, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.148327589035034, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7408246397972107, + "num_tokens": 434727742.0, + "step": 16814 + }, + { + "epoch": 1.8465846694487151, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.178705930709839, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7230296730995178, + "num_tokens": 434753236.0, + "step": 16815 + }, + { + "epoch": 1.846694487151329, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1457769870758057, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7104675769805908, + "num_tokens": 434780497.0, + "step": 16816 + }, + { + "epoch": 1.8468043048539424, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.482449531555176, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7415757775306702, + "num_tokens": 434801305.0, + "step": 16817 + }, + { + "epoch": 1.846914122556556, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.342787742614746, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7366307973861694, + "num_tokens": 434825040.0, + "step": 16818 + }, + { + "epoch": 1.8470239402591697, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.434547185897827, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7112833261489868, + "num_tokens": 434848252.0, + "step": 16819 + }, + { + "epoch": 1.8471337579617835, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3351941108703613, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7140125036239624, + "num_tokens": 434874096.0, + "step": 16820 + }, + { + "epoch": 1.8472435756643972, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.453835964202881, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7424507737159729, + "num_tokens": 434894708.0, + "step": 16821 + }, + { + "epoch": 1.8473533933670108, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.047816276550293, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7159075736999512, + "num_tokens": 434923205.0, + "step": 16822 + }, + { + "epoch": 1.8474632110696243, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5677073001861572, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7415038347244263, + "num_tokens": 434944794.0, + "step": 16823 + }, + { + "epoch": 1.847573028772238, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2218124866485596, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7153096795082092, + "num_tokens": 434969538.0, + "step": 16824 + }, + { + "epoch": 1.8476828464748518, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3894734382629395, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7184216380119324, + "num_tokens": 434991412.0, + "step": 16825 + }, + { + "epoch": 1.8477926641774656, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1450319290161133, + "learning_rate": 1e-06, + "loss": 0.9953, + "mean_token_accuracy": 0.6993287801742554, + "num_tokens": 435021016.0, + "step": 16826 + }, + { + "epoch": 1.847902481880079, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.363433599472046, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7090910077095032, + "num_tokens": 435044767.0, + "step": 16827 + }, + { + "epoch": 1.8480122995826926, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2696197032928467, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7252428531646729, + "num_tokens": 435068460.0, + "step": 16828 + }, + { + "epoch": 1.8481221172853064, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2141225337982178, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7237595319747925, + "num_tokens": 435092176.0, + "step": 16829 + }, + { + "epoch": 1.8482319349879202, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.050191879272461, + "learning_rate": 1e-06, + "loss": 0.9559, + "mean_token_accuracy": 0.7092219591140747, + "num_tokens": 435122503.0, + "step": 16830 + }, + { + "epoch": 1.8483417526905337, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.379302978515625, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7220205068588257, + "num_tokens": 435147829.0, + "step": 16831 + }, + { + "epoch": 1.8484515703931472, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3386504650115967, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7067579627037048, + "num_tokens": 435170486.0, + "step": 16832 + }, + { + "epoch": 1.848561388095761, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.169156551361084, + "learning_rate": 1e-06, + "loss": 1.0058, + "mean_token_accuracy": 0.7149794101715088, + "num_tokens": 435198505.0, + "step": 16833 + }, + { + "epoch": 1.8486712057983747, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5472710132598877, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7179626226425171, + "num_tokens": 435220991.0, + "step": 16834 + }, + { + "epoch": 1.8487810235009885, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3413679599761963, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.714232325553894, + "num_tokens": 435247749.0, + "step": 16835 + }, + { + "epoch": 1.848890841203602, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.689023017883301, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7160888314247131, + "num_tokens": 435266982.0, + "step": 16836 + }, + { + "epoch": 1.8490006589062156, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.613783836364746, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.742522120475769, + "num_tokens": 435286124.0, + "step": 16837 + }, + { + "epoch": 1.8491104766088293, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2654106616973877, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7440265417098999, + "num_tokens": 435308497.0, + "step": 16838 + }, + { + "epoch": 1.849220294311443, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.080650806427002, + "learning_rate": 1e-06, + "loss": 1.0186, + "mean_token_accuracy": 0.6848188638687134, + "num_tokens": 435337722.0, + "step": 16839 + }, + { + "epoch": 1.8493301120140566, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4522762298583984, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7381911873817444, + "num_tokens": 435358826.0, + "step": 16840 + }, + { + "epoch": 1.8494399297166704, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4126875400543213, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7055389881134033, + "num_tokens": 435383016.0, + "step": 16841 + }, + { + "epoch": 1.849549747419284, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1944708824157715, + "learning_rate": 1e-06, + "loss": 1.0042, + "mean_token_accuracy": 0.7020895481109619, + "num_tokens": 435411040.0, + "step": 16842 + }, + { + "epoch": 1.8496595651218977, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.227123498916626, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7319043874740601, + "num_tokens": 435437714.0, + "step": 16843 + }, + { + "epoch": 1.8497693828245114, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.223217487335205, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.723118245601654, + "num_tokens": 435463892.0, + "step": 16844 + }, + { + "epoch": 1.849879200527125, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.031648874282837, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7038582563400269, + "num_tokens": 435494700.0, + "step": 16845 + }, + { + "epoch": 1.8499890182297385, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.6736772060394287, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7307538986206055, + "num_tokens": 435514755.0, + "step": 16846 + }, + { + "epoch": 1.8500988359323522, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.602895498275757, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7475190162658691, + "num_tokens": 435534376.0, + "step": 16847 + }, + { + "epoch": 1.850208653634966, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4386465549468994, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7168580293655396, + "num_tokens": 435557910.0, + "step": 16848 + }, + { + "epoch": 1.8503184713375798, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.8581184148788452, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.7033188939094543, + "num_tokens": 435593600.0, + "step": 16849 + }, + { + "epoch": 1.8504282890401933, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.431597948074341, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.6990774869918823, + "num_tokens": 435615993.0, + "step": 16850 + }, + { + "epoch": 1.8505381067428068, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2078022956848145, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.717016339302063, + "num_tokens": 435640678.0, + "step": 16851 + }, + { + "epoch": 1.8506479244454206, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.246089458465576, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7074089646339417, + "num_tokens": 435666368.0, + "step": 16852 + }, + { + "epoch": 1.8507577421480343, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4765636920928955, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7306199073791504, + "num_tokens": 435686611.0, + "step": 16853 + }, + { + "epoch": 1.8508675598506479, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.126246690750122, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7170931100845337, + "num_tokens": 435715453.0, + "step": 16854 + }, + { + "epoch": 1.8509773775532616, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4228196144104004, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7020549774169922, + "num_tokens": 435739985.0, + "step": 16855 + }, + { + "epoch": 1.8510871952558752, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.137540817260742, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7121046781539917, + "num_tokens": 435767221.0, + "step": 16856 + }, + { + "epoch": 1.851197012958489, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4865996837615967, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7458809018135071, + "num_tokens": 435790554.0, + "step": 16857 + }, + { + "epoch": 1.8513068306611027, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1667749881744385, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7288663983345032, + "num_tokens": 435817734.0, + "step": 16858 + }, + { + "epoch": 1.8514166483637162, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.12835693359375, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7051641345024109, + "num_tokens": 435845409.0, + "step": 16859 + }, + { + "epoch": 1.8515264660663298, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.46820068359375, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.725025475025177, + "num_tokens": 435867146.0, + "step": 16860 + }, + { + "epoch": 1.8516362837689435, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.271268367767334, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.715015172958374, + "num_tokens": 435896313.0, + "step": 16861 + }, + { + "epoch": 1.8517461014715573, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.6304714679718018, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7129477262496948, + "num_tokens": 435916873.0, + "step": 16862 + }, + { + "epoch": 1.851855919174171, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2186577320098877, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7349432706832886, + "num_tokens": 435940352.0, + "step": 16863 + }, + { + "epoch": 1.8519657368767846, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.268049955368042, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7314246892929077, + "num_tokens": 435965409.0, + "step": 16864 + }, + { + "epoch": 1.852075554579398, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3470730781555176, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.737067461013794, + "num_tokens": 435989123.0, + "step": 16865 + }, + { + "epoch": 1.8521853722820119, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3821001052856445, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7377010583877563, + "num_tokens": 436010896.0, + "step": 16866 + }, + { + "epoch": 1.8522951899846256, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.214496374130249, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7103240489959717, + "num_tokens": 436037438.0, + "step": 16867 + }, + { + "epoch": 1.8524050076872391, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5838053226470947, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7126275300979614, + "num_tokens": 436056645.0, + "step": 16868 + }, + { + "epoch": 1.8525148253898527, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2054073810577393, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7026183605194092, + "num_tokens": 436084261.0, + "step": 16869 + }, + { + "epoch": 1.8526246430924664, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0718719959259033, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7102688550949097, + "num_tokens": 436115170.0, + "step": 16870 + }, + { + "epoch": 1.8527344607950802, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4137821197509766, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7243611812591553, + "num_tokens": 436138017.0, + "step": 16871 + }, + { + "epoch": 1.852844278497694, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2790539264678955, + "learning_rate": 1e-06, + "loss": 0.9681, + "mean_token_accuracy": 0.7091041803359985, + "num_tokens": 436164826.0, + "step": 16872 + }, + { + "epoch": 1.8529540962003075, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.4854965209960938, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7379541397094727, + "num_tokens": 436186950.0, + "step": 16873 + }, + { + "epoch": 1.853063913902921, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0547280311584473, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.700343132019043, + "num_tokens": 436218554.0, + "step": 16874 + }, + { + "epoch": 1.8531737316055348, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.228459358215332, + "learning_rate": 1e-06, + "loss": 1.0481, + "mean_token_accuracy": 0.6887063980102539, + "num_tokens": 436245732.0, + "step": 16875 + }, + { + "epoch": 1.8532835493081485, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.017446994781494, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7303236722946167, + "num_tokens": 436275907.0, + "step": 16876 + }, + { + "epoch": 1.8533933670107623, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3815927505493164, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.7518224120140076, + "num_tokens": 436298559.0, + "step": 16877 + }, + { + "epoch": 1.8535031847133758, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.082502841949463, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7091634273529053, + "num_tokens": 436328711.0, + "step": 16878 + }, + { + "epoch": 1.8536130024159894, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.405362606048584, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7230991125106812, + "num_tokens": 436351002.0, + "step": 16879 + }, + { + "epoch": 1.8537228201186031, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.223740339279175, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7095205187797546, + "num_tokens": 436376844.0, + "step": 16880 + }, + { + "epoch": 1.8538326378212169, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2055442333221436, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.6977838277816772, + "num_tokens": 436403455.0, + "step": 16881 + }, + { + "epoch": 1.8539424555238304, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.319740056991577, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7166610360145569, + "num_tokens": 436427524.0, + "step": 16882 + }, + { + "epoch": 1.854052273226444, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.145659923553467, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7297153472900391, + "num_tokens": 436454649.0, + "step": 16883 + }, + { + "epoch": 1.8541620909290577, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3415029048919678, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7354685068130493, + "num_tokens": 436477195.0, + "step": 16884 + }, + { + "epoch": 1.8542719086316715, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0915215015411377, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7197762727737427, + "num_tokens": 436505904.0, + "step": 16885 + }, + { + "epoch": 1.8543817263342852, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.317237615585327, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7190953493118286, + "num_tokens": 436531585.0, + "step": 16886 + }, + { + "epoch": 1.8544915440368988, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0163638591766357, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.7026199102401733, + "num_tokens": 436563798.0, + "step": 16887 + }, + { + "epoch": 1.8546013617395123, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.078791379928589, + "learning_rate": 1e-06, + "loss": 0.9467, + "mean_token_accuracy": 0.7186756134033203, + "num_tokens": 436592580.0, + "step": 16888 + }, + { + "epoch": 1.854711179442126, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.2548604011535645, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7343659400939941, + "num_tokens": 436615814.0, + "step": 16889 + }, + { + "epoch": 1.8548209971447398, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.150315761566162, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7255374789237976, + "num_tokens": 436642617.0, + "step": 16890 + }, + { + "epoch": 1.8549308148473536, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.3504443168640137, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7115428447723389, + "num_tokens": 436666971.0, + "step": 16891 + }, + { + "epoch": 1.855040632549967, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.112318992614746, + "learning_rate": 1e-06, + "loss": 0.9445, + "mean_token_accuracy": 0.7075623869895935, + "num_tokens": 436695065.0, + "step": 16892 + }, + { + "epoch": 1.8551504502525806, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0047976970672607, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7129427194595337, + "num_tokens": 436724658.0, + "step": 16893 + }, + { + "epoch": 1.8552602679551944, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1908130645751953, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7186128497123718, + "num_tokens": 436750097.0, + "step": 16894 + }, + { + "epoch": 1.8553700856578081, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.245067834854126, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7322320342063904, + "num_tokens": 436774023.0, + "step": 16895 + }, + { + "epoch": 1.8554799033604217, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1729588508605957, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7196213603019714, + "num_tokens": 436800885.0, + "step": 16896 + }, + { + "epoch": 1.8555897210630352, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.5878231525421143, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7129043340682983, + "num_tokens": 436824212.0, + "step": 16897 + }, + { + "epoch": 1.855699538765649, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.630194664001465, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.741055428981781, + "num_tokens": 436843050.0, + "step": 16898 + }, + { + "epoch": 1.8558093564682627, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 1.9374371767044067, + "learning_rate": 1e-06, + "loss": 0.9984, + "mean_token_accuracy": 0.7011910676956177, + "num_tokens": 436880926.0, + "step": 16899 + }, + { + "epoch": 1.8559191741708765, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.1649208068847656, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7158029675483704, + "num_tokens": 436908418.0, + "step": 16900 + }, + { + "epoch": 1.85602899187349, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.252349376678467, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7407821416854858, + "num_tokens": 436933281.0, + "step": 16901 + }, + { + "epoch": 1.8561388095761036, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.0983307361602783, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7186033725738525, + "num_tokens": 436959755.0, + "step": 16902 + }, + { + "epoch": 1.8562486272787173, + "ewc_loss": 9.953975677490234e-06, + "grad_norm": 2.252084493637085, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7512723803520203, + "num_tokens": 436983183.0, + "step": 16903 + }, + { + "epoch": 1.856358444981331, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.630232095718384, + "learning_rate": 1e-06, + "loss": 0.7393, + "mean_token_accuracy": 0.7619285583496094, + "num_tokens": 437000635.0, + "step": 16904 + }, + { + "epoch": 1.8564682626839446, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.1750166416168213, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7140454053878784, + "num_tokens": 437025503.0, + "step": 16905 + }, + { + "epoch": 1.8565780803865584, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.279752731323242, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7234023213386536, + "num_tokens": 437051630.0, + "step": 16906 + }, + { + "epoch": 1.856687898089172, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.343168258666992, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7419617176055908, + "num_tokens": 437076770.0, + "step": 16907 + }, + { + "epoch": 1.8567977157917857, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.0130746364593506, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7273129224777222, + "num_tokens": 437108400.0, + "step": 16908 + }, + { + "epoch": 1.8569075334943994, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.0163965225219727, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7099502086639404, + "num_tokens": 437140555.0, + "step": 16909 + }, + { + "epoch": 1.857017351197013, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.4632270336151123, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7237879633903503, + "num_tokens": 437162231.0, + "step": 16910 + }, + { + "epoch": 1.8571271688996265, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.1328816413879395, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7134384512901306, + "num_tokens": 437188757.0, + "step": 16911 + }, + { + "epoch": 1.8572369866022402, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.0970797538757324, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7444393634796143, + "num_tokens": 437215715.0, + "step": 16912 + }, + { + "epoch": 1.857346804304854, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2498302459716797, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7221556305885315, + "num_tokens": 437240293.0, + "step": 16913 + }, + { + "epoch": 1.8574566220074678, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.030482292175293, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7270780205726624, + "num_tokens": 437271806.0, + "step": 16914 + }, + { + "epoch": 1.8575664397100813, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2781152725219727, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7364996671676636, + "num_tokens": 437295395.0, + "step": 16915 + }, + { + "epoch": 1.8576762574126948, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.54557204246521, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7191600203514099, + "num_tokens": 437316129.0, + "step": 16916 + }, + { + "epoch": 1.8577860751153086, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.21602463722229, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7026749849319458, + "num_tokens": 437342790.0, + "step": 16917 + }, + { + "epoch": 1.8578958928179223, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.108405828475952, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.7004788517951965, + "num_tokens": 437372755.0, + "step": 16918 + }, + { + "epoch": 1.8580057105205359, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 1.9053757190704346, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7174601554870605, + "num_tokens": 437406550.0, + "step": 16919 + }, + { + "epoch": 1.8581155282231496, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.3356378078460693, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7449891567230225, + "num_tokens": 437428676.0, + "step": 16920 + }, + { + "epoch": 1.8582253459257632, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2138638496398926, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.704633891582489, + "num_tokens": 437455018.0, + "step": 16921 + }, + { + "epoch": 1.858335163628377, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.3802809715270996, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.7050772309303284, + "num_tokens": 437479847.0, + "step": 16922 + }, + { + "epoch": 1.8584449813309907, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2192087173461914, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7079081535339355, + "num_tokens": 437507113.0, + "step": 16923 + }, + { + "epoch": 1.8585547990336042, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.4499382972717285, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7065098881721497, + "num_tokens": 437529316.0, + "step": 16924 + }, + { + "epoch": 1.8586646167362177, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.0259745121002197, + "learning_rate": 1e-06, + "loss": 0.9896, + "mean_token_accuracy": 0.691765546798706, + "num_tokens": 437560522.0, + "step": 16925 + }, + { + "epoch": 1.8587744344388315, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.017885446548462, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.6986489295959473, + "num_tokens": 437590821.0, + "step": 16926 + }, + { + "epoch": 1.8588842521414453, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2314822673797607, + "learning_rate": 1e-06, + "loss": 1.0348, + "mean_token_accuracy": 0.7027386426925659, + "num_tokens": 437619416.0, + "step": 16927 + }, + { + "epoch": 1.858994069844059, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.380575180053711, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7040758728981018, + "num_tokens": 437643616.0, + "step": 16928 + }, + { + "epoch": 1.8591038875466726, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.145373821258545, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.729274570941925, + "num_tokens": 437670853.0, + "step": 16929 + }, + { + "epoch": 1.859213705249286, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.4756667613983154, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7274371981620789, + "num_tokens": 437692695.0, + "step": 16930 + }, + { + "epoch": 1.8593235229518998, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.114248275756836, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7268158793449402, + "num_tokens": 437719371.0, + "step": 16931 + }, + { + "epoch": 1.8594333406545136, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2003087997436523, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7335690259933472, + "num_tokens": 437743999.0, + "step": 16932 + }, + { + "epoch": 1.8595431583571271, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.3221285343170166, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7116589546203613, + "num_tokens": 437766725.0, + "step": 16933 + }, + { + "epoch": 1.8596529760597407, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.2968647480010986, + "learning_rate": 1e-06, + "loss": 0.7792, + "mean_token_accuracy": 0.7469645142555237, + "num_tokens": 437790488.0, + "step": 16934 + }, + { + "epoch": 1.8597627937623544, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.3031277656555176, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7152523398399353, + "num_tokens": 437814868.0, + "step": 16935 + }, + { + "epoch": 1.8598726114649682, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.4787235260009766, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7297723889350891, + "num_tokens": 437834285.0, + "step": 16936 + }, + { + "epoch": 1.859982429167582, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.039525270462036, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.697317361831665, + "num_tokens": 437862652.0, + "step": 16937 + }, + { + "epoch": 1.8600922468701955, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.256831645965576, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7176880240440369, + "num_tokens": 437886820.0, + "step": 16938 + }, + { + "epoch": 1.860202064572809, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.003796339035034, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7297115325927734, + "num_tokens": 437919256.0, + "step": 16939 + }, + { + "epoch": 1.8603118822754228, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.141233205795288, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7025258541107178, + "num_tokens": 437947549.0, + "step": 16940 + }, + { + "epoch": 1.8604216999780365, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.254931688308716, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7075005769729614, + "num_tokens": 437972843.0, + "step": 16941 + }, + { + "epoch": 1.8605315176806503, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.301811695098877, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7304945588111877, + "num_tokens": 437998600.0, + "step": 16942 + }, + { + "epoch": 1.8606413353832638, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.0039217472076416, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7200241088867188, + "num_tokens": 438028102.0, + "step": 16943 + }, + { + "epoch": 1.8607511530858774, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.3191728591918945, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7054867744445801, + "num_tokens": 438052822.0, + "step": 16944 + }, + { + "epoch": 1.860860970788491, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.128709554672241, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7119631767272949, + "num_tokens": 438082844.0, + "step": 16945 + }, + { + "epoch": 1.8609707884911049, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 2.222994565963745, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7198519706726074, + "num_tokens": 438109890.0, + "step": 16946 + }, + { + "epoch": 1.8610806061937184, + "ewc_loss": 1.0013580322265625e-05, + "grad_norm": 1.9217609167099, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7173935174942017, + "num_tokens": 438140895.0, + "step": 16947 + }, + { + "epoch": 1.861190423896332, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2729570865631104, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7346702218055725, + "num_tokens": 438165806.0, + "step": 16948 + }, + { + "epoch": 1.8613002415989457, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3367204666137695, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7329238057136536, + "num_tokens": 438190317.0, + "step": 16949 + }, + { + "epoch": 1.8614100593015594, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.143052577972412, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7184447646141052, + "num_tokens": 438215540.0, + "step": 16950 + }, + { + "epoch": 1.8615198770041732, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.288691997528076, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7338584661483765, + "num_tokens": 438239349.0, + "step": 16951 + }, + { + "epoch": 1.8616296947067867, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.942531704902649, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.7029286623001099, + "num_tokens": 438272407.0, + "step": 16952 + }, + { + "epoch": 1.8617395124094003, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.5338120460510254, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7544891238212585, + "num_tokens": 438292060.0, + "step": 16953 + }, + { + "epoch": 1.861849330112014, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9808815717697144, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7043598890304565, + "num_tokens": 438323163.0, + "step": 16954 + }, + { + "epoch": 1.8619591478146278, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.214806079864502, + "learning_rate": 1e-06, + "loss": 1.0, + "mean_token_accuracy": 0.699062705039978, + "num_tokens": 438349393.0, + "step": 16955 + }, + { + "epoch": 1.8620689655172413, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2443652153015137, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7225669622421265, + "num_tokens": 438374889.0, + "step": 16956 + }, + { + "epoch": 1.862178783219855, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.114532470703125, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7061635255813599, + "num_tokens": 438404116.0, + "step": 16957 + }, + { + "epoch": 1.8622886009224686, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.265137195587158, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.717790961265564, + "num_tokens": 438428559.0, + "step": 16958 + }, + { + "epoch": 1.8623984186250824, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.408151149749756, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7073018550872803, + "num_tokens": 438452106.0, + "step": 16959 + }, + { + "epoch": 1.8625082363276961, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.278642177581787, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.715478241443634, + "num_tokens": 438476888.0, + "step": 16960 + }, + { + "epoch": 1.8626180540303097, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9606906175613403, + "learning_rate": 1e-06, + "loss": 0.9788, + "mean_token_accuracy": 0.6960024833679199, + "num_tokens": 438509799.0, + "step": 16961 + }, + { + "epoch": 1.8627278717329232, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2573485374450684, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7229772806167603, + "num_tokens": 438533819.0, + "step": 16962 + }, + { + "epoch": 1.862837689435537, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1853251457214355, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7086020708084106, + "num_tokens": 438559396.0, + "step": 16963 + }, + { + "epoch": 1.8629475071381507, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2796990871429443, + "learning_rate": 1e-06, + "loss": 0.7755, + "mean_token_accuracy": 0.7518607378005981, + "num_tokens": 438581493.0, + "step": 16964 + }, + { + "epoch": 1.8630573248407645, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3872954845428467, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.741701602935791, + "num_tokens": 438603969.0, + "step": 16965 + }, + { + "epoch": 1.863167142543378, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1808745861053467, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.734758734703064, + "num_tokens": 438629259.0, + "step": 16966 + }, + { + "epoch": 1.8632769602459915, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0254480838775635, + "learning_rate": 1e-06, + "loss": 1.0234, + "mean_token_accuracy": 0.6882800459861755, + "num_tokens": 438659759.0, + "step": 16967 + }, + { + "epoch": 1.8633867779486053, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9032738208770752, + "learning_rate": 1e-06, + "loss": 0.9908, + "mean_token_accuracy": 0.6945891976356506, + "num_tokens": 438693993.0, + "step": 16968 + }, + { + "epoch": 1.863496595651219, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.336388349533081, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7322474718093872, + "num_tokens": 438717877.0, + "step": 16969 + }, + { + "epoch": 1.8636064133538326, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2962818145751953, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7195379734039307, + "num_tokens": 438741555.0, + "step": 16970 + }, + { + "epoch": 1.8637162310564463, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1207945346832275, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7098570466041565, + "num_tokens": 438772483.0, + "step": 16971 + }, + { + "epoch": 1.8638260487590599, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.268847942352295, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7368636131286621, + "num_tokens": 438796946.0, + "step": 16972 + }, + { + "epoch": 1.8639358664616736, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9693044424057007, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.70667564868927, + "num_tokens": 438828630.0, + "step": 16973 + }, + { + "epoch": 1.8640456841642874, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3202531337738037, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7256424427032471, + "num_tokens": 438854638.0, + "step": 16974 + }, + { + "epoch": 1.864155501866901, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.197209596633911, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.734768271446228, + "num_tokens": 438880277.0, + "step": 16975 + }, + { + "epoch": 1.8642653195695145, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.965516209602356, + "learning_rate": 1e-06, + "loss": 0.9601, + "mean_token_accuracy": 0.707284688949585, + "num_tokens": 438913785.0, + "step": 16976 + }, + { + "epoch": 1.8643751372721282, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.532595634460449, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7288118004798889, + "num_tokens": 438935297.0, + "step": 16977 + }, + { + "epoch": 1.864484954974742, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1647579669952393, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7272152900695801, + "num_tokens": 438960935.0, + "step": 16978 + }, + { + "epoch": 1.8645947726773557, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.239684820175171, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.725563645362854, + "num_tokens": 438985536.0, + "step": 16979 + }, + { + "epoch": 1.8647045903799693, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.196485757827759, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.6981593370437622, + "num_tokens": 439014239.0, + "step": 16980 + }, + { + "epoch": 1.8648144080825828, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2677009105682373, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7201204299926758, + "num_tokens": 439039532.0, + "step": 16981 + }, + { + "epoch": 1.8649242257851966, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.008479118347168, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.720403790473938, + "num_tokens": 439070929.0, + "step": 16982 + }, + { + "epoch": 1.8650340434878103, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.339435577392578, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7251198291778564, + "num_tokens": 439093825.0, + "step": 16983 + }, + { + "epoch": 1.8651438611904239, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3978538513183594, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7105741500854492, + "num_tokens": 439117753.0, + "step": 16984 + }, + { + "epoch": 1.8652536788930374, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1321115493774414, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7510157823562622, + "num_tokens": 439144930.0, + "step": 16985 + }, + { + "epoch": 1.8653634965956511, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.602435827255249, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7212439179420471, + "num_tokens": 439165820.0, + "step": 16986 + }, + { + "epoch": 1.865473314298265, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.367478847503662, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7423316240310669, + "num_tokens": 439188198.0, + "step": 16987 + }, + { + "epoch": 1.8655831320008787, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0769314765930176, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7335426807403564, + "num_tokens": 439215580.0, + "step": 16988 + }, + { + "epoch": 1.8656929497034922, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.25101375579834, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7289285659790039, + "num_tokens": 439238300.0, + "step": 16989 + }, + { + "epoch": 1.8658027674061057, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2514681816101074, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.709796130657196, + "num_tokens": 439263492.0, + "step": 16990 + }, + { + "epoch": 1.8659125851087195, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.302738666534424, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7165062427520752, + "num_tokens": 439290625.0, + "step": 16991 + }, + { + "epoch": 1.8660224028113332, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.341669797897339, + "learning_rate": 1e-06, + "loss": 1.0048, + "mean_token_accuracy": 0.694307267665863, + "num_tokens": 439316988.0, + "step": 16992 + }, + { + "epoch": 1.866132220513947, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.425004005432129, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7288212776184082, + "num_tokens": 439338441.0, + "step": 16993 + }, + { + "epoch": 1.8662420382165605, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.6105520725250244, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7319619059562683, + "num_tokens": 439359117.0, + "step": 16994 + }, + { + "epoch": 1.866351855919174, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3234715461730957, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7111285924911499, + "num_tokens": 439383259.0, + "step": 16995 + }, + { + "epoch": 1.8664616736217878, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.218477964401245, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7158504128456116, + "num_tokens": 439407786.0, + "step": 16996 + }, + { + "epoch": 1.8665714913244016, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.021057367324829, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7124913930892944, + "num_tokens": 439437130.0, + "step": 16997 + }, + { + "epoch": 1.8666813090270151, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0506999492645264, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7348290681838989, + "num_tokens": 439465442.0, + "step": 16998 + }, + { + "epoch": 1.8667911267296287, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1211345195770264, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7140613794326782, + "num_tokens": 439494648.0, + "step": 16999 + }, + { + "epoch": 1.8669009444322424, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2128946781158447, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7251512408256531, + "num_tokens": 439520459.0, + "step": 17000 + }, + { + "epoch": 1.8670107621348562, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3041841983795166, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.709580659866333, + "num_tokens": 439546501.0, + "step": 17001 + }, + { + "epoch": 1.86712057983747, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0288894176483154, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7245989441871643, + "num_tokens": 439576103.0, + "step": 17002 + }, + { + "epoch": 1.8672303975400835, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.5765769481658936, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7153948545455933, + "num_tokens": 439597541.0, + "step": 17003 + }, + { + "epoch": 1.867340215242697, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0956313610076904, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7272166013717651, + "num_tokens": 439627409.0, + "step": 17004 + }, + { + "epoch": 1.8674500329453108, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3837027549743652, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7157102823257446, + "num_tokens": 439649990.0, + "step": 17005 + }, + { + "epoch": 1.8675598506479245, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2724664211273193, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7276977896690369, + "num_tokens": 439676100.0, + "step": 17006 + }, + { + "epoch": 1.8676696683505383, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.265001058578491, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7062695026397705, + "num_tokens": 439701954.0, + "step": 17007 + }, + { + "epoch": 1.8677794860531518, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1582870483398438, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.721455991268158, + "num_tokens": 439729123.0, + "step": 17008 + }, + { + "epoch": 1.8678893037557653, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2597250938415527, + "learning_rate": 1e-06, + "loss": 0.9817, + "mean_token_accuracy": 0.7023629546165466, + "num_tokens": 439755615.0, + "step": 17009 + }, + { + "epoch": 1.867999121458379, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0407958030700684, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7091257572174072, + "num_tokens": 439786232.0, + "step": 17010 + }, + { + "epoch": 1.8681089391609929, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.293306350708008, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7256653308868408, + "num_tokens": 439810890.0, + "step": 17011 + }, + { + "epoch": 1.8682187568636064, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.418219804763794, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7170112133026123, + "num_tokens": 439833482.0, + "step": 17012 + }, + { + "epoch": 1.86832857456622, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1494133472442627, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.728590726852417, + "num_tokens": 439859092.0, + "step": 17013 + }, + { + "epoch": 1.8684383922688337, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0945212841033936, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7228555679321289, + "num_tokens": 439886906.0, + "step": 17014 + }, + { + "epoch": 1.8685482099714474, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2309839725494385, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7058804035186768, + "num_tokens": 439912722.0, + "step": 17015 + }, + { + "epoch": 1.8686580276740612, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.4045920372009277, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7331078052520752, + "num_tokens": 439936368.0, + "step": 17016 + }, + { + "epoch": 1.8687678453766747, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3617103099823, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7246795892715454, + "num_tokens": 439961316.0, + "step": 17017 + }, + { + "epoch": 1.8688776630792883, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.232861280441284, + "learning_rate": 1e-06, + "loss": 0.978, + "mean_token_accuracy": 0.697953462600708, + "num_tokens": 439989532.0, + "step": 17018 + }, + { + "epoch": 1.868987480781902, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2161056995391846, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.716933012008667, + "num_tokens": 440016373.0, + "step": 17019 + }, + { + "epoch": 1.8690972984845158, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.146162748336792, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7400898933410645, + "num_tokens": 440041652.0, + "step": 17020 + }, + { + "epoch": 1.8692071161871293, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.256533145904541, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7153835296630859, + "num_tokens": 440065463.0, + "step": 17021 + }, + { + "epoch": 1.869316933889743, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2096004486083984, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.70350182056427, + "num_tokens": 440092608.0, + "step": 17022 + }, + { + "epoch": 1.8694267515923566, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9961638450622559, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7176415324211121, + "num_tokens": 440122046.0, + "step": 17023 + }, + { + "epoch": 1.8695365692949704, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9450401067733765, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7308952808380127, + "num_tokens": 440152237.0, + "step": 17024 + }, + { + "epoch": 1.8696463869975841, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1246588230133057, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7062112092971802, + "num_tokens": 440179011.0, + "step": 17025 + }, + { + "epoch": 1.8697562047001977, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2547967433929443, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7297306060791016, + "num_tokens": 440202428.0, + "step": 17026 + }, + { + "epoch": 1.8698660224028112, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.3030459880828857, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7011861801147461, + "num_tokens": 440228411.0, + "step": 17027 + }, + { + "epoch": 1.869975840105425, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.211019992828369, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7266325354576111, + "num_tokens": 440253868.0, + "step": 17028 + }, + { + "epoch": 1.8700856578080387, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2353031635284424, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7287583351135254, + "num_tokens": 440276775.0, + "step": 17029 + }, + { + "epoch": 1.8701954755106525, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.100020408630371, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7126104235649109, + "num_tokens": 440304245.0, + "step": 17030 + }, + { + "epoch": 1.870305293213266, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.360187292098999, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7242175340652466, + "num_tokens": 440328836.0, + "step": 17031 + }, + { + "epoch": 1.8704151109158795, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.430462598800659, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7248718738555908, + "num_tokens": 440350971.0, + "step": 17032 + }, + { + "epoch": 1.8705249286184933, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.152057409286499, + "learning_rate": 1e-06, + "loss": 1.0393, + "mean_token_accuracy": 0.6793932914733887, + "num_tokens": 440377766.0, + "step": 17033 + }, + { + "epoch": 1.870634746321107, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9917917251586914, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7397780418395996, + "num_tokens": 440407060.0, + "step": 17034 + }, + { + "epoch": 1.8707445640237206, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0808513164520264, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7037011981010437, + "num_tokens": 440433321.0, + "step": 17035 + }, + { + "epoch": 1.8708543817263343, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0922396183013916, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7275481224060059, + "num_tokens": 440461191.0, + "step": 17036 + }, + { + "epoch": 1.8709641994289479, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2553024291992188, + "learning_rate": 1e-06, + "loss": 0.9942, + "mean_token_accuracy": 0.6988656520843506, + "num_tokens": 440485817.0, + "step": 17037 + }, + { + "epoch": 1.8710740171315616, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0840811729431152, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7187989354133606, + "num_tokens": 440512880.0, + "step": 17038 + }, + { + "epoch": 1.8711838348341754, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.0202152729034424, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7275180816650391, + "num_tokens": 440543790.0, + "step": 17039 + }, + { + "epoch": 1.871293652536789, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2999565601348877, + "learning_rate": 1e-06, + "loss": 0.9169, + "mean_token_accuracy": 0.7114766240119934, + "num_tokens": 440568853.0, + "step": 17040 + }, + { + "epoch": 1.8714034702394025, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.010906457901001, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7048134803771973, + "num_tokens": 440599282.0, + "step": 17041 + }, + { + "epoch": 1.8715132879420162, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.005683660507202, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7497676610946655, + "num_tokens": 440629140.0, + "step": 17042 + }, + { + "epoch": 1.87162310564463, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.2779154777526855, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.71962571144104, + "num_tokens": 440653455.0, + "step": 17043 + }, + { + "epoch": 1.8717329233472437, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.1814470291137695, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7250678539276123, + "num_tokens": 440680585.0, + "step": 17044 + }, + { + "epoch": 1.8718427410498573, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.486180305480957, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7261765003204346, + "num_tokens": 440703091.0, + "step": 17045 + }, + { + "epoch": 1.8719525587524708, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 2.201432943344116, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.72324538230896, + "num_tokens": 440729516.0, + "step": 17046 + }, + { + "epoch": 1.8720623764550846, + "ewc_loss": 1.0073184967041016e-05, + "grad_norm": 1.9368250370025635, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.727530837059021, + "num_tokens": 440764333.0, + "step": 17047 + }, + { + "epoch": 1.8721721941576983, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.006864547729492, + "learning_rate": 1e-06, + "loss": 0.9627, + "mean_token_accuracy": 0.706354558467865, + "num_tokens": 440794112.0, + "step": 17048 + }, + { + "epoch": 1.8722820118603118, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.342089891433716, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.736722469329834, + "num_tokens": 440816068.0, + "step": 17049 + }, + { + "epoch": 1.8723918295629254, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3767759799957275, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7311693429946899, + "num_tokens": 440839606.0, + "step": 17050 + }, + { + "epoch": 1.8725016472655391, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1474690437316895, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7075478434562683, + "num_tokens": 440867155.0, + "step": 17051 + }, + { + "epoch": 1.872611464968153, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1262495517730713, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.715709924697876, + "num_tokens": 440892677.0, + "step": 17052 + }, + { + "epoch": 1.8727212826707667, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.160475969314575, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7108843326568604, + "num_tokens": 440920143.0, + "step": 17053 + }, + { + "epoch": 1.8728311003733802, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.39424729347229, + "learning_rate": 1e-06, + "loss": 0.7186, + "mean_token_accuracy": 0.7674002647399902, + "num_tokens": 440940554.0, + "step": 17054 + }, + { + "epoch": 1.8729409180759937, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0196454524993896, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7344785332679749, + "num_tokens": 440971347.0, + "step": 17055 + }, + { + "epoch": 1.8730507357786075, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.247067928314209, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7151637077331543, + "num_tokens": 440997428.0, + "step": 17056 + }, + { + "epoch": 1.8731605534812212, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2541418075561523, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7311648726463318, + "num_tokens": 441020541.0, + "step": 17057 + }, + { + "epoch": 1.873270371183835, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1924779415130615, + "learning_rate": 1e-06, + "loss": 0.7527, + "mean_token_accuracy": 0.7606629729270935, + "num_tokens": 441044714.0, + "step": 17058 + }, + { + "epoch": 1.8733801888864485, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.234447717666626, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7320601940155029, + "num_tokens": 441068532.0, + "step": 17059 + }, + { + "epoch": 1.873490006589062, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.455599546432495, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7486592531204224, + "num_tokens": 441091202.0, + "step": 17060 + }, + { + "epoch": 1.8735998242916758, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5378007888793945, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7162456512451172, + "num_tokens": 441112498.0, + "step": 17061 + }, + { + "epoch": 1.8737096419942896, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0988078117370605, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7127853631973267, + "num_tokens": 441142109.0, + "step": 17062 + }, + { + "epoch": 1.8738194596969031, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.702371597290039, + "learning_rate": 1e-06, + "loss": 0.9768, + "mean_token_accuracy": 0.7013869881629944, + "num_tokens": 441164003.0, + "step": 17063 + }, + { + "epoch": 1.8739292773995166, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1958234310150146, + "learning_rate": 1e-06, + "loss": 0.9773, + "mean_token_accuracy": 0.696267306804657, + "num_tokens": 441189905.0, + "step": 17064 + }, + { + "epoch": 1.8740390951021304, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.217545986175537, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7219852209091187, + "num_tokens": 441215805.0, + "step": 17065 + }, + { + "epoch": 1.8741489128047442, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0989625453948975, + "learning_rate": 1e-06, + "loss": 0.9482, + "mean_token_accuracy": 0.7089983820915222, + "num_tokens": 441246601.0, + "step": 17066 + }, + { + "epoch": 1.874258730507358, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0406792163848877, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7317361235618591, + "num_tokens": 441274797.0, + "step": 17067 + }, + { + "epoch": 1.8743685482099715, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.287390947341919, + "learning_rate": 1e-06, + "loss": 0.9382, + "mean_token_accuracy": 0.705229640007019, + "num_tokens": 441301370.0, + "step": 17068 + }, + { + "epoch": 1.874478365912585, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.465445041656494, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7263281345367432, + "num_tokens": 441323660.0, + "step": 17069 + }, + { + "epoch": 1.8745881836151987, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0892832279205322, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7127891182899475, + "num_tokens": 441352209.0, + "step": 17070 + }, + { + "epoch": 1.8746980013178125, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9529873132705688, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7181581258773804, + "num_tokens": 441383721.0, + "step": 17071 + }, + { + "epoch": 1.8748078190204263, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2164292335510254, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7111617922782898, + "num_tokens": 441409299.0, + "step": 17072 + }, + { + "epoch": 1.8749176367230398, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2232532501220703, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7193354368209839, + "num_tokens": 441434302.0, + "step": 17073 + }, + { + "epoch": 1.8750274544256533, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.116417407989502, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7204654812812805, + "num_tokens": 441461110.0, + "step": 17074 + }, + { + "epoch": 1.875137272128267, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9021011590957642, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7188456654548645, + "num_tokens": 441493487.0, + "step": 17075 + }, + { + "epoch": 1.8752470898308808, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.189117670059204, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7460735440254211, + "num_tokens": 441518779.0, + "step": 17076 + }, + { + "epoch": 1.8753569075334944, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.282151460647583, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7430974245071411, + "num_tokens": 441542646.0, + "step": 17077 + }, + { + "epoch": 1.875466725236108, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.89133882522583, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7242133617401123, + "num_tokens": 441573331.0, + "step": 17078 + }, + { + "epoch": 1.8755765429387217, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.072549343109131, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7314021587371826, + "num_tokens": 441602259.0, + "step": 17079 + }, + { + "epoch": 1.8756863606413354, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9979784488677979, + "learning_rate": 1e-06, + "loss": 1.0155, + "mean_token_accuracy": 0.696277379989624, + "num_tokens": 441634513.0, + "step": 17080 + }, + { + "epoch": 1.8757961783439492, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3750648498535156, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.726134717464447, + "num_tokens": 441657125.0, + "step": 17081 + }, + { + "epoch": 1.8759059960465627, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1464221477508545, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7131073474884033, + "num_tokens": 441683874.0, + "step": 17082 + }, + { + "epoch": 1.8760158137491763, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.205376148223877, + "learning_rate": 1e-06, + "loss": 0.9632, + "mean_token_accuracy": 0.6989010572433472, + "num_tokens": 441710732.0, + "step": 17083 + }, + { + "epoch": 1.87612563145179, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.309591054916382, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7323933839797974, + "num_tokens": 441736131.0, + "step": 17084 + }, + { + "epoch": 1.8762354491544038, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.433368682861328, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7154039740562439, + "num_tokens": 441758858.0, + "step": 17085 + }, + { + "epoch": 1.8763452668570173, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3142123222351074, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7293678522109985, + "num_tokens": 441782241.0, + "step": 17086 + }, + { + "epoch": 1.876455084559631, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2476115226745605, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.730815589427948, + "num_tokens": 441805904.0, + "step": 17087 + }, + { + "epoch": 1.8765649022622446, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.571262836456299, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.729840874671936, + "num_tokens": 441826314.0, + "step": 17088 + }, + { + "epoch": 1.8766747199648584, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2997326850891113, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.719816267490387, + "num_tokens": 441852105.0, + "step": 17089 + }, + { + "epoch": 1.876784537667472, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.016230821609497, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7108734250068665, + "num_tokens": 441882518.0, + "step": 17090 + }, + { + "epoch": 1.8768943553700856, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.534538984298706, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7064733505249023, + "num_tokens": 441905581.0, + "step": 17091 + }, + { + "epoch": 1.8770041730726992, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.8283286094665527, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7506425380706787, + "num_tokens": 441921785.0, + "step": 17092 + }, + { + "epoch": 1.877113990775313, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.29219126701355, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7359232902526855, + "num_tokens": 441944956.0, + "step": 17093 + }, + { + "epoch": 1.8772238084779267, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.424952745437622, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.72556471824646, + "num_tokens": 441967212.0, + "step": 17094 + }, + { + "epoch": 1.8773336261805404, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9930020570755005, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7154548764228821, + "num_tokens": 441997960.0, + "step": 17095 + }, + { + "epoch": 1.877443443883154, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.266380548477173, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7143019437789917, + "num_tokens": 442022187.0, + "step": 17096 + }, + { + "epoch": 1.8775532615857675, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.172959804534912, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.6993362903594971, + "num_tokens": 442050708.0, + "step": 17097 + }, + { + "epoch": 1.8776630792883813, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2241830825805664, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7450657486915588, + "num_tokens": 442075492.0, + "step": 17098 + }, + { + "epoch": 1.877772896990995, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.162245988845825, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7025004625320435, + "num_tokens": 442102548.0, + "step": 17099 + }, + { + "epoch": 1.8778827146936086, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.306854248046875, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7112146615982056, + "num_tokens": 442130805.0, + "step": 17100 + }, + { + "epoch": 1.8779925323962223, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.448786497116089, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.7106043696403503, + "num_tokens": 442154279.0, + "step": 17101 + }, + { + "epoch": 1.8781023500988359, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.145648241043091, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.715388298034668, + "num_tokens": 442182025.0, + "step": 17102 + }, + { + "epoch": 1.8782121678014496, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.356471300125122, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7328898310661316, + "num_tokens": 442205503.0, + "step": 17103 + }, + { + "epoch": 1.8783219855040634, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3235738277435303, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7331472039222717, + "num_tokens": 442228249.0, + "step": 17104 + }, + { + "epoch": 1.878431803206677, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4076335430145264, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7194926738739014, + "num_tokens": 442249551.0, + "step": 17105 + }, + { + "epoch": 1.8785416209092904, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3711419105529785, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7314363718032837, + "num_tokens": 442271194.0, + "step": 17106 + }, + { + "epoch": 1.8786514386119042, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1070430278778076, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7222698330879211, + "num_tokens": 442300042.0, + "step": 17107 + }, + { + "epoch": 1.878761256314518, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2465574741363525, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7062174081802368, + "num_tokens": 442328394.0, + "step": 17108 + }, + { + "epoch": 1.8788710740171317, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0074334144592285, + "learning_rate": 1e-06, + "loss": 0.973, + "mean_token_accuracy": 0.700607419013977, + "num_tokens": 442360783.0, + "step": 17109 + }, + { + "epoch": 1.8789808917197452, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.288336992263794, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7190730571746826, + "num_tokens": 442385450.0, + "step": 17110 + }, + { + "epoch": 1.8790907094223588, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.257622718811035, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7137864232063293, + "num_tokens": 442410895.0, + "step": 17111 + }, + { + "epoch": 1.8792005271249725, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.297494888305664, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7401787042617798, + "num_tokens": 442433978.0, + "step": 17112 + }, + { + "epoch": 1.8793103448275863, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3012092113494873, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7164486646652222, + "num_tokens": 442459487.0, + "step": 17113 + }, + { + "epoch": 1.8794201625301998, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.18740177154541, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7166815400123596, + "num_tokens": 442488748.0, + "step": 17114 + }, + { + "epoch": 1.8795299802328134, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.128171920776367, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7389587759971619, + "num_tokens": 442513277.0, + "step": 17115 + }, + { + "epoch": 1.8796397979354271, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3558356761932373, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7443245649337769, + "num_tokens": 442535248.0, + "step": 17116 + }, + { + "epoch": 1.8797496156380409, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.008267641067505, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7197076678276062, + "num_tokens": 442565238.0, + "step": 17117 + }, + { + "epoch": 1.8798594333406546, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1527926921844482, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7316336631774902, + "num_tokens": 442592048.0, + "step": 17118 + }, + { + "epoch": 1.8799692510432682, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.170548439025879, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7064684629440308, + "num_tokens": 442619496.0, + "step": 17119 + }, + { + "epoch": 1.8800790687458817, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2874221801757812, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7156774997711182, + "num_tokens": 442644476.0, + "step": 17120 + }, + { + "epoch": 1.8801888864484955, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2334091663360596, + "learning_rate": 1e-06, + "loss": 0.9569, + "mean_token_accuracy": 0.7084532976150513, + "num_tokens": 442670967.0, + "step": 17121 + }, + { + "epoch": 1.8802987041511092, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.326467752456665, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7196928262710571, + "num_tokens": 442694055.0, + "step": 17122 + }, + { + "epoch": 1.880408521853723, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.34108567237854, + "learning_rate": 1e-06, + "loss": 0.9819, + "mean_token_accuracy": 0.6986020803451538, + "num_tokens": 442719608.0, + "step": 17123 + }, + { + "epoch": 1.8805183395563365, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2154996395111084, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7178425788879395, + "num_tokens": 442745522.0, + "step": 17124 + }, + { + "epoch": 1.88062815725895, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9659892320632935, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7448431253433228, + "num_tokens": 442774483.0, + "step": 17125 + }, + { + "epoch": 1.8807379749615638, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.34999418258667, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7228751182556152, + "num_tokens": 442799053.0, + "step": 17126 + }, + { + "epoch": 1.8808477926641776, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9452521800994873, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.7009880542755127, + "num_tokens": 442830422.0, + "step": 17127 + }, + { + "epoch": 1.880957610366791, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.142909049987793, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7081157565116882, + "num_tokens": 442858114.0, + "step": 17128 + }, + { + "epoch": 1.8810674280694046, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2478175163269043, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.6988397240638733, + "num_tokens": 442885656.0, + "step": 17129 + }, + { + "epoch": 1.8811772457720184, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9304412603378296, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7348666787147522, + "num_tokens": 442915149.0, + "step": 17130 + }, + { + "epoch": 1.8812870634746321, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2099714279174805, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7204189896583557, + "num_tokens": 442940307.0, + "step": 17131 + }, + { + "epoch": 1.881396881177246, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9501630067825317, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7155808210372925, + "num_tokens": 442971327.0, + "step": 17132 + }, + { + "epoch": 1.8815066988798594, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4624505043029785, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7269233465194702, + "num_tokens": 442993456.0, + "step": 17133 + }, + { + "epoch": 1.881616516582473, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0609710216522217, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7121872901916504, + "num_tokens": 443022173.0, + "step": 17134 + }, + { + "epoch": 1.8817263342850867, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.190300941467285, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.730985164642334, + "num_tokens": 443048031.0, + "step": 17135 + }, + { + "epoch": 1.8818361519877005, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6032071113586426, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7248015403747559, + "num_tokens": 443068998.0, + "step": 17136 + }, + { + "epoch": 1.881945969690314, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.414252519607544, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7056992650032043, + "num_tokens": 443090374.0, + "step": 17137 + }, + { + "epoch": 1.8820557873929278, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2159571647644043, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7291179895401001, + "num_tokens": 443116477.0, + "step": 17138 + }, + { + "epoch": 1.8821656050955413, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2387802600860596, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7452125549316406, + "num_tokens": 443142580.0, + "step": 17139 + }, + { + "epoch": 1.882275422798155, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.485771894454956, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7191808819770813, + "num_tokens": 443165982.0, + "step": 17140 + }, + { + "epoch": 1.8823852405007688, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9845281839370728, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7284208536148071, + "num_tokens": 443197490.0, + "step": 17141 + }, + { + "epoch": 1.8824950582033824, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.522228717803955, + "learning_rate": 1e-06, + "loss": 0.7817, + "mean_token_accuracy": 0.7507582306861877, + "num_tokens": 443216578.0, + "step": 17142 + }, + { + "epoch": 1.882604875905996, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.041851043701172, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.734990119934082, + "num_tokens": 443246303.0, + "step": 17143 + }, + { + "epoch": 1.8827146936086097, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.242347478866577, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7155051231384277, + "num_tokens": 443272511.0, + "step": 17144 + }, + { + "epoch": 1.8828245113112234, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2337396144866943, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7131820321083069, + "num_tokens": 443297898.0, + "step": 17145 + }, + { + "epoch": 1.8829343290138372, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3485312461853027, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7286146879196167, + "num_tokens": 443320756.0, + "step": 17146 + }, + { + "epoch": 1.8830441467164507, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.308633804321289, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7155733704566956, + "num_tokens": 443344590.0, + "step": 17147 + }, + { + "epoch": 1.8831539644190642, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.174175500869751, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7065262198448181, + "num_tokens": 443372794.0, + "step": 17148 + }, + { + "epoch": 1.883263782121678, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2361900806427, + "learning_rate": 1e-06, + "loss": 0.9774, + "mean_token_accuracy": 0.7029574513435364, + "num_tokens": 443401270.0, + "step": 17149 + }, + { + "epoch": 1.8833735998242918, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9530510902404785, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7111228108406067, + "num_tokens": 443432763.0, + "step": 17150 + }, + { + "epoch": 1.8834834175269053, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0752596855163574, + "learning_rate": 1e-06, + "loss": 0.9614, + "mean_token_accuracy": 0.7034855484962463, + "num_tokens": 443462146.0, + "step": 17151 + }, + { + "epoch": 1.883593235229519, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0941011905670166, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7404515743255615, + "num_tokens": 443489265.0, + "step": 17152 + }, + { + "epoch": 1.8837030529321326, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0576553344726562, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7105540037155151, + "num_tokens": 443521043.0, + "step": 17153 + }, + { + "epoch": 1.8838128706347463, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.140639066696167, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7221187353134155, + "num_tokens": 443548868.0, + "step": 17154 + }, + { + "epoch": 1.88392268833736, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.247957229614258, + "learning_rate": 1e-06, + "loss": 0.7486, + "mean_token_accuracy": 0.7572178840637207, + "num_tokens": 443572380.0, + "step": 17155 + }, + { + "epoch": 1.8840325060399736, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.353743553161621, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7244786024093628, + "num_tokens": 443595791.0, + "step": 17156 + }, + { + "epoch": 1.8841423237425872, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.115144729614258, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.711990475654602, + "num_tokens": 443623415.0, + "step": 17157 + }, + { + "epoch": 1.884252141445201, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0881266593933105, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7141244411468506, + "num_tokens": 443653345.0, + "step": 17158 + }, + { + "epoch": 1.8843619591478147, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1954691410064697, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7301374673843384, + "num_tokens": 443681577.0, + "step": 17159 + }, + { + "epoch": 1.8844717768504284, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2484774589538574, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7246444821357727, + "num_tokens": 443707466.0, + "step": 17160 + }, + { + "epoch": 1.884581594553042, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3374972343444824, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.716566264629364, + "num_tokens": 443730996.0, + "step": 17161 + }, + { + "epoch": 1.8846914122556555, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4165072441101074, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7099785804748535, + "num_tokens": 443754314.0, + "step": 17162 + }, + { + "epoch": 1.8848012299582693, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4634788036346436, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7301821708679199, + "num_tokens": 443775221.0, + "step": 17163 + }, + { + "epoch": 1.884911047660883, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3310694694519043, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.699006199836731, + "num_tokens": 443799455.0, + "step": 17164 + }, + { + "epoch": 1.8850208653634966, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.307363748550415, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7078702449798584, + "num_tokens": 443823852.0, + "step": 17165 + }, + { + "epoch": 1.88513068306611, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.172232151031494, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.6976132988929749, + "num_tokens": 443850175.0, + "step": 17166 + }, + { + "epoch": 1.8852405007687238, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3519461154937744, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7335309386253357, + "num_tokens": 443874583.0, + "step": 17167 + }, + { + "epoch": 1.8853503184713376, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5443825721740723, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7394152879714966, + "num_tokens": 443894730.0, + "step": 17168 + }, + { + "epoch": 1.8854601361739514, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6299517154693604, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7469481229782104, + "num_tokens": 443914815.0, + "step": 17169 + }, + { + "epoch": 1.885569953876565, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9961282014846802, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6939486265182495, + "num_tokens": 443946817.0, + "step": 17170 + }, + { + "epoch": 1.8856797715791784, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.281534433364868, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7114019393920898, + "num_tokens": 443972576.0, + "step": 17171 + }, + { + "epoch": 1.8857895892817922, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1874101161956787, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7311620116233826, + "num_tokens": 443997637.0, + "step": 17172 + }, + { + "epoch": 1.885899406984406, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6165804862976074, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7295489311218262, + "num_tokens": 444016602.0, + "step": 17173 + }, + { + "epoch": 1.8860092246870197, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.171633005142212, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7270158529281616, + "num_tokens": 444042596.0, + "step": 17174 + }, + { + "epoch": 1.8861190423896332, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9626891613006592, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.73948073387146, + "num_tokens": 444071027.0, + "step": 17175 + }, + { + "epoch": 1.8862288600922468, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2143821716308594, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7072530388832092, + "num_tokens": 444097511.0, + "step": 17176 + }, + { + "epoch": 1.8863386777948605, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2127432823181152, + "learning_rate": 1e-06, + "loss": 0.988, + "mean_token_accuracy": 0.710865318775177, + "num_tokens": 444123105.0, + "step": 17177 + }, + { + "epoch": 1.8864484954974743, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0064332485198975, + "learning_rate": 1e-06, + "loss": 0.9995, + "mean_token_accuracy": 0.6977061629295349, + "num_tokens": 444152880.0, + "step": 17178 + }, + { + "epoch": 1.8865583132000878, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2284629344940186, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7118898630142212, + "num_tokens": 444177159.0, + "step": 17179 + }, + { + "epoch": 1.8866681309027014, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0203773975372314, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.729800820350647, + "num_tokens": 444204765.0, + "step": 17180 + }, + { + "epoch": 1.8867779486053151, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.8022689819335938, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7069001793861389, + "num_tokens": 444240019.0, + "step": 17181 + }, + { + "epoch": 1.8868877663079289, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1095142364501953, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7263258695602417, + "num_tokens": 444265462.0, + "step": 17182 + }, + { + "epoch": 1.8869975840105426, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2127647399902344, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7459173202514648, + "num_tokens": 444290496.0, + "step": 17183 + }, + { + "epoch": 1.8871074017131562, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1448075771331787, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7355220317840576, + "num_tokens": 444316974.0, + "step": 17184 + }, + { + "epoch": 1.8872172194157697, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.143785238265991, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7184165716171265, + "num_tokens": 444347182.0, + "step": 17185 + }, + { + "epoch": 1.8873270371183835, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.265047788619995, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7088472843170166, + "num_tokens": 444373363.0, + "step": 17186 + }, + { + "epoch": 1.8874368548209972, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5657765865325928, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7093295454978943, + "num_tokens": 444395418.0, + "step": 17187 + }, + { + "epoch": 1.887546672523611, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0064284801483154, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.743377685546875, + "num_tokens": 444423375.0, + "step": 17188 + }, + { + "epoch": 1.8876564902262245, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3963022232055664, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7289533019065857, + "num_tokens": 444444920.0, + "step": 17189 + }, + { + "epoch": 1.887766307928838, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1439151763916016, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.7091619968414307, + "num_tokens": 444474529.0, + "step": 17190 + }, + { + "epoch": 1.8878761256314518, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.283384084701538, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7077245116233826, + "num_tokens": 444500093.0, + "step": 17191 + }, + { + "epoch": 1.8879859433340656, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2081568241119385, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7058897018432617, + "num_tokens": 444527583.0, + "step": 17192 + }, + { + "epoch": 1.888095761036679, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.113013505935669, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7462188005447388, + "num_tokens": 444554468.0, + "step": 17193 + }, + { + "epoch": 1.8882055787392926, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.473649263381958, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7122721672058105, + "num_tokens": 444576911.0, + "step": 17194 + }, + { + "epoch": 1.8883153964419064, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1817848682403564, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7104883193969727, + "num_tokens": 444606474.0, + "step": 17195 + }, + { + "epoch": 1.8884252141445201, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1445441246032715, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.6979894638061523, + "num_tokens": 444634777.0, + "step": 17196 + }, + { + "epoch": 1.888535031847134, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.208139657974243, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7195295691490173, + "num_tokens": 444659687.0, + "step": 17197 + }, + { + "epoch": 1.8886448495497474, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3208625316619873, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.73476642370224, + "num_tokens": 444684014.0, + "step": 17198 + }, + { + "epoch": 1.888754667252361, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3815441131591797, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7054648399353027, + "num_tokens": 444706289.0, + "step": 17199 + }, + { + "epoch": 1.8888644849549747, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.468447685241699, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7200542688369751, + "num_tokens": 444727952.0, + "step": 17200 + }, + { + "epoch": 1.8889743026575885, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0511579513549805, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7223336696624756, + "num_tokens": 444757089.0, + "step": 17201 + }, + { + "epoch": 1.889084120360202, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9762976169586182, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7028003931045532, + "num_tokens": 444787948.0, + "step": 17202 + }, + { + "epoch": 1.8891939380628158, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0158958435058594, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.6931107044219971, + "num_tokens": 444821121.0, + "step": 17203 + }, + { + "epoch": 1.8893037557654293, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4357893466949463, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7238966226577759, + "num_tokens": 444843954.0, + "step": 17204 + }, + { + "epoch": 1.889413573468043, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.389416456222534, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7045720219612122, + "num_tokens": 444865200.0, + "step": 17205 + }, + { + "epoch": 1.8895233911706568, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.338156223297119, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7272229194641113, + "num_tokens": 444888083.0, + "step": 17206 + }, + { + "epoch": 1.8896332088732704, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2976341247558594, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7372789978981018, + "num_tokens": 444912017.0, + "step": 17207 + }, + { + "epoch": 1.8897430265758839, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3866708278656006, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7339489459991455, + "num_tokens": 444934303.0, + "step": 17208 + }, + { + "epoch": 1.8898528442784976, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4774677753448486, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7225659489631653, + "num_tokens": 444956792.0, + "step": 17209 + }, + { + "epoch": 1.8899626619811114, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.411710500717163, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7246080636978149, + "num_tokens": 444978512.0, + "step": 17210 + }, + { + "epoch": 1.8900724796837252, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.178313970565796, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7157078981399536, + "num_tokens": 445007847.0, + "step": 17211 + }, + { + "epoch": 1.8901822973863387, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1944549083709717, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7156263589859009, + "num_tokens": 445034565.0, + "step": 17212 + }, + { + "epoch": 1.8902921150889522, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9866925477981567, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7099255323410034, + "num_tokens": 445066671.0, + "step": 17213 + }, + { + "epoch": 1.890401932791566, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9519661664962769, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.705491840839386, + "num_tokens": 445100563.0, + "step": 17214 + }, + { + "epoch": 1.8905117504941797, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0487821102142334, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7305945158004761, + "num_tokens": 445129719.0, + "step": 17215 + }, + { + "epoch": 1.8906215681967933, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4520227909088135, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.712681770324707, + "num_tokens": 445152273.0, + "step": 17216 + }, + { + "epoch": 1.890731385899407, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2159461975097656, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7328564524650574, + "num_tokens": 445176937.0, + "step": 17217 + }, + { + "epoch": 1.8908412036020206, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9721943140029907, + "learning_rate": 1e-06, + "loss": 0.96, + "mean_token_accuracy": 0.7001407146453857, + "num_tokens": 445207189.0, + "step": 17218 + }, + { + "epoch": 1.8909510213046343, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.418654203414917, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7197607755661011, + "num_tokens": 445229436.0, + "step": 17219 + }, + { + "epoch": 1.891060839007248, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2590510845184326, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7212934494018555, + "num_tokens": 445257816.0, + "step": 17220 + }, + { + "epoch": 1.8911706567098616, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1996803283691406, + "learning_rate": 1e-06, + "loss": 0.9514, + "mean_token_accuracy": 0.7133216261863708, + "num_tokens": 445282658.0, + "step": 17221 + }, + { + "epoch": 1.8912804744124752, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.102170467376709, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7118731141090393, + "num_tokens": 445309648.0, + "step": 17222 + }, + { + "epoch": 1.891390292115089, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.247089147567749, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7217801809310913, + "num_tokens": 445336103.0, + "step": 17223 + }, + { + "epoch": 1.8915001098177027, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9877187013626099, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.718966543674469, + "num_tokens": 445366873.0, + "step": 17224 + }, + { + "epoch": 1.8916099275203164, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1922080516815186, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7023165225982666, + "num_tokens": 445394047.0, + "step": 17225 + }, + { + "epoch": 1.89171974522293, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.388411045074463, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7162160277366638, + "num_tokens": 445418118.0, + "step": 17226 + }, + { + "epoch": 1.8918295629255435, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0769290924072266, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7172986268997192, + "num_tokens": 445446647.0, + "step": 17227 + }, + { + "epoch": 1.8919393806281573, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2724015712738037, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7275006771087646, + "num_tokens": 445470464.0, + "step": 17228 + }, + { + "epoch": 1.892049198330771, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.40932035446167, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7380416393280029, + "num_tokens": 445492287.0, + "step": 17229 + }, + { + "epoch": 1.8921590160333845, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.309913396835327, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7144811153411865, + "num_tokens": 445516898.0, + "step": 17230 + }, + { + "epoch": 1.892268833735998, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2694504261016846, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7114136219024658, + "num_tokens": 445541226.0, + "step": 17231 + }, + { + "epoch": 1.8923786514386118, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.38116192817688, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7313615679740906, + "num_tokens": 445564130.0, + "step": 17232 + }, + { + "epoch": 1.8924884691412256, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.200404644012451, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.728878378868103, + "num_tokens": 445589156.0, + "step": 17233 + }, + { + "epoch": 1.8925982868438394, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2711572647094727, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7207162976264954, + "num_tokens": 445613269.0, + "step": 17234 + }, + { + "epoch": 1.8927081045464529, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.132575273513794, + "learning_rate": 1e-06, + "loss": 0.9734, + "mean_token_accuracy": 0.7073633670806885, + "num_tokens": 445642926.0, + "step": 17235 + }, + { + "epoch": 1.8928179222490664, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4301459789276123, + "learning_rate": 1e-06, + "loss": 0.775, + "mean_token_accuracy": 0.7509452104568481, + "num_tokens": 445663012.0, + "step": 17236 + }, + { + "epoch": 1.8929277399516802, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.139329433441162, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7264431715011597, + "num_tokens": 445690161.0, + "step": 17237 + }, + { + "epoch": 1.893037557654294, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.18790864944458, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7096540927886963, + "num_tokens": 445716987.0, + "step": 17238 + }, + { + "epoch": 1.8931473753569077, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2217557430267334, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.7445909380912781, + "num_tokens": 445741219.0, + "step": 17239 + }, + { + "epoch": 1.8932571930595212, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3336939811706543, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7070291638374329, + "num_tokens": 445766150.0, + "step": 17240 + }, + { + "epoch": 1.8933670107621348, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0753509998321533, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7083337306976318, + "num_tokens": 445793915.0, + "step": 17241 + }, + { + "epoch": 1.8934768284647485, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.168747901916504, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7098623514175415, + "num_tokens": 445819999.0, + "step": 17242 + }, + { + "epoch": 1.8935866461673623, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1154415607452393, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7129030227661133, + "num_tokens": 445849330.0, + "step": 17243 + }, + { + "epoch": 1.8936964638699758, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1836636066436768, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7054863572120667, + "num_tokens": 445876150.0, + "step": 17244 + }, + { + "epoch": 1.8938062815725893, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0664610862731934, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7194184064865112, + "num_tokens": 445906024.0, + "step": 17245 + }, + { + "epoch": 1.893916099275203, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.080267906188965, + "learning_rate": 1e-06, + "loss": 0.959, + "mean_token_accuracy": 0.7021079063415527, + "num_tokens": 445934276.0, + "step": 17246 + }, + { + "epoch": 1.8940259169778169, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1498656272888184, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7074344158172607, + "num_tokens": 445961624.0, + "step": 17247 + }, + { + "epoch": 1.8941357346804306, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1275277137756348, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7231819033622742, + "num_tokens": 445991363.0, + "step": 17248 + }, + { + "epoch": 1.8942455523830442, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4424633979797363, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.7098894119262695, + "num_tokens": 446014534.0, + "step": 17249 + }, + { + "epoch": 1.8943553700856577, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.490781784057617, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7389500737190247, + "num_tokens": 446036606.0, + "step": 17250 + }, + { + "epoch": 1.8944651877882714, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5743916034698486, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7256966829299927, + "num_tokens": 446056241.0, + "step": 17251 + }, + { + "epoch": 1.8945750054908852, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.883225440979004, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7276105880737305, + "num_tokens": 446073179.0, + "step": 17252 + }, + { + "epoch": 1.894684823193499, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2460339069366455, + "learning_rate": 1e-06, + "loss": 0.9875, + "mean_token_accuracy": 0.7043954133987427, + "num_tokens": 446102627.0, + "step": 17253 + }, + { + "epoch": 1.8947946408961125, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.174295663833618, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.728069543838501, + "num_tokens": 446129432.0, + "step": 17254 + }, + { + "epoch": 1.894904458598726, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3510921001434326, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7317692637443542, + "num_tokens": 446151081.0, + "step": 17255 + }, + { + "epoch": 1.8950142763013398, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3782997131347656, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7144045829772949, + "num_tokens": 446176354.0, + "step": 17256 + }, + { + "epoch": 1.8951240940039535, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2218098640441895, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7190455198287964, + "num_tokens": 446202852.0, + "step": 17257 + }, + { + "epoch": 1.895233911706567, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.444406032562256, + "learning_rate": 1e-06, + "loss": 0.9915, + "mean_token_accuracy": 0.6931113004684448, + "num_tokens": 446229433.0, + "step": 17258 + }, + { + "epoch": 1.8953437294091806, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4195451736450195, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7337778806686401, + "num_tokens": 446250044.0, + "step": 17259 + }, + { + "epoch": 1.8954535471117944, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.318265199661255, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7132329940795898, + "num_tokens": 446274982.0, + "step": 17260 + }, + { + "epoch": 1.8955633648144081, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4568593502044678, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7283084988594055, + "num_tokens": 446296475.0, + "step": 17261 + }, + { + "epoch": 1.8956731825170219, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.918959379196167, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7114317417144775, + "num_tokens": 446329801.0, + "step": 17262 + }, + { + "epoch": 1.8957830002196354, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.225677728652954, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7049480676651001, + "num_tokens": 446354878.0, + "step": 17263 + }, + { + "epoch": 1.895892817922249, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9949249029159546, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7214229106903076, + "num_tokens": 446388505.0, + "step": 17264 + }, + { + "epoch": 1.8960026356248627, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.472215414047241, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7107736468315125, + "num_tokens": 446409303.0, + "step": 17265 + }, + { + "epoch": 1.8961124533274765, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.082427978515625, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7143797874450684, + "num_tokens": 446437388.0, + "step": 17266 + }, + { + "epoch": 1.89622227103009, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0671794414520264, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7054396271705627, + "num_tokens": 446467633.0, + "step": 17267 + }, + { + "epoch": 1.8963320887327038, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.18088436126709, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7135841846466064, + "num_tokens": 446495235.0, + "step": 17268 + }, + { + "epoch": 1.8964419064353173, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.145225763320923, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7079132199287415, + "num_tokens": 446526089.0, + "step": 17269 + }, + { + "epoch": 1.896551724137931, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6158745288848877, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7247062921524048, + "num_tokens": 446546020.0, + "step": 17270 + }, + { + "epoch": 1.8966615418405448, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0777242183685303, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.724217414855957, + "num_tokens": 446573748.0, + "step": 17271 + }, + { + "epoch": 1.8967713595431583, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2256081104278564, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7286787629127502, + "num_tokens": 446600106.0, + "step": 17272 + }, + { + "epoch": 1.8968811772457719, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5981571674346924, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7356637120246887, + "num_tokens": 446621102.0, + "step": 17273 + }, + { + "epoch": 1.8969909949483856, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2018213272094727, + "learning_rate": 1e-06, + "loss": 1.0062, + "mean_token_accuracy": 0.6915521621704102, + "num_tokens": 446647799.0, + "step": 17274 + }, + { + "epoch": 1.8971008126509994, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.162989854812622, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7179003357887268, + "num_tokens": 446673387.0, + "step": 17275 + }, + { + "epoch": 1.8972106303536131, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.259506940841675, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7201966047286987, + "num_tokens": 446698085.0, + "step": 17276 + }, + { + "epoch": 1.8973204480562267, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.239161729812622, + "learning_rate": 1e-06, + "loss": 0.986, + "mean_token_accuracy": 0.6904854774475098, + "num_tokens": 446724016.0, + "step": 17277 + }, + { + "epoch": 1.8974302657588402, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2012887001037598, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7314473986625671, + "num_tokens": 446749350.0, + "step": 17278 + }, + { + "epoch": 1.897540083461454, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.032169818878174, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7269411087036133, + "num_tokens": 446779779.0, + "step": 17279 + }, + { + "epoch": 1.8976499011640677, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1951522827148438, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7298693656921387, + "num_tokens": 446805939.0, + "step": 17280 + }, + { + "epoch": 1.8977597188666813, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.110661506652832, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7127662897109985, + "num_tokens": 446837087.0, + "step": 17281 + }, + { + "epoch": 1.897869536569295, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2201988697052, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.726715087890625, + "num_tokens": 446862282.0, + "step": 17282 + }, + { + "epoch": 1.8979793542719086, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0052828788757324, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7316355109214783, + "num_tokens": 446892472.0, + "step": 17283 + }, + { + "epoch": 1.8980891719745223, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2237143516540527, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7346075773239136, + "num_tokens": 446917216.0, + "step": 17284 + }, + { + "epoch": 1.898198989677136, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.113426923751831, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7245156764984131, + "num_tokens": 446942707.0, + "step": 17285 + }, + { + "epoch": 1.8983088073797496, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1111934185028076, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7179163694381714, + "num_tokens": 446970127.0, + "step": 17286 + }, + { + "epoch": 1.8984186250823631, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.620267391204834, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7156655788421631, + "num_tokens": 446990487.0, + "step": 17287 + }, + { + "epoch": 1.898528442784977, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.18847393989563, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7313738465309143, + "num_tokens": 447016971.0, + "step": 17288 + }, + { + "epoch": 1.8986382604875907, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9807313680648804, + "learning_rate": 1e-06, + "loss": 1.0386, + "mean_token_accuracy": 0.6878267526626587, + "num_tokens": 447049599.0, + "step": 17289 + }, + { + "epoch": 1.8987480781902044, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.155503511428833, + "learning_rate": 1e-06, + "loss": 0.9534, + "mean_token_accuracy": 0.7193507552146912, + "num_tokens": 447077909.0, + "step": 17290 + }, + { + "epoch": 1.898857895892818, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2570834159851074, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.745887279510498, + "num_tokens": 447102460.0, + "step": 17291 + }, + { + "epoch": 1.8989677135954315, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.239365816116333, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.724844217300415, + "num_tokens": 447127190.0, + "step": 17292 + }, + { + "epoch": 1.8990775312980452, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.23185396194458, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7367368936538696, + "num_tokens": 447152851.0, + "step": 17293 + }, + { + "epoch": 1.899187349000659, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.093852996826172, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7124518156051636, + "num_tokens": 447181968.0, + "step": 17294 + }, + { + "epoch": 1.8992971667032725, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3791356086730957, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7306888699531555, + "num_tokens": 447204277.0, + "step": 17295 + }, + { + "epoch": 1.899406984405886, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.399014711380005, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7326757907867432, + "num_tokens": 447226752.0, + "step": 17296 + }, + { + "epoch": 1.8995168021084998, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.36842942237854, + "learning_rate": 1e-06, + "loss": 0.797, + "mean_token_accuracy": 0.7502280473709106, + "num_tokens": 447248056.0, + "step": 17297 + }, + { + "epoch": 1.8996266198111136, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2750890254974365, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7120279669761658, + "num_tokens": 447273500.0, + "step": 17298 + }, + { + "epoch": 1.8997364375137273, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.172555923461914, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7180614471435547, + "num_tokens": 447302646.0, + "step": 17299 + }, + { + "epoch": 1.8998462552163409, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.213808059692383, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7201822996139526, + "num_tokens": 447328778.0, + "step": 17300 + }, + { + "epoch": 1.8999560729189544, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2980387210845947, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7257824540138245, + "num_tokens": 447354432.0, + "step": 17301 + }, + { + "epoch": 1.9000658906215682, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.176055908203125, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7293721437454224, + "num_tokens": 447381128.0, + "step": 17302 + }, + { + "epoch": 1.900175708324182, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.396754264831543, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7182238101959229, + "num_tokens": 447406027.0, + "step": 17303 + }, + { + "epoch": 1.9002855260267957, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.177624225616455, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.750805139541626, + "num_tokens": 447430713.0, + "step": 17304 + }, + { + "epoch": 1.9003953437294092, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3448402881622314, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7293278574943542, + "num_tokens": 447455069.0, + "step": 17305 + }, + { + "epoch": 1.9005051614320227, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1183300018310547, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7122609615325928, + "num_tokens": 447483421.0, + "step": 17306 + }, + { + "epoch": 1.9006149791346365, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.029994249343872, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7104980945587158, + "num_tokens": 447511348.0, + "step": 17307 + }, + { + "epoch": 1.9007247968372503, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.418527364730835, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7350454330444336, + "num_tokens": 447533441.0, + "step": 17308 + }, + { + "epoch": 1.9008346145398638, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1737289428710938, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7237058877944946, + "num_tokens": 447558387.0, + "step": 17309 + }, + { + "epoch": 1.9009444322424773, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9201542139053345, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7050418853759766, + "num_tokens": 447592242.0, + "step": 17310 + }, + { + "epoch": 1.901054249945091, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3871254920959473, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7062121629714966, + "num_tokens": 447615537.0, + "step": 17311 + }, + { + "epoch": 1.9011640676477048, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.065410852432251, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7264449596405029, + "num_tokens": 447645715.0, + "step": 17312 + }, + { + "epoch": 1.9012738853503186, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0867087841033936, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.730985164642334, + "num_tokens": 447674361.0, + "step": 17313 + }, + { + "epoch": 1.9013837030529321, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1496713161468506, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7107852697372437, + "num_tokens": 447700851.0, + "step": 17314 + }, + { + "epoch": 1.9014935207555457, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.061863422393799, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7195202708244324, + "num_tokens": 447729638.0, + "step": 17315 + }, + { + "epoch": 1.9016033384581594, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9932080507278442, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7051011323928833, + "num_tokens": 447760890.0, + "step": 17316 + }, + { + "epoch": 1.9017131561607732, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0692567825317383, + "learning_rate": 1e-06, + "loss": 0.7865, + "mean_token_accuracy": 0.7442444562911987, + "num_tokens": 447785658.0, + "step": 17317 + }, + { + "epoch": 1.9018229738633867, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2802674770355225, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7103350758552551, + "num_tokens": 447811159.0, + "step": 17318 + }, + { + "epoch": 1.9019327915660005, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.114610195159912, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7312840223312378, + "num_tokens": 447838044.0, + "step": 17319 + }, + { + "epoch": 1.902042609268614, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.262709856033325, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7315039038658142, + "num_tokens": 447865087.0, + "step": 17320 + }, + { + "epoch": 1.9021524269712278, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2101311683654785, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7124515771865845, + "num_tokens": 447890684.0, + "step": 17321 + }, + { + "epoch": 1.9022622446738415, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.372680187225342, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.724609375, + "num_tokens": 447915223.0, + "step": 17322 + }, + { + "epoch": 1.902372062376455, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.130779981613159, + "learning_rate": 1e-06, + "loss": 0.9704, + "mean_token_accuracy": 0.7085066437721252, + "num_tokens": 447943562.0, + "step": 17323 + }, + { + "epoch": 1.9024818800790686, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4074323177337646, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7276552319526672, + "num_tokens": 447967291.0, + "step": 17324 + }, + { + "epoch": 1.9025916977816824, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.218653440475464, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7105164527893066, + "num_tokens": 447993169.0, + "step": 17325 + }, + { + "epoch": 1.9027015154842961, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.064215898513794, + "learning_rate": 1e-06, + "loss": 0.9767, + "mean_token_accuracy": 0.708834171295166, + "num_tokens": 448021760.0, + "step": 17326 + }, + { + "epoch": 1.9028113331869099, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9552206993103027, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6983664035797119, + "num_tokens": 448054045.0, + "step": 17327 + }, + { + "epoch": 1.9029211508895234, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0166211128234863, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7219569683074951, + "num_tokens": 448083280.0, + "step": 17328 + }, + { + "epoch": 1.903030968592137, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.13622784614563, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.720220685005188, + "num_tokens": 448110152.0, + "step": 17329 + }, + { + "epoch": 1.9031407862947507, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0283005237579346, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.698589563369751, + "num_tokens": 448141465.0, + "step": 17330 + }, + { + "epoch": 1.9032506039973645, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1730377674102783, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7380781173706055, + "num_tokens": 448165574.0, + "step": 17331 + }, + { + "epoch": 1.903360421699978, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0579168796539307, + "learning_rate": 1e-06, + "loss": 0.9746, + "mean_token_accuracy": 0.7021997570991516, + "num_tokens": 448193919.0, + "step": 17332 + }, + { + "epoch": 1.9034702394025917, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.21508526802063, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7288256883621216, + "num_tokens": 448219763.0, + "step": 17333 + }, + { + "epoch": 1.9035800571052053, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0111019611358643, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6938276290893555, + "num_tokens": 448251785.0, + "step": 17334 + }, + { + "epoch": 1.903689874807819, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1934328079223633, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7223461866378784, + "num_tokens": 448278044.0, + "step": 17335 + }, + { + "epoch": 1.9037996925104328, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2023425102233887, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7305547595024109, + "num_tokens": 448303866.0, + "step": 17336 + }, + { + "epoch": 1.9039095102130463, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9990484714508057, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.709400475025177, + "num_tokens": 448334191.0, + "step": 17337 + }, + { + "epoch": 1.9040193279156599, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2252800464630127, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7145548462867737, + "num_tokens": 448361400.0, + "step": 17338 + }, + { + "epoch": 1.9041291456182736, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.637178659439087, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7133030295372009, + "num_tokens": 448381714.0, + "step": 17339 + }, + { + "epoch": 1.9042389633208874, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2206311225891113, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7152124643325806, + "num_tokens": 448407440.0, + "step": 17340 + }, + { + "epoch": 1.9043487810235011, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.469231128692627, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7242506742477417, + "num_tokens": 448429954.0, + "step": 17341 + }, + { + "epoch": 1.9044585987261147, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2574126720428467, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7371270656585693, + "num_tokens": 448455485.0, + "step": 17342 + }, + { + "epoch": 1.9045684164287282, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.395888090133667, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7436316013336182, + "num_tokens": 448480074.0, + "step": 17343 + }, + { + "epoch": 1.904678234131342, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1419248580932617, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.737845242023468, + "num_tokens": 448507473.0, + "step": 17344 + }, + { + "epoch": 1.9047880518339557, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.443192481994629, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.7194080948829651, + "num_tokens": 448529085.0, + "step": 17345 + }, + { + "epoch": 1.9048978695365693, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3802597522735596, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.719290554523468, + "num_tokens": 448553224.0, + "step": 17346 + }, + { + "epoch": 1.9050076872391828, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.350463390350342, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7054264545440674, + "num_tokens": 448577920.0, + "step": 17347 + }, + { + "epoch": 1.9051175049417965, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1782712936401367, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7265719175338745, + "num_tokens": 448604254.0, + "step": 17348 + }, + { + "epoch": 1.9052273226444103, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.412336587905884, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7213298678398132, + "num_tokens": 448626666.0, + "step": 17349 + }, + { + "epoch": 1.905337140347024, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.8879636526107788, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7105560898780823, + "num_tokens": 448661643.0, + "step": 17350 + }, + { + "epoch": 1.9054469580496376, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.18715763092041, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7164892554283142, + "num_tokens": 448688672.0, + "step": 17351 + }, + { + "epoch": 1.9055567757522511, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.8554866313934326, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7262780666351318, + "num_tokens": 448707192.0, + "step": 17352 + }, + { + "epoch": 1.9056665934548649, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.322227954864502, + "learning_rate": 1e-06, + "loss": 0.9765, + "mean_token_accuracy": 0.7032873630523682, + "num_tokens": 448733459.0, + "step": 17353 + }, + { + "epoch": 1.9057764111574786, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.143510341644287, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7329727411270142, + "num_tokens": 448758398.0, + "step": 17354 + }, + { + "epoch": 1.9058862288600924, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.450819253921509, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7203431725502014, + "num_tokens": 448779754.0, + "step": 17355 + }, + { + "epoch": 1.905996046562706, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.436927556991577, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7105438709259033, + "num_tokens": 448804077.0, + "step": 17356 + }, + { + "epoch": 1.9061058642653195, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6344850063323975, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7282841801643372, + "num_tokens": 448825247.0, + "step": 17357 + }, + { + "epoch": 1.9062156819679332, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.169038772583008, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7265831828117371, + "num_tokens": 448850877.0, + "step": 17358 + }, + { + "epoch": 1.906325499670547, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.351830005645752, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7095698714256287, + "num_tokens": 448873759.0, + "step": 17359 + }, + { + "epoch": 1.9064353173731605, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0915474891662598, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.701701283454895, + "num_tokens": 448902244.0, + "step": 17360 + }, + { + "epoch": 1.906545135075774, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.887012004852295, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7217402458190918, + "num_tokens": 448937166.0, + "step": 17361 + }, + { + "epoch": 1.9066549527783878, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3048813343048096, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7204751372337341, + "num_tokens": 448961982.0, + "step": 17362 + }, + { + "epoch": 1.9067647704810016, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1178791522979736, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.707211434841156, + "num_tokens": 448993537.0, + "step": 17363 + }, + { + "epoch": 1.9068745881836153, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.195861339569092, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7237796187400818, + "num_tokens": 449020416.0, + "step": 17364 + }, + { + "epoch": 1.9069844058862289, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.197404623031616, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7388182282447815, + "num_tokens": 449046315.0, + "step": 17365 + }, + { + "epoch": 1.9070942235888424, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4060375690460205, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7053514719009399, + "num_tokens": 449072172.0, + "step": 17366 + }, + { + "epoch": 1.9072040412914562, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4438416957855225, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7293034195899963, + "num_tokens": 449094238.0, + "step": 17367 + }, + { + "epoch": 1.90731385899407, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.314152240753174, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7200876474380493, + "num_tokens": 449118465.0, + "step": 17368 + }, + { + "epoch": 1.9074236766966837, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4193952083587646, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7082526087760925, + "num_tokens": 449140523.0, + "step": 17369 + }, + { + "epoch": 1.9075334943992972, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3208889961242676, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7101381421089172, + "num_tokens": 449167458.0, + "step": 17370 + }, + { + "epoch": 1.9076433121019107, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5418996810913086, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7316817045211792, + "num_tokens": 449187811.0, + "step": 17371 + }, + { + "epoch": 1.9077531298045245, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.207498073577881, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7386418581008911, + "num_tokens": 449212999.0, + "step": 17372 + }, + { + "epoch": 1.9078629475071383, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2375128269195557, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7016037702560425, + "num_tokens": 449238024.0, + "step": 17373 + }, + { + "epoch": 1.9079727652097518, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0454859733581543, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7179880738258362, + "num_tokens": 449266862.0, + "step": 17374 + }, + { + "epoch": 1.9080825829123653, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0557713508605957, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.734582781791687, + "num_tokens": 449296136.0, + "step": 17375 + }, + { + "epoch": 1.908192400614979, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.215182304382324, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7062037587165833, + "num_tokens": 449322698.0, + "step": 17376 + }, + { + "epoch": 1.9083022183175928, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1420371532440186, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.6986050605773926, + "num_tokens": 449351229.0, + "step": 17377 + }, + { + "epoch": 1.9084120360202066, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3358044624328613, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7338326573371887, + "num_tokens": 449374457.0, + "step": 17378 + }, + { + "epoch": 1.9085218537228201, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1354117393493652, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7168368697166443, + "num_tokens": 449401738.0, + "step": 17379 + }, + { + "epoch": 1.9086316714254337, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1844911575317383, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.71498703956604, + "num_tokens": 449428375.0, + "step": 17380 + }, + { + "epoch": 1.9087414891280474, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0647706985473633, + "learning_rate": 1e-06, + "loss": 1.059, + "mean_token_accuracy": 0.6848366856575012, + "num_tokens": 449460367.0, + "step": 17381 + }, + { + "epoch": 1.9088513068306612, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.415147066116333, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7363715171813965, + "num_tokens": 449481939.0, + "step": 17382 + }, + { + "epoch": 1.9089611245332747, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2485110759735107, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.715110182762146, + "num_tokens": 449507779.0, + "step": 17383 + }, + { + "epoch": 1.9090709422358885, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2571468353271484, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7267649173736572, + "num_tokens": 449536049.0, + "step": 17384 + }, + { + "epoch": 1.909180759938502, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5469205379486084, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7364712357521057, + "num_tokens": 449556604.0, + "step": 17385 + }, + { + "epoch": 1.9092905776411158, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.073988437652588, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7236806154251099, + "num_tokens": 449583690.0, + "step": 17386 + }, + { + "epoch": 1.9094003953437295, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.184234619140625, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7348570823669434, + "num_tokens": 449609306.0, + "step": 17387 + }, + { + "epoch": 1.909510213046343, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.079787015914917, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7240723371505737, + "num_tokens": 449637102.0, + "step": 17388 + }, + { + "epoch": 1.9096200307489566, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2751009464263916, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7358027696609497, + "num_tokens": 449659902.0, + "step": 17389 + }, + { + "epoch": 1.9097298484515703, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.22955322265625, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7247340083122253, + "num_tokens": 449685570.0, + "step": 17390 + }, + { + "epoch": 1.909839666154184, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.248504877090454, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.72892165184021, + "num_tokens": 449709286.0, + "step": 17391 + }, + { + "epoch": 1.9099494838567979, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.169434070587158, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.71690434217453, + "num_tokens": 449735615.0, + "step": 17392 + }, + { + "epoch": 1.9100593015594114, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1212921142578125, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7135995626449585, + "num_tokens": 449762331.0, + "step": 17393 + }, + { + "epoch": 1.910169119262025, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.9790599346160889, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7123028635978699, + "num_tokens": 449795113.0, + "step": 17394 + }, + { + "epoch": 1.9102789369646387, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2687532901763916, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7293093800544739, + "num_tokens": 449821302.0, + "step": 17395 + }, + { + "epoch": 1.9103887546672524, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2017734050750732, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.713569164276123, + "num_tokens": 449847719.0, + "step": 17396 + }, + { + "epoch": 1.910498572369866, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1982905864715576, + "learning_rate": 1e-06, + "loss": 0.961, + "mean_token_accuracy": 0.7023900747299194, + "num_tokens": 449874132.0, + "step": 17397 + }, + { + "epoch": 1.9106083900724797, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.6559338569641113, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7321982383728027, + "num_tokens": 449892049.0, + "step": 17398 + }, + { + "epoch": 1.9107182077750933, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.1154730319976807, + "learning_rate": 1e-06, + "loss": 0.9184, + "mean_token_accuracy": 0.7168184518814087, + "num_tokens": 449918827.0, + "step": 17399 + }, + { + "epoch": 1.910828025477707, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.5647528171539307, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7135640382766724, + "num_tokens": 449940306.0, + "step": 17400 + }, + { + "epoch": 1.9109378431803208, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.0638203620910645, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7067182064056396, + "num_tokens": 449968701.0, + "step": 17401 + }, + { + "epoch": 1.9110476608829343, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.3051936626434326, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.6996219158172607, + "num_tokens": 449994837.0, + "step": 17402 + }, + { + "epoch": 1.9111574785855479, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2505600452423096, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7067583799362183, + "num_tokens": 450021966.0, + "step": 17403 + }, + { + "epoch": 1.9112672962881616, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.113355875015259, + "learning_rate": 1e-06, + "loss": 0.9428, + "mean_token_accuracy": 0.707089900970459, + "num_tokens": 450050687.0, + "step": 17404 + }, + { + "epoch": 1.9113771139907754, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.242807626724243, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7220225930213928, + "num_tokens": 450076238.0, + "step": 17405 + }, + { + "epoch": 1.9114869316933891, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.4880475997924805, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.720839262008667, + "num_tokens": 450096935.0, + "step": 17406 + }, + { + "epoch": 1.9115967493960027, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.355731248855591, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7208753228187561, + "num_tokens": 450120989.0, + "step": 17407 + }, + { + "epoch": 1.9117065670986162, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 2.2276594638824463, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7261251211166382, + "num_tokens": 450148570.0, + "step": 17408 + }, + { + "epoch": 1.91181638480123, + "ewc_loss": 1.0132789611816406e-05, + "grad_norm": 1.8411023616790771, + "learning_rate": 1e-06, + "loss": 0.7546, + "mean_token_accuracy": 0.7587372064590454, + "num_tokens": 450180517.0, + "step": 17409 + }, + { + "epoch": 1.9119262025038437, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3718364238739014, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7353131771087646, + "num_tokens": 450204352.0, + "step": 17410 + }, + { + "epoch": 1.9120360202064572, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.5340797901153564, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7204583883285522, + "num_tokens": 450223829.0, + "step": 17411 + }, + { + "epoch": 1.9121458379090708, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.329789161682129, + "learning_rate": 1e-06, + "loss": 0.9269, + "mean_token_accuracy": 0.7134193181991577, + "num_tokens": 450249659.0, + "step": 17412 + }, + { + "epoch": 1.9122556556116845, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3631768226623535, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7414703965187073, + "num_tokens": 450271616.0, + "step": 17413 + }, + { + "epoch": 1.9123654733142983, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3460190296173096, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7407280802726746, + "num_tokens": 450293969.0, + "step": 17414 + }, + { + "epoch": 1.912475291016912, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4438059329986572, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7296586036682129, + "num_tokens": 450317278.0, + "step": 17415 + }, + { + "epoch": 1.9125851087195256, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3443920612335205, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7104389667510986, + "num_tokens": 450343184.0, + "step": 17416 + }, + { + "epoch": 1.9126949264221391, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.6748721599578857, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7481944561004639, + "num_tokens": 450360937.0, + "step": 17417 + }, + { + "epoch": 1.9128047441247529, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.232240676879883, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7194423675537109, + "num_tokens": 450384911.0, + "step": 17418 + }, + { + "epoch": 1.9129145618273666, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.8337562084198, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7485047578811646, + "num_tokens": 450403271.0, + "step": 17419 + }, + { + "epoch": 1.9130243795299804, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3536019325256348, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7416282296180725, + "num_tokens": 450425048.0, + "step": 17420 + }, + { + "epoch": 1.913134197232594, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.171198844909668, + "learning_rate": 1e-06, + "loss": 0.7664, + "mean_token_accuracy": 0.7566871643066406, + "num_tokens": 450449798.0, + "step": 17421 + }, + { + "epoch": 1.9132440149352075, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4422528743743896, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7267152667045593, + "num_tokens": 450471716.0, + "step": 17422 + }, + { + "epoch": 1.9133538326378212, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3427720069885254, + "learning_rate": 1e-06, + "loss": 1.0278, + "mean_token_accuracy": 0.6935936212539673, + "num_tokens": 450499535.0, + "step": 17423 + }, + { + "epoch": 1.913463650340435, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.263176918029785, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7231372594833374, + "num_tokens": 450523739.0, + "step": 17424 + }, + { + "epoch": 1.9135734680430485, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.133495569229126, + "learning_rate": 1e-06, + "loss": 1.0525, + "mean_token_accuracy": 0.6791558265686035, + "num_tokens": 450556393.0, + "step": 17425 + }, + { + "epoch": 1.913683285745662, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4255290031433105, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7286030054092407, + "num_tokens": 450577971.0, + "step": 17426 + }, + { + "epoch": 1.9137931034482758, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 1.991451382637024, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7106143236160278, + "num_tokens": 450609518.0, + "step": 17427 + }, + { + "epoch": 1.9139029211508896, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.5677075386047363, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7372176647186279, + "num_tokens": 450629031.0, + "step": 17428 + }, + { + "epoch": 1.9140127388535033, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.1150875091552734, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7258482575416565, + "num_tokens": 450656033.0, + "step": 17429 + }, + { + "epoch": 1.9141225565561168, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2493627071380615, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7132565379142761, + "num_tokens": 450681630.0, + "step": 17430 + }, + { + "epoch": 1.9142323742587304, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4947144985198975, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7509146928787231, + "num_tokens": 450701588.0, + "step": 17431 + }, + { + "epoch": 1.9143421919613441, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.359327554702759, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7303512692451477, + "num_tokens": 450724736.0, + "step": 17432 + }, + { + "epoch": 1.914452009663958, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3392953872680664, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7266944646835327, + "num_tokens": 450748266.0, + "step": 17433 + }, + { + "epoch": 1.9145618273665717, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.098500967025757, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7253085374832153, + "num_tokens": 450777544.0, + "step": 17434 + }, + { + "epoch": 1.9146716450691852, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2876546382904053, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7251942753791809, + "num_tokens": 450802881.0, + "step": 17435 + }, + { + "epoch": 1.9147814627717987, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.154787540435791, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7072751522064209, + "num_tokens": 450832711.0, + "step": 17436 + }, + { + "epoch": 1.9148912804744125, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.071711301803589, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7106425762176514, + "num_tokens": 450862156.0, + "step": 17437 + }, + { + "epoch": 1.9150010981770262, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.1860511302948, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7024109363555908, + "num_tokens": 450887857.0, + "step": 17438 + }, + { + "epoch": 1.9151109158796398, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.1116621494293213, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7173708081245422, + "num_tokens": 450916891.0, + "step": 17439 + }, + { + "epoch": 1.9152207335822533, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.178286552429199, + "learning_rate": 1e-06, + "loss": 0.9343, + "mean_token_accuracy": 0.7018280029296875, + "num_tokens": 450943925.0, + "step": 17440 + }, + { + "epoch": 1.915330551284867, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.421644687652588, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.748796284198761, + "num_tokens": 450964413.0, + "step": 17441 + }, + { + "epoch": 1.9154403689874808, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2624928951263428, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7212470769882202, + "num_tokens": 450989912.0, + "step": 17442 + }, + { + "epoch": 1.9155501866900946, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.345813512802124, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.714759349822998, + "num_tokens": 451014695.0, + "step": 17443 + }, + { + "epoch": 1.9156600043927081, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.4787654876708984, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.706688642501831, + "num_tokens": 451035290.0, + "step": 17444 + }, + { + "epoch": 1.9157698220953217, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.5346274375915527, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7292666435241699, + "num_tokens": 451057017.0, + "step": 17445 + }, + { + "epoch": 1.9158796397979354, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2894437313079834, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7307648658752441, + "num_tokens": 451081191.0, + "step": 17446 + }, + { + "epoch": 1.9159894575005492, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.1864731311798096, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7288628816604614, + "num_tokens": 451107356.0, + "step": 17447 + }, + { + "epoch": 1.9160992752031627, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 1.934901237487793, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.707944393157959, + "num_tokens": 451142762.0, + "step": 17448 + }, + { + "epoch": 1.9162090929057765, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.1593515872955322, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7258690595626831, + "num_tokens": 451170603.0, + "step": 17449 + }, + { + "epoch": 1.91631891060839, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.134463310241699, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7062507271766663, + "num_tokens": 451200087.0, + "step": 17450 + }, + { + "epoch": 1.9164287283110037, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3326258659362793, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.70917809009552, + "num_tokens": 451223252.0, + "step": 17451 + }, + { + "epoch": 1.9165385460136175, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3800370693206787, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.7149250507354736, + "num_tokens": 451245659.0, + "step": 17452 + }, + { + "epoch": 1.916648363716231, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.378314733505249, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7372406125068665, + "num_tokens": 451267037.0, + "step": 17453 + }, + { + "epoch": 1.9167581814188446, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.5551788806915283, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7260111570358276, + "num_tokens": 451288055.0, + "step": 17454 + }, + { + "epoch": 1.9168679991214583, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.5806047916412354, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.707552433013916, + "num_tokens": 451309635.0, + "step": 17455 + }, + { + "epoch": 1.916977816824072, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.009799003601074, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7201046943664551, + "num_tokens": 451341564.0, + "step": 17456 + }, + { + "epoch": 1.9170876345266858, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.097505569458008, + "learning_rate": 1e-06, + "loss": 1.0168, + "mean_token_accuracy": 0.6952619552612305, + "num_tokens": 451369371.0, + "step": 17457 + }, + { + "epoch": 1.9171974522292994, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2557222843170166, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7282257080078125, + "num_tokens": 451393862.0, + "step": 17458 + }, + { + "epoch": 1.917307269931913, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.2203450202941895, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.715523362159729, + "num_tokens": 451421284.0, + "step": 17459 + }, + { + "epoch": 1.9174170876345267, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.3597519397735596, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7104151248931885, + "num_tokens": 451445744.0, + "step": 17460 + }, + { + "epoch": 1.9175269053371404, + "ewc_loss": 1.0192394256591797e-05, + "grad_norm": 2.272305488586426, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7262860536575317, + "num_tokens": 451469195.0, + "step": 17461 + }, + { + "epoch": 1.917636723039754, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3650600910186768, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7219275236129761, + "num_tokens": 451492045.0, + "step": 17462 + }, + { + "epoch": 1.9177465407423677, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4564027786254883, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7203510403633118, + "num_tokens": 451514767.0, + "step": 17463 + }, + { + "epoch": 1.9178563584449813, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1940557956695557, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.724001407623291, + "num_tokens": 451541390.0, + "step": 17464 + }, + { + "epoch": 1.917966176147595, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2544069290161133, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7248390913009644, + "num_tokens": 451565160.0, + "step": 17465 + }, + { + "epoch": 1.9180759938502088, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4343323707580566, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7093381881713867, + "num_tokens": 451587544.0, + "step": 17466 + }, + { + "epoch": 1.9181858115528223, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3735806941986084, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7460631132125854, + "num_tokens": 451612067.0, + "step": 17467 + }, + { + "epoch": 1.9182956292554358, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1723031997680664, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7371381521224976, + "num_tokens": 451638093.0, + "step": 17468 + }, + { + "epoch": 1.9184054469580496, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1410090923309326, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7223803997039795, + "num_tokens": 451665022.0, + "step": 17469 + }, + { + "epoch": 1.9185152646606634, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.227551221847534, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7100303173065186, + "num_tokens": 451692246.0, + "step": 17470 + }, + { + "epoch": 1.9186250823632771, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.317600965499878, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7459805607795715, + "num_tokens": 451715934.0, + "step": 17471 + }, + { + "epoch": 1.9187349000658906, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9530071020126343, + "learning_rate": 1e-06, + "loss": 1.0032, + "mean_token_accuracy": 0.6942018270492554, + "num_tokens": 451751060.0, + "step": 17472 + }, + { + "epoch": 1.9188447177685042, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.969704031944275, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7114561796188354, + "num_tokens": 451782865.0, + "step": 17473 + }, + { + "epoch": 1.918954535471118, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0518434047698975, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7265349626541138, + "num_tokens": 451812176.0, + "step": 17474 + }, + { + "epoch": 1.9190643531737317, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2329001426696777, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7395034432411194, + "num_tokens": 451835676.0, + "step": 17475 + }, + { + "epoch": 1.9191741708763452, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0646533966064453, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7151170969009399, + "num_tokens": 451864951.0, + "step": 17476 + }, + { + "epoch": 1.9192839885789588, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.284581422805786, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7186295986175537, + "num_tokens": 451888418.0, + "step": 17477 + }, + { + "epoch": 1.9193938062815725, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.067664861679077, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7031087875366211, + "num_tokens": 451917245.0, + "step": 17478 + }, + { + "epoch": 1.9195036239841863, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2895314693450928, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7516288757324219, + "num_tokens": 451940714.0, + "step": 17479 + }, + { + "epoch": 1.9196134416868, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1077709197998047, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7106688022613525, + "num_tokens": 451971952.0, + "step": 17480 + }, + { + "epoch": 1.9197232593894136, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.116987705230713, + "learning_rate": 1e-06, + "loss": 0.9776, + "mean_token_accuracy": 0.7030125260353088, + "num_tokens": 452002024.0, + "step": 17481 + }, + { + "epoch": 1.919833077092027, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.175874948501587, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7216985821723938, + "num_tokens": 452029407.0, + "step": 17482 + }, + { + "epoch": 1.9199428947946409, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0861730575561523, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7150100469589233, + "num_tokens": 452056850.0, + "step": 17483 + }, + { + "epoch": 1.9200527124972546, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1187055110931396, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7154639959335327, + "num_tokens": 452084598.0, + "step": 17484 + }, + { + "epoch": 1.9201625301998684, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1114094257354736, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7294994592666626, + "num_tokens": 452112922.0, + "step": 17485 + }, + { + "epoch": 1.920272347902482, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0366597175598145, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7095410823822021, + "num_tokens": 452142665.0, + "step": 17486 + }, + { + "epoch": 1.9203821656050954, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.319040298461914, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7236106395721436, + "num_tokens": 452165606.0, + "step": 17487 + }, + { + "epoch": 1.9204919833077092, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.5224006175994873, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7474747896194458, + "num_tokens": 452184198.0, + "step": 17488 + }, + { + "epoch": 1.920601801010323, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4634006023406982, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7202631831169128, + "num_tokens": 452206770.0, + "step": 17489 + }, + { + "epoch": 1.9207116187129365, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2341480255126953, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7167198061943054, + "num_tokens": 452233802.0, + "step": 17490 + }, + { + "epoch": 1.92082143641555, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3929603099823, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7245514392852783, + "num_tokens": 452257788.0, + "step": 17491 + }, + { + "epoch": 1.9209312541181638, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.057337522506714, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.714246392250061, + "num_tokens": 452285869.0, + "step": 17492 + }, + { + "epoch": 1.9210410718207775, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.937561273574829, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7261379957199097, + "num_tokens": 452317624.0, + "step": 17493 + }, + { + "epoch": 1.9211508895233913, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2834134101867676, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7249009609222412, + "num_tokens": 452341393.0, + "step": 17494 + }, + { + "epoch": 1.9212607072260048, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3246517181396484, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7206476926803589, + "num_tokens": 452366505.0, + "step": 17495 + }, + { + "epoch": 1.9213705249286184, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.474299669265747, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7431261539459229, + "num_tokens": 452387653.0, + "step": 17496 + }, + { + "epoch": 1.9214803426312321, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.344667911529541, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7395370006561279, + "num_tokens": 452410836.0, + "step": 17497 + }, + { + "epoch": 1.9215901603338459, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9813714027404785, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7144858837127686, + "num_tokens": 452441639.0, + "step": 17498 + }, + { + "epoch": 1.9216999780364594, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.486659288406372, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7180256843566895, + "num_tokens": 452462803.0, + "step": 17499 + }, + { + "epoch": 1.9218097957390732, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.32430362701416, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7371946573257446, + "num_tokens": 452485567.0, + "step": 17500 + }, + { + "epoch": 1.9219196134416867, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.245417833328247, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7049821615219116, + "num_tokens": 452512177.0, + "step": 17501 + }, + { + "epoch": 1.9220294311443005, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.068049669265747, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.726408839225769, + "num_tokens": 452540671.0, + "step": 17502 + }, + { + "epoch": 1.9221392488469142, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9996449947357178, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7203559875488281, + "num_tokens": 452572954.0, + "step": 17503 + }, + { + "epoch": 1.9222490665495278, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4514665603637695, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7167288661003113, + "num_tokens": 452595664.0, + "step": 17504 + }, + { + "epoch": 1.9223588842521413, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.5083212852478027, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.73005211353302, + "num_tokens": 452616085.0, + "step": 17505 + }, + { + "epoch": 1.922468701954755, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.483328342437744, + "learning_rate": 1e-06, + "loss": 0.954, + "mean_token_accuracy": 0.7037777304649353, + "num_tokens": 452641245.0, + "step": 17506 + }, + { + "epoch": 1.9225785196573688, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2714016437530518, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7117782831192017, + "num_tokens": 452666302.0, + "step": 17507 + }, + { + "epoch": 1.9226883373599826, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.295813798904419, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7270538210868835, + "num_tokens": 452689351.0, + "step": 17508 + }, + { + "epoch": 1.922798155062596, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.610450029373169, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7298498749732971, + "num_tokens": 452709981.0, + "step": 17509 + }, + { + "epoch": 1.9229079727652096, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0263278484344482, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7168123722076416, + "num_tokens": 452739967.0, + "step": 17510 + }, + { + "epoch": 1.9230177904678234, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.239131450653076, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7259410619735718, + "num_tokens": 452766530.0, + "step": 17511 + }, + { + "epoch": 1.9231276081704372, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1328866481781006, + "learning_rate": 1e-06, + "loss": 1.0078, + "mean_token_accuracy": 0.6966404914855957, + "num_tokens": 452795231.0, + "step": 17512 + }, + { + "epoch": 1.9232374258730507, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1162092685699463, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7163535356521606, + "num_tokens": 452823578.0, + "step": 17513 + }, + { + "epoch": 1.9233472435756644, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0090267658233643, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7104218006134033, + "num_tokens": 452855213.0, + "step": 17514 + }, + { + "epoch": 1.923457061278278, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2012135982513428, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6917709112167358, + "num_tokens": 452885761.0, + "step": 17515 + }, + { + "epoch": 1.9235668789808917, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2862725257873535, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7343676686286926, + "num_tokens": 452909656.0, + "step": 17516 + }, + { + "epoch": 1.9236766966835055, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9317245483398438, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7212927937507629, + "num_tokens": 452945783.0, + "step": 17517 + }, + { + "epoch": 1.923786514386119, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.341736078262329, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7262002229690552, + "num_tokens": 452967176.0, + "step": 17518 + }, + { + "epoch": 1.9238963320887326, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.350041151046753, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.7018788456916809, + "num_tokens": 452991130.0, + "step": 17519 + }, + { + "epoch": 1.9240061497913463, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1321139335632324, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7320773005485535, + "num_tokens": 453017318.0, + "step": 17520 + }, + { + "epoch": 1.92411596749396, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3718740940093994, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7055045366287231, + "num_tokens": 453042097.0, + "step": 17521 + }, + { + "epoch": 1.9242257851965738, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1363749504089355, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7105825543403625, + "num_tokens": 453070639.0, + "step": 17522 + }, + { + "epoch": 1.9243356028991874, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3557515144348145, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7433542013168335, + "num_tokens": 453092593.0, + "step": 17523 + }, + { + "epoch": 1.924445420601801, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9570788145065308, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7440534234046936, + "num_tokens": 453121695.0, + "step": 17524 + }, + { + "epoch": 1.9245552383044147, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.072479248046875, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7079349756240845, + "num_tokens": 453150979.0, + "step": 17525 + }, + { + "epoch": 1.9246650560070284, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.909409761428833, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7055824995040894, + "num_tokens": 453184250.0, + "step": 17526 + }, + { + "epoch": 1.924774873709642, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.027318239212036, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6891523003578186, + "num_tokens": 453215976.0, + "step": 17527 + }, + { + "epoch": 1.9248846914122557, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.322817325592041, + "learning_rate": 1e-06, + "loss": 0.7699, + "mean_token_accuracy": 0.7508933544158936, + "num_tokens": 453239722.0, + "step": 17528 + }, + { + "epoch": 1.9249945091148692, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1435317993164062, + "learning_rate": 1e-06, + "loss": 0.955, + "mean_token_accuracy": 0.703545331954956, + "num_tokens": 453267887.0, + "step": 17529 + }, + { + "epoch": 1.925104326817483, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.069296360015869, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7084903120994568, + "num_tokens": 453297401.0, + "step": 17530 + }, + { + "epoch": 1.9252141445200968, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9501479864120483, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.722161591053009, + "num_tokens": 453328276.0, + "step": 17531 + }, + { + "epoch": 1.9253239622227103, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0171561241149902, + "learning_rate": 1e-06, + "loss": 0.9894, + "mean_token_accuracy": 0.7040454745292664, + "num_tokens": 453361354.0, + "step": 17532 + }, + { + "epoch": 1.9254337799253238, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2699999809265137, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7249597311019897, + "num_tokens": 453385604.0, + "step": 17533 + }, + { + "epoch": 1.9255435976279376, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.185070753097534, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7289183139801025, + "num_tokens": 453411241.0, + "step": 17534 + }, + { + "epoch": 1.9256534153305513, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.385453224182129, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.751203179359436, + "num_tokens": 453432239.0, + "step": 17535 + }, + { + "epoch": 1.925763233033165, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.06172513961792, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7173372507095337, + "num_tokens": 453461818.0, + "step": 17536 + }, + { + "epoch": 1.9258730507357786, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.6654913425445557, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7302573919296265, + "num_tokens": 453482038.0, + "step": 17537 + }, + { + "epoch": 1.9259828684383922, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1578123569488525, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7282809019088745, + "num_tokens": 453508784.0, + "step": 17538 + }, + { + "epoch": 1.926092686141006, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.019505262374878, + "learning_rate": 1e-06, + "loss": 1.0187, + "mean_token_accuracy": 0.6964076161384583, + "num_tokens": 453538687.0, + "step": 17539 + }, + { + "epoch": 1.9262025038436197, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2176425457000732, + "learning_rate": 1e-06, + "loss": 0.971, + "mean_token_accuracy": 0.7002868056297302, + "num_tokens": 453564911.0, + "step": 17540 + }, + { + "epoch": 1.9263123215462332, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9282668828964233, + "learning_rate": 1e-06, + "loss": 0.9822, + "mean_token_accuracy": 0.6952396035194397, + "num_tokens": 453599779.0, + "step": 17541 + }, + { + "epoch": 1.9264221392488468, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2512471675872803, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.710866391658783, + "num_tokens": 453624549.0, + "step": 17542 + }, + { + "epoch": 1.9265319569514605, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3180127143859863, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7220690250396729, + "num_tokens": 453648337.0, + "step": 17543 + }, + { + "epoch": 1.9266417746540743, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1706044673919678, + "learning_rate": 1e-06, + "loss": 0.7621, + "mean_token_accuracy": 0.752569317817688, + "num_tokens": 453675954.0, + "step": 17544 + }, + { + "epoch": 1.926751592356688, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3818347454071045, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.6990417242050171, + "num_tokens": 453702509.0, + "step": 17545 + }, + { + "epoch": 1.9268614100593016, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0686779022216797, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7373045086860657, + "num_tokens": 453730629.0, + "step": 17546 + }, + { + "epoch": 1.926971227761915, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4407365322113037, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7293789982795715, + "num_tokens": 453752699.0, + "step": 17547 + }, + { + "epoch": 1.9270810454645289, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.285926342010498, + "learning_rate": 1e-06, + "loss": 0.9909, + "mean_token_accuracy": 0.6946567296981812, + "num_tokens": 453779843.0, + "step": 17548 + }, + { + "epoch": 1.9271908631671426, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1633312702178955, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7263238430023193, + "num_tokens": 453805125.0, + "step": 17549 + }, + { + "epoch": 1.9273006808697564, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0473520755767822, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7069534659385681, + "num_tokens": 453834983.0, + "step": 17550 + }, + { + "epoch": 1.92741049857237, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1927192211151123, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7332068681716919, + "num_tokens": 453861241.0, + "step": 17551 + }, + { + "epoch": 1.9275203162749834, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2705130577087402, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7146327495574951, + "num_tokens": 453887612.0, + "step": 17552 + }, + { + "epoch": 1.9276301339775972, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4713480472564697, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7443315982818604, + "num_tokens": 453909501.0, + "step": 17553 + }, + { + "epoch": 1.927739951680211, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.5790576934814453, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7337819337844849, + "num_tokens": 453930561.0, + "step": 17554 + }, + { + "epoch": 1.9278497693828245, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.073234796524048, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.6934642791748047, + "num_tokens": 453960082.0, + "step": 17555 + }, + { + "epoch": 1.927959587085438, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.8308908939361572, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7179611325263977, + "num_tokens": 453994622.0, + "step": 17556 + }, + { + "epoch": 1.9280694047880518, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1991662979125977, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7371948957443237, + "num_tokens": 454019594.0, + "step": 17557 + }, + { + "epoch": 1.9281792224906655, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3722355365753174, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7510802745819092, + "num_tokens": 454042410.0, + "step": 17558 + }, + { + "epoch": 1.9282890401932793, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9544146060943604, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7287046313285828, + "num_tokens": 454074885.0, + "step": 17559 + }, + { + "epoch": 1.9283988578958928, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3060224056243896, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7052795886993408, + "num_tokens": 454099407.0, + "step": 17560 + }, + { + "epoch": 1.9285086755985064, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1796538829803467, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7230392098426819, + "num_tokens": 454125813.0, + "step": 17561 + }, + { + "epoch": 1.9286184933011201, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.102297067642212, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7479887008666992, + "num_tokens": 454152954.0, + "step": 17562 + }, + { + "epoch": 1.9287283110037339, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2061474323272705, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7268561124801636, + "num_tokens": 454179351.0, + "step": 17563 + }, + { + "epoch": 1.9288381287063474, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2015347480773926, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7147922515869141, + "num_tokens": 454205510.0, + "step": 17564 + }, + { + "epoch": 1.9289479464089612, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3823046684265137, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7334526777267456, + "num_tokens": 454228764.0, + "step": 17565 + }, + { + "epoch": 1.9290577641115747, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.564218521118164, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7353656888008118, + "num_tokens": 454248196.0, + "step": 17566 + }, + { + "epoch": 1.9291675818141885, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.361923933029175, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7341050505638123, + "num_tokens": 454269712.0, + "step": 17567 + }, + { + "epoch": 1.9292773995168022, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1701481342315674, + "learning_rate": 1e-06, + "loss": 0.9628, + "mean_token_accuracy": 0.7010427117347717, + "num_tokens": 454297607.0, + "step": 17568 + }, + { + "epoch": 1.9293872172194158, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4037158489227295, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7224225997924805, + "num_tokens": 454321168.0, + "step": 17569 + }, + { + "epoch": 1.9294970349220293, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9908279180526733, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7188948392868042, + "num_tokens": 454352864.0, + "step": 17570 + }, + { + "epoch": 1.929606852624643, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2341156005859375, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7335407137870789, + "num_tokens": 454377003.0, + "step": 17571 + }, + { + "epoch": 1.9297166703272568, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3050196170806885, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7094682455062866, + "num_tokens": 454403443.0, + "step": 17572 + }, + { + "epoch": 1.9298264880298706, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1788134574890137, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7260297536849976, + "num_tokens": 454428688.0, + "step": 17573 + }, + { + "epoch": 1.929936305732484, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.422506809234619, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.7482061386108398, + "num_tokens": 454448632.0, + "step": 17574 + }, + { + "epoch": 1.9300461234350976, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0605530738830566, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7286628484725952, + "num_tokens": 454478079.0, + "step": 17575 + }, + { + "epoch": 1.9301559411377114, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2262189388275146, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7313175201416016, + "num_tokens": 454503560.0, + "step": 17576 + }, + { + "epoch": 1.9302657588403251, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.148469924926758, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.706911563873291, + "num_tokens": 454532357.0, + "step": 17577 + }, + { + "epoch": 1.9303755765429387, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.073120594024658, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7129770517349243, + "num_tokens": 454561608.0, + "step": 17578 + }, + { + "epoch": 1.9304853942455524, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1679303646087646, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.7140146493911743, + "num_tokens": 454590535.0, + "step": 17579 + }, + { + "epoch": 1.930595211948166, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3650784492492676, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7402059435844421, + "num_tokens": 454613021.0, + "step": 17580 + }, + { + "epoch": 1.9307050296507797, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.282116651535034, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7041107416152954, + "num_tokens": 454638183.0, + "step": 17581 + }, + { + "epoch": 1.9308148473533935, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1717147827148438, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7252991199493408, + "num_tokens": 454664254.0, + "step": 17582 + }, + { + "epoch": 1.930924665056007, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.864824891090393, + "learning_rate": 1e-06, + "loss": 1.0036, + "mean_token_accuracy": 0.698131799697876, + "num_tokens": 454702713.0, + "step": 17583 + }, + { + "epoch": 1.9310344827586206, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.115506172180176, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.714751124382019, + "num_tokens": 454729770.0, + "step": 17584 + }, + { + "epoch": 1.9311443004612343, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.163184881210327, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7229286432266235, + "num_tokens": 454754170.0, + "step": 17585 + }, + { + "epoch": 1.931254118163848, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.277536392211914, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7137365341186523, + "num_tokens": 454779132.0, + "step": 17586 + }, + { + "epoch": 1.9313639358664618, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2931792736053467, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7091578245162964, + "num_tokens": 454803807.0, + "step": 17587 + }, + { + "epoch": 1.9314737535690754, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.22503662109375, + "learning_rate": 1e-06, + "loss": 0.9871, + "mean_token_accuracy": 0.703689455986023, + "num_tokens": 454831878.0, + "step": 17588 + }, + { + "epoch": 1.931583571271689, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1892735958099365, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7122644782066345, + "num_tokens": 454859770.0, + "step": 17589 + }, + { + "epoch": 1.9316933889743026, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.640398383140564, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7072357535362244, + "num_tokens": 454905089.0, + "step": 17590 + }, + { + "epoch": 1.9318032066769164, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.101681709289551, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7190603613853455, + "num_tokens": 454934153.0, + "step": 17591 + }, + { + "epoch": 1.93191302437953, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0136659145355225, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6891294717788696, + "num_tokens": 454965708.0, + "step": 17592 + }, + { + "epoch": 1.9320228420821435, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.12514591217041, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7019845247268677, + "num_tokens": 454994316.0, + "step": 17593 + }, + { + "epoch": 1.9321326597847572, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3960516452789307, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7342446446418762, + "num_tokens": 455016511.0, + "step": 17594 + }, + { + "epoch": 1.932242477487371, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4748191833496094, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7536061406135559, + "num_tokens": 455036822.0, + "step": 17595 + }, + { + "epoch": 1.9323522951899847, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.110624313354492, + "learning_rate": 1e-06, + "loss": 0.9259, + "mean_token_accuracy": 0.7223713397979736, + "num_tokens": 455065624.0, + "step": 17596 + }, + { + "epoch": 1.9324621128925983, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2602741718292236, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7176787853240967, + "num_tokens": 455092259.0, + "step": 17597 + }, + { + "epoch": 1.9325719305952118, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.512960910797119, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7213833332061768, + "num_tokens": 455114492.0, + "step": 17598 + }, + { + "epoch": 1.9326817482978256, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.411071538925171, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7385778427124023, + "num_tokens": 455135634.0, + "step": 17599 + }, + { + "epoch": 1.9327915660004393, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0604610443115234, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7177375555038452, + "num_tokens": 455167182.0, + "step": 17600 + }, + { + "epoch": 1.932901383703053, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.305231809616089, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7182313203811646, + "num_tokens": 455191465.0, + "step": 17601 + }, + { + "epoch": 1.9330112014056666, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9239990711212158, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7210001945495605, + "num_tokens": 455223966.0, + "step": 17602 + }, + { + "epoch": 1.9331210191082802, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.091787815093994, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7185947895050049, + "num_tokens": 455250006.0, + "step": 17603 + }, + { + "epoch": 1.933230836810894, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3878445625305176, + "learning_rate": 1e-06, + "loss": 0.9158, + "mean_token_accuracy": 0.7132973670959473, + "num_tokens": 455273183.0, + "step": 17604 + }, + { + "epoch": 1.9333406545135077, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2478952407836914, + "learning_rate": 1e-06, + "loss": 0.9649, + "mean_token_accuracy": 0.705285906791687, + "num_tokens": 455302408.0, + "step": 17605 + }, + { + "epoch": 1.9334504722161212, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.182859420776367, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7152299284934998, + "num_tokens": 455330213.0, + "step": 17606 + }, + { + "epoch": 1.9335602899187347, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4170355796813965, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.6999485492706299, + "num_tokens": 455354541.0, + "step": 17607 + }, + { + "epoch": 1.9336701076213485, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.128490447998047, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7087787389755249, + "num_tokens": 455382122.0, + "step": 17608 + }, + { + "epoch": 1.9337799253239623, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1875789165496826, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7122886180877686, + "num_tokens": 455408109.0, + "step": 17609 + }, + { + "epoch": 1.933889743026576, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.458759069442749, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7317860126495361, + "num_tokens": 455429011.0, + "step": 17610 + }, + { + "epoch": 1.9339995607291895, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.290423631668091, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7227512001991272, + "num_tokens": 455453465.0, + "step": 17611 + }, + { + "epoch": 1.934109378431803, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9791566133499146, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7141779661178589, + "num_tokens": 455484298.0, + "step": 17612 + }, + { + "epoch": 1.9342191961344168, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1878767013549805, + "learning_rate": 1e-06, + "loss": 0.9432, + "mean_token_accuracy": 0.706870973110199, + "num_tokens": 455510085.0, + "step": 17613 + }, + { + "epoch": 1.9343290138370306, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.273064136505127, + "learning_rate": 1e-06, + "loss": 0.981, + "mean_token_accuracy": 0.7043904662132263, + "num_tokens": 455534673.0, + "step": 17614 + }, + { + "epoch": 1.9344388315396444, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.277484178543091, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7153393030166626, + "num_tokens": 455559873.0, + "step": 17615 + }, + { + "epoch": 1.934548649242258, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.089210271835327, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7165724039077759, + "num_tokens": 455588007.0, + "step": 17616 + }, + { + "epoch": 1.9346584669448714, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2102346420288086, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7403063774108887, + "num_tokens": 455612967.0, + "step": 17617 + }, + { + "epoch": 1.9347682846474852, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.157550096511841, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.715351939201355, + "num_tokens": 455641419.0, + "step": 17618 + }, + { + "epoch": 1.934878102350099, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.407618761062622, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7258386015892029, + "num_tokens": 455663530.0, + "step": 17619 + }, + { + "epoch": 1.9349879200527125, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1501221656799316, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7218738198280334, + "num_tokens": 455690929.0, + "step": 17620 + }, + { + "epoch": 1.935097737755326, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2279653549194336, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7237160205841064, + "num_tokens": 455716597.0, + "step": 17621 + }, + { + "epoch": 1.9352075554579398, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1233060359954834, + "learning_rate": 1e-06, + "loss": 0.9605, + "mean_token_accuracy": 0.7072385549545288, + "num_tokens": 455745358.0, + "step": 17622 + }, + { + "epoch": 1.9353173731605535, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.047947883605957, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7042762637138367, + "num_tokens": 455773826.0, + "step": 17623 + }, + { + "epoch": 1.9354271908631673, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.063804864883423, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7015661597251892, + "num_tokens": 455802065.0, + "step": 17624 + }, + { + "epoch": 1.9355370085657808, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.0226123332977295, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.717902660369873, + "num_tokens": 455831596.0, + "step": 17625 + }, + { + "epoch": 1.9356468262683943, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.048785448074341, + "learning_rate": 1e-06, + "loss": 0.9867, + "mean_token_accuracy": 0.6994075775146484, + "num_tokens": 455860969.0, + "step": 17626 + }, + { + "epoch": 1.935756643971008, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3746540546417236, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7200742959976196, + "num_tokens": 455884957.0, + "step": 17627 + }, + { + "epoch": 1.9358664616736219, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.1566014289855957, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.6986057162284851, + "num_tokens": 455912229.0, + "step": 17628 + }, + { + "epoch": 1.9359762793762354, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.175614356994629, + "learning_rate": 1e-06, + "loss": 0.9885, + "mean_token_accuracy": 0.7011898756027222, + "num_tokens": 455939141.0, + "step": 17629 + }, + { + "epoch": 1.9360860970788492, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 1.9997460842132568, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7066700458526611, + "num_tokens": 455970042.0, + "step": 17630 + }, + { + "epoch": 1.9361959147814627, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.4313406944274902, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7379367351531982, + "num_tokens": 455991574.0, + "step": 17631 + }, + { + "epoch": 1.9363057324840764, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.3273873329162598, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.728147029876709, + "num_tokens": 456014313.0, + "step": 17632 + }, + { + "epoch": 1.9364155501866902, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2966256141662598, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7355655431747437, + "num_tokens": 456038844.0, + "step": 17633 + }, + { + "epoch": 1.9365253678893037, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.145230770111084, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7092623710632324, + "num_tokens": 456065470.0, + "step": 17634 + }, + { + "epoch": 1.9366351855919173, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.073596239089966, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7266927361488342, + "num_tokens": 456094501.0, + "step": 17635 + }, + { + "epoch": 1.936745003294531, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.2237443923950195, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.718111515045166, + "num_tokens": 456121203.0, + "step": 17636 + }, + { + "epoch": 1.9368548209971448, + "ewc_loss": 1.0251998901367188e-05, + "grad_norm": 2.169909954071045, + "learning_rate": 1e-06, + "loss": 0.9876, + "mean_token_accuracy": 0.700354814529419, + "num_tokens": 456147979.0, + "step": 17637 + }, + { + "epoch": 1.9369646386997585, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.3422305583953857, + "learning_rate": 1e-06, + "loss": 0.7666, + "mean_token_accuracy": 0.7488055229187012, + "num_tokens": 456169219.0, + "step": 17638 + }, + { + "epoch": 1.937074456402372, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1057567596435547, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7082551717758179, + "num_tokens": 456197959.0, + "step": 17639 + }, + { + "epoch": 1.9371842741049856, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1423916816711426, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7176574468612671, + "num_tokens": 456225101.0, + "step": 17640 + }, + { + "epoch": 1.9372940918075994, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2488770484924316, + "learning_rate": 1e-06, + "loss": 0.9847, + "mean_token_accuracy": 0.70462965965271, + "num_tokens": 456251271.0, + "step": 17641 + }, + { + "epoch": 1.9374039095102131, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1770613193511963, + "learning_rate": 1e-06, + "loss": 0.9618, + "mean_token_accuracy": 0.7025631666183472, + "num_tokens": 456279573.0, + "step": 17642 + }, + { + "epoch": 1.9375137272128267, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.265582799911499, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7218527793884277, + "num_tokens": 456304106.0, + "step": 17643 + }, + { + "epoch": 1.9376235449154404, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.4552066326141357, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7142081260681152, + "num_tokens": 456326000.0, + "step": 17644 + }, + { + "epoch": 1.937733362618054, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.076352119445801, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.7034989595413208, + "num_tokens": 456357308.0, + "step": 17645 + }, + { + "epoch": 1.9378431803206677, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.332667589187622, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7154827117919922, + "num_tokens": 456379764.0, + "step": 17646 + }, + { + "epoch": 1.9379529980232815, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.216453790664673, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7319562435150146, + "num_tokens": 456404692.0, + "step": 17647 + }, + { + "epoch": 1.938062815725895, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.272465229034424, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7049760222434998, + "num_tokens": 456429190.0, + "step": 17648 + }, + { + "epoch": 1.9381726334285085, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.0100514888763428, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7187480926513672, + "num_tokens": 456459855.0, + "step": 17649 + }, + { + "epoch": 1.9382824511311223, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.9516741037368774, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7157804369926453, + "num_tokens": 456489508.0, + "step": 17650 + }, + { + "epoch": 1.938392268833736, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2373740673065186, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7067214250564575, + "num_tokens": 456514346.0, + "step": 17651 + }, + { + "epoch": 1.9385020865363498, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.9365200996398926, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.6939053535461426, + "num_tokens": 456547450.0, + "step": 17652 + }, + { + "epoch": 1.9386119042389633, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.662024974822998, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7467833757400513, + "num_tokens": 456565621.0, + "step": 17653 + }, + { + "epoch": 1.9387217219415769, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.5331544876098633, + "learning_rate": 1e-06, + "loss": 0.9361, + "mean_token_accuracy": 0.7121312618255615, + "num_tokens": 456585911.0, + "step": 17654 + }, + { + "epoch": 1.9388315396441906, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1439454555511475, + "learning_rate": 1e-06, + "loss": 1.0071, + "mean_token_accuracy": 0.7023340463638306, + "num_tokens": 456614068.0, + "step": 17655 + }, + { + "epoch": 1.9389413573468044, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.9174622297286987, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.713799238204956, + "num_tokens": 456646528.0, + "step": 17656 + }, + { + "epoch": 1.939051175049418, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.243758201599121, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.7032949924468994, + "num_tokens": 456672194.0, + "step": 17657 + }, + { + "epoch": 1.9391609927520315, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.4652092456817627, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7197368144989014, + "num_tokens": 456692698.0, + "step": 17658 + }, + { + "epoch": 1.9392708104546452, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.305516242980957, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7165853381156921, + "num_tokens": 456715664.0, + "step": 17659 + }, + { + "epoch": 1.939380628157259, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1229562759399414, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.724382221698761, + "num_tokens": 456741854.0, + "step": 17660 + }, + { + "epoch": 1.9394904458598727, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.111898422241211, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7108458280563354, + "num_tokens": 456768364.0, + "step": 17661 + }, + { + "epoch": 1.9396002635624863, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.057375431060791, + "learning_rate": 1e-06, + "loss": 1.0162, + "mean_token_accuracy": 0.6950404644012451, + "num_tokens": 456799532.0, + "step": 17662 + }, + { + "epoch": 1.9397100812650998, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2835261821746826, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7420389652252197, + "num_tokens": 456823868.0, + "step": 17663 + }, + { + "epoch": 1.9398198989677136, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.681396484375, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7503852844238281, + "num_tokens": 456843968.0, + "step": 17664 + }, + { + "epoch": 1.9399297166703273, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.6702284812927246, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7332233190536499, + "num_tokens": 456862883.0, + "step": 17665 + }, + { + "epoch": 1.940039534372941, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.339992046356201, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7173094749450684, + "num_tokens": 456886445.0, + "step": 17666 + }, + { + "epoch": 1.9401493520755546, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.4761037826538086, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7221814393997192, + "num_tokens": 456908085.0, + "step": 17667 + }, + { + "epoch": 1.9402591697781681, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1430885791778564, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.723048985004425, + "num_tokens": 456937381.0, + "step": 17668 + }, + { + "epoch": 1.940368987480782, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.27571177482605, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7308200597763062, + "num_tokens": 456961903.0, + "step": 17669 + }, + { + "epoch": 1.9404788051833957, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.300571918487549, + "learning_rate": 1e-06, + "loss": 1.0096, + "mean_token_accuracy": 0.696121096611023, + "num_tokens": 456986959.0, + "step": 17670 + }, + { + "epoch": 1.9405886228860092, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.547098398208618, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7146631479263306, + "num_tokens": 457006516.0, + "step": 17671 + }, + { + "epoch": 1.9406984405886227, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.8890403509140015, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7315462827682495, + "num_tokens": 457038982.0, + "step": 17672 + }, + { + "epoch": 1.9408082582912365, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.944517731666565, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7296477556228638, + "num_tokens": 457070470.0, + "step": 17673 + }, + { + "epoch": 1.9409180759938502, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.5618736743927, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7404120564460754, + "num_tokens": 457090603.0, + "step": 17674 + }, + { + "epoch": 1.941027893696464, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.318244695663452, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7208932042121887, + "num_tokens": 457116720.0, + "step": 17675 + }, + { + "epoch": 1.9411377113990775, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.4307706356048584, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7392553687095642, + "num_tokens": 457138482.0, + "step": 17676 + }, + { + "epoch": 1.941247529101691, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2226927280426025, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7293886542320251, + "num_tokens": 457162724.0, + "step": 17677 + }, + { + "epoch": 1.9413573468043048, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.261535882949829, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7331188917160034, + "num_tokens": 457187631.0, + "step": 17678 + }, + { + "epoch": 1.9414671645069186, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.094911575317383, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7168827652931213, + "num_tokens": 457217336.0, + "step": 17679 + }, + { + "epoch": 1.9415769822095323, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.328340768814087, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7382147312164307, + "num_tokens": 457239708.0, + "step": 17680 + }, + { + "epoch": 1.9416867999121459, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1374399662017822, + "learning_rate": 1e-06, + "loss": 1.0091, + "mean_token_accuracy": 0.6941540837287903, + "num_tokens": 457273370.0, + "step": 17681 + }, + { + "epoch": 1.9417966176147594, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 1.997497797012329, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7216411232948303, + "num_tokens": 457303652.0, + "step": 17682 + }, + { + "epoch": 1.9419064353173732, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.293229341506958, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7177746295928955, + "num_tokens": 457329030.0, + "step": 17683 + }, + { + "epoch": 1.942016253019987, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.790109872817993, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7357316613197327, + "num_tokens": 457346283.0, + "step": 17684 + }, + { + "epoch": 1.9421260707226005, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.739877939224243, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7310830354690552, + "num_tokens": 457366083.0, + "step": 17685 + }, + { + "epoch": 1.942235888425214, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.104058265686035, + "learning_rate": 1e-06, + "loss": 0.9595, + "mean_token_accuracy": 0.7025537490844727, + "num_tokens": 457397897.0, + "step": 17686 + }, + { + "epoch": 1.9423457061278278, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2040679454803467, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7181512713432312, + "num_tokens": 457422345.0, + "step": 17687 + }, + { + "epoch": 1.9424555238304415, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.1467220783233643, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7274217009544373, + "num_tokens": 457448229.0, + "step": 17688 + }, + { + "epoch": 1.9425653415330553, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.026895523071289, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7140680551528931, + "num_tokens": 457477440.0, + "step": 17689 + }, + { + "epoch": 1.9426751592356688, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.0943193435668945, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7059729099273682, + "num_tokens": 457505166.0, + "step": 17690 + }, + { + "epoch": 1.9427849769382823, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.036435842514038, + "learning_rate": 1e-06, + "loss": 1.0026, + "mean_token_accuracy": 0.7009353637695312, + "num_tokens": 457536226.0, + "step": 17691 + }, + { + "epoch": 1.942894794640896, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.0372819900512695, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.719131588935852, + "num_tokens": 457565023.0, + "step": 17692 + }, + { + "epoch": 1.9430046123435099, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.121506690979004, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7262673377990723, + "num_tokens": 457592425.0, + "step": 17693 + }, + { + "epoch": 1.9431144300461234, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.20359468460083, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7091752886772156, + "num_tokens": 457621475.0, + "step": 17694 + }, + { + "epoch": 1.9432242477487371, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.0531535148620605, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7068783044815063, + "num_tokens": 457652399.0, + "step": 17695 + }, + { + "epoch": 1.9433340654513507, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2006187438964844, + "learning_rate": 1e-06, + "loss": 1.0312, + "mean_token_accuracy": 0.6837896108627319, + "num_tokens": 457680403.0, + "step": 17696 + }, + { + "epoch": 1.9434438831539644, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2195072174072266, + "learning_rate": 1e-06, + "loss": 0.9578, + "mean_token_accuracy": 0.7154934406280518, + "num_tokens": 457706330.0, + "step": 17697 + }, + { + "epoch": 1.9435537008565782, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.357588768005371, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.698161244392395, + "num_tokens": 457728851.0, + "step": 17698 + }, + { + "epoch": 1.9436635185591917, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.540142297744751, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7273542881011963, + "num_tokens": 457749633.0, + "step": 17699 + }, + { + "epoch": 1.9437733362618053, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.2702577114105225, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7157963514328003, + "num_tokens": 457774304.0, + "step": 17700 + }, + { + "epoch": 1.943883153964419, + "ewc_loss": 1.0311603546142578e-05, + "grad_norm": 2.11490535736084, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7419098615646362, + "num_tokens": 457801445.0, + "step": 17701 + }, + { + "epoch": 1.9439929716670328, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.223715305328369, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.729727029800415, + "num_tokens": 457825155.0, + "step": 17702 + }, + { + "epoch": 1.9441027893696465, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1551566123962402, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7182320952415466, + "num_tokens": 457852221.0, + "step": 17703 + }, + { + "epoch": 1.94421260707226, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.152299165725708, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7136129140853882, + "num_tokens": 457879946.0, + "step": 17704 + }, + { + "epoch": 1.9443224247748736, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.078385353088379, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7169325351715088, + "num_tokens": 457910546.0, + "step": 17705 + }, + { + "epoch": 1.9444322424774874, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3944294452667236, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7317063212394714, + "num_tokens": 457933478.0, + "step": 17706 + }, + { + "epoch": 1.9445420601801011, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0810859203338623, + "learning_rate": 1e-06, + "loss": 0.9348, + "mean_token_accuracy": 0.7142986059188843, + "num_tokens": 457964229.0, + "step": 17707 + }, + { + "epoch": 1.9446518778827147, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.285984516143799, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7075912952423096, + "num_tokens": 457988781.0, + "step": 17708 + }, + { + "epoch": 1.9447616955853284, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1805920600891113, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7069360017776489, + "num_tokens": 458016238.0, + "step": 17709 + }, + { + "epoch": 1.944871513287942, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.015491008758545, + "learning_rate": 1e-06, + "loss": 0.9962, + "mean_token_accuracy": 0.6960250735282898, + "num_tokens": 458047596.0, + "step": 17710 + }, + { + "epoch": 1.9449813309905557, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.076561212539673, + "learning_rate": 1e-06, + "loss": 1.0025, + "mean_token_accuracy": 0.6980003118515015, + "num_tokens": 458077736.0, + "step": 17711 + }, + { + "epoch": 1.9450911486931695, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.391119956970215, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7159476280212402, + "num_tokens": 458102156.0, + "step": 17712 + }, + { + "epoch": 1.945200966395783, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1405296325683594, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7239353060722351, + "num_tokens": 458132768.0, + "step": 17713 + }, + { + "epoch": 1.9453107840983965, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.175227642059326, + "learning_rate": 1e-06, + "loss": 0.9549, + "mean_token_accuracy": 0.706376850605011, + "num_tokens": 458159677.0, + "step": 17714 + }, + { + "epoch": 1.9454206018010103, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.27490496635437, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7179545164108276, + "num_tokens": 458184722.0, + "step": 17715 + }, + { + "epoch": 1.945530419503624, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.9141454696655273, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.719058632850647, + "num_tokens": 458217156.0, + "step": 17716 + }, + { + "epoch": 1.9456402372062378, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4382927417755127, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7208930253982544, + "num_tokens": 458238832.0, + "step": 17717 + }, + { + "epoch": 1.9457500549088513, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3411049842834473, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7100481986999512, + "num_tokens": 458262370.0, + "step": 17718 + }, + { + "epoch": 1.9458598726114649, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.645526647567749, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7395741939544678, + "num_tokens": 458281145.0, + "step": 17719 + }, + { + "epoch": 1.9459696903140786, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1658997535705566, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7207337617874146, + "num_tokens": 458308876.0, + "step": 17720 + }, + { + "epoch": 1.9460795080166924, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.008948564529419, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7022168636322021, + "num_tokens": 458338453.0, + "step": 17721 + }, + { + "epoch": 1.946189325719306, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1010544300079346, + "learning_rate": 1e-06, + "loss": 0.9385, + "mean_token_accuracy": 0.7074621915817261, + "num_tokens": 458364645.0, + "step": 17722 + }, + { + "epoch": 1.9462991434219195, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3506088256835938, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7080708146095276, + "num_tokens": 458389238.0, + "step": 17723 + }, + { + "epoch": 1.9464089611245332, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.270747184753418, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7179672718048096, + "num_tokens": 458412718.0, + "step": 17724 + }, + { + "epoch": 1.946518778827147, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.177957773208618, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7366947531700134, + "num_tokens": 458440883.0, + "step": 17725 + }, + { + "epoch": 1.9466285965297607, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.008999824523926, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7117065191268921, + "num_tokens": 458471234.0, + "step": 17726 + }, + { + "epoch": 1.9467384142323743, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.5984885692596436, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7126904129981995, + "num_tokens": 458491313.0, + "step": 17727 + }, + { + "epoch": 1.9468482319349878, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.9718438386917114, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7295976877212524, + "num_tokens": 458521269.0, + "step": 17728 + }, + { + "epoch": 1.9469580496376016, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.220580816268921, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.7058298587799072, + "num_tokens": 458549404.0, + "step": 17729 + }, + { + "epoch": 1.9470678673402153, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.256364583969116, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7041504979133606, + "num_tokens": 458576406.0, + "step": 17730 + }, + { + "epoch": 1.947177685042829, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3650577068328857, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7238528728485107, + "num_tokens": 458601780.0, + "step": 17731 + }, + { + "epoch": 1.9472875027454426, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2246387004852295, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7368981242179871, + "num_tokens": 458626965.0, + "step": 17732 + }, + { + "epoch": 1.9473973204480561, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.206160545349121, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7349875569343567, + "num_tokens": 458652567.0, + "step": 17733 + }, + { + "epoch": 1.94750713815067, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3742971420288086, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7120009660720825, + "num_tokens": 458678063.0, + "step": 17734 + }, + { + "epoch": 1.9476169558532836, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2114577293395996, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7292281985282898, + "num_tokens": 458703002.0, + "step": 17735 + }, + { + "epoch": 1.9477267735558972, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3001227378845215, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7278976440429688, + "num_tokens": 458726273.0, + "step": 17736 + }, + { + "epoch": 1.9478365912585107, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3487987518310547, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7216809988021851, + "num_tokens": 458748947.0, + "step": 17737 + }, + { + "epoch": 1.9479464089611245, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3464436531066895, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7204694151878357, + "num_tokens": 458773159.0, + "step": 17738 + }, + { + "epoch": 1.9480562266637382, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0816256999969482, + "learning_rate": 1e-06, + "loss": 0.9192, + "mean_token_accuracy": 0.7123003602027893, + "num_tokens": 458802385.0, + "step": 17739 + }, + { + "epoch": 1.948166044366352, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2035939693450928, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7221100330352783, + "num_tokens": 458828374.0, + "step": 17740 + }, + { + "epoch": 1.9482758620689655, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2937700748443604, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7057427167892456, + "num_tokens": 458852194.0, + "step": 17741 + }, + { + "epoch": 1.948385679771579, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1267240047454834, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7387839555740356, + "num_tokens": 458879585.0, + "step": 17742 + }, + { + "epoch": 1.9484954974741928, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.036442518234253, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7207369804382324, + "num_tokens": 458910244.0, + "step": 17743 + }, + { + "epoch": 1.9486053151768066, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1296298503875732, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7175473570823669, + "num_tokens": 458937185.0, + "step": 17744 + }, + { + "epoch": 1.94871513287942, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0955379009246826, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7382164001464844, + "num_tokens": 458963727.0, + "step": 17745 + }, + { + "epoch": 1.9488249505820339, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.95798921585083, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7050643563270569, + "num_tokens": 458996251.0, + "step": 17746 + }, + { + "epoch": 1.9489347682846474, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0082640647888184, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7146006226539612, + "num_tokens": 459027137.0, + "step": 17747 + }, + { + "epoch": 1.9490445859872612, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3043954372406006, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7357194423675537, + "num_tokens": 459050086.0, + "step": 17748 + }, + { + "epoch": 1.949154403689875, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.273714542388916, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.73504638671875, + "num_tokens": 459073889.0, + "step": 17749 + }, + { + "epoch": 1.9492642213924884, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.187044382095337, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7122217416763306, + "num_tokens": 459100458.0, + "step": 17750 + }, + { + "epoch": 1.949374039095102, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.571274995803833, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7359856963157654, + "num_tokens": 459119890.0, + "step": 17751 + }, + { + "epoch": 1.9494838567977157, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.21592116355896, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7065718770027161, + "num_tokens": 459147450.0, + "step": 17752 + }, + { + "epoch": 1.9495936745003295, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.137303352355957, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7497116923332214, + "num_tokens": 459172636.0, + "step": 17753 + }, + { + "epoch": 1.9497034922029433, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.213845729827881, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7068794369697571, + "num_tokens": 459200885.0, + "step": 17754 + }, + { + "epoch": 1.9498133099055568, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0326645374298096, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7291935086250305, + "num_tokens": 459229408.0, + "step": 17755 + }, + { + "epoch": 1.9499231276081703, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4235644340515137, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7420486211776733, + "num_tokens": 459254265.0, + "step": 17756 + }, + { + "epoch": 1.950032945310784, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.220381259918213, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7110298871994019, + "num_tokens": 459280332.0, + "step": 17757 + }, + { + "epoch": 1.9501427630133978, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.213320255279541, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.726078987121582, + "num_tokens": 459305595.0, + "step": 17758 + }, + { + "epoch": 1.9502525807160114, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.605563163757324, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7443080544471741, + "num_tokens": 459325299.0, + "step": 17759 + }, + { + "epoch": 1.9503623984186251, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2598958015441895, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7187392115592957, + "num_tokens": 459352836.0, + "step": 17760 + }, + { + "epoch": 1.9504722161212387, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2722368240356445, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7063348889350891, + "num_tokens": 459379227.0, + "step": 17761 + }, + { + "epoch": 1.9505820338238524, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0566656589508057, + "learning_rate": 1e-06, + "loss": 0.9771, + "mean_token_accuracy": 0.6970216035842896, + "num_tokens": 459411692.0, + "step": 17762 + }, + { + "epoch": 1.9506918515264662, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.876572847366333, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7256008982658386, + "num_tokens": 459445915.0, + "step": 17763 + }, + { + "epoch": 1.9508016692290797, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3639094829559326, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.7039827108383179, + "num_tokens": 459471396.0, + "step": 17764 + }, + { + "epoch": 1.9509114869316933, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.497260570526123, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.7213144302368164, + "num_tokens": 459493047.0, + "step": 17765 + }, + { + "epoch": 1.951021304634307, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4307804107666016, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7300122380256653, + "num_tokens": 459515684.0, + "step": 17766 + }, + { + "epoch": 1.9511311223369208, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3024003505706787, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7285543084144592, + "num_tokens": 459540503.0, + "step": 17767 + }, + { + "epoch": 1.9512409400395345, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.079697847366333, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7149311304092407, + "num_tokens": 459569352.0, + "step": 17768 + }, + { + "epoch": 1.951350757742148, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.491061210632324, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7448136806488037, + "num_tokens": 459591727.0, + "step": 17769 + }, + { + "epoch": 1.9514605754447616, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 3.032578706741333, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7350752949714661, + "num_tokens": 459613008.0, + "step": 17770 + }, + { + "epoch": 1.9515703931473753, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.11868953704834, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7229297161102295, + "num_tokens": 459640519.0, + "step": 17771 + }, + { + "epoch": 1.951680210849989, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2226409912109375, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7163822650909424, + "num_tokens": 459668946.0, + "step": 17772 + }, + { + "epoch": 1.9517900285526026, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.9111956357955933, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7243847846984863, + "num_tokens": 459701833.0, + "step": 17773 + }, + { + "epoch": 1.9518998462552162, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1522140502929688, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7288558483123779, + "num_tokens": 459729092.0, + "step": 17774 + }, + { + "epoch": 1.95200966395783, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3133959770202637, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7070342302322388, + "num_tokens": 459755210.0, + "step": 17775 + }, + { + "epoch": 1.9521194816604437, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.034752368927002, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6963151097297668, + "num_tokens": 459787978.0, + "step": 17776 + }, + { + "epoch": 1.9522292993630574, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.142454147338867, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7311928868293762, + "num_tokens": 459814198.0, + "step": 17777 + }, + { + "epoch": 1.952339117065671, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 1.9912818670272827, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7487375736236572, + "num_tokens": 459843631.0, + "step": 17778 + }, + { + "epoch": 1.9524489347682845, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3903706073760986, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7297082543373108, + "num_tokens": 459865903.0, + "step": 17779 + }, + { + "epoch": 1.9525587524708983, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4350342750549316, + "learning_rate": 1e-06, + "loss": 0.9545, + "mean_token_accuracy": 0.7026671171188354, + "num_tokens": 459890911.0, + "step": 17780 + }, + { + "epoch": 1.952668570173512, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.200313091278076, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7327694892883301, + "num_tokens": 459916634.0, + "step": 17781 + }, + { + "epoch": 1.9527783878761258, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.057490587234497, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7070143222808838, + "num_tokens": 459945596.0, + "step": 17782 + }, + { + "epoch": 1.9528882055787393, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.191467046737671, + "learning_rate": 1e-06, + "loss": 1.0125, + "mean_token_accuracy": 0.6875122785568237, + "num_tokens": 459971509.0, + "step": 17783 + }, + { + "epoch": 1.9529980232813529, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.264608860015869, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7148047685623169, + "num_tokens": 459995866.0, + "step": 17784 + }, + { + "epoch": 1.9531078409839666, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4089813232421875, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7190545797348022, + "num_tokens": 460019422.0, + "step": 17785 + }, + { + "epoch": 1.9532176586865804, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0755062103271484, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7114044427871704, + "num_tokens": 460051120.0, + "step": 17786 + }, + { + "epoch": 1.953327476389194, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2339491844177246, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7293094396591187, + "num_tokens": 460075998.0, + "step": 17787 + }, + { + "epoch": 1.9534372940918074, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3067853450775146, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7410600185394287, + "num_tokens": 460102115.0, + "step": 17788 + }, + { + "epoch": 1.9535471117944212, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.362238883972168, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7190947532653809, + "num_tokens": 460124918.0, + "step": 17789 + }, + { + "epoch": 1.953656929497035, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1367733478546143, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7152283191680908, + "num_tokens": 460151773.0, + "step": 17790 + }, + { + "epoch": 1.9537667471996487, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.0537219047546387, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.71370530128479, + "num_tokens": 460180962.0, + "step": 17791 + }, + { + "epoch": 1.9538765649022622, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3334951400756836, + "learning_rate": 1e-06, + "loss": 1.0052, + "mean_token_accuracy": 0.6946847438812256, + "num_tokens": 460204462.0, + "step": 17792 + }, + { + "epoch": 1.9539863826048758, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.1703078746795654, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7046156525611877, + "num_tokens": 460231154.0, + "step": 17793 + }, + { + "epoch": 1.9540962003074895, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4913978576660156, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.736410915851593, + "num_tokens": 460251325.0, + "step": 17794 + }, + { + "epoch": 1.9542060180101033, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.3621842861175537, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7127150297164917, + "num_tokens": 460275225.0, + "step": 17795 + }, + { + "epoch": 1.954315835712717, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.2282721996307373, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7336478233337402, + "num_tokens": 460300141.0, + "step": 17796 + }, + { + "epoch": 1.9544256534153306, + "ewc_loss": 1.0371208190917969e-05, + "grad_norm": 2.4298434257507324, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7328637838363647, + "num_tokens": 460321850.0, + "step": 17797 + }, + { + "epoch": 1.9545354711179441, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 1.905709147453308, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.738003134727478, + "num_tokens": 460353093.0, + "step": 17798 + }, + { + "epoch": 1.9546452888205579, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.214151382446289, + "learning_rate": 1e-06, + "loss": 1.0522, + "mean_token_accuracy": 0.6871812343597412, + "num_tokens": 460380598.0, + "step": 17799 + }, + { + "epoch": 1.9547551065231716, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.174612522125244, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.713453471660614, + "num_tokens": 460405826.0, + "step": 17800 + }, + { + "epoch": 1.9548649242257852, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.281686305999756, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.723720133304596, + "num_tokens": 460429248.0, + "step": 17801 + }, + { + "epoch": 1.9549747419283987, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2281336784362793, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7415052652359009, + "num_tokens": 460453280.0, + "step": 17802 + }, + { + "epoch": 1.9550845596310125, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.353701114654541, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7254467606544495, + "num_tokens": 460476265.0, + "step": 17803 + }, + { + "epoch": 1.9551943773336262, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 1.9576244354248047, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7157357931137085, + "num_tokens": 460506821.0, + "step": 17804 + }, + { + "epoch": 1.95530419503624, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.089533567428589, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7052724361419678, + "num_tokens": 460535198.0, + "step": 17805 + }, + { + "epoch": 1.9554140127388535, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.326925277709961, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7072085738182068, + "num_tokens": 460559697.0, + "step": 17806 + }, + { + "epoch": 1.955523830441467, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1627540588378906, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7067406177520752, + "num_tokens": 460587873.0, + "step": 17807 + }, + { + "epoch": 1.9556336481440808, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2181179523468018, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6978839635848999, + "num_tokens": 460613492.0, + "step": 17808 + }, + { + "epoch": 1.9557434658466946, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.466696262359619, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7381141185760498, + "num_tokens": 460633636.0, + "step": 17809 + }, + { + "epoch": 1.955853283549308, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1013081073760986, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7329045534133911, + "num_tokens": 460660312.0, + "step": 17810 + }, + { + "epoch": 1.9559631012519219, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.6875274181365967, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7227720022201538, + "num_tokens": 460678619.0, + "step": 17811 + }, + { + "epoch": 1.9560729189545354, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.4811978340148926, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7348229885101318, + "num_tokens": 460699511.0, + "step": 17812 + }, + { + "epoch": 1.9561827366571491, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.150473117828369, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7222514152526855, + "num_tokens": 460726563.0, + "step": 17813 + }, + { + "epoch": 1.956292554359763, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.025489091873169, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7354904413223267, + "num_tokens": 460755421.0, + "step": 17814 + }, + { + "epoch": 1.9564023720623764, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1384756565093994, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.721109926700592, + "num_tokens": 460781504.0, + "step": 17815 + }, + { + "epoch": 1.95651218976499, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.189819574356079, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7251968383789062, + "num_tokens": 460808805.0, + "step": 17816 + }, + { + "epoch": 1.9566220074676037, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.106126308441162, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.7111591696739197, + "num_tokens": 460836943.0, + "step": 17817 + }, + { + "epoch": 1.9567318251702175, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2170333862304688, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7088497877120972, + "num_tokens": 460863149.0, + "step": 17818 + }, + { + "epoch": 1.9568416428728312, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.114245891571045, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7328847646713257, + "num_tokens": 460891102.0, + "step": 17819 + }, + { + "epoch": 1.9569514605754448, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.141273021697998, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7096516489982605, + "num_tokens": 460917830.0, + "step": 17820 + }, + { + "epoch": 1.9570612782780583, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.119727373123169, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7117616534233093, + "num_tokens": 460943706.0, + "step": 17821 + }, + { + "epoch": 1.957171095980672, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.0653350353240967, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7362142205238342, + "num_tokens": 460970926.0, + "step": 17822 + }, + { + "epoch": 1.9572809136832858, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.161867141723633, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.729971170425415, + "num_tokens": 460995967.0, + "step": 17823 + }, + { + "epoch": 1.9573907313858994, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.009920597076416, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7122257947921753, + "num_tokens": 461026643.0, + "step": 17824 + }, + { + "epoch": 1.9575005490885131, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2465901374816895, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7092665433883667, + "num_tokens": 461052393.0, + "step": 17825 + }, + { + "epoch": 1.9576103667911267, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 1.9870558977127075, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7131840586662292, + "num_tokens": 461083688.0, + "step": 17826 + }, + { + "epoch": 1.9577201844937404, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.3511054515838623, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7164202928543091, + "num_tokens": 461106965.0, + "step": 17827 + }, + { + "epoch": 1.9578300021963542, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1947691440582275, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.6911725997924805, + "num_tokens": 461134676.0, + "step": 17828 + }, + { + "epoch": 1.9579398198989677, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.184209108352661, + "learning_rate": 1e-06, + "loss": 0.993, + "mean_token_accuracy": 0.6996793746948242, + "num_tokens": 461163256.0, + "step": 17829 + }, + { + "epoch": 1.9580496376015812, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2768445014953613, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7373727560043335, + "num_tokens": 461186258.0, + "step": 17830 + }, + { + "epoch": 1.958159455304195, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.016627550125122, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.705631673336029, + "num_tokens": 461218074.0, + "step": 17831 + }, + { + "epoch": 1.9582692730068088, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.154757499694824, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7280442714691162, + "num_tokens": 461243234.0, + "step": 17832 + }, + { + "epoch": 1.9583790907094225, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.0324273109436035, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.712715744972229, + "num_tokens": 461272954.0, + "step": 17833 + }, + { + "epoch": 1.958488908412036, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.320484161376953, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7415828108787537, + "num_tokens": 461295956.0, + "step": 17834 + }, + { + "epoch": 1.9585987261146496, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.127070665359497, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7200309038162231, + "num_tokens": 461323378.0, + "step": 17835 + }, + { + "epoch": 1.9587085438172633, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.140653371810913, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.713117778301239, + "num_tokens": 461349861.0, + "step": 17836 + }, + { + "epoch": 1.958818361519877, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.55037784576416, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.731277346611023, + "num_tokens": 461370899.0, + "step": 17837 + }, + { + "epoch": 1.9589281792224906, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.256401538848877, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7347270250320435, + "num_tokens": 461397994.0, + "step": 17838 + }, + { + "epoch": 1.9590379969251042, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.416534185409546, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.722791850566864, + "num_tokens": 461421247.0, + "step": 17839 + }, + { + "epoch": 1.959147814627718, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.3895344734191895, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7153917551040649, + "num_tokens": 461443922.0, + "step": 17840 + }, + { + "epoch": 1.9592576323303317, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.4153974056243896, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7228359580039978, + "num_tokens": 461466502.0, + "step": 17841 + }, + { + "epoch": 1.9593674500329454, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 1.9502865076065063, + "learning_rate": 1e-06, + "loss": 0.9658, + "mean_token_accuracy": 0.7009309530258179, + "num_tokens": 461502352.0, + "step": 17842 + }, + { + "epoch": 1.959477267735559, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.2298989295959473, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.72750324010849, + "num_tokens": 461526845.0, + "step": 17843 + }, + { + "epoch": 1.9595870854381725, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1349034309387207, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7387043237686157, + "num_tokens": 461552144.0, + "step": 17844 + }, + { + "epoch": 1.9596969031407863, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.0335657596588135, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7242349982261658, + "num_tokens": 461581473.0, + "step": 17845 + }, + { + "epoch": 1.9598067208434, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.17954683303833, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7307862043380737, + "num_tokens": 461607343.0, + "step": 17846 + }, + { + "epoch": 1.9599165385460138, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.1153295040130615, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.715685248374939, + "num_tokens": 461634676.0, + "step": 17847 + }, + { + "epoch": 1.9600263562486273, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 1.9784473180770874, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.7082353830337524, + "num_tokens": 461667066.0, + "step": 17848 + }, + { + "epoch": 1.9601361739512408, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.5341739654541016, + "learning_rate": 1e-06, + "loss": 0.7164, + "mean_token_accuracy": 0.7716960310935974, + "num_tokens": 461687263.0, + "step": 17849 + }, + { + "epoch": 1.9602459916538546, + "ewc_loss": 1.043081283569336e-05, + "grad_norm": 2.126251220703125, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7390750646591187, + "num_tokens": 461712768.0, + "step": 17850 + }, + { + "epoch": 1.9603558093564684, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.244694709777832, + "learning_rate": 1e-06, + "loss": 0.9652, + "mean_token_accuracy": 0.6988568902015686, + "num_tokens": 461738675.0, + "step": 17851 + }, + { + "epoch": 1.960465627059082, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2552504539489746, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7167608737945557, + "num_tokens": 461764686.0, + "step": 17852 + }, + { + "epoch": 1.9605754447616954, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1194300651550293, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.72059166431427, + "num_tokens": 461792000.0, + "step": 17853 + }, + { + "epoch": 1.9606852624643092, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.4072482585906982, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7189595699310303, + "num_tokens": 461815714.0, + "step": 17854 + }, + { + "epoch": 1.960795080166923, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.238025665283203, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7165864706039429, + "num_tokens": 461840276.0, + "step": 17855 + }, + { + "epoch": 1.9609048978695367, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9778369665145874, + "learning_rate": 1e-06, + "loss": 0.9812, + "mean_token_accuracy": 0.6953891515731812, + "num_tokens": 461871940.0, + "step": 17856 + }, + { + "epoch": 1.9610147155721502, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1492843627929688, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7368775606155396, + "num_tokens": 461900224.0, + "step": 17857 + }, + { + "epoch": 1.9611245332747638, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2116568088531494, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7195193767547607, + "num_tokens": 461925930.0, + "step": 17858 + }, + { + "epoch": 1.9612343509773775, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2662999629974365, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7235345840454102, + "num_tokens": 461951480.0, + "step": 17859 + }, + { + "epoch": 1.9613441686799913, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.233295202255249, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7184110879898071, + "num_tokens": 461975796.0, + "step": 17860 + }, + { + "epoch": 1.961453986382605, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.496420383453369, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7189125418663025, + "num_tokens": 461996344.0, + "step": 17861 + }, + { + "epoch": 1.9615638040852186, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1550791263580322, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7328479886054993, + "num_tokens": 462023553.0, + "step": 17862 + }, + { + "epoch": 1.9616736217878321, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.088036060333252, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7064487934112549, + "num_tokens": 462049915.0, + "step": 17863 + }, + { + "epoch": 1.9617834394904459, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1012823581695557, + "learning_rate": 1e-06, + "loss": 0.9641, + "mean_token_accuracy": 0.7040337324142456, + "num_tokens": 462079966.0, + "step": 17864 + }, + { + "epoch": 1.9618932571930596, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.390817165374756, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7161484956741333, + "num_tokens": 462103348.0, + "step": 17865 + }, + { + "epoch": 1.9620030748956732, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1951706409454346, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.720085620880127, + "num_tokens": 462131788.0, + "step": 17866 + }, + { + "epoch": 1.9621128925982867, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.4659910202026367, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7201324701309204, + "num_tokens": 462154304.0, + "step": 17867 + }, + { + "epoch": 1.9622227103009005, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.150092124938965, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7335389256477356, + "num_tokens": 462180987.0, + "step": 17868 + }, + { + "epoch": 1.9623325280035142, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0618703365325928, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7182741165161133, + "num_tokens": 462210465.0, + "step": 17869 + }, + { + "epoch": 1.962442345706128, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.200718402862549, + "learning_rate": 1e-06, + "loss": 0.9507, + "mean_token_accuracy": 0.7159664630889893, + "num_tokens": 462236827.0, + "step": 17870 + }, + { + "epoch": 1.9625521634087415, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0879011154174805, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.722789466381073, + "num_tokens": 462264662.0, + "step": 17871 + }, + { + "epoch": 1.962661981111355, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.3291268348693848, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7263486981391907, + "num_tokens": 462288633.0, + "step": 17872 + }, + { + "epoch": 1.9627717988139688, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1479952335357666, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7099916338920593, + "num_tokens": 462315207.0, + "step": 17873 + }, + { + "epoch": 1.9628816165165826, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2172164916992188, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7077456116676331, + "num_tokens": 462343361.0, + "step": 17874 + }, + { + "epoch": 1.962991434219196, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.4223709106445312, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.716062605381012, + "num_tokens": 462365597.0, + "step": 17875 + }, + { + "epoch": 1.9631012519218098, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2404088973999023, + "learning_rate": 1e-06, + "loss": 0.953, + "mean_token_accuracy": 0.7037933468818665, + "num_tokens": 462391846.0, + "step": 17876 + }, + { + "epoch": 1.9632110696244234, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.110725164413452, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.717735767364502, + "num_tokens": 462418944.0, + "step": 17877 + }, + { + "epoch": 1.9633208873270371, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.257581949234009, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7142767906188965, + "num_tokens": 462446632.0, + "step": 17878 + }, + { + "epoch": 1.963430705029651, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.3050239086151123, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7314108610153198, + "num_tokens": 462471488.0, + "step": 17879 + }, + { + "epoch": 1.9635405227322644, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1228456497192383, + "learning_rate": 1e-06, + "loss": 0.9777, + "mean_token_accuracy": 0.694631814956665, + "num_tokens": 462499777.0, + "step": 17880 + }, + { + "epoch": 1.963650340434878, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.4322593212127686, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7184150218963623, + "num_tokens": 462521390.0, + "step": 17881 + }, + { + "epoch": 1.9637601581374917, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2136309146881104, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7254197597503662, + "num_tokens": 462548402.0, + "step": 17882 + }, + { + "epoch": 1.9638699758401055, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1056666374206543, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7218477725982666, + "num_tokens": 462579022.0, + "step": 17883 + }, + { + "epoch": 1.9639797935427192, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.5048396587371826, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7381901144981384, + "num_tokens": 462600350.0, + "step": 17884 + }, + { + "epoch": 1.9640896112453328, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.080608606338501, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7161587476730347, + "num_tokens": 462628836.0, + "step": 17885 + }, + { + "epoch": 1.9641994289479463, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.6132802963256836, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7291892170906067, + "num_tokens": 462648404.0, + "step": 17886 + }, + { + "epoch": 1.96430924665056, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.037968397140503, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7333641052246094, + "num_tokens": 462675662.0, + "step": 17887 + }, + { + "epoch": 1.9644190643531738, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1445844173431396, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7084616422653198, + "num_tokens": 462702373.0, + "step": 17888 + }, + { + "epoch": 1.9645288820557874, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.423302412033081, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7266753911972046, + "num_tokens": 462724181.0, + "step": 17889 + }, + { + "epoch": 1.964638699758401, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1329901218414307, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.712303876876831, + "num_tokens": 462751807.0, + "step": 17890 + }, + { + "epoch": 1.9647485174610146, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.002746105194092, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7170227766036987, + "num_tokens": 462783331.0, + "step": 17891 + }, + { + "epoch": 1.9648583351636284, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2240726947784424, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7324978113174438, + "num_tokens": 462812266.0, + "step": 17892 + }, + { + "epoch": 1.9649681528662422, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2000036239624023, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7236263751983643, + "num_tokens": 462841811.0, + "step": 17893 + }, + { + "epoch": 1.9650779705688557, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0265161991119385, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7210956811904907, + "num_tokens": 462870788.0, + "step": 17894 + }, + { + "epoch": 1.9651877882714692, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.167203903198242, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6962258219718933, + "num_tokens": 462899120.0, + "step": 17895 + }, + { + "epoch": 1.965297605974083, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9656665325164795, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7315256595611572, + "num_tokens": 462928270.0, + "step": 17896 + }, + { + "epoch": 1.9654074236766967, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.930591344833374, + "learning_rate": 1e-06, + "loss": 1.0074, + "mean_token_accuracy": 0.7029982209205627, + "num_tokens": 462960055.0, + "step": 17897 + }, + { + "epoch": 1.9655172413793105, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1175577640533447, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7222204208374023, + "num_tokens": 462989831.0, + "step": 17898 + }, + { + "epoch": 1.965627059081924, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.5546810626983643, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7011562585830688, + "num_tokens": 463011219.0, + "step": 17899 + }, + { + "epoch": 1.9657368767845376, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.8415617942810059, + "learning_rate": 1e-06, + "loss": 1.0082, + "mean_token_accuracy": 0.6952118277549744, + "num_tokens": 463047470.0, + "step": 17900 + }, + { + "epoch": 1.9658466944871513, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.5446174144744873, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7249955534934998, + "num_tokens": 463068746.0, + "step": 17901 + }, + { + "epoch": 1.965956512189765, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.269947052001953, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7281766533851624, + "num_tokens": 463095132.0, + "step": 17902 + }, + { + "epoch": 1.9660663298923786, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.918057918548584, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7334175109863281, + "num_tokens": 463124970.0, + "step": 17903 + }, + { + "epoch": 1.9661761475949922, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.110316753387451, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.716981053352356, + "num_tokens": 463154144.0, + "step": 17904 + }, + { + "epoch": 1.966285965297606, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.411210060119629, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.717100977897644, + "num_tokens": 463179158.0, + "step": 17905 + }, + { + "epoch": 1.9663957830002197, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0508878231048584, + "learning_rate": 1e-06, + "loss": 0.7769, + "mean_token_accuracy": 0.752068042755127, + "num_tokens": 463206314.0, + "step": 17906 + }, + { + "epoch": 1.9665056007028334, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.5911362171173096, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7202557325363159, + "num_tokens": 463228266.0, + "step": 17907 + }, + { + "epoch": 1.966615418405447, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1846847534179688, + "learning_rate": 1e-06, + "loss": 0.9351, + "mean_token_accuracy": 0.7055649161338806, + "num_tokens": 463255744.0, + "step": 17908 + }, + { + "epoch": 1.9667252361080605, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9904098510742188, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7118610143661499, + "num_tokens": 463287159.0, + "step": 17909 + }, + { + "epoch": 1.9668350538106742, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.10968279838562, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7462780475616455, + "num_tokens": 463313671.0, + "step": 17910 + }, + { + "epoch": 1.966944871513288, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.310020923614502, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7163912057876587, + "num_tokens": 463338766.0, + "step": 17911 + }, + { + "epoch": 1.9670546892159018, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1142139434814453, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.727218747138977, + "num_tokens": 463365966.0, + "step": 17912 + }, + { + "epoch": 1.9671645069185153, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.262193441390991, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7164919972419739, + "num_tokens": 463389649.0, + "step": 17913 + }, + { + "epoch": 1.9672743246211288, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1032652854919434, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7143725752830505, + "num_tokens": 463418898.0, + "step": 17914 + }, + { + "epoch": 1.9673841423237426, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.416912078857422, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.713253915309906, + "num_tokens": 463442296.0, + "step": 17915 + }, + { + "epoch": 1.9674939600263563, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0830156803131104, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7163535952568054, + "num_tokens": 463470572.0, + "step": 17916 + }, + { + "epoch": 1.9676037777289699, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0177524089813232, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7152571678161621, + "num_tokens": 463501542.0, + "step": 17917 + }, + { + "epoch": 1.9677135954315834, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.272580146789551, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7003225088119507, + "num_tokens": 463528562.0, + "step": 17918 + }, + { + "epoch": 1.9678234131341972, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1138761043548584, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7218703627586365, + "num_tokens": 463557046.0, + "step": 17919 + }, + { + "epoch": 1.967933230836811, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2599048614501953, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7244031429290771, + "num_tokens": 463581800.0, + "step": 17920 + }, + { + "epoch": 1.9680430485394247, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.20413875579834, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7246531248092651, + "num_tokens": 463607472.0, + "step": 17921 + }, + { + "epoch": 1.9681528662420382, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.160081148147583, + "learning_rate": 1e-06, + "loss": 0.9536, + "mean_token_accuracy": 0.7002926468849182, + "num_tokens": 463635685.0, + "step": 17922 + }, + { + "epoch": 1.9682626839446518, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0273146629333496, + "learning_rate": 1e-06, + "loss": 1.0276, + "mean_token_accuracy": 0.6956064701080322, + "num_tokens": 463667783.0, + "step": 17923 + }, + { + "epoch": 1.9683725016472655, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0376875400543213, + "learning_rate": 1e-06, + "loss": 0.9421, + "mean_token_accuracy": 0.7181119322776794, + "num_tokens": 463699725.0, + "step": 17924 + }, + { + "epoch": 1.9684823193498793, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.399068593978882, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7358620166778564, + "num_tokens": 463721407.0, + "step": 17925 + }, + { + "epoch": 1.9685921370524928, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.213428497314453, + "learning_rate": 1e-06, + "loss": 0.992, + "mean_token_accuracy": 0.7094508409500122, + "num_tokens": 463749097.0, + "step": 17926 + }, + { + "epoch": 1.9687019547551066, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1052775382995605, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7248116731643677, + "num_tokens": 463777892.0, + "step": 17927 + }, + { + "epoch": 1.96881177245772, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2663204669952393, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7233400344848633, + "num_tokens": 463802494.0, + "step": 17928 + }, + { + "epoch": 1.9689215901603339, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.179955005645752, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7116110324859619, + "num_tokens": 463829345.0, + "step": 17929 + }, + { + "epoch": 1.9690314078629476, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.332214117050171, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.714228630065918, + "num_tokens": 463852750.0, + "step": 17930 + }, + { + "epoch": 1.9691412255655611, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2628931999206543, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7378413677215576, + "num_tokens": 463876341.0, + "step": 17931 + }, + { + "epoch": 1.9692510432681747, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1840832233428955, + "learning_rate": 1e-06, + "loss": 0.9782, + "mean_token_accuracy": 0.6958122253417969, + "num_tokens": 463903977.0, + "step": 17932 + }, + { + "epoch": 1.9693608609707884, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.967408537864685, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7050254344940186, + "num_tokens": 463935305.0, + "step": 17933 + }, + { + "epoch": 1.9694706786734022, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0993077754974365, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7233400940895081, + "num_tokens": 463964995.0, + "step": 17934 + }, + { + "epoch": 1.969580496376016, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1153860092163086, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7166153192520142, + "num_tokens": 463995158.0, + "step": 17935 + }, + { + "epoch": 1.9696903140786295, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1697134971618652, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7182438969612122, + "num_tokens": 464022617.0, + "step": 17936 + }, + { + "epoch": 1.969800131781243, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.46450138092041, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.727534830570221, + "num_tokens": 464044568.0, + "step": 17937 + }, + { + "epoch": 1.9699099494838568, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.353647470474243, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7321174144744873, + "num_tokens": 464066711.0, + "step": 17938 + }, + { + "epoch": 1.9700197671864705, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.026343584060669, + "learning_rate": 1e-06, + "loss": 0.9504, + "mean_token_accuracy": 0.7142715454101562, + "num_tokens": 464096915.0, + "step": 17939 + }, + { + "epoch": 1.970129584889084, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1502373218536377, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7166881561279297, + "num_tokens": 464123541.0, + "step": 17940 + }, + { + "epoch": 1.9702394025916978, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1321535110473633, + "learning_rate": 1e-06, + "loss": 0.9232, + "mean_token_accuracy": 0.7116372585296631, + "num_tokens": 464151300.0, + "step": 17941 + }, + { + "epoch": 1.9703492202943114, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.143502712249756, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7123379707336426, + "num_tokens": 464178356.0, + "step": 17942 + }, + { + "epoch": 1.9704590379969251, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2385590076446533, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6943328380584717, + "num_tokens": 464206079.0, + "step": 17943 + }, + { + "epoch": 1.9705688556995389, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9483309984207153, + "learning_rate": 1e-06, + "loss": 1.0001, + "mean_token_accuracy": 0.6901708841323853, + "num_tokens": 464238899.0, + "step": 17944 + }, + { + "epoch": 1.9706786734021524, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0653860569000244, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7093855738639832, + "num_tokens": 464269439.0, + "step": 17945 + }, + { + "epoch": 1.970788491104766, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.341815710067749, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7266006469726562, + "num_tokens": 464292768.0, + "step": 17946 + }, + { + "epoch": 1.9708983088073797, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0447850227355957, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7187641859054565, + "num_tokens": 464322796.0, + "step": 17947 + }, + { + "epoch": 1.9710081265099935, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9986553192138672, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.716414749622345, + "num_tokens": 464353679.0, + "step": 17948 + }, + { + "epoch": 1.9711179442126072, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.6855340003967285, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7296598553657532, + "num_tokens": 464372732.0, + "step": 17949 + }, + { + "epoch": 1.9712277619152208, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9198988676071167, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.731421172618866, + "num_tokens": 464405373.0, + "step": 17950 + }, + { + "epoch": 1.9713375796178343, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.3726916313171387, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7107694149017334, + "num_tokens": 464430443.0, + "step": 17951 + }, + { + "epoch": 1.971447397320448, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1449711322784424, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.7034711241722107, + "num_tokens": 464457515.0, + "step": 17952 + }, + { + "epoch": 1.9715572150230618, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.667168617248535, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7161281108856201, + "num_tokens": 464477028.0, + "step": 17953 + }, + { + "epoch": 1.9716670327256753, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.173621416091919, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.717158317565918, + "num_tokens": 464502140.0, + "step": 17954 + }, + { + "epoch": 1.9717768504282889, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.048616647720337, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7267223596572876, + "num_tokens": 464530992.0, + "step": 17955 + }, + { + "epoch": 1.9718866681309026, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9994632005691528, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6928805112838745, + "num_tokens": 464565271.0, + "step": 17956 + }, + { + "epoch": 1.9719964858335164, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1838502883911133, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7444406151771545, + "num_tokens": 464592186.0, + "step": 17957 + }, + { + "epoch": 1.9721063035361301, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2469046115875244, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7297253608703613, + "num_tokens": 464616573.0, + "step": 17958 + }, + { + "epoch": 1.9722161212387437, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.344758987426758, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7010663747787476, + "num_tokens": 464641020.0, + "step": 17959 + }, + { + "epoch": 1.9723259389413572, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0549917221069336, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7381188869476318, + "num_tokens": 464671457.0, + "step": 17960 + }, + { + "epoch": 1.972435756643971, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9572477340698242, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7329562902450562, + "num_tokens": 464701479.0, + "step": 17961 + }, + { + "epoch": 1.9725455743465847, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0784568786621094, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7086568474769592, + "num_tokens": 464730943.0, + "step": 17962 + }, + { + "epoch": 1.9726553920491985, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.46162748336792, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7120788097381592, + "num_tokens": 464753566.0, + "step": 17963 + }, + { + "epoch": 1.972765209751812, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.254011392593384, + "learning_rate": 1e-06, + "loss": 0.9906, + "mean_token_accuracy": 0.6980846524238586, + "num_tokens": 464781410.0, + "step": 17964 + }, + { + "epoch": 1.9728750274544256, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.3358652591705322, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7270262837409973, + "num_tokens": 464806484.0, + "step": 17965 + }, + { + "epoch": 1.9729848451570393, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1834449768066406, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7120380997657776, + "num_tokens": 464833348.0, + "step": 17966 + }, + { + "epoch": 1.973094662859653, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.406881332397461, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7108864784240723, + "num_tokens": 464856986.0, + "step": 17967 + }, + { + "epoch": 1.9732044805622666, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9879941940307617, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7430049180984497, + "num_tokens": 464887986.0, + "step": 17968 + }, + { + "epoch": 1.9733142982648801, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.5779035091400146, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7357572913169861, + "num_tokens": 464907957.0, + "step": 17969 + }, + { + "epoch": 1.973424115967494, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.088104486465454, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7371985912322998, + "num_tokens": 464933222.0, + "step": 17970 + }, + { + "epoch": 1.9735339336701077, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.202019214630127, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7178528904914856, + "num_tokens": 464959829.0, + "step": 17971 + }, + { + "epoch": 1.9736437513727214, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.328246831893921, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7247726917266846, + "num_tokens": 464983039.0, + "step": 17972 + }, + { + "epoch": 1.973753569075335, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.033555507659912, + "learning_rate": 1e-06, + "loss": 0.9623, + "mean_token_accuracy": 0.7001661062240601, + "num_tokens": 465012388.0, + "step": 17973 + }, + { + "epoch": 1.9738633867779485, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.021677017211914, + "learning_rate": 1e-06, + "loss": 0.9688, + "mean_token_accuracy": 0.7048945426940918, + "num_tokens": 465042863.0, + "step": 17974 + }, + { + "epoch": 1.9739732044805622, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2844135761260986, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7296713590621948, + "num_tokens": 465066140.0, + "step": 17975 + }, + { + "epoch": 1.974083022183176, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.2017440795898438, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7312017679214478, + "num_tokens": 465091795.0, + "step": 17976 + }, + { + "epoch": 1.9741928398857898, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.416015625, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7318754196166992, + "num_tokens": 465115843.0, + "step": 17977 + }, + { + "epoch": 1.9743026575884033, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9951376914978027, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7189395427703857, + "num_tokens": 465143631.0, + "step": 17978 + }, + { + "epoch": 1.9744124752910168, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1752161979675293, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6984415054321289, + "num_tokens": 465170006.0, + "step": 17979 + }, + { + "epoch": 1.9745222929936306, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.1073555946350098, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7213460206985474, + "num_tokens": 465196759.0, + "step": 17980 + }, + { + "epoch": 1.9746321106962443, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0762805938720703, + "learning_rate": 1e-06, + "loss": 0.7937, + "mean_token_accuracy": 0.7462893724441528, + "num_tokens": 465223186.0, + "step": 17981 + }, + { + "epoch": 1.9747419283988579, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.4240407943725586, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7162727117538452, + "num_tokens": 465244970.0, + "step": 17982 + }, + { + "epoch": 1.9748517461014714, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.325256586074829, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7289539575576782, + "num_tokens": 465267666.0, + "step": 17983 + }, + { + "epoch": 1.9749615638040852, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 1.9370486736297607, + "learning_rate": 1e-06, + "loss": 0.9873, + "mean_token_accuracy": 0.694137454032898, + "num_tokens": 465300502.0, + "step": 17984 + }, + { + "epoch": 1.975071381506699, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.10559344291687, + "learning_rate": 1e-06, + "loss": 1.0088, + "mean_token_accuracy": 0.6901425719261169, + "num_tokens": 465330538.0, + "step": 17985 + }, + { + "epoch": 1.9751811992093127, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.130082368850708, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.722231388092041, + "num_tokens": 465356579.0, + "step": 17986 + }, + { + "epoch": 1.9752910169119262, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.549069404602051, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7353808283805847, + "num_tokens": 465375455.0, + "step": 17987 + }, + { + "epoch": 1.9754008346145397, + "ewc_loss": 1.049041748046875e-05, + "grad_norm": 2.0630621910095215, + "learning_rate": 1e-06, + "loss": 0.998, + "mean_token_accuracy": 0.6890416145324707, + "num_tokens": 465406089.0, + "step": 17988 + }, + { + "epoch": 1.9755106523171535, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3840768337249756, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7419949173927307, + "num_tokens": 465428073.0, + "step": 17989 + }, + { + "epoch": 1.9756204700197673, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1173906326293945, + "learning_rate": 1e-06, + "loss": 0.9393, + "mean_token_accuracy": 0.7156496047973633, + "num_tokens": 465454586.0, + "step": 17990 + }, + { + "epoch": 1.9757302877223808, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2174313068389893, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7177905440330505, + "num_tokens": 465479020.0, + "step": 17991 + }, + { + "epoch": 1.9758401054249946, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4745616912841797, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.729904055595398, + "num_tokens": 465499620.0, + "step": 17992 + }, + { + "epoch": 1.975949923127608, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3025059700012207, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.729891300201416, + "num_tokens": 465523489.0, + "step": 17993 + }, + { + "epoch": 1.9760597408302218, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2774269580841064, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7284015417098999, + "num_tokens": 465550394.0, + "step": 17994 + }, + { + "epoch": 1.9761695585328356, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.360398530960083, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7360439300537109, + "num_tokens": 465574074.0, + "step": 17995 + }, + { + "epoch": 1.9762793762354491, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.800428867340088, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7379181385040283, + "num_tokens": 465610552.0, + "step": 17996 + }, + { + "epoch": 1.9763891939380627, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.249490976333618, + "learning_rate": 1e-06, + "loss": 0.951, + "mean_token_accuracy": 0.708802342414856, + "num_tokens": 465638687.0, + "step": 17997 + }, + { + "epoch": 1.9764990116406764, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.6595094203948975, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7214000821113586, + "num_tokens": 465659469.0, + "step": 17998 + }, + { + "epoch": 1.9766088293432902, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0056347846984863, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7150701284408569, + "num_tokens": 465690690.0, + "step": 17999 + }, + { + "epoch": 1.976718647045904, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4001400470733643, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7349649667739868, + "num_tokens": 465713521.0, + "step": 18000 + }, + { + "epoch": 1.9768284647485175, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.085078477859497, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7086554765701294, + "num_tokens": 465741454.0, + "step": 18001 + }, + { + "epoch": 1.976938282451131, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.194798231124878, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7214555740356445, + "num_tokens": 465767905.0, + "step": 18002 + }, + { + "epoch": 1.9770481001537448, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.174792766571045, + "learning_rate": 1e-06, + "loss": 1.0038, + "mean_token_accuracy": 0.6939458847045898, + "num_tokens": 465795347.0, + "step": 18003 + }, + { + "epoch": 1.9771579178563585, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.8029439449310303, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7289429903030396, + "num_tokens": 465812978.0, + "step": 18004 + }, + { + "epoch": 1.977267735558972, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2487571239471436, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7117061614990234, + "num_tokens": 465838833.0, + "step": 18005 + }, + { + "epoch": 1.9773775532615858, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1784017086029053, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7403126358985901, + "num_tokens": 465865017.0, + "step": 18006 + }, + { + "epoch": 1.9774873709641994, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.307166814804077, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7105503082275391, + "num_tokens": 465887636.0, + "step": 18007 + }, + { + "epoch": 1.977597188666813, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0096161365509033, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6996702551841736, + "num_tokens": 465917893.0, + "step": 18008 + }, + { + "epoch": 1.9777070063694269, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.055485248565674, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7267018556594849, + "num_tokens": 465946956.0, + "step": 18009 + }, + { + "epoch": 1.9778168240720404, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3030500411987305, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7263230681419373, + "num_tokens": 465971921.0, + "step": 18010 + }, + { + "epoch": 1.977926641774654, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.182812452316284, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.718472957611084, + "num_tokens": 465998156.0, + "step": 18011 + }, + { + "epoch": 1.9780364594772677, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.039032220840454, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7161372303962708, + "num_tokens": 466029346.0, + "step": 18012 + }, + { + "epoch": 1.9781462771798815, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.98850679397583, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7312314510345459, + "num_tokens": 466059698.0, + "step": 18013 + }, + { + "epoch": 1.9782560948824952, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.328977584838867, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.7013359665870667, + "num_tokens": 466084530.0, + "step": 18014 + }, + { + "epoch": 1.9783659125851087, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9191094636917114, + "learning_rate": 1e-06, + "loss": 1.003, + "mean_token_accuracy": 0.6934247016906738, + "num_tokens": 466118536.0, + "step": 18015 + }, + { + "epoch": 1.9784757302877223, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2712767124176025, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7316409349441528, + "num_tokens": 466142907.0, + "step": 18016 + }, + { + "epoch": 1.978585547990336, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9647691249847412, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7252035140991211, + "num_tokens": 466175850.0, + "step": 18017 + }, + { + "epoch": 1.9786953656929498, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.408477544784546, + "learning_rate": 1e-06, + "loss": 0.969, + "mean_token_accuracy": 0.6984913349151611, + "num_tokens": 466201036.0, + "step": 18018 + }, + { + "epoch": 1.9788051833955633, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4249155521392822, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7312804460525513, + "num_tokens": 466221700.0, + "step": 18019 + }, + { + "epoch": 1.9789150010981769, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.6345760822296143, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7222170829772949, + "num_tokens": 466241325.0, + "step": 18020 + }, + { + "epoch": 1.9790248188007906, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9864155054092407, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7138495445251465, + "num_tokens": 466272510.0, + "step": 18021 + }, + { + "epoch": 1.9791346365034044, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1945621967315674, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7172733545303345, + "num_tokens": 466297912.0, + "step": 18022 + }, + { + "epoch": 1.9792444542060181, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2353880405426025, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7145558595657349, + "num_tokens": 466324169.0, + "step": 18023 + }, + { + "epoch": 1.9793542719086317, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.091778516769409, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7117575407028198, + "num_tokens": 466351938.0, + "step": 18024 + }, + { + "epoch": 1.9794640896112452, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1168289184570312, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7120039463043213, + "num_tokens": 466381266.0, + "step": 18025 + }, + { + "epoch": 1.979573907313859, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1047680377960205, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7312067151069641, + "num_tokens": 466408555.0, + "step": 18026 + }, + { + "epoch": 1.9796837250164727, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.491969347000122, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7211689352989197, + "num_tokens": 466429737.0, + "step": 18027 + }, + { + "epoch": 1.9797935427190865, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2556557655334473, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7212505340576172, + "num_tokens": 466454995.0, + "step": 18028 + }, + { + "epoch": 1.9799033604217, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0337462425231934, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7073990106582642, + "num_tokens": 466485326.0, + "step": 18029 + }, + { + "epoch": 1.9800131781243135, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9525952339172363, + "learning_rate": 1e-06, + "loss": 0.9633, + "mean_token_accuracy": 0.7005367279052734, + "num_tokens": 466517105.0, + "step": 18030 + }, + { + "epoch": 1.9801229958269273, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.278535842895508, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7311667799949646, + "num_tokens": 466542822.0, + "step": 18031 + }, + { + "epoch": 1.980232813529541, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.413710594177246, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7252693176269531, + "num_tokens": 466566406.0, + "step": 18032 + }, + { + "epoch": 1.9803426312321546, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.138442277908325, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7218420505523682, + "num_tokens": 466594203.0, + "step": 18033 + }, + { + "epoch": 1.9804524489347681, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.014348268508911, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7154019474983215, + "num_tokens": 466624685.0, + "step": 18034 + }, + { + "epoch": 1.9805622666373819, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.074286699295044, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7206643223762512, + "num_tokens": 466652801.0, + "step": 18035 + }, + { + "epoch": 1.9806720843399956, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.229428768157959, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7187663316726685, + "num_tokens": 466678910.0, + "step": 18036 + }, + { + "epoch": 1.9807819020426094, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1489946842193604, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7249166965484619, + "num_tokens": 466707742.0, + "step": 18037 + }, + { + "epoch": 1.980891719745223, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.542938709259033, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7105409502983093, + "num_tokens": 466730741.0, + "step": 18038 + }, + { + "epoch": 1.9810015374478365, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.138162136077881, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7114708423614502, + "num_tokens": 466757928.0, + "step": 18039 + }, + { + "epoch": 1.9811113551504502, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3031482696533203, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7161507606506348, + "num_tokens": 466781909.0, + "step": 18040 + }, + { + "epoch": 1.981221172853064, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.343715190887451, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.7095543146133423, + "num_tokens": 466807173.0, + "step": 18041 + }, + { + "epoch": 1.9813309905556777, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.325974225997925, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7301774024963379, + "num_tokens": 466829869.0, + "step": 18042 + }, + { + "epoch": 1.9814408082582913, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0590925216674805, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7263821959495544, + "num_tokens": 466857696.0, + "step": 18043 + }, + { + "epoch": 1.9815506259609048, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3984808921813965, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7336134910583496, + "num_tokens": 466879323.0, + "step": 18044 + }, + { + "epoch": 1.9816604436635186, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2670228481292725, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7179756164550781, + "num_tokens": 466904558.0, + "step": 18045 + }, + { + "epoch": 1.9817702613661323, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5242972373962402, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7224958539009094, + "num_tokens": 466924408.0, + "step": 18046 + }, + { + "epoch": 1.9818800790687459, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3127264976501465, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7458680272102356, + "num_tokens": 466945415.0, + "step": 18047 + }, + { + "epoch": 1.9819898967713594, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.136821746826172, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7251200675964355, + "num_tokens": 466971936.0, + "step": 18048 + }, + { + "epoch": 1.9820997144739732, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1895503997802734, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.7120957374572754, + "num_tokens": 466999690.0, + "step": 18049 + }, + { + "epoch": 1.982209532176587, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5743257999420166, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7426225543022156, + "num_tokens": 467019527.0, + "step": 18050 + }, + { + "epoch": 1.9823193498792007, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0553042888641357, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7291859984397888, + "num_tokens": 467047899.0, + "step": 18051 + }, + { + "epoch": 1.9824291675818142, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5041558742523193, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7320631742477417, + "num_tokens": 467069050.0, + "step": 18052 + }, + { + "epoch": 1.9825389852844277, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3130435943603516, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7434043884277344, + "num_tokens": 467091624.0, + "step": 18053 + }, + { + "epoch": 1.9826488029870415, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.194789409637451, + "learning_rate": 1e-06, + "loss": 0.9542, + "mean_token_accuracy": 0.6981150507926941, + "num_tokens": 467119004.0, + "step": 18054 + }, + { + "epoch": 1.9827586206896552, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0792229175567627, + "learning_rate": 1e-06, + "loss": 1.0196, + "mean_token_accuracy": 0.6854846477508545, + "num_tokens": 467150163.0, + "step": 18055 + }, + { + "epoch": 1.9828684383922688, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.024531602859497, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7426855564117432, + "num_tokens": 467180353.0, + "step": 18056 + }, + { + "epoch": 1.9829782560948825, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3753931522369385, + "learning_rate": 1e-06, + "loss": 0.7582, + "mean_token_accuracy": 0.7570975422859192, + "num_tokens": 467200651.0, + "step": 18057 + }, + { + "epoch": 1.983088073797496, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.21233868598938, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7395730018615723, + "num_tokens": 467225799.0, + "step": 18058 + }, + { + "epoch": 1.9831978915001098, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.8341987133026123, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7282007932662964, + "num_tokens": 467243075.0, + "step": 18059 + }, + { + "epoch": 1.9833077092027236, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0787928104400635, + "learning_rate": 1e-06, + "loss": 0.9882, + "mean_token_accuracy": 0.6988124847412109, + "num_tokens": 467272737.0, + "step": 18060 + }, + { + "epoch": 1.9834175269053371, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9455361366271973, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7322301864624023, + "num_tokens": 467301245.0, + "step": 18061 + }, + { + "epoch": 1.9835273446079507, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9887118339538574, + "learning_rate": 1e-06, + "loss": 0.9478, + "mean_token_accuracy": 0.7055496573448181, + "num_tokens": 467331176.0, + "step": 18062 + }, + { + "epoch": 1.9836371623105644, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.461214303970337, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7484715580940247, + "num_tokens": 467350422.0, + "step": 18063 + }, + { + "epoch": 1.9837469800131782, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.118518352508545, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7317332625389099, + "num_tokens": 467377058.0, + "step": 18064 + }, + { + "epoch": 1.983856797715792, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.142362356185913, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.7041783928871155, + "num_tokens": 467405378.0, + "step": 18065 + }, + { + "epoch": 1.9839666154184055, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.160336971282959, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7185578346252441, + "num_tokens": 467433938.0, + "step": 18066 + }, + { + "epoch": 1.984076433121019, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4497361183166504, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.699127197265625, + "num_tokens": 467456400.0, + "step": 18067 + }, + { + "epoch": 1.9841862508236328, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0707993507385254, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7137212157249451, + "num_tokens": 467485415.0, + "step": 18068 + }, + { + "epoch": 1.9842960685262465, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.043867826461792, + "learning_rate": 1e-06, + "loss": 1.0239, + "mean_token_accuracy": 0.6893336772918701, + "num_tokens": 467516613.0, + "step": 18069 + }, + { + "epoch": 1.98440588622886, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1658735275268555, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7235857248306274, + "num_tokens": 467541554.0, + "step": 18070 + }, + { + "epoch": 1.9845157039314738, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.475445032119751, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7363455295562744, + "num_tokens": 467562551.0, + "step": 18071 + }, + { + "epoch": 1.9846255216340873, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.283576726913452, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7151215076446533, + "num_tokens": 467587875.0, + "step": 18072 + }, + { + "epoch": 1.984735339336701, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.409773111343384, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7385856509208679, + "num_tokens": 467609473.0, + "step": 18073 + }, + { + "epoch": 1.9848451570393149, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0687801837921143, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7315223813056946, + "num_tokens": 467638318.0, + "step": 18074 + }, + { + "epoch": 1.9849549747419284, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.052502155303955, + "learning_rate": 1e-06, + "loss": 0.9972, + "mean_token_accuracy": 0.6940192580223083, + "num_tokens": 467669113.0, + "step": 18075 + }, + { + "epoch": 1.985064792444542, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.287640333175659, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7019492387771606, + "num_tokens": 467695282.0, + "step": 18076 + }, + { + "epoch": 1.9851746101471557, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.420300006866455, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7235194444656372, + "num_tokens": 467718225.0, + "step": 18077 + }, + { + "epoch": 1.9852844278497694, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4983670711517334, + "learning_rate": 1e-06, + "loss": 0.9635, + "mean_token_accuracy": 0.6977289915084839, + "num_tokens": 467741915.0, + "step": 18078 + }, + { + "epoch": 1.9853942455523832, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4395289421081543, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7306554317474365, + "num_tokens": 467762792.0, + "step": 18079 + }, + { + "epoch": 1.9855040632549967, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4570422172546387, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7193905711174011, + "num_tokens": 467784467.0, + "step": 18080 + }, + { + "epoch": 1.9856138809576103, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.222167730331421, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7319625616073608, + "num_tokens": 467809037.0, + "step": 18081 + }, + { + "epoch": 1.985723698660224, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.35916805267334, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7362722754478455, + "num_tokens": 467833099.0, + "step": 18082 + }, + { + "epoch": 1.9858335163628378, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.195882558822632, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7148540019989014, + "num_tokens": 467861147.0, + "step": 18083 + }, + { + "epoch": 1.9859433340654513, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2810232639312744, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.711532711982727, + "num_tokens": 467886383.0, + "step": 18084 + }, + { + "epoch": 1.9860531517680649, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.6948022842407227, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7470084428787231, + "num_tokens": 467904883.0, + "step": 18085 + }, + { + "epoch": 1.9861629694706786, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.476248264312744, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7200310230255127, + "num_tokens": 467927722.0, + "step": 18086 + }, + { + "epoch": 1.9862727871732924, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.208299160003662, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7401788234710693, + "num_tokens": 467954533.0, + "step": 18087 + }, + { + "epoch": 1.9863826048759061, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.084327220916748, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7495272755622864, + "num_tokens": 467980580.0, + "step": 18088 + }, + { + "epoch": 1.9864924225785197, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.515291690826416, + "learning_rate": 1e-06, + "loss": 0.7936, + "mean_token_accuracy": 0.7429181337356567, + "num_tokens": 468001046.0, + "step": 18089 + }, + { + "epoch": 1.9866022402811332, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4895081520080566, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7349915504455566, + "num_tokens": 468022546.0, + "step": 18090 + }, + { + "epoch": 1.986712057983747, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4522414207458496, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7201442718505859, + "num_tokens": 468045996.0, + "step": 18091 + }, + { + "epoch": 1.9868218756863607, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1251935958862305, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7375386953353882, + "num_tokens": 468073492.0, + "step": 18092 + }, + { + "epoch": 1.9869316933889745, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.45342755317688, + "learning_rate": 1e-06, + "loss": 0.9693, + "mean_token_accuracy": 0.7073320746421814, + "num_tokens": 468097229.0, + "step": 18093 + }, + { + "epoch": 1.987041511091588, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1127889156341553, + "learning_rate": 1e-06, + "loss": 0.9558, + "mean_token_accuracy": 0.6991395950317383, + "num_tokens": 468124449.0, + "step": 18094 + }, + { + "epoch": 1.9871513287942015, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.108214855194092, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7040461897850037, + "num_tokens": 468153002.0, + "step": 18095 + }, + { + "epoch": 1.9872611464968153, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.375218629837036, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7274025082588196, + "num_tokens": 468175728.0, + "step": 18096 + }, + { + "epoch": 1.987370964199429, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.08310866355896, + "learning_rate": 1e-06, + "loss": 1.0093, + "mean_token_accuracy": 0.6898171901702881, + "num_tokens": 468208067.0, + "step": 18097 + }, + { + "epoch": 1.9874807819020426, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5188181400299072, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7349885702133179, + "num_tokens": 468229495.0, + "step": 18098 + }, + { + "epoch": 1.9875905996046561, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1924216747283936, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7308534979820251, + "num_tokens": 468256501.0, + "step": 18099 + }, + { + "epoch": 1.9877004173072699, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.308703899383545, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7404274940490723, + "num_tokens": 468280381.0, + "step": 18100 + }, + { + "epoch": 1.9878102350098836, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9983303546905518, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7134581208229065, + "num_tokens": 468311940.0, + "step": 18101 + }, + { + "epoch": 1.9879200527124974, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.502652168273926, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7214713096618652, + "num_tokens": 468335383.0, + "step": 18102 + }, + { + "epoch": 1.988029870415111, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.358499050140381, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7076190114021301, + "num_tokens": 468358251.0, + "step": 18103 + }, + { + "epoch": 1.9881396881177245, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3724732398986816, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7492825984954834, + "num_tokens": 468381560.0, + "step": 18104 + }, + { + "epoch": 1.9882495058203382, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.436809539794922, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7064359784126282, + "num_tokens": 468405235.0, + "step": 18105 + }, + { + "epoch": 1.988359323522952, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1278531551361084, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.728186845779419, + "num_tokens": 468433779.0, + "step": 18106 + }, + { + "epoch": 1.9884691412255655, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9218343496322632, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.732353925704956, + "num_tokens": 468465278.0, + "step": 18107 + }, + { + "epoch": 1.9885789589281793, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4299657344818115, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7375949621200562, + "num_tokens": 468489088.0, + "step": 18108 + }, + { + "epoch": 1.9886887766307928, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0829460620880127, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7105783224105835, + "num_tokens": 468521256.0, + "step": 18109 + }, + { + "epoch": 1.9887985943334066, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4743831157684326, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.735734224319458, + "num_tokens": 468541124.0, + "step": 18110 + }, + { + "epoch": 1.9889084120360203, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3655240535736084, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7258647680282593, + "num_tokens": 468565961.0, + "step": 18111 + }, + { + "epoch": 1.9890182297386338, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.005284309387207, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7111486792564392, + "num_tokens": 468595579.0, + "step": 18112 + }, + { + "epoch": 1.9891280474412474, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2387442588806152, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7243788242340088, + "num_tokens": 468621118.0, + "step": 18113 + }, + { + "epoch": 1.9892378651438611, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5156631469726562, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7317399978637695, + "num_tokens": 468641545.0, + "step": 18114 + }, + { + "epoch": 1.989347682846475, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1388444900512695, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7194691896438599, + "num_tokens": 468668558.0, + "step": 18115 + }, + { + "epoch": 1.9894575005490887, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1026415824890137, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7011824250221252, + "num_tokens": 468697598.0, + "step": 18116 + }, + { + "epoch": 1.9895673182517022, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.08935809135437, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7094516754150391, + "num_tokens": 468726401.0, + "step": 18117 + }, + { + "epoch": 1.9896771359543157, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4672389030456543, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7083219289779663, + "num_tokens": 468748943.0, + "step": 18118 + }, + { + "epoch": 1.9897869536569295, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0144636631011963, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7167117595672607, + "num_tokens": 468777685.0, + "step": 18119 + }, + { + "epoch": 1.9898967713595432, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.391549825668335, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7234658598899841, + "num_tokens": 468800301.0, + "step": 18120 + }, + { + "epoch": 1.9900065890621568, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0937910079956055, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7259691953659058, + "num_tokens": 468826461.0, + "step": 18121 + }, + { + "epoch": 1.9901164067647705, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.93743896484375, + "learning_rate": 1e-06, + "loss": 1.0432, + "mean_token_accuracy": 0.6957740783691406, + "num_tokens": 468858553.0, + "step": 18122 + }, + { + "epoch": 1.990226224467384, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9909366369247437, + "learning_rate": 1e-06, + "loss": 0.9724, + "mean_token_accuracy": 0.7006270289421082, + "num_tokens": 468888908.0, + "step": 18123 + }, + { + "epoch": 1.9903360421699978, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.981881856918335, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6916638016700745, + "num_tokens": 468922491.0, + "step": 18124 + }, + { + "epoch": 1.9904458598726116, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9830117225646973, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.709709644317627, + "num_tokens": 468954283.0, + "step": 18125 + }, + { + "epoch": 1.9905556775752251, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.378082513809204, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7132529616355896, + "num_tokens": 468977253.0, + "step": 18126 + }, + { + "epoch": 1.9906654952778386, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.453040599822998, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7319715619087219, + "num_tokens": 468997728.0, + "step": 18127 + }, + { + "epoch": 1.9907753129804524, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.287280559539795, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7227574586868286, + "num_tokens": 469022017.0, + "step": 18128 + }, + { + "epoch": 1.9908851306830662, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2642815113067627, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7242193222045898, + "num_tokens": 469045797.0, + "step": 18129 + }, + { + "epoch": 1.99099494838568, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1063432693481445, + "learning_rate": 1e-06, + "loss": 0.9785, + "mean_token_accuracy": 0.6967289447784424, + "num_tokens": 469077174.0, + "step": 18130 + }, + { + "epoch": 1.9911047660882935, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.096789598464966, + "learning_rate": 1e-06, + "loss": 1.0059, + "mean_token_accuracy": 0.6952475905418396, + "num_tokens": 469104610.0, + "step": 18131 + }, + { + "epoch": 1.991214583790907, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9265623092651367, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7133193016052246, + "num_tokens": 469136746.0, + "step": 18132 + }, + { + "epoch": 1.9913244014935207, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.013573169708252, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.708716630935669, + "num_tokens": 469167149.0, + "step": 18133 + }, + { + "epoch": 1.9914342191961345, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0737764835357666, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7147008776664734, + "num_tokens": 469194948.0, + "step": 18134 + }, + { + "epoch": 1.991544036898748, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.135924816131592, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7161484956741333, + "num_tokens": 469220938.0, + "step": 18135 + }, + { + "epoch": 1.9916538546013618, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.87718665599823, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7242760062217712, + "num_tokens": 469257117.0, + "step": 18136 + }, + { + "epoch": 1.9917636723039753, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1585605144500732, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7179864645004272, + "num_tokens": 469284383.0, + "step": 18137 + }, + { + "epoch": 1.991873490006589, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3682808876037598, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7271289229393005, + "num_tokens": 469307692.0, + "step": 18138 + }, + { + "epoch": 1.9919833077092028, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.371924638748169, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7170333862304688, + "num_tokens": 469330147.0, + "step": 18139 + }, + { + "epoch": 1.9920931254118164, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.205430030822754, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7223356366157532, + "num_tokens": 469355986.0, + "step": 18140 + }, + { + "epoch": 1.99220294311443, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.173370122909546, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7231049537658691, + "num_tokens": 469384324.0, + "step": 18141 + }, + { + "epoch": 1.9923127608170437, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.9378050565719604, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7248409986495972, + "num_tokens": 469414795.0, + "step": 18142 + }, + { + "epoch": 1.9924225785196574, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4165990352630615, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.712067723274231, + "num_tokens": 469437569.0, + "step": 18143 + }, + { + "epoch": 1.9925323962222712, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3283956050872803, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7424514293670654, + "num_tokens": 469461788.0, + "step": 18144 + }, + { + "epoch": 1.9926422139248847, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.127457857131958, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7131216526031494, + "num_tokens": 469489058.0, + "step": 18145 + }, + { + "epoch": 1.9927520316274983, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0708508491516113, + "learning_rate": 1e-06, + "loss": 1.0111, + "mean_token_accuracy": 0.6930287480354309, + "num_tokens": 469518606.0, + "step": 18146 + }, + { + "epoch": 1.992861849330112, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2686548233032227, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7236953973770142, + "num_tokens": 469542224.0, + "step": 18147 + }, + { + "epoch": 1.9929716670327258, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.081139087677002, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7154175043106079, + "num_tokens": 469571020.0, + "step": 18148 + }, + { + "epoch": 1.9930814847353393, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0657801628112793, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.726651132106781, + "num_tokens": 469598688.0, + "step": 18149 + }, + { + "epoch": 1.9931913024379528, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.003147840499878, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7271695137023926, + "num_tokens": 469628287.0, + "step": 18150 + }, + { + "epoch": 1.9933011201405666, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 1.940799593925476, + "learning_rate": 1e-06, + "loss": 0.9666, + "mean_token_accuracy": 0.7070227861404419, + "num_tokens": 469661848.0, + "step": 18151 + }, + { + "epoch": 1.9934109378431804, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0544469356536865, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7282222509384155, + "num_tokens": 469691358.0, + "step": 18152 + }, + { + "epoch": 1.993520755545794, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.22404146194458, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7222598791122437, + "num_tokens": 469719859.0, + "step": 18153 + }, + { + "epoch": 1.9936305732484076, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2516777515411377, + "learning_rate": 1e-06, + "loss": 0.947, + "mean_token_accuracy": 0.7064282894134521, + "num_tokens": 469744208.0, + "step": 18154 + }, + { + "epoch": 1.9937403909510212, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0378568172454834, + "learning_rate": 1e-06, + "loss": 0.9594, + "mean_token_accuracy": 0.7043502330780029, + "num_tokens": 469775148.0, + "step": 18155 + }, + { + "epoch": 1.993850208653635, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5574951171875, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7556843757629395, + "num_tokens": 469796709.0, + "step": 18156 + }, + { + "epoch": 1.9939600263562487, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4970037937164307, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7168911099433899, + "num_tokens": 469819795.0, + "step": 18157 + }, + { + "epoch": 1.9940698440588625, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.235234022140503, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7419270277023315, + "num_tokens": 469844011.0, + "step": 18158 + }, + { + "epoch": 1.994179661761476, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.325777292251587, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7208157777786255, + "num_tokens": 469867553.0, + "step": 18159 + }, + { + "epoch": 1.9942894794640895, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.166874408721924, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7255910038948059, + "num_tokens": 469896905.0, + "step": 18160 + }, + { + "epoch": 1.9943992971667033, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3898000717163086, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.704699695110321, + "num_tokens": 469920200.0, + "step": 18161 + }, + { + "epoch": 1.994509114869317, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.4731791019439697, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7193430662155151, + "num_tokens": 469944573.0, + "step": 18162 + }, + { + "epoch": 1.9946189325719306, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.285834312438965, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.722650408744812, + "num_tokens": 469967220.0, + "step": 18163 + }, + { + "epoch": 1.994728750274544, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.5546977519989014, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7311740517616272, + "num_tokens": 469987942.0, + "step": 18164 + }, + { + "epoch": 1.9948385679771579, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1032726764678955, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7204211354255676, + "num_tokens": 470017012.0, + "step": 18165 + }, + { + "epoch": 1.9949483856797716, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.207927942276001, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7261308431625366, + "num_tokens": 470042843.0, + "step": 18166 + }, + { + "epoch": 1.9950582033823854, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.145578384399414, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7435758709907532, + "num_tokens": 470069002.0, + "step": 18167 + }, + { + "epoch": 1.995168021084999, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.43749737739563, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.732254683971405, + "num_tokens": 470089815.0, + "step": 18168 + }, + { + "epoch": 1.9952778387876124, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.0121572017669678, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7078129053115845, + "num_tokens": 470121345.0, + "step": 18169 + }, + { + "epoch": 1.9953876564902262, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.103020668029785, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7331758737564087, + "num_tokens": 470148531.0, + "step": 18170 + }, + { + "epoch": 1.99549747419284, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.3024208545684814, + "learning_rate": 1e-06, + "loss": 1.0009, + "mean_token_accuracy": 0.6994706988334656, + "num_tokens": 470174290.0, + "step": 18171 + }, + { + "epoch": 1.9956072918954535, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.1980197429656982, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7153609395027161, + "num_tokens": 470201166.0, + "step": 18172 + }, + { + "epoch": 1.9957171095980673, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.47857403755188, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7225285768508911, + "num_tokens": 470222186.0, + "step": 18173 + }, + { + "epoch": 1.9958269273006808, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 1.964518427848816, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7359427213668823, + "num_tokens": 470253168.0, + "step": 18174 + }, + { + "epoch": 1.9959367450032945, + "ewc_loss": 1.055002212524414e-05, + "grad_norm": 2.2493584156036377, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7173081636428833, + "num_tokens": 470281135.0, + "step": 18175 + }, + { + "epoch": 1.9960465627059083, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.2211358547210693, + "learning_rate": 1e-06, + "loss": 0.9856, + "mean_token_accuracy": 0.7150324583053589, + "num_tokens": 470307947.0, + "step": 18176 + }, + { + "epoch": 1.9961563804085218, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.0969202518463135, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.751173734664917, + "num_tokens": 470335331.0, + "step": 18177 + }, + { + "epoch": 1.9962661981111354, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.047790288925171, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7121115922927856, + "num_tokens": 470365413.0, + "step": 18178 + }, + { + "epoch": 1.9963760158137491, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.025494337081909, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7029188275337219, + "num_tokens": 470396737.0, + "step": 18179 + }, + { + "epoch": 1.9964858335163629, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.2946929931640625, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7118260860443115, + "num_tokens": 470422654.0, + "step": 18180 + }, + { + "epoch": 1.9965956512189766, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.12050724029541, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7282140254974365, + "num_tokens": 470449058.0, + "step": 18181 + }, + { + "epoch": 1.9967054689215902, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.3649826049804688, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7409217357635498, + "num_tokens": 470471733.0, + "step": 18182 + }, + { + "epoch": 1.9968152866242037, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.4200081825256348, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7194709777832031, + "num_tokens": 470494067.0, + "step": 18183 + }, + { + "epoch": 1.9969251043268175, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.1516926288604736, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.717918872833252, + "num_tokens": 470521583.0, + "step": 18184 + }, + { + "epoch": 1.9970349220294312, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.2803263664245605, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7171491384506226, + "num_tokens": 470545232.0, + "step": 18185 + }, + { + "epoch": 1.9971447397320448, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.0703341960906982, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7250380516052246, + "num_tokens": 470573581.0, + "step": 18186 + }, + { + "epoch": 1.9972545574346585, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.5091800689697266, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7274298071861267, + "num_tokens": 470594915.0, + "step": 18187 + }, + { + "epoch": 1.997364375137272, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.2567718029022217, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7255687713623047, + "num_tokens": 470619696.0, + "step": 18188 + }, + { + "epoch": 1.9974741928398858, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.1339352130889893, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6961519718170166, + "num_tokens": 470648134.0, + "step": 18189 + }, + { + "epoch": 1.9975840105424996, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.289569616317749, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.732520341873169, + "num_tokens": 470672694.0, + "step": 18190 + }, + { + "epoch": 1.997693828245113, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.145228624343872, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7565696239471436, + "num_tokens": 470698698.0, + "step": 18191 + }, + { + "epoch": 1.9978036459477266, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.1197502613067627, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7298706769943237, + "num_tokens": 470725993.0, + "step": 18192 + }, + { + "epoch": 1.9979134636503404, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.192289113998413, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.7153307199478149, + "num_tokens": 470751735.0, + "step": 18193 + }, + { + "epoch": 1.9980232813529542, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 3.0983526706695557, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7388318181037903, + "num_tokens": 470768977.0, + "step": 18194 + }, + { + "epoch": 1.998133099055568, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.0090324878692627, + "learning_rate": 1e-06, + "loss": 0.9165, + "mean_token_accuracy": 0.714354395866394, + "num_tokens": 470797383.0, + "step": 18195 + }, + { + "epoch": 1.9982429167581814, + "ewc_loss": 1.0609626770019531e-05, + "grad_norm": 2.4256365299224854, + "learning_rate": 1e-06, + "loss": 0.7727, + "mean_token_accuracy": 0.7456895709037781, + "num_tokens": 470818262.0, + "step": 18196 + }, + { + "epoch": 1.998352734460795, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.2065541744232178, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7091392278671265, + "num_tokens": 470845484.0, + "step": 18197 + }, + { + "epoch": 1.9984625521634087, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.3357019424438477, + "learning_rate": 1e-06, + "loss": 1.0002, + "mean_token_accuracy": 0.700406551361084, + "num_tokens": 470870283.0, + "step": 18198 + }, + { + "epoch": 1.9985723698660225, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.455326557159424, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7311717867851257, + "num_tokens": 470891092.0, + "step": 18199 + }, + { + "epoch": 1.998682187568636, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.18160343170166, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7273160219192505, + "num_tokens": 470916592.0, + "step": 18200 + }, + { + "epoch": 1.9987920052712496, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.156841516494751, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7392761707305908, + "num_tokens": 470943251.0, + "step": 18201 + }, + { + "epoch": 1.9989018229738633, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.3077776432037354, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.735693097114563, + "num_tokens": 470970150.0, + "step": 18202 + }, + { + "epoch": 1.999011640676477, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 1.9574558734893799, + "learning_rate": 1e-06, + "loss": 0.9886, + "mean_token_accuracy": 0.6938722133636475, + "num_tokens": 471001184.0, + "step": 18203 + }, + { + "epoch": 1.9991214583790908, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.325165033340454, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7245347499847412, + "num_tokens": 471025144.0, + "step": 18204 + }, + { + "epoch": 1.9992312760817044, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 1.9325858354568481, + "learning_rate": 1e-06, + "loss": 0.9825, + "mean_token_accuracy": 0.7057605981826782, + "num_tokens": 471059042.0, + "step": 18205 + }, + { + "epoch": 1.999341093784318, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 1.9942443370819092, + "learning_rate": 1e-06, + "loss": 1.0124, + "mean_token_accuracy": 0.6931479573249817, + "num_tokens": 471090984.0, + "step": 18206 + }, + { + "epoch": 1.9994509114869317, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.418107509613037, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.733526349067688, + "num_tokens": 471111508.0, + "step": 18207 + }, + { + "epoch": 1.9995607291895454, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.251080274581909, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7094980478286743, + "num_tokens": 471136394.0, + "step": 18208 + }, + { + "epoch": 1.9996705468921592, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.207872152328491, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7167853713035583, + "num_tokens": 471164069.0, + "step": 18209 + }, + { + "epoch": 1.9997803645947727, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.312894105911255, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7199431657791138, + "num_tokens": 471189023.0, + "step": 18210 + }, + { + "epoch": 1.9998901822973862, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.493185043334961, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7351716160774231, + "num_tokens": 471209443.0, + "step": 18211 + }, + { + "epoch": 2.0, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.23978328704834, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.707793653011322, + "num_tokens": 471235014.0, + "step": 18212 + }, + { + "epoch": 2.0001098177026138, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.186476469039917, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7308070659637451, + "num_tokens": 471260654.0, + "step": 18213 + }, + { + "epoch": 2.0002196354052275, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.067966938018799, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7292884588241577, + "num_tokens": 471286421.0, + "step": 18214 + }, + { + "epoch": 2.000329453107841, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.1988203525543213, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.730724573135376, + "num_tokens": 471310258.0, + "step": 18215 + }, + { + "epoch": 2.0004392708104546, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.0777978897094727, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7241271734237671, + "num_tokens": 471338723.0, + "step": 18216 + }, + { + "epoch": 2.0005490885130683, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.0513124465942383, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7300291061401367, + "num_tokens": 471369007.0, + "step": 18217 + }, + { + "epoch": 2.000658906215682, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4199323654174805, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7449621558189392, + "num_tokens": 471390448.0, + "step": 18218 + }, + { + "epoch": 2.0007687239182954, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.0154247283935547, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.719764232635498, + "num_tokens": 471420002.0, + "step": 18219 + }, + { + "epoch": 2.000878541620909, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4998600482940674, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7256611585617065, + "num_tokens": 471442498.0, + "step": 18220 + }, + { + "epoch": 2.000988359323523, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.5209712982177734, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7356828451156616, + "num_tokens": 471463307.0, + "step": 18221 + }, + { + "epoch": 2.0010981770261367, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.287076234817505, + "learning_rate": 1e-06, + "loss": 0.7411, + "mean_token_accuracy": 0.7549124956130981, + "num_tokens": 471487919.0, + "step": 18222 + }, + { + "epoch": 2.0012079947287504, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.189697504043579, + "learning_rate": 1e-06, + "loss": 0.7686, + "mean_token_accuracy": 0.7578709125518799, + "num_tokens": 471514828.0, + "step": 18223 + }, + { + "epoch": 2.0013178124313638, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4879586696624756, + "learning_rate": 1e-06, + "loss": 0.7502, + "mean_token_accuracy": 0.7576148509979248, + "num_tokens": 471537025.0, + "step": 18224 + }, + { + "epoch": 2.0014276301339775, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.5108492374420166, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7271302938461304, + "num_tokens": 471557930.0, + "step": 18225 + }, + { + "epoch": 2.0015374478365913, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.2452151775360107, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7222776412963867, + "num_tokens": 471582641.0, + "step": 18226 + }, + { + "epoch": 2.001647265539205, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.4233975410461426, + "learning_rate": 1e-06, + "loss": 0.7251, + "mean_token_accuracy": 0.7580405473709106, + "num_tokens": 471604738.0, + "step": 18227 + }, + { + "epoch": 2.001757083241819, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.2202696800231934, + "learning_rate": 1e-06, + "loss": 0.937, + "mean_token_accuracy": 0.7028297185897827, + "num_tokens": 471634425.0, + "step": 18228 + }, + { + "epoch": 2.001866900944432, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.23266863822937, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7307177782058716, + "num_tokens": 471660871.0, + "step": 18229 + }, + { + "epoch": 2.001976718647046, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.1759519577026367, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7303972244262695, + "num_tokens": 471688335.0, + "step": 18230 + }, + { + "epoch": 2.0020865363496596, + "ewc_loss": 1.0669231414794922e-05, + "grad_norm": 2.1007165908813477, + "learning_rate": 1e-06, + "loss": 0.9328, + "mean_token_accuracy": 0.7094002962112427, + "num_tokens": 471718337.0, + "step": 18231 + }, + { + "epoch": 2.0021963540522734, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.9732669591903687, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7154520750045776, + "num_tokens": 471751966.0, + "step": 18232 + }, + { + "epoch": 2.0023061717548867, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5065441131591797, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.7511072158813477, + "num_tokens": 471773386.0, + "step": 18233 + }, + { + "epoch": 2.0024159894575004, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.465053081512451, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7356410026550293, + "num_tokens": 471795477.0, + "step": 18234 + }, + { + "epoch": 2.002525807160114, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.047492742538452, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7251440286636353, + "num_tokens": 471825957.0, + "step": 18235 + }, + { + "epoch": 2.002635624862728, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2212023735046387, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7229952216148376, + "num_tokens": 471853767.0, + "step": 18236 + }, + { + "epoch": 2.0027454425653417, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2282557487487793, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7190385460853577, + "num_tokens": 471883082.0, + "step": 18237 + }, + { + "epoch": 2.002855260267955, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.20819091796875, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7377181649208069, + "num_tokens": 471910218.0, + "step": 18238 + }, + { + "epoch": 2.0029650779705688, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2669756412506104, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7271148562431335, + "num_tokens": 471936993.0, + "step": 18239 + }, + { + "epoch": 2.0030748956731825, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2859561443328857, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.703425407409668, + "num_tokens": 471964226.0, + "step": 18240 + }, + { + "epoch": 2.0031847133757963, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.413508892059326, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.731135368347168, + "num_tokens": 471986803.0, + "step": 18241 + }, + { + "epoch": 2.00329453107841, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4671318531036377, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7317598462104797, + "num_tokens": 472010234.0, + "step": 18242 + }, + { + "epoch": 2.0034043487810234, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.6036629676818848, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7356000542640686, + "num_tokens": 472032126.0, + "step": 18243 + }, + { + "epoch": 2.003514166483637, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.373239755630493, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7257825136184692, + "num_tokens": 472057020.0, + "step": 18244 + }, + { + "epoch": 2.003623984186251, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5405845642089844, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7214484214782715, + "num_tokens": 472079754.0, + "step": 18245 + }, + { + "epoch": 2.0037338018888646, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0979158878326416, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7496870756149292, + "num_tokens": 472107357.0, + "step": 18246 + }, + { + "epoch": 2.003843619591478, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3909783363342285, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7403638362884521, + "num_tokens": 472132314.0, + "step": 18247 + }, + { + "epoch": 2.0039534372940917, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3063597679138184, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7204006910324097, + "num_tokens": 472159056.0, + "step": 18248 + }, + { + "epoch": 2.0040632549967055, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1803884506225586, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7485488057136536, + "num_tokens": 472187610.0, + "step": 18249 + }, + { + "epoch": 2.004173072699319, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4430463314056396, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7353249788284302, + "num_tokens": 472211027.0, + "step": 18250 + }, + { + "epoch": 2.004282890401933, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.153938055038452, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7275130748748779, + "num_tokens": 472240380.0, + "step": 18251 + }, + { + "epoch": 2.0043927081045463, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0656375885009766, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7506652474403381, + "num_tokens": 472270893.0, + "step": 18252 + }, + { + "epoch": 2.00450252580716, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1988227367401123, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7480717897415161, + "num_tokens": 472296851.0, + "step": 18253 + }, + { + "epoch": 2.004612343509774, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.255004405975342, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7300735712051392, + "num_tokens": 472324270.0, + "step": 18254 + }, + { + "epoch": 2.0047221612123876, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2323737144470215, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.740349292755127, + "num_tokens": 472350052.0, + "step": 18255 + }, + { + "epoch": 2.0048319789150013, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.212714433670044, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7319315671920776, + "num_tokens": 472377760.0, + "step": 18256 + }, + { + "epoch": 2.0049417966176146, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.9408648014068604, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7225598096847534, + "num_tokens": 472413478.0, + "step": 18257 + }, + { + "epoch": 2.0050516143202284, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.263317584991455, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7038744688034058, + "num_tokens": 472441679.0, + "step": 18258 + }, + { + "epoch": 2.005161432022842, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2593493461608887, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7463539838790894, + "num_tokens": 472467174.0, + "step": 18259 + }, + { + "epoch": 2.005271249725456, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.291313886642456, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7411954998970032, + "num_tokens": 472491450.0, + "step": 18260 + }, + { + "epoch": 2.005381067428069, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.54445219039917, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7514766454696655, + "num_tokens": 472513867.0, + "step": 18261 + }, + { + "epoch": 2.005490885130683, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.300032138824463, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7404289245605469, + "num_tokens": 472538909.0, + "step": 18262 + }, + { + "epoch": 2.0056007028332967, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.8934577703475952, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7228308916091919, + "num_tokens": 472573516.0, + "step": 18263 + }, + { + "epoch": 2.0057105205359105, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1748785972595215, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7165003418922424, + "num_tokens": 472604040.0, + "step": 18264 + }, + { + "epoch": 2.0058203382385242, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4053478240966797, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7394550442695618, + "num_tokens": 472628842.0, + "step": 18265 + }, + { + "epoch": 2.0059301559411375, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4444189071655273, + "learning_rate": 1e-06, + "loss": 0.9417, + "mean_token_accuracy": 0.7099360227584839, + "num_tokens": 472653793.0, + "step": 18266 + }, + { + "epoch": 2.0060399736437513, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2537405490875244, + "learning_rate": 1e-06, + "loss": 0.9426, + "mean_token_accuracy": 0.7058662176132202, + "num_tokens": 472682750.0, + "step": 18267 + }, + { + "epoch": 2.006149791346365, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2737603187561035, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7061156630516052, + "num_tokens": 472708280.0, + "step": 18268 + }, + { + "epoch": 2.006259609048979, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3328018188476562, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7203912734985352, + "num_tokens": 472732535.0, + "step": 18269 + }, + { + "epoch": 2.0063694267515926, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1338322162628174, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7110434174537659, + "num_tokens": 472761119.0, + "step": 18270 + }, + { + "epoch": 2.006479244454206, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.70124888420105, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7436884641647339, + "num_tokens": 472782492.0, + "step": 18271 + }, + { + "epoch": 2.0065890621568196, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.250931978225708, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7241591215133667, + "num_tokens": 472809592.0, + "step": 18272 + }, + { + "epoch": 2.0066988798594334, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1607346534729004, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7262840270996094, + "num_tokens": 472838346.0, + "step": 18273 + }, + { + "epoch": 2.006808697562047, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.171044111251831, + "learning_rate": 1e-06, + "loss": 0.9363, + "mean_token_accuracy": 0.7140432596206665, + "num_tokens": 472868408.0, + "step": 18274 + }, + { + "epoch": 2.0069185152646605, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4599454402923584, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7444345355033875, + "num_tokens": 472891999.0, + "step": 18275 + }, + { + "epoch": 2.0070283329672742, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.7734761238098145, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7491053938865662, + "num_tokens": 472912478.0, + "step": 18276 + }, + { + "epoch": 2.007138150669888, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.372725248336792, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7213315367698669, + "num_tokens": 472937871.0, + "step": 18277 + }, + { + "epoch": 2.0072479683725017, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1033170223236084, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7127413153648376, + "num_tokens": 472968653.0, + "step": 18278 + }, + { + "epoch": 2.0073577860751155, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.278412342071533, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7338457107543945, + "num_tokens": 472993966.0, + "step": 18279 + }, + { + "epoch": 2.007467603777729, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4908013343811035, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7279822826385498, + "num_tokens": 473019198.0, + "step": 18280 + }, + { + "epoch": 2.0075774214803426, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.226011037826538, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7228102684020996, + "num_tokens": 473047784.0, + "step": 18281 + }, + { + "epoch": 2.0076872391829563, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.360684633255005, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7208026647567749, + "num_tokens": 473073103.0, + "step": 18282 + }, + { + "epoch": 2.00779705688557, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.389446973800659, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7462136745452881, + "num_tokens": 473098209.0, + "step": 18283 + }, + { + "epoch": 2.0079068745881834, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.367729425430298, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.6995086669921875, + "num_tokens": 473123539.0, + "step": 18284 + }, + { + "epoch": 2.008016692290797, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.9816944599151611, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7125421762466431, + "num_tokens": 473158145.0, + "step": 18285 + }, + { + "epoch": 2.008126509993411, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1933224201202393, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7212225198745728, + "num_tokens": 473185679.0, + "step": 18286 + }, + { + "epoch": 2.0082363276960247, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.6725990772247314, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7409748435020447, + "num_tokens": 473205197.0, + "step": 18287 + }, + { + "epoch": 2.0083461453986384, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.305424213409424, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.739084005355835, + "num_tokens": 473230948.0, + "step": 18288 + }, + { + "epoch": 2.0084559631012517, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5700345039367676, + "learning_rate": 1e-06, + "loss": 0.7429, + "mean_token_accuracy": 0.7597724795341492, + "num_tokens": 473251055.0, + "step": 18289 + }, + { + "epoch": 2.0085657808038655, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.474498748779297, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7354758977890015, + "num_tokens": 473274792.0, + "step": 18290 + }, + { + "epoch": 2.0086755985064793, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0202531814575195, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7533628940582275, + "num_tokens": 473304591.0, + "step": 18291 + }, + { + "epoch": 2.008785416209093, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2063498497009277, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7360572218894958, + "num_tokens": 473333610.0, + "step": 18292 + }, + { + "epoch": 2.0088952339117068, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.749861478805542, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7387377619743347, + "num_tokens": 473352052.0, + "step": 18293 + }, + { + "epoch": 2.00900505161432, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0551931858062744, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7136374711990356, + "num_tokens": 473384444.0, + "step": 18294 + }, + { + "epoch": 2.009114869316934, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0370309352874756, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7229598164558411, + "num_tokens": 473415544.0, + "step": 18295 + }, + { + "epoch": 2.0092246870195476, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.337531328201294, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7260673642158508, + "num_tokens": 473443069.0, + "step": 18296 + }, + { + "epoch": 2.0093345047221614, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.314988136291504, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7415651679039001, + "num_tokens": 473467650.0, + "step": 18297 + }, + { + "epoch": 2.0094443224247747, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.118438482284546, + "learning_rate": 1e-06, + "loss": 0.7547, + "mean_token_accuracy": 0.740338146686554, + "num_tokens": 473495979.0, + "step": 18298 + }, + { + "epoch": 2.0095541401273884, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.20576810836792, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7424167990684509, + "num_tokens": 473521867.0, + "step": 18299 + }, + { + "epoch": 2.009663957830002, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.13254714012146, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7366459369659424, + "num_tokens": 473551775.0, + "step": 18300 + }, + { + "epoch": 2.009773775532616, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.8050527572631836, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7550919651985168, + "num_tokens": 473570580.0, + "step": 18301 + }, + { + "epoch": 2.0098835932352297, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.7378273010253906, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7406516671180725, + "num_tokens": 473589849.0, + "step": 18302 + }, + { + "epoch": 2.009993410937843, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.028553009033203, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7271484136581421, + "num_tokens": 473621943.0, + "step": 18303 + }, + { + "epoch": 2.0101032286404568, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.463520050048828, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7390653491020203, + "num_tokens": 473644412.0, + "step": 18304 + }, + { + "epoch": 2.0102130463430705, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 1.893390417098999, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7429754734039307, + "num_tokens": 473681233.0, + "step": 18305 + }, + { + "epoch": 2.0103228640456843, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5477583408355713, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7427175045013428, + "num_tokens": 473703760.0, + "step": 18306 + }, + { + "epoch": 2.010432681748298, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3222501277923584, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7297487258911133, + "num_tokens": 473728373.0, + "step": 18307 + }, + { + "epoch": 2.0105424994509113, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2278082370758057, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7287329435348511, + "num_tokens": 473754675.0, + "step": 18308 + }, + { + "epoch": 2.010652317153525, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4516379833221436, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7382887601852417, + "num_tokens": 473776730.0, + "step": 18309 + }, + { + "epoch": 2.010762134856139, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.2309250831604004, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7244798541069031, + "num_tokens": 473806682.0, + "step": 18310 + }, + { + "epoch": 2.0108719525587526, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1541144847869873, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7092375159263611, + "num_tokens": 473837602.0, + "step": 18311 + }, + { + "epoch": 2.010981770261366, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3376500606536865, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7376710176467896, + "num_tokens": 473861516.0, + "step": 18312 + }, + { + "epoch": 2.0110915879639797, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.237107038497925, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7492448091506958, + "num_tokens": 473887367.0, + "step": 18313 + }, + { + "epoch": 2.0112014056665934, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.243018865585327, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.726718544960022, + "num_tokens": 473914911.0, + "step": 18314 + }, + { + "epoch": 2.011311223369207, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4041695594787598, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7254197597503662, + "num_tokens": 473941329.0, + "step": 18315 + }, + { + "epoch": 2.011421041071821, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.72985577583313, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7265530824661255, + "num_tokens": 473962147.0, + "step": 18316 + }, + { + "epoch": 2.0115308587744343, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.293663263320923, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7226502895355225, + "num_tokens": 473989895.0, + "step": 18317 + }, + { + "epoch": 2.011640676477048, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4317715167999268, + "learning_rate": 1e-06, + "loss": 0.7246, + "mean_token_accuracy": 0.7659417390823364, + "num_tokens": 474010572.0, + "step": 18318 + }, + { + "epoch": 2.011750494179662, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.6228551864624023, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7312033176422119, + "num_tokens": 474032697.0, + "step": 18319 + }, + { + "epoch": 2.0118603118822755, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0255768299102783, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7524557709693909, + "num_tokens": 474063383.0, + "step": 18320 + }, + { + "epoch": 2.0119701295848893, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.238865852355957, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7285696268081665, + "num_tokens": 474091568.0, + "step": 18321 + }, + { + "epoch": 2.0120799472875026, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.380572557449341, + "learning_rate": 1e-06, + "loss": 0.7507, + "mean_token_accuracy": 0.7588304877281189, + "num_tokens": 474113882.0, + "step": 18322 + }, + { + "epoch": 2.0121897649901164, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1694345474243164, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7327202558517456, + "num_tokens": 474143063.0, + "step": 18323 + }, + { + "epoch": 2.01229958269273, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.279970645904541, + "learning_rate": 1e-06, + "loss": 0.7183, + "mean_token_accuracy": 0.7712731957435608, + "num_tokens": 474167338.0, + "step": 18324 + }, + { + "epoch": 2.012409400395344, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.544658899307251, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7462625503540039, + "num_tokens": 474191108.0, + "step": 18325 + }, + { + "epoch": 2.012519218097957, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.317430019378662, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.732525110244751, + "num_tokens": 474216320.0, + "step": 18326 + }, + { + "epoch": 2.012629035800571, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0628068447113037, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7296690940856934, + "num_tokens": 474248732.0, + "step": 18327 + }, + { + "epoch": 2.0127388535031847, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4419915676116943, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7547908425331116, + "num_tokens": 474270619.0, + "step": 18328 + }, + { + "epoch": 2.0128486712057985, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.510969877243042, + "learning_rate": 1e-06, + "loss": 0.715, + "mean_token_accuracy": 0.76781165599823, + "num_tokens": 474293168.0, + "step": 18329 + }, + { + "epoch": 2.0129584889084122, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3722383975982666, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7319809198379517, + "num_tokens": 474318595.0, + "step": 18330 + }, + { + "epoch": 2.0130683066110255, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.237015962600708, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.7362003326416016, + "num_tokens": 474346507.0, + "step": 18331 + }, + { + "epoch": 2.0131781243136393, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.270535469055176, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7156233787536621, + "num_tokens": 474374731.0, + "step": 18332 + }, + { + "epoch": 2.013287942016253, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5650761127471924, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7490955591201782, + "num_tokens": 474396677.0, + "step": 18333 + }, + { + "epoch": 2.013397759718867, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1905357837677, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7212953567504883, + "num_tokens": 474424538.0, + "step": 18334 + }, + { + "epoch": 2.01350757742148, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.6912944316864014, + "learning_rate": 1e-06, + "loss": 0.7613, + "mean_token_accuracy": 0.7568835020065308, + "num_tokens": 474444738.0, + "step": 18335 + }, + { + "epoch": 2.013617395124094, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5833706855773926, + "learning_rate": 1e-06, + "loss": 0.7296, + "mean_token_accuracy": 0.7608805894851685, + "num_tokens": 474464517.0, + "step": 18336 + }, + { + "epoch": 2.0137272128267076, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.133553981781006, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.725277841091156, + "num_tokens": 474495613.0, + "step": 18337 + }, + { + "epoch": 2.0138370305293214, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3660521507263184, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.709713339805603, + "num_tokens": 474521657.0, + "step": 18338 + }, + { + "epoch": 2.013946848231935, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.226698875427246, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7274688482284546, + "num_tokens": 474549732.0, + "step": 18339 + }, + { + "epoch": 2.0140566659345485, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.098280191421509, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7296133041381836, + "num_tokens": 474578853.0, + "step": 18340 + }, + { + "epoch": 2.014166483637162, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.0849084854125977, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7249940633773804, + "num_tokens": 474609898.0, + "step": 18341 + }, + { + "epoch": 2.014276301339776, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.693094491958618, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7536717653274536, + "num_tokens": 474630096.0, + "step": 18342 + }, + { + "epoch": 2.0143861190423897, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.375596761703491, + "learning_rate": 1e-06, + "loss": 0.7647, + "mean_token_accuracy": 0.752402126789093, + "num_tokens": 474655330.0, + "step": 18343 + }, + { + "epoch": 2.0144959367450035, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.457376480102539, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7261080145835876, + "num_tokens": 474679402.0, + "step": 18344 + }, + { + "epoch": 2.014605754447617, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.1518115997314453, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7413479089736938, + "num_tokens": 474709822.0, + "step": 18345 + }, + { + "epoch": 2.0147155721502306, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.3129916191101074, + "learning_rate": 1e-06, + "loss": 0.7576, + "mean_token_accuracy": 0.7598872184753418, + "num_tokens": 474734308.0, + "step": 18346 + }, + { + "epoch": 2.0148253898528443, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.4560978412628174, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7503182888031006, + "num_tokens": 474757181.0, + "step": 18347 + }, + { + "epoch": 2.014935207555458, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5191986560821533, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7293788194656372, + "num_tokens": 474779652.0, + "step": 18348 + }, + { + "epoch": 2.0150450252580714, + "ewc_loss": 1.0788440704345703e-05, + "grad_norm": 2.5787394046783447, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7301154136657715, + "num_tokens": 474802679.0, + "step": 18349 + }, + { + "epoch": 2.015154842960685, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.524075746536255, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7293321490287781, + "num_tokens": 474826802.0, + "step": 18350 + }, + { + "epoch": 2.015264660663299, + "ewc_loss": 1.0848045349121094e-05, + "grad_norm": 2.340926170349121, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7310866117477417, + "num_tokens": 474852591.0, + "step": 18351 + }, + { + "epoch": 2.0153744783659127, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.129690408706665, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7427197098731995, + "num_tokens": 474881174.0, + "step": 18352 + }, + { + "epoch": 2.0154842960685264, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.036741256713867, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7380139827728271, + "num_tokens": 474912081.0, + "step": 18353 + }, + { + "epoch": 2.0155941137711397, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.507075309753418, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7167534828186035, + "num_tokens": 474935198.0, + "step": 18354 + }, + { + "epoch": 2.0157039314737535, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.500521183013916, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7209739685058594, + "num_tokens": 474958871.0, + "step": 18355 + }, + { + "epoch": 2.0158137491763672, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.396134853363037, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.732359766960144, + "num_tokens": 474982605.0, + "step": 18356 + }, + { + "epoch": 2.015923566878981, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.4818058013916016, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.725786566734314, + "num_tokens": 475006223.0, + "step": 18357 + }, + { + "epoch": 2.0160333845815948, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.191689968109131, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7310695052146912, + "num_tokens": 475034253.0, + "step": 18358 + }, + { + "epoch": 2.016143202284208, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.183819055557251, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7227945327758789, + "num_tokens": 475064327.0, + "step": 18359 + }, + { + "epoch": 2.016253019986822, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3421998023986816, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7374820709228516, + "num_tokens": 475088498.0, + "step": 18360 + }, + { + "epoch": 2.0163628376894356, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.281254768371582, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7341275811195374, + "num_tokens": 475113631.0, + "step": 18361 + }, + { + "epoch": 2.0164726553920493, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.119391918182373, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7235276699066162, + "num_tokens": 475144330.0, + "step": 18362 + }, + { + "epoch": 2.0165824730946627, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3458445072174072, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7265385389328003, + "num_tokens": 475173602.0, + "step": 18363 + }, + { + "epoch": 2.0166922907972764, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.4675023555755615, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7394803762435913, + "num_tokens": 475197199.0, + "step": 18364 + }, + { + "epoch": 2.01680210849989, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 1.9836074113845825, + "learning_rate": 1e-06, + "loss": 0.7629, + "mean_token_accuracy": 0.7538686394691467, + "num_tokens": 475228261.0, + "step": 18365 + }, + { + "epoch": 2.016911926202504, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2741341590881348, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7213531136512756, + "num_tokens": 475255246.0, + "step": 18366 + }, + { + "epoch": 2.0170217439051177, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.4866783618927, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7567649483680725, + "num_tokens": 475278080.0, + "step": 18367 + }, + { + "epoch": 2.017131561607731, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.825488328933716, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7480592727661133, + "num_tokens": 475297692.0, + "step": 18368 + }, + { + "epoch": 2.0172413793103448, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.348832130432129, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7071108818054199, + "num_tokens": 475323342.0, + "step": 18369 + }, + { + "epoch": 2.0173511970129585, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3634912967681885, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7367442846298218, + "num_tokens": 475346994.0, + "step": 18370 + }, + { + "epoch": 2.0174610147155723, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3201634883880615, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7324509024620056, + "num_tokens": 475372315.0, + "step": 18371 + }, + { + "epoch": 2.017570832418186, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.336315631866455, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7472567558288574, + "num_tokens": 475396438.0, + "step": 18372 + }, + { + "epoch": 2.0176806501207993, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3495144844055176, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7180233001708984, + "num_tokens": 475422077.0, + "step": 18373 + }, + { + "epoch": 2.017790467823413, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.276186943054199, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7381892204284668, + "num_tokens": 475447617.0, + "step": 18374 + }, + { + "epoch": 2.017900285526027, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.506615400314331, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7422314882278442, + "num_tokens": 475471261.0, + "step": 18375 + }, + { + "epoch": 2.0180101032286406, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.6720473766326904, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7398484945297241, + "num_tokens": 475492817.0, + "step": 18376 + }, + { + "epoch": 2.018119920931254, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3009936809539795, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7255182266235352, + "num_tokens": 475519515.0, + "step": 18377 + }, + { + "epoch": 2.0182297386338677, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.294870138168335, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7333288788795471, + "num_tokens": 475546017.0, + "step": 18378 + }, + { + "epoch": 2.0183395563364814, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2004451751708984, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.7136937379837036, + "num_tokens": 475576814.0, + "step": 18379 + }, + { + "epoch": 2.018449374039095, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.286707878112793, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7177481651306152, + "num_tokens": 475604395.0, + "step": 18380 + }, + { + "epoch": 2.018559191741709, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.215973377227783, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7157164812088013, + "num_tokens": 475632406.0, + "step": 18381 + }, + { + "epoch": 2.0186690094443223, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.049683094024658, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7225099802017212, + "num_tokens": 475663063.0, + "step": 18382 + }, + { + "epoch": 2.018778827146936, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.6539604663848877, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7406120896339417, + "num_tokens": 475683086.0, + "step": 18383 + }, + { + "epoch": 2.0188886448495498, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.7743823528289795, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7439550161361694, + "num_tokens": 475701641.0, + "step": 18384 + }, + { + "epoch": 2.0189984625521635, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.455533504486084, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7206649780273438, + "num_tokens": 475725651.0, + "step": 18385 + }, + { + "epoch": 2.0191082802547773, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.081348419189453, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7320053577423096, + "num_tokens": 475755963.0, + "step": 18386 + }, + { + "epoch": 2.0192180979573906, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3584256172180176, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7170311212539673, + "num_tokens": 475782500.0, + "step": 18387 + }, + { + "epoch": 2.0193279156600044, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.297945976257324, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.724561870098114, + "num_tokens": 475810085.0, + "step": 18388 + }, + { + "epoch": 2.019437733362618, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.60304594039917, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7378604412078857, + "num_tokens": 475831769.0, + "step": 18389 + }, + { + "epoch": 2.019547551065232, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2321784496307373, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7197285890579224, + "num_tokens": 475860288.0, + "step": 18390 + }, + { + "epoch": 2.019657368767845, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.4550766944885254, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.7569947838783264, + "num_tokens": 475881887.0, + "step": 18391 + }, + { + "epoch": 2.019767186470459, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.5343921184539795, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7195481657981873, + "num_tokens": 475907087.0, + "step": 18392 + }, + { + "epoch": 2.0198770041730727, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.5587384700775146, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7281066179275513, + "num_tokens": 475930532.0, + "step": 18393 + }, + { + "epoch": 2.0199868218756865, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3361470699310303, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7460002303123474, + "num_tokens": 475956755.0, + "step": 18394 + }, + { + "epoch": 2.0200966395783, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.1137208938598633, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7330724000930786, + "num_tokens": 475988953.0, + "step": 18395 + }, + { + "epoch": 2.0202064572809135, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.6504266262054443, + "learning_rate": 1e-06, + "loss": 0.7841, + "mean_token_accuracy": 0.7457733154296875, + "num_tokens": 476010678.0, + "step": 18396 + }, + { + "epoch": 2.0203162749835273, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.271601915359497, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7414807081222534, + "num_tokens": 476038599.0, + "step": 18397 + }, + { + "epoch": 2.020426092686141, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.185361385345459, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7199540138244629, + "num_tokens": 476067170.0, + "step": 18398 + }, + { + "epoch": 2.020535910388755, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3574366569519043, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7377078533172607, + "num_tokens": 476093353.0, + "step": 18399 + }, + { + "epoch": 2.020645728091368, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.1937148571014404, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7285172939300537, + "num_tokens": 476120892.0, + "step": 18400 + }, + { + "epoch": 2.020755545793982, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.240180730819702, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7281183004379272, + "num_tokens": 476148830.0, + "step": 18401 + }, + { + "epoch": 2.0208653634965956, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.252380847930908, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7062152624130249, + "num_tokens": 476176269.0, + "step": 18402 + }, + { + "epoch": 2.0209751811992094, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.27634334564209, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.70676189661026, + "num_tokens": 476203739.0, + "step": 18403 + }, + { + "epoch": 2.021084998901823, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.558955192565918, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.741305947303772, + "num_tokens": 476227125.0, + "step": 18404 + }, + { + "epoch": 2.0211948166044365, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3381779193878174, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.72535240650177, + "num_tokens": 476253997.0, + "step": 18405 + }, + { + "epoch": 2.02130463430705, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3920459747314453, + "learning_rate": 1e-06, + "loss": 0.6847, + "mean_token_accuracy": 0.7778387665748596, + "num_tokens": 476276686.0, + "step": 18406 + }, + { + "epoch": 2.021414452009664, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.0170843601226807, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7224357724189758, + "num_tokens": 476307945.0, + "step": 18407 + }, + { + "epoch": 2.0215242697122777, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 1.922082781791687, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7182421088218689, + "num_tokens": 476344153.0, + "step": 18408 + }, + { + "epoch": 2.0216340874148915, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2370996475219727, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7498815655708313, + "num_tokens": 476372534.0, + "step": 18409 + }, + { + "epoch": 2.021743905117505, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.256777286529541, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7215846180915833, + "num_tokens": 476400143.0, + "step": 18410 + }, + { + "epoch": 2.0218537228201185, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.5691871643066406, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.7545452117919922, + "num_tokens": 476421887.0, + "step": 18411 + }, + { + "epoch": 2.0219635405227323, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2470619678497314, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7322957515716553, + "num_tokens": 476448408.0, + "step": 18412 + }, + { + "epoch": 2.022073358225346, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.8462491035461426, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7451640367507935, + "num_tokens": 476467945.0, + "step": 18413 + }, + { + "epoch": 2.0221831759279594, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 1.990563988685608, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.715430498123169, + "num_tokens": 476504064.0, + "step": 18414 + }, + { + "epoch": 2.022292993630573, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.411022663116455, + "learning_rate": 1e-06, + "loss": 0.8019, + "mean_token_accuracy": 0.7406188249588013, + "num_tokens": 476528563.0, + "step": 18415 + }, + { + "epoch": 2.022402811333187, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.5780928134918213, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7337173819541931, + "num_tokens": 476551278.0, + "step": 18416 + }, + { + "epoch": 2.0225126290358006, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.686187267303467, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7534488439559937, + "num_tokens": 476571059.0, + "step": 18417 + }, + { + "epoch": 2.0226224467384144, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.198054075241089, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7362918257713318, + "num_tokens": 476601332.0, + "step": 18418 + }, + { + "epoch": 2.0227322644410277, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.767007827758789, + "learning_rate": 1e-06, + "loss": 0.7461, + "mean_token_accuracy": 0.7576292157173157, + "num_tokens": 476620721.0, + "step": 18419 + }, + { + "epoch": 2.0228420821436415, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3954784870147705, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7349785566329956, + "num_tokens": 476647984.0, + "step": 18420 + }, + { + "epoch": 2.0229518998462552, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2933831214904785, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7385929822921753, + "num_tokens": 476675219.0, + "step": 18421 + }, + { + "epoch": 2.023061717548869, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.1021711826324463, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7165999412536621, + "num_tokens": 476707059.0, + "step": 18422 + }, + { + "epoch": 2.0231715352514827, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.18259334564209, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.747024655342102, + "num_tokens": 476734933.0, + "step": 18423 + }, + { + "epoch": 2.023281352954096, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.416893482208252, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7481428384780884, + "num_tokens": 476758364.0, + "step": 18424 + }, + { + "epoch": 2.02339117065671, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.212409734725952, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7168801426887512, + "num_tokens": 476788896.0, + "step": 18425 + }, + { + "epoch": 2.0235009883593236, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.42571759223938, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7105087637901306, + "num_tokens": 476814130.0, + "step": 18426 + }, + { + "epoch": 2.0236108060619373, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3020050525665283, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7431440353393555, + "num_tokens": 476839446.0, + "step": 18427 + }, + { + "epoch": 2.0237206237645506, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.1321122646331787, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.714546799659729, + "num_tokens": 476871147.0, + "step": 18428 + }, + { + "epoch": 2.0238304414671644, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.2769949436187744, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7250676155090332, + "num_tokens": 476897450.0, + "step": 18429 + }, + { + "epoch": 2.023940259169778, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.532773733139038, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7206447124481201, + "num_tokens": 476921078.0, + "step": 18430 + }, + { + "epoch": 2.024050076872392, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.3742713928222656, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7317000031471252, + "num_tokens": 476944763.0, + "step": 18431 + }, + { + "epoch": 2.0241598945750057, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.7541322708129883, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7450729012489319, + "num_tokens": 476964532.0, + "step": 18432 + }, + { + "epoch": 2.024269712277619, + "ewc_loss": 1.0907649993896484e-05, + "grad_norm": 2.6454918384552, + "learning_rate": 1e-06, + "loss": 0.7935, + "mean_token_accuracy": 0.7456167936325073, + "num_tokens": 476984769.0, + "step": 18433 + }, + { + "epoch": 2.0243795299802327, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.353193998336792, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7218223214149475, + "num_tokens": 477010814.0, + "step": 18434 + }, + { + "epoch": 2.0244893476828465, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.2811694145202637, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7322739958763123, + "num_tokens": 477037902.0, + "step": 18435 + }, + { + "epoch": 2.0245991653854603, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.370915651321411, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7367799282073975, + "num_tokens": 477062536.0, + "step": 18436 + }, + { + "epoch": 2.024708983088074, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.400700330734253, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7388759255409241, + "num_tokens": 477088327.0, + "step": 18437 + }, + { + "epoch": 2.0248188007906873, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.3804144859313965, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7568107843399048, + "num_tokens": 477111255.0, + "step": 18438 + }, + { + "epoch": 2.024928618493301, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.1048834323883057, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7372353076934814, + "num_tokens": 477140742.0, + "step": 18439 + }, + { + "epoch": 2.025038436195915, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.3583202362060547, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7264441251754761, + "num_tokens": 477167026.0, + "step": 18440 + }, + { + "epoch": 2.0251482538985286, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.5694401264190674, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7309178113937378, + "num_tokens": 477189984.0, + "step": 18441 + }, + { + "epoch": 2.025258071601142, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.342977523803711, + "learning_rate": 1e-06, + "loss": 0.9846, + "mean_token_accuracy": 0.6980788707733154, + "num_tokens": 477216833.0, + "step": 18442 + }, + { + "epoch": 2.0253678893037557, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.604445457458496, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7263621687889099, + "num_tokens": 477239306.0, + "step": 18443 + }, + { + "epoch": 2.0254777070063694, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.5963656902313232, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7333754301071167, + "num_tokens": 477261024.0, + "step": 18444 + }, + { + "epoch": 2.025587524708983, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.282576084136963, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7372267246246338, + "num_tokens": 477287270.0, + "step": 18445 + }, + { + "epoch": 2.025697342411597, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.281252145767212, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7345775365829468, + "num_tokens": 477315531.0, + "step": 18446 + }, + { + "epoch": 2.0258071601142102, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.30979585647583, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.738420844078064, + "num_tokens": 477340295.0, + "step": 18447 + }, + { + "epoch": 2.025916977816824, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.5000250339508057, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7456624507904053, + "num_tokens": 477362513.0, + "step": 18448 + }, + { + "epoch": 2.0260267955194378, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.2046146392822266, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7054842710494995, + "num_tokens": 477390622.0, + "step": 18449 + }, + { + "epoch": 2.0261366132220515, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.342749834060669, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7354881763458252, + "num_tokens": 477415906.0, + "step": 18450 + }, + { + "epoch": 2.0262464309246653, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.25730037689209, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7344850301742554, + "num_tokens": 477442984.0, + "step": 18451 + }, + { + "epoch": 2.0263562486272786, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.2422022819519043, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7423012852668762, + "num_tokens": 477467160.0, + "step": 18452 + }, + { + "epoch": 2.0264660663298923, + "ewc_loss": 1.1086463928222656e-05, + "grad_norm": 2.133772134780884, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7390706539154053, + "num_tokens": 477494952.0, + "step": 18453 + }, + { + "epoch": 2.026575884032506, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.26069712638855, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7236186265945435, + "num_tokens": 477523578.0, + "step": 18454 + }, + { + "epoch": 2.02668570173512, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.3373029232025146, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7393130660057068, + "num_tokens": 477547502.0, + "step": 18455 + }, + { + "epoch": 2.026795519437733, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.6986396312713623, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7193198204040527, + "num_tokens": 477568762.0, + "step": 18456 + }, + { + "epoch": 2.026905337140347, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.5087192058563232, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.7444372177124023, + "num_tokens": 477591095.0, + "step": 18457 + }, + { + "epoch": 2.0270151548429607, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.3756327629089355, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7358113527297974, + "num_tokens": 477615159.0, + "step": 18458 + }, + { + "epoch": 2.0271249725455744, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.13971209526062, + "learning_rate": 1e-06, + "loss": 0.952, + "mean_token_accuracy": 0.7025029063224792, + "num_tokens": 477646241.0, + "step": 18459 + }, + { + "epoch": 2.027234790248188, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.025932550430298, + "learning_rate": 1e-06, + "loss": 0.9636, + "mean_token_accuracy": 0.7014056444168091, + "num_tokens": 477678701.0, + "step": 18460 + }, + { + "epoch": 2.0273446079508015, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.2072010040283203, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7091010808944702, + "num_tokens": 477708276.0, + "step": 18461 + }, + { + "epoch": 2.0274544256534153, + "ewc_loss": 1.1026859283447266e-05, + "grad_norm": 2.1986374855041504, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7288334369659424, + "num_tokens": 477734701.0, + "step": 18462 + }, + { + "epoch": 2.027564243356029, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.2278549671173096, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7256625890731812, + "num_tokens": 477761108.0, + "step": 18463 + }, + { + "epoch": 2.027674061058643, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.235548973083496, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7479279637336731, + "num_tokens": 477787164.0, + "step": 18464 + }, + { + "epoch": 2.027783878761256, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4628429412841797, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.729699969291687, + "num_tokens": 477811238.0, + "step": 18465 + }, + { + "epoch": 2.02789369646387, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.1386160850524902, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7284661531448364, + "num_tokens": 477840378.0, + "step": 18466 + }, + { + "epoch": 2.0280035141664836, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.340585470199585, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7185710072517395, + "num_tokens": 477866312.0, + "step": 18467 + }, + { + "epoch": 2.0281133318690974, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4093990325927734, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.748138964176178, + "num_tokens": 477889462.0, + "step": 18468 + }, + { + "epoch": 2.028223149571711, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.3458023071289062, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7320811748504639, + "num_tokens": 477912523.0, + "step": 18469 + }, + { + "epoch": 2.0283329672743244, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.3116466999053955, + "learning_rate": 1e-06, + "loss": 0.8154, + "mean_token_accuracy": 0.7468249797821045, + "num_tokens": 477938584.0, + "step": 18470 + }, + { + "epoch": 2.028442784976938, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.1204984188079834, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7240756750106812, + "num_tokens": 477969511.0, + "step": 18471 + }, + { + "epoch": 2.028552602679552, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.3582818508148193, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7146351933479309, + "num_tokens": 477996399.0, + "step": 18472 + }, + { + "epoch": 2.0286624203821657, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.514788866043091, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7398655414581299, + "num_tokens": 478019229.0, + "step": 18473 + }, + { + "epoch": 2.0287722380847795, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.600558280944824, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7463647127151489, + "num_tokens": 478040209.0, + "step": 18474 + }, + { + "epoch": 2.028882055787393, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.384800672531128, + "learning_rate": 1e-06, + "loss": 0.7848, + "mean_token_accuracy": 0.7489991784095764, + "num_tokens": 478062881.0, + "step": 18475 + }, + { + "epoch": 2.0289918734900065, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4711110591888428, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7503695487976074, + "num_tokens": 478085248.0, + "step": 18476 + }, + { + "epoch": 2.0291016911926203, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.2829556465148926, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.729017972946167, + "num_tokens": 478111876.0, + "step": 18477 + }, + { + "epoch": 2.029211508895234, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.23572039604187, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7344245314598083, + "num_tokens": 478140060.0, + "step": 18478 + }, + { + "epoch": 2.0293213265978474, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.029621124267578, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7133288979530334, + "num_tokens": 478173027.0, + "step": 18479 + }, + { + "epoch": 2.029431144300461, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.445128917694092, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7333850860595703, + "num_tokens": 478197596.0, + "step": 18480 + }, + { + "epoch": 2.029540962003075, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.223687171936035, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.7442843317985535, + "num_tokens": 478227789.0, + "step": 18481 + }, + { + "epoch": 2.0296507797056886, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.134998321533203, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.721807599067688, + "num_tokens": 478259865.0, + "step": 18482 + }, + { + "epoch": 2.0297605974083024, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.706624746322632, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7413944005966187, + "num_tokens": 478280288.0, + "step": 18483 + }, + { + "epoch": 2.0298704151109157, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.636319160461426, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7319093942642212, + "num_tokens": 478301159.0, + "step": 18484 + }, + { + "epoch": 2.0299802328135295, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4089581966400146, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.737790584564209, + "num_tokens": 478326072.0, + "step": 18485 + }, + { + "epoch": 2.030090050516143, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.0909290313720703, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6946413516998291, + "num_tokens": 478362978.0, + "step": 18486 + }, + { + "epoch": 2.030199868218757, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.0914649963378906, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.733735203742981, + "num_tokens": 478395259.0, + "step": 18487 + }, + { + "epoch": 2.0303096859213707, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.214864492416382, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7219827771186829, + "num_tokens": 478424449.0, + "step": 18488 + }, + { + "epoch": 2.030419503623984, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.0923473834991455, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7187716364860535, + "num_tokens": 478454349.0, + "step": 18489 + }, + { + "epoch": 2.030529321326598, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.6499292850494385, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7222139835357666, + "num_tokens": 478476519.0, + "step": 18490 + }, + { + "epoch": 2.0306391390292116, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4392755031585693, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7286481857299805, + "num_tokens": 478501584.0, + "step": 18491 + }, + { + "epoch": 2.0307489567318253, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.348393440246582, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7272605299949646, + "num_tokens": 478527028.0, + "step": 18492 + }, + { + "epoch": 2.0308587744344386, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.148017406463623, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7293864488601685, + "num_tokens": 478556497.0, + "step": 18493 + }, + { + "epoch": 2.0309685921370524, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.3655872344970703, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.736736536026001, + "num_tokens": 478581600.0, + "step": 18494 + }, + { + "epoch": 2.031078409839666, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.355154037475586, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7194013595581055, + "num_tokens": 478608224.0, + "step": 18495 + }, + { + "epoch": 2.03118822754228, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.588479518890381, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.745593786239624, + "num_tokens": 478630025.0, + "step": 18496 + }, + { + "epoch": 2.0312980452448937, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.2690298557281494, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7277563810348511, + "num_tokens": 478655944.0, + "step": 18497 + }, + { + "epoch": 2.031407862947507, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.298330545425415, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7196260690689087, + "num_tokens": 478682722.0, + "step": 18498 + }, + { + "epoch": 2.0315176806501207, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.212730646133423, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7227040529251099, + "num_tokens": 478712650.0, + "step": 18499 + }, + { + "epoch": 2.0316274983527345, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 1.8146439790725708, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7169357538223267, + "num_tokens": 478750751.0, + "step": 18500 + }, + { + "epoch": 2.0317373160553482, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.2002551555633545, + "learning_rate": 1e-06, + "loss": 0.9577, + "mean_token_accuracy": 0.7042251229286194, + "num_tokens": 478782213.0, + "step": 18501 + }, + { + "epoch": 2.031847133757962, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.604027509689331, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7261864542961121, + "num_tokens": 478806595.0, + "step": 18502 + }, + { + "epoch": 2.0319569514605753, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 1.979168176651001, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7085762619972229, + "num_tokens": 478842045.0, + "step": 18503 + }, + { + "epoch": 2.032066769163189, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.2666501998901367, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7258606553077698, + "num_tokens": 478870674.0, + "step": 18504 + }, + { + "epoch": 2.032176586865803, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.661159038543701, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.73772132396698, + "num_tokens": 478891731.0, + "step": 18505 + }, + { + "epoch": 2.0322864045684166, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.1241462230682373, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7256435751914978, + "num_tokens": 478920241.0, + "step": 18506 + }, + { + "epoch": 2.03239622227103, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.0648372173309326, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7109459638595581, + "num_tokens": 478951766.0, + "step": 18507 + }, + { + "epoch": 2.0325060399736437, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.077470302581787, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7214297652244568, + "num_tokens": 478984028.0, + "step": 18508 + }, + { + "epoch": 2.0326158576762574, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.1874148845672607, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7183772325515747, + "num_tokens": 479013463.0, + "step": 18509 + }, + { + "epoch": 2.032725675378871, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.358462333679199, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.7101748585700989, + "num_tokens": 479038882.0, + "step": 18510 + }, + { + "epoch": 2.032835493081485, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.4964599609375, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7185341119766235, + "num_tokens": 479063048.0, + "step": 18511 + }, + { + "epoch": 2.0329453107840982, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.204780101776123, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7164148688316345, + "num_tokens": 479092793.0, + "step": 18512 + }, + { + "epoch": 2.033055128486712, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.530689239501953, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7394202351570129, + "num_tokens": 479114094.0, + "step": 18513 + }, + { + "epoch": 2.0331649461893258, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.281977891921997, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7454277276992798, + "num_tokens": 479139288.0, + "step": 18514 + }, + { + "epoch": 2.0332747638919395, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.455825090408325, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7377802133560181, + "num_tokens": 479161789.0, + "step": 18515 + }, + { + "epoch": 2.033384581594553, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.322165012359619, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7222832441329956, + "num_tokens": 479187593.0, + "step": 18516 + }, + { + "epoch": 2.0334943992971666, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.233877420425415, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7350383400917053, + "num_tokens": 479215416.0, + "step": 18517 + }, + { + "epoch": 2.0336042169997803, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.246797800064087, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7272275686264038, + "num_tokens": 479242278.0, + "step": 18518 + }, + { + "epoch": 2.033714034702394, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.163475751876831, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7190120220184326, + "num_tokens": 479270308.0, + "step": 18519 + }, + { + "epoch": 2.033823852405008, + "ewc_loss": 1.1146068572998047e-05, + "grad_norm": 2.257622480392456, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7435885667800903, + "num_tokens": 479297337.0, + "step": 18520 + }, + { + "epoch": 2.033933670107621, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.198625326156616, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7200866937637329, + "num_tokens": 479326319.0, + "step": 18521 + }, + { + "epoch": 2.034043487810235, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2655036449432373, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7344601154327393, + "num_tokens": 479353862.0, + "step": 18522 + }, + { + "epoch": 2.0341533055128487, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.257615566253662, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7311042547225952, + "num_tokens": 479380234.0, + "step": 18523 + }, + { + "epoch": 2.0342631232154624, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.613253593444824, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7432460784912109, + "num_tokens": 479400222.0, + "step": 18524 + }, + { + "epoch": 2.034372940918076, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2859690189361572, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7343038320541382, + "num_tokens": 479425746.0, + "step": 18525 + }, + { + "epoch": 2.0344827586206895, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4579780101776123, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7427082061767578, + "num_tokens": 479449819.0, + "step": 18526 + }, + { + "epoch": 2.0345925763233033, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.3539066314697266, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7249810695648193, + "num_tokens": 479475323.0, + "step": 18527 + }, + { + "epoch": 2.034702394025917, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1348440647125244, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7101079225540161, + "num_tokens": 479505738.0, + "step": 18528 + }, + { + "epoch": 2.0348122117285308, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5701959133148193, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.739159107208252, + "num_tokens": 479527435.0, + "step": 18529 + }, + { + "epoch": 2.034922029431144, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.829942226409912, + "learning_rate": 1e-06, + "loss": 0.7175, + "mean_token_accuracy": 0.7669069766998291, + "num_tokens": 479545792.0, + "step": 18530 + }, + { + "epoch": 2.035031847133758, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2967278957366943, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7390021085739136, + "num_tokens": 479570974.0, + "step": 18531 + }, + { + "epoch": 2.0351416648363716, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.7497618198394775, + "learning_rate": 1e-06, + "loss": 0.7266, + "mean_token_accuracy": 0.7642325162887573, + "num_tokens": 479588853.0, + "step": 18532 + }, + { + "epoch": 2.0352514825389854, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.7378861904144287, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7425499558448792, + "num_tokens": 479609539.0, + "step": 18533 + }, + { + "epoch": 2.035361300241599, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2846457958221436, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7401673793792725, + "num_tokens": 479635985.0, + "step": 18534 + }, + { + "epoch": 2.0354711179442124, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2678802013397217, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.726366400718689, + "num_tokens": 479662486.0, + "step": 18535 + }, + { + "epoch": 2.035580935646826, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5002923011779785, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7283215522766113, + "num_tokens": 479686596.0, + "step": 18536 + }, + { + "epoch": 2.03569075334944, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.007145881652832, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7402007579803467, + "num_tokens": 479718032.0, + "step": 18537 + }, + { + "epoch": 2.0358005710520537, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4724738597869873, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7316069006919861, + "num_tokens": 479743117.0, + "step": 18538 + }, + { + "epoch": 2.0359103887546675, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1975789070129395, + "learning_rate": 1e-06, + "loss": 0.7843, + "mean_token_accuracy": 0.754180371761322, + "num_tokens": 479772035.0, + "step": 18539 + }, + { + "epoch": 2.0360202064572808, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4259438514709473, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7321764826774597, + "num_tokens": 479797667.0, + "step": 18540 + }, + { + "epoch": 2.0361300241598945, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1230390071868896, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7038843035697937, + "num_tokens": 479825969.0, + "step": 18541 + }, + { + "epoch": 2.0362398418625083, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4328739643096924, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7333439588546753, + "num_tokens": 479851139.0, + "step": 18542 + }, + { + "epoch": 2.036349659565122, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.320551872253418, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7234063744544983, + "num_tokens": 479876346.0, + "step": 18543 + }, + { + "epoch": 2.0364594772677354, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.868110179901123, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7323529720306396, + "num_tokens": 479897002.0, + "step": 18544 + }, + { + "epoch": 2.036569294970349, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5122411251068115, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7346097230911255, + "num_tokens": 479919893.0, + "step": 18545 + }, + { + "epoch": 2.036679112672963, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1624035835266113, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7321925759315491, + "num_tokens": 479949082.0, + "step": 18546 + }, + { + "epoch": 2.0367889303755766, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.581894636154175, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7286416888237, + "num_tokens": 479972061.0, + "step": 18547 + }, + { + "epoch": 2.0368987480781904, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1856741905212402, + "learning_rate": 1e-06, + "loss": 0.7662, + "mean_token_accuracy": 0.7587230801582336, + "num_tokens": 480000751.0, + "step": 18548 + }, + { + "epoch": 2.0370085657808037, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1935067176818848, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7298450469970703, + "num_tokens": 480029222.0, + "step": 18549 + }, + { + "epoch": 2.0371183834834174, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.484354257583618, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7508634328842163, + "num_tokens": 480053003.0, + "step": 18550 + }, + { + "epoch": 2.037228201186031, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2083606719970703, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7379442453384399, + "num_tokens": 480080069.0, + "step": 18551 + }, + { + "epoch": 2.037338018888645, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4770596027374268, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7330845594406128, + "num_tokens": 480103242.0, + "step": 18552 + }, + { + "epoch": 2.0374478365912587, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.415438652038574, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7296939492225647, + "num_tokens": 480128365.0, + "step": 18553 + }, + { + "epoch": 2.037557654293872, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5869393348693848, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7517121434211731, + "num_tokens": 480150859.0, + "step": 18554 + }, + { + "epoch": 2.037667471996486, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4334497451782227, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7358064651489258, + "num_tokens": 480176966.0, + "step": 18555 + }, + { + "epoch": 2.0377772896990995, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.724135398864746, + "learning_rate": 1e-06, + "loss": 0.7865, + "mean_token_accuracy": 0.7469001412391663, + "num_tokens": 480196250.0, + "step": 18556 + }, + { + "epoch": 2.0378871074017133, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4028587341308594, + "learning_rate": 1e-06, + "loss": 0.764, + "mean_token_accuracy": 0.7508620619773865, + "num_tokens": 480218428.0, + "step": 18557 + }, + { + "epoch": 2.0379969251043266, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1528244018554688, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.744412899017334, + "num_tokens": 480247440.0, + "step": 18558 + }, + { + "epoch": 2.0381067428069404, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5197293758392334, + "learning_rate": 1e-06, + "loss": 0.7599, + "mean_token_accuracy": 0.754229724407196, + "num_tokens": 480270897.0, + "step": 18559 + }, + { + "epoch": 2.038216560509554, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1920604705810547, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.745248556137085, + "num_tokens": 480299102.0, + "step": 18560 + }, + { + "epoch": 2.038326378212168, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.6411688327789307, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7209087014198303, + "num_tokens": 480322035.0, + "step": 18561 + }, + { + "epoch": 2.0384361959147816, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.328693151473999, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7475827932357788, + "num_tokens": 480346328.0, + "step": 18562 + }, + { + "epoch": 2.038546013617395, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5359716415405273, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7384527921676636, + "num_tokens": 480370514.0, + "step": 18563 + }, + { + "epoch": 2.0386558313200087, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4836089611053467, + "learning_rate": 1e-06, + "loss": 0.773, + "mean_token_accuracy": 0.7493175268173218, + "num_tokens": 480391909.0, + "step": 18564 + }, + { + "epoch": 2.0387656490226225, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5385727882385254, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7396049499511719, + "num_tokens": 480415314.0, + "step": 18565 + }, + { + "epoch": 2.0388754667252362, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1780645847320557, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7209787368774414, + "num_tokens": 480444693.0, + "step": 18566 + }, + { + "epoch": 2.03898528442785, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.3028066158294678, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7251139879226685, + "num_tokens": 480472164.0, + "step": 18567 + }, + { + "epoch": 2.0390951021304633, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.016301393508911, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7581376433372498, + "num_tokens": 480504905.0, + "step": 18568 + }, + { + "epoch": 2.039204919833077, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.460308074951172, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.726974606513977, + "num_tokens": 480530076.0, + "step": 18569 + }, + { + "epoch": 2.039314737535691, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.0209052562713623, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7240158319473267, + "num_tokens": 480560966.0, + "step": 18570 + }, + { + "epoch": 2.0394245552383046, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.283146619796753, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7104746699333191, + "num_tokens": 480586988.0, + "step": 18571 + }, + { + "epoch": 2.039534372940918, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2982826232910156, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7073790431022644, + "num_tokens": 480615259.0, + "step": 18572 + }, + { + "epoch": 2.0396441906435316, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.0577094554901123, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7179349660873413, + "num_tokens": 480646526.0, + "step": 18573 + }, + { + "epoch": 2.0397540083461454, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.7915420532226562, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.728522777557373, + "num_tokens": 480667027.0, + "step": 18574 + }, + { + "epoch": 2.039863826048759, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.326310873031616, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7368185520172119, + "num_tokens": 480690476.0, + "step": 18575 + }, + { + "epoch": 2.039973643751373, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2564172744750977, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7137999534606934, + "num_tokens": 480719978.0, + "step": 18576 + }, + { + "epoch": 2.0400834614539862, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.0679636001586914, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7388895750045776, + "num_tokens": 480749163.0, + "step": 18577 + }, + { + "epoch": 2.0401932791566, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.6300439834594727, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7525591850280762, + "num_tokens": 480769114.0, + "step": 18578 + }, + { + "epoch": 2.0403030968592137, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.3943417072296143, + "learning_rate": 1e-06, + "loss": 0.7238, + "mean_token_accuracy": 0.7665854692459106, + "num_tokens": 480792241.0, + "step": 18579 + }, + { + "epoch": 2.0404129145618275, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.381242513656616, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7160022258758545, + "num_tokens": 480817932.0, + "step": 18580 + }, + { + "epoch": 2.0405227322644413, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5953903198242188, + "learning_rate": 1e-06, + "loss": 0.7905, + "mean_token_accuracy": 0.7450848817825317, + "num_tokens": 480839867.0, + "step": 18581 + }, + { + "epoch": 2.0406325499670546, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.389039993286133, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7472142577171326, + "num_tokens": 480863797.0, + "step": 18582 + }, + { + "epoch": 2.0407423676696683, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.134395122528076, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7405149936676025, + "num_tokens": 480894772.0, + "step": 18583 + }, + { + "epoch": 2.040852185372282, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.557365655899048, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.735562264919281, + "num_tokens": 480917888.0, + "step": 18584 + }, + { + "epoch": 2.040962003074896, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5623672008514404, + "learning_rate": 1e-06, + "loss": 0.734, + "mean_token_accuracy": 0.7570999264717102, + "num_tokens": 480937887.0, + "step": 18585 + }, + { + "epoch": 2.041071820777509, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.58319354057312, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7483872771263123, + "num_tokens": 480959462.0, + "step": 18586 + }, + { + "epoch": 2.041181638480123, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2353897094726562, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7184473276138306, + "num_tokens": 480987958.0, + "step": 18587 + }, + { + "epoch": 2.0412914561827367, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4335930347442627, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7211860418319702, + "num_tokens": 481011290.0, + "step": 18588 + }, + { + "epoch": 2.0414012738853504, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.377908229827881, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7433427572250366, + "num_tokens": 481036578.0, + "step": 18589 + }, + { + "epoch": 2.041511091587964, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.208132266998291, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7168616056442261, + "num_tokens": 481065866.0, + "step": 18590 + }, + { + "epoch": 2.0416209092905775, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4382271766662598, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7032359838485718, + "num_tokens": 481092331.0, + "step": 18591 + }, + { + "epoch": 2.0417307269931912, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5449044704437256, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7264572381973267, + "num_tokens": 481116641.0, + "step": 18592 + }, + { + "epoch": 2.041840544695805, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.131002187728882, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7190189361572266, + "num_tokens": 481147880.0, + "step": 18593 + }, + { + "epoch": 2.0419503623984188, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1288137435913086, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7303210496902466, + "num_tokens": 481177967.0, + "step": 18594 + }, + { + "epoch": 2.042060180101032, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.615070343017578, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7359416484832764, + "num_tokens": 481199113.0, + "step": 18595 + }, + { + "epoch": 2.042169997803646, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.253122329711914, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7351540327072144, + "num_tokens": 481224513.0, + "step": 18596 + }, + { + "epoch": 2.0422798155062596, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5634381771087646, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.733496904373169, + "num_tokens": 481247538.0, + "step": 18597 + }, + { + "epoch": 2.0423896332088733, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.7305567264556885, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7430083751678467, + "num_tokens": 481268354.0, + "step": 18598 + }, + { + "epoch": 2.042499450911487, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.410287857055664, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.716242790222168, + "num_tokens": 481292886.0, + "step": 18599 + }, + { + "epoch": 2.0426092686141004, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.3140041828155518, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.73430335521698, + "num_tokens": 481318875.0, + "step": 18600 + }, + { + "epoch": 2.042719086316714, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.0045793056488037, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7161361575126648, + "num_tokens": 481350672.0, + "step": 18601 + }, + { + "epoch": 2.042828904019328, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.7206871509552, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7114460468292236, + "num_tokens": 481373624.0, + "step": 18602 + }, + { + "epoch": 2.0429387217219417, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.777202606201172, + "learning_rate": 1e-06, + "loss": 0.7047, + "mean_token_accuracy": 0.7696042656898499, + "num_tokens": 481392991.0, + "step": 18603 + }, + { + "epoch": 2.0430485394245554, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.688577651977539, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7344013452529907, + "num_tokens": 481413218.0, + "step": 18604 + }, + { + "epoch": 2.0431583571271688, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.158611297607422, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7144145369529724, + "num_tokens": 481441656.0, + "step": 18605 + }, + { + "epoch": 2.0432681748297825, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.8360490798950195, + "learning_rate": 1e-06, + "loss": 0.7722, + "mean_token_accuracy": 0.7442368268966675, + "num_tokens": 481459766.0, + "step": 18606 + }, + { + "epoch": 2.0433779925323963, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1046791076660156, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7380601167678833, + "num_tokens": 481486822.0, + "step": 18607 + }, + { + "epoch": 2.04348781023501, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5329408645629883, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7467458248138428, + "num_tokens": 481508810.0, + "step": 18608 + }, + { + "epoch": 2.0435976279376233, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2525787353515625, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.718633234500885, + "num_tokens": 481535761.0, + "step": 18609 + }, + { + "epoch": 2.043707445640237, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 1.9560515880584717, + "learning_rate": 1e-06, + "loss": 0.7287, + "mean_token_accuracy": 0.759687066078186, + "num_tokens": 481567869.0, + "step": 18610 + }, + { + "epoch": 2.043817263342851, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.209070920944214, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7143599987030029, + "num_tokens": 481596756.0, + "step": 18611 + }, + { + "epoch": 2.0439270810454646, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2487645149230957, + "learning_rate": 1e-06, + "loss": 0.9764, + "mean_token_accuracy": 0.700017511844635, + "num_tokens": 481625764.0, + "step": 18612 + }, + { + "epoch": 2.0440368987480784, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1867198944091797, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7213418483734131, + "num_tokens": 481654419.0, + "step": 18613 + }, + { + "epoch": 2.0441467164506917, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.6171817779541016, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7432576417922974, + "num_tokens": 481673690.0, + "step": 18614 + }, + { + "epoch": 2.0442565341533054, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.2849156856536865, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7346278429031372, + "num_tokens": 481700537.0, + "step": 18615 + }, + { + "epoch": 2.044366351855919, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.559370994567871, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7476228475570679, + "num_tokens": 481721716.0, + "step": 18616 + }, + { + "epoch": 2.044476169558533, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5255534648895264, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7364956736564636, + "num_tokens": 481744694.0, + "step": 18617 + }, + { + "epoch": 2.0445859872611467, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.1987714767456055, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7338035702705383, + "num_tokens": 481773619.0, + "step": 18618 + }, + { + "epoch": 2.04469580496376, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.5268235206604004, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7447168827056885, + "num_tokens": 481796860.0, + "step": 18619 + }, + { + "epoch": 2.044805622666374, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.156635046005249, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7317569851875305, + "num_tokens": 481824971.0, + "step": 18620 + }, + { + "epoch": 2.0449154403689875, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4224705696105957, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7348966598510742, + "num_tokens": 481851380.0, + "step": 18621 + }, + { + "epoch": 2.0450252580716013, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.488044023513794, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7473775148391724, + "num_tokens": 481875181.0, + "step": 18622 + }, + { + "epoch": 2.0451350757742146, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.4264185428619385, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.710879921913147, + "num_tokens": 481901916.0, + "step": 18623 + }, + { + "epoch": 2.0452448934768284, + "ewc_loss": 1.1265277862548828e-05, + "grad_norm": 2.338538408279419, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7201614379882812, + "num_tokens": 481929392.0, + "step": 18624 + }, + { + "epoch": 2.045354711179442, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.47292423248291, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7324919700622559, + "num_tokens": 481955024.0, + "step": 18625 + }, + { + "epoch": 2.045464528882056, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.281935930252075, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7358617782592773, + "num_tokens": 481982868.0, + "step": 18626 + }, + { + "epoch": 2.0455743465846696, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.199787139892578, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7167186737060547, + "num_tokens": 482010241.0, + "step": 18627 + }, + { + "epoch": 2.045684164287283, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.187962532043457, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7128859758377075, + "num_tokens": 482038554.0, + "step": 18628 + }, + { + "epoch": 2.0457939819898967, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.041910171508789, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7183606624603271, + "num_tokens": 482071571.0, + "step": 18629 + }, + { + "epoch": 2.0459037996925105, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.419435739517212, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7134847640991211, + "num_tokens": 482094918.0, + "step": 18630 + }, + { + "epoch": 2.046013617395124, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.8991811275482178, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7237783670425415, + "num_tokens": 482113080.0, + "step": 18631 + }, + { + "epoch": 2.046123435097738, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.2667016983032227, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7415834665298462, + "num_tokens": 482139261.0, + "step": 18632 + }, + { + "epoch": 2.0462332528003513, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.7377912998199463, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7416935563087463, + "num_tokens": 482159933.0, + "step": 18633 + }, + { + "epoch": 2.046343070502965, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.5202834606170654, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7265512943267822, + "num_tokens": 482184511.0, + "step": 18634 + }, + { + "epoch": 2.046452888205579, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.1726672649383545, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7253676652908325, + "num_tokens": 482214353.0, + "step": 18635 + }, + { + "epoch": 2.0465627059081926, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.337965250015259, + "learning_rate": 1e-06, + "loss": 0.7741, + "mean_token_accuracy": 0.7583044171333313, + "num_tokens": 482238019.0, + "step": 18636 + }, + { + "epoch": 2.046672523610806, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.4290993213653564, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7315715551376343, + "num_tokens": 482263259.0, + "step": 18637 + }, + { + "epoch": 2.0467823413134196, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.0668153762817383, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7325085997581482, + "num_tokens": 482297651.0, + "step": 18638 + }, + { + "epoch": 2.0468921590160334, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.29642653465271, + "learning_rate": 1e-06, + "loss": 0.9755, + "mean_token_accuracy": 0.6975563764572144, + "num_tokens": 482326127.0, + "step": 18639 + }, + { + "epoch": 2.047001976718647, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.132779121398926, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.730812668800354, + "num_tokens": 482355648.0, + "step": 18640 + }, + { + "epoch": 2.047111794421261, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 1.8806202411651611, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7394088506698608, + "num_tokens": 482393914.0, + "step": 18641 + }, + { + "epoch": 2.047221612123874, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.3527567386627197, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7202101349830627, + "num_tokens": 482419669.0, + "step": 18642 + }, + { + "epoch": 2.047331429826488, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.5575644969940186, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7325035333633423, + "num_tokens": 482441513.0, + "step": 18643 + }, + { + "epoch": 2.0474412475291017, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.4306015968322754, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7164497375488281, + "num_tokens": 482467869.0, + "step": 18644 + }, + { + "epoch": 2.0475510652317155, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.222316265106201, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7391664981842041, + "num_tokens": 482496206.0, + "step": 18645 + }, + { + "epoch": 2.047660882934329, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.395145893096924, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7172268629074097, + "num_tokens": 482521093.0, + "step": 18646 + }, + { + "epoch": 2.0477707006369426, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.5121419429779053, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7352752685546875, + "num_tokens": 482545337.0, + "step": 18647 + }, + { + "epoch": 2.0478805183395563, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.231997489929199, + "learning_rate": 1e-06, + "loss": 0.7775, + "mean_token_accuracy": 0.7593817710876465, + "num_tokens": 482572245.0, + "step": 18648 + }, + { + "epoch": 2.04799033604217, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.1412298679351807, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7425493597984314, + "num_tokens": 482600882.0, + "step": 18649 + }, + { + "epoch": 2.048100153744784, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.3903510570526123, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7443618774414062, + "num_tokens": 482626059.0, + "step": 18650 + }, + { + "epoch": 2.048209971447397, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.2098920345306396, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7095234394073486, + "num_tokens": 482654852.0, + "step": 18651 + }, + { + "epoch": 2.048319789150011, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.33494234085083, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7508533000946045, + "num_tokens": 482681155.0, + "step": 18652 + }, + { + "epoch": 2.0484296068526247, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.7348592281341553, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7510601282119751, + "num_tokens": 482699738.0, + "step": 18653 + }, + { + "epoch": 2.0485394245552384, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.1790380477905273, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.721198320388794, + "num_tokens": 482729568.0, + "step": 18654 + }, + { + "epoch": 2.048649242257852, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.4234235286712646, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.746009349822998, + "num_tokens": 482752127.0, + "step": 18655 + }, + { + "epoch": 2.0487590599604655, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.487352132797241, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.736297607421875, + "num_tokens": 482775429.0, + "step": 18656 + }, + { + "epoch": 2.0488688776630792, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.2320055961608887, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.7170361280441284, + "num_tokens": 482803518.0, + "step": 18657 + }, + { + "epoch": 2.048978695365693, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.309032917022705, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7243691682815552, + "num_tokens": 482829798.0, + "step": 18658 + }, + { + "epoch": 2.0490885130683067, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.281890630722046, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7368322610855103, + "num_tokens": 482858683.0, + "step": 18659 + }, + { + "epoch": 2.04919833077092, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.22377610206604, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.725508451461792, + "num_tokens": 482884621.0, + "step": 18660 + }, + { + "epoch": 2.049308148473534, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.308034658432007, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7348644137382507, + "num_tokens": 482910212.0, + "step": 18661 + }, + { + "epoch": 2.0494179661761476, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.123835802078247, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7457667589187622, + "num_tokens": 482938617.0, + "step": 18662 + }, + { + "epoch": 2.0495277838787613, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.135422706604004, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7330003976821899, + "num_tokens": 482968113.0, + "step": 18663 + }, + { + "epoch": 2.049637601581375, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.2285757064819336, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7107130885124207, + "num_tokens": 482998324.0, + "step": 18664 + }, + { + "epoch": 2.0497474192839884, + "ewc_loss": 1.1324882507324219e-05, + "grad_norm": 2.2072203159332275, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7176605463027954, + "num_tokens": 483025136.0, + "step": 18665 + }, + { + "epoch": 2.049857236986602, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4508767127990723, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7396882772445679, + "num_tokens": 483048026.0, + "step": 18666 + }, + { + "epoch": 2.049967054689216, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.6665523052215576, + "learning_rate": 1e-06, + "loss": 0.7614, + "mean_token_accuracy": 0.7565421462059021, + "num_tokens": 483069609.0, + "step": 18667 + }, + { + "epoch": 2.0500768723918297, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.464895725250244, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7391806840896606, + "num_tokens": 483092995.0, + "step": 18668 + }, + { + "epoch": 2.0501866900944434, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2268316745758057, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.727787971496582, + "num_tokens": 483119666.0, + "step": 18669 + }, + { + "epoch": 2.0502965077970567, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.0982158184051514, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7608926296234131, + "num_tokens": 483150595.0, + "step": 18670 + }, + { + "epoch": 2.0504063254996705, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.439678430557251, + "learning_rate": 1e-06, + "loss": 0.7852, + "mean_token_accuracy": 0.7425732612609863, + "num_tokens": 483173866.0, + "step": 18671 + }, + { + "epoch": 2.0505161432022843, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.6495630741119385, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.724755585193634, + "num_tokens": 483195512.0, + "step": 18672 + }, + { + "epoch": 2.050625960904898, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.1838154792785645, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7482415437698364, + "num_tokens": 483224278.0, + "step": 18673 + }, + { + "epoch": 2.0507357786075113, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.228386163711548, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.735988974571228, + "num_tokens": 483251606.0, + "step": 18674 + }, + { + "epoch": 2.050845596310125, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4240996837615967, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7250592708587646, + "num_tokens": 483276901.0, + "step": 18675 + }, + { + "epoch": 2.050955414012739, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.193682909011841, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7437096238136292, + "num_tokens": 483304570.0, + "step": 18676 + }, + { + "epoch": 2.0510652317153526, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2130165100097656, + "learning_rate": 1e-06, + "loss": 0.758, + "mean_token_accuracy": 0.7570288777351379, + "num_tokens": 483330328.0, + "step": 18677 + }, + { + "epoch": 2.0511750494179664, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3712658882141113, + "learning_rate": 1e-06, + "loss": 0.9642, + "mean_token_accuracy": 0.7053318619728088, + "num_tokens": 483356872.0, + "step": 18678 + }, + { + "epoch": 2.0512848671205797, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.433018922805786, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7219778299331665, + "num_tokens": 483382817.0, + "step": 18679 + }, + { + "epoch": 2.0513946848231934, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.1822621822357178, + "learning_rate": 1e-06, + "loss": 0.7654, + "mean_token_accuracy": 0.7541681528091431, + "num_tokens": 483411726.0, + "step": 18680 + }, + { + "epoch": 2.051504502525807, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3174827098846436, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7344740629196167, + "num_tokens": 483440186.0, + "step": 18681 + }, + { + "epoch": 2.051614320228421, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5059852600097656, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7396529912948608, + "num_tokens": 483463864.0, + "step": 18682 + }, + { + "epoch": 2.0517241379310347, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.303091287612915, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7328381538391113, + "num_tokens": 483488716.0, + "step": 18683 + }, + { + "epoch": 2.051833955633648, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2025198936462402, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7157138586044312, + "num_tokens": 483516897.0, + "step": 18684 + }, + { + "epoch": 2.0519437733362618, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.584080934524536, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7387596368789673, + "num_tokens": 483539481.0, + "step": 18685 + }, + { + "epoch": 2.0520535910388755, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.481576919555664, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.750067412853241, + "num_tokens": 483562160.0, + "step": 18686 + }, + { + "epoch": 2.0521634087414893, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5006017684936523, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.725713849067688, + "num_tokens": 483587201.0, + "step": 18687 + }, + { + "epoch": 2.0522732264441026, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.492711305618286, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7395685911178589, + "num_tokens": 483610822.0, + "step": 18688 + }, + { + "epoch": 2.0523830441467164, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5696370601654053, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7289955615997314, + "num_tokens": 483632577.0, + "step": 18689 + }, + { + "epoch": 2.05249286184933, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2392430305480957, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7208414077758789, + "num_tokens": 483659643.0, + "step": 18690 + }, + { + "epoch": 2.052602679551944, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.421700954437256, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.731858491897583, + "num_tokens": 483685053.0, + "step": 18691 + }, + { + "epoch": 2.0527124972545576, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5601370334625244, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7284321784973145, + "num_tokens": 483708837.0, + "step": 18692 + }, + { + "epoch": 2.052822314957171, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.396591901779175, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7336065769195557, + "num_tokens": 483732545.0, + "step": 18693 + }, + { + "epoch": 2.0529321326597847, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4362871646881104, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7490094900131226, + "num_tokens": 483755112.0, + "step": 18694 + }, + { + "epoch": 2.0530419503623984, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.567237615585327, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7410266399383545, + "num_tokens": 483777305.0, + "step": 18695 + }, + { + "epoch": 2.053151768065012, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.421377182006836, + "learning_rate": 1e-06, + "loss": 0.7319, + "mean_token_accuracy": 0.7609773874282837, + "num_tokens": 483800500.0, + "step": 18696 + }, + { + "epoch": 2.0532615857676255, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3084869384765625, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7352632284164429, + "num_tokens": 483826162.0, + "step": 18697 + }, + { + "epoch": 2.0533714034702393, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.056074380874634, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.734859824180603, + "num_tokens": 483855322.0, + "step": 18698 + }, + { + "epoch": 2.053481221172853, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.6252281665802, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7378641366958618, + "num_tokens": 483877944.0, + "step": 18699 + }, + { + "epoch": 2.053591038875467, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.260322332382202, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7361460328102112, + "num_tokens": 483904372.0, + "step": 18700 + }, + { + "epoch": 2.0537008565780805, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.389373540878296, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7047794461250305, + "num_tokens": 483933983.0, + "step": 18701 + }, + { + "epoch": 2.053810674280694, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.385220766067505, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7236396074295044, + "num_tokens": 483960201.0, + "step": 18702 + }, + { + "epoch": 2.0539204919833076, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.114656448364258, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7158728837966919, + "num_tokens": 483990681.0, + "step": 18703 + }, + { + "epoch": 2.0540303096859214, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2062745094299316, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7363318204879761, + "num_tokens": 484019494.0, + "step": 18704 + }, + { + "epoch": 2.054140127388535, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4275474548339844, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7320486903190613, + "num_tokens": 484044720.0, + "step": 18705 + }, + { + "epoch": 2.054249945091149, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2341549396514893, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7303935885429382, + "num_tokens": 484071022.0, + "step": 18706 + }, + { + "epoch": 2.054359762793762, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.6011013984680176, + "learning_rate": 1e-06, + "loss": 0.7376, + "mean_token_accuracy": 0.7616604566574097, + "num_tokens": 484089798.0, + "step": 18707 + }, + { + "epoch": 2.054469580496376, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3535420894622803, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7423920631408691, + "num_tokens": 484115280.0, + "step": 18708 + }, + { + "epoch": 2.0545793981989897, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.555180549621582, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7253888845443726, + "num_tokens": 484139932.0, + "step": 18709 + }, + { + "epoch": 2.0546892159016035, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.0315237045288086, + "learning_rate": 1e-06, + "loss": 0.7702, + "mean_token_accuracy": 0.7524808049201965, + "num_tokens": 484170720.0, + "step": 18710 + }, + { + "epoch": 2.054799033604217, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.1306350231170654, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7246274948120117, + "num_tokens": 484199463.0, + "step": 18711 + }, + { + "epoch": 2.0549088513068305, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.534367084503174, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7464173436164856, + "num_tokens": 484223309.0, + "step": 18712 + }, + { + "epoch": 2.0550186690094443, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3477892875671387, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7145984768867493, + "num_tokens": 484252023.0, + "step": 18713 + }, + { + "epoch": 2.055128486712058, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.536716938018799, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7346245050430298, + "num_tokens": 484276029.0, + "step": 18714 + }, + { + "epoch": 2.055238304414672, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.244882583618164, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7146919965744019, + "num_tokens": 484305587.0, + "step": 18715 + }, + { + "epoch": 2.055348122117285, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.8730382919311523, + "learning_rate": 1e-06, + "loss": 0.7484, + "mean_token_accuracy": 0.7591559886932373, + "num_tokens": 484324145.0, + "step": 18716 + }, + { + "epoch": 2.055457939819899, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.040330410003662, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7257649898529053, + "num_tokens": 484356125.0, + "step": 18717 + }, + { + "epoch": 2.0555677575225126, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2582271099090576, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.726094663143158, + "num_tokens": 484385600.0, + "step": 18718 + }, + { + "epoch": 2.0556775752251264, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.431089401245117, + "learning_rate": 1e-06, + "loss": 0.7739, + "mean_token_accuracy": 0.7560358643531799, + "num_tokens": 484410154.0, + "step": 18719 + }, + { + "epoch": 2.05578739292774, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5324254035949707, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7474654912948608, + "num_tokens": 484433103.0, + "step": 18720 + }, + { + "epoch": 2.0558972106303535, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.2160420417785645, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7303866147994995, + "num_tokens": 484461954.0, + "step": 18721 + }, + { + "epoch": 2.0560070283329672, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3514535427093506, + "learning_rate": 1e-06, + "loss": 0.7329, + "mean_token_accuracy": 0.7620244026184082, + "num_tokens": 484484988.0, + "step": 18722 + }, + { + "epoch": 2.056116846035581, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.121535539627075, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7278522253036499, + "num_tokens": 484516484.0, + "step": 18723 + }, + { + "epoch": 2.0562266637381947, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4617977142333984, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7276119589805603, + "num_tokens": 484539532.0, + "step": 18724 + }, + { + "epoch": 2.056336481440808, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.5719921588897705, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7291647791862488, + "num_tokens": 484564198.0, + "step": 18725 + }, + { + "epoch": 2.056446299143422, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3269214630126953, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7236205339431763, + "num_tokens": 484590622.0, + "step": 18726 + }, + { + "epoch": 2.0565561168460356, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.227898120880127, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7012812495231628, + "num_tokens": 484619910.0, + "step": 18727 + }, + { + "epoch": 2.0566659345486493, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.651155948638916, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.738183319568634, + "num_tokens": 484640093.0, + "step": 18728 + }, + { + "epoch": 2.056775752251263, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.501699924468994, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7212655544281006, + "num_tokens": 484669591.0, + "step": 18729 + }, + { + "epoch": 2.0568855699538764, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.4015424251556396, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7324101328849792, + "num_tokens": 484693323.0, + "step": 18730 + }, + { + "epoch": 2.05699538765649, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.6226394176483154, + "learning_rate": 1e-06, + "loss": 0.7418, + "mean_token_accuracy": 0.7626070976257324, + "num_tokens": 484713895.0, + "step": 18731 + }, + { + "epoch": 2.057105205359104, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.617609977722168, + "learning_rate": 1e-06, + "loss": 0.7029, + "mean_token_accuracy": 0.7720239758491516, + "num_tokens": 484734883.0, + "step": 18732 + }, + { + "epoch": 2.0572150230617177, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.482625722885132, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7362276911735535, + "num_tokens": 484757425.0, + "step": 18733 + }, + { + "epoch": 2.0573248407643314, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.286980628967285, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7357066869735718, + "num_tokens": 484782516.0, + "step": 18734 + }, + { + "epoch": 2.0574346584669447, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.679532527923584, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7453300952911377, + "num_tokens": 484805520.0, + "step": 18735 + }, + { + "epoch": 2.0575444761695585, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.3115901947021484, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7074261903762817, + "num_tokens": 484833321.0, + "step": 18736 + }, + { + "epoch": 2.0576542938721722, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.0744986534118652, + "learning_rate": 1e-06, + "loss": 0.975, + "mean_token_accuracy": 0.6957526803016663, + "num_tokens": 484868316.0, + "step": 18737 + }, + { + "epoch": 2.057764111574786, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.756021738052368, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7389230728149414, + "num_tokens": 484892628.0, + "step": 18738 + }, + { + "epoch": 2.0578739292773993, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.206937074661255, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7279538512229919, + "num_tokens": 484919840.0, + "step": 18739 + }, + { + "epoch": 2.057983746980013, + "ewc_loss": 1.138448715209961e-05, + "grad_norm": 2.110973834991455, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.73919677734375, + "num_tokens": 484949046.0, + "step": 18740 + }, + { + "epoch": 2.058093564682627, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.468489408493042, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7231019139289856, + "num_tokens": 484973166.0, + "step": 18741 + }, + { + "epoch": 2.0582033823852406, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4124834537506104, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7227773070335388, + "num_tokens": 484997745.0, + "step": 18742 + }, + { + "epoch": 2.0583132000878543, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.100789785385132, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7331827878952026, + "num_tokens": 485030912.0, + "step": 18743 + }, + { + "epoch": 2.0584230177904677, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2145955562591553, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7339089512825012, + "num_tokens": 485059141.0, + "step": 18744 + }, + { + "epoch": 2.0585328354930814, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4034583568573, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.745032548904419, + "num_tokens": 485083742.0, + "step": 18745 + }, + { + "epoch": 2.058642653195695, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.6862096786499023, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7436546683311462, + "num_tokens": 485105674.0, + "step": 18746 + }, + { + "epoch": 2.058752470898309, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2026782035827637, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7332294583320618, + "num_tokens": 485132428.0, + "step": 18747 + }, + { + "epoch": 2.0588622886009227, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.504502773284912, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7421122789382935, + "num_tokens": 485155653.0, + "step": 18748 + }, + { + "epoch": 2.058972106303536, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3055665493011475, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7308362722396851, + "num_tokens": 485183472.0, + "step": 18749 + }, + { + "epoch": 2.0590819240061498, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.296342372894287, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7265765070915222, + "num_tokens": 485208921.0, + "step": 18750 + }, + { + "epoch": 2.0591917417087635, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.025989532470703, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7211376428604126, + "num_tokens": 485243201.0, + "step": 18751 + }, + { + "epoch": 2.0593015594113773, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.146064519882202, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7317854166030884, + "num_tokens": 485274135.0, + "step": 18752 + }, + { + "epoch": 2.0594113771139906, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2809579372406006, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.725800096988678, + "num_tokens": 485303106.0, + "step": 18753 + }, + { + "epoch": 2.0595211948166043, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4039180278778076, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7383854389190674, + "num_tokens": 485326573.0, + "step": 18754 + }, + { + "epoch": 2.059631012519218, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2680931091308594, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.735317587852478, + "num_tokens": 485354403.0, + "step": 18755 + }, + { + "epoch": 2.059740830221832, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3904948234558105, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7236629724502563, + "num_tokens": 485381116.0, + "step": 18756 + }, + { + "epoch": 2.0598506479244456, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.418003559112549, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.748543918132782, + "num_tokens": 485404852.0, + "step": 18757 + }, + { + "epoch": 2.059960465627059, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3496546745300293, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7462276220321655, + "num_tokens": 485429368.0, + "step": 18758 + }, + { + "epoch": 2.0600702833296727, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.0369327068328857, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7199080586433411, + "num_tokens": 485462570.0, + "step": 18759 + }, + { + "epoch": 2.0601801010322864, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3107376098632812, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7511255741119385, + "num_tokens": 485486324.0, + "step": 18760 + }, + { + "epoch": 2.0602899187349, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.381136178970337, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7391724586486816, + "num_tokens": 485514079.0, + "step": 18761 + }, + { + "epoch": 2.060399736437514, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3412013053894043, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7166692018508911, + "num_tokens": 485539712.0, + "step": 18762 + }, + { + "epoch": 2.0605095541401273, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.359429359436035, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.719103217124939, + "num_tokens": 485565998.0, + "step": 18763 + }, + { + "epoch": 2.060619371842741, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.559617280960083, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7228389382362366, + "num_tokens": 485590371.0, + "step": 18764 + }, + { + "epoch": 2.060729189545355, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4802424907684326, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7347051501274109, + "num_tokens": 485613661.0, + "step": 18765 + }, + { + "epoch": 2.0608390072479685, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3414528369903564, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7357445955276489, + "num_tokens": 485640582.0, + "step": 18766 + }, + { + "epoch": 2.060948824950582, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3705060482025146, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7161232829093933, + "num_tokens": 485667177.0, + "step": 18767 + }, + { + "epoch": 2.0610586426531956, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2151577472686768, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7391524314880371, + "num_tokens": 485692337.0, + "step": 18768 + }, + { + "epoch": 2.0611684603558094, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4642791748046875, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7230157852172852, + "num_tokens": 485715895.0, + "step": 18769 + }, + { + "epoch": 2.061278278058423, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.44201922416687, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7244811058044434, + "num_tokens": 485740056.0, + "step": 18770 + }, + { + "epoch": 2.061388095761037, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.301069974899292, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7374712824821472, + "num_tokens": 485765531.0, + "step": 18771 + }, + { + "epoch": 2.06149791346365, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.274960994720459, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.735629677772522, + "num_tokens": 485790828.0, + "step": 18772 + }, + { + "epoch": 2.061607731166264, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.1676223278045654, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7151883840560913, + "num_tokens": 485821307.0, + "step": 18773 + }, + { + "epoch": 2.0617175488688777, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.489871025085449, + "learning_rate": 1e-06, + "loss": 0.7775, + "mean_token_accuracy": 0.7489719986915588, + "num_tokens": 485842368.0, + "step": 18774 + }, + { + "epoch": 2.0618273665714915, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2668941020965576, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7324228286743164, + "num_tokens": 485869451.0, + "step": 18775 + }, + { + "epoch": 2.0619371842741048, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3635149002075195, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7277311682701111, + "num_tokens": 485894433.0, + "step": 18776 + }, + { + "epoch": 2.0620470019767185, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.659605026245117, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7458068132400513, + "num_tokens": 485915047.0, + "step": 18777 + }, + { + "epoch": 2.0621568196793323, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.306396484375, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7282627820968628, + "num_tokens": 485942877.0, + "step": 18778 + }, + { + "epoch": 2.062266637381946, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.455329418182373, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7329832911491394, + "num_tokens": 485966534.0, + "step": 18779 + }, + { + "epoch": 2.06237645508456, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.822246551513672, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7472422122955322, + "num_tokens": 485985634.0, + "step": 18780 + }, + { + "epoch": 2.062486272787173, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.5908148288726807, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7340526580810547, + "num_tokens": 486012183.0, + "step": 18781 + }, + { + "epoch": 2.062596090489787, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2453787326812744, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7242096066474915, + "num_tokens": 486041381.0, + "step": 18782 + }, + { + "epoch": 2.0627059081924006, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.307652473449707, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.740218997001648, + "num_tokens": 486067899.0, + "step": 18783 + }, + { + "epoch": 2.0628157258950144, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.500582456588745, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7353976964950562, + "num_tokens": 486090785.0, + "step": 18784 + }, + { + "epoch": 2.062925543597628, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.255075216293335, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7291245460510254, + "num_tokens": 486117247.0, + "step": 18785 + }, + { + "epoch": 2.0630353613002415, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.44327712059021, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7340702414512634, + "num_tokens": 486142138.0, + "step": 18786 + }, + { + "epoch": 2.063145179002855, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.8833160400390625, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7482340335845947, + "num_tokens": 486162436.0, + "step": 18787 + }, + { + "epoch": 2.063254996705469, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.666083574295044, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7548487782478333, + "num_tokens": 486183170.0, + "step": 18788 + }, + { + "epoch": 2.0633648144080827, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.079282283782959, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7161375284194946, + "num_tokens": 486219250.0, + "step": 18789 + }, + { + "epoch": 2.063474632110696, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3838021755218506, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7217313051223755, + "num_tokens": 486248913.0, + "step": 18790 + }, + { + "epoch": 2.06358444981331, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.606397867202759, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7417678236961365, + "num_tokens": 486270415.0, + "step": 18791 + }, + { + "epoch": 2.0636942675159236, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2691233158111572, + "learning_rate": 1e-06, + "loss": 0.903, + "mean_token_accuracy": 0.7124271392822266, + "num_tokens": 486298050.0, + "step": 18792 + }, + { + "epoch": 2.0638040852185373, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.420208215713501, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7242640852928162, + "num_tokens": 486321495.0, + "step": 18793 + }, + { + "epoch": 2.063913902921151, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3489720821380615, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7417353391647339, + "num_tokens": 486346931.0, + "step": 18794 + }, + { + "epoch": 2.0640237206237644, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.409404754638672, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7201327681541443, + "num_tokens": 486374258.0, + "step": 18795 + }, + { + "epoch": 2.064133538326378, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.2975656986236572, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.723069965839386, + "num_tokens": 486401074.0, + "step": 18796 + }, + { + "epoch": 2.064243356028992, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.3980484008789062, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7399731874465942, + "num_tokens": 486425411.0, + "step": 18797 + }, + { + "epoch": 2.0643531737316057, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.423495292663574, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7314152717590332, + "num_tokens": 486450354.0, + "step": 18798 + }, + { + "epoch": 2.0644629914342194, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.172027111053467, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7446898221969604, + "num_tokens": 486479627.0, + "step": 18799 + }, + { + "epoch": 2.0645728091368327, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.4851081371307373, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7353184223175049, + "num_tokens": 486502869.0, + "step": 18800 + }, + { + "epoch": 2.0646826268394465, + "ewc_loss": 1.1444091796875e-05, + "grad_norm": 2.0503146648406982, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7181357145309448, + "num_tokens": 486534327.0, + "step": 18801 + }, + { + "epoch": 2.0647924445420602, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.406493902206421, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7182793617248535, + "num_tokens": 486559490.0, + "step": 18802 + }, + { + "epoch": 2.064902262244674, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.634943962097168, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7356671094894409, + "num_tokens": 486579601.0, + "step": 18803 + }, + { + "epoch": 2.0650120799472873, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.606809139251709, + "learning_rate": 1e-06, + "loss": 0.7731, + "mean_token_accuracy": 0.7584236860275269, + "num_tokens": 486600957.0, + "step": 18804 + }, + { + "epoch": 2.065121897649901, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3674917221069336, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7375510334968567, + "num_tokens": 486627356.0, + "step": 18805 + }, + { + "epoch": 2.065231715352515, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3206799030303955, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7366253733634949, + "num_tokens": 486653002.0, + "step": 18806 + }, + { + "epoch": 2.0653415330551286, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.367978096008301, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7467851638793945, + "num_tokens": 486677465.0, + "step": 18807 + }, + { + "epoch": 2.0654513507577423, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3371098041534424, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7252068519592285, + "num_tokens": 486704398.0, + "step": 18808 + }, + { + "epoch": 2.0655611684603556, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.199094533920288, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7278524041175842, + "num_tokens": 486735319.0, + "step": 18809 + }, + { + "epoch": 2.0656709861629694, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.6225473880767822, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7270472049713135, + "num_tokens": 486756914.0, + "step": 18810 + }, + { + "epoch": 2.065780803865583, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.319232225418091, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7413502931594849, + "num_tokens": 486781933.0, + "step": 18811 + }, + { + "epoch": 2.065890621568197, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3065433502197266, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.729920506477356, + "num_tokens": 486808660.0, + "step": 18812 + }, + { + "epoch": 2.0660004392708107, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.061314105987549, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.715502142906189, + "num_tokens": 486844039.0, + "step": 18813 + }, + { + "epoch": 2.066110256973424, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 1.9889402389526367, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7243611812591553, + "num_tokens": 486878633.0, + "step": 18814 + }, + { + "epoch": 2.0662200746760377, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.223862886428833, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7401329278945923, + "num_tokens": 486907779.0, + "step": 18815 + }, + { + "epoch": 2.0663298923786515, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.2903716564178467, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7200597524642944, + "num_tokens": 486938039.0, + "step": 18816 + }, + { + "epoch": 2.0664397100812653, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.215869188308716, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7120202779769897, + "num_tokens": 486967462.0, + "step": 18817 + }, + { + "epoch": 2.0665495277838786, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.233466386795044, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7465718984603882, + "num_tokens": 486994806.0, + "step": 18818 + }, + { + "epoch": 2.0666593454864923, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3599772453308105, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7133623361587524, + "num_tokens": 487021312.0, + "step": 18819 + }, + { + "epoch": 2.066769163189106, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.2849655151367188, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7351924180984497, + "num_tokens": 487048531.0, + "step": 18820 + }, + { + "epoch": 2.06687898089172, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3787412643432617, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7358229160308838, + "num_tokens": 487074614.0, + "step": 18821 + }, + { + "epoch": 2.0669887985943336, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.496080160140991, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7337249517440796, + "num_tokens": 487097613.0, + "step": 18822 + }, + { + "epoch": 2.067098616296947, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3507943153381348, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7267017364501953, + "num_tokens": 487124192.0, + "step": 18823 + }, + { + "epoch": 2.0672084339995607, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.06246280670166, + "learning_rate": 1e-06, + "loss": 0.9575, + "mean_token_accuracy": 0.6996673345565796, + "num_tokens": 487157659.0, + "step": 18824 + }, + { + "epoch": 2.0673182517021744, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.183647394180298, + "learning_rate": 1e-06, + "loss": 0.754, + "mean_token_accuracy": 0.7555254101753235, + "num_tokens": 487186665.0, + "step": 18825 + }, + { + "epoch": 2.067428069404788, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.257117986679077, + "learning_rate": 1e-06, + "loss": 0.7617, + "mean_token_accuracy": 0.7553576231002808, + "num_tokens": 487213021.0, + "step": 18826 + }, + { + "epoch": 2.0675378871074015, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4580109119415283, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7285088300704956, + "num_tokens": 487236826.0, + "step": 18827 + }, + { + "epoch": 2.0676477048100153, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.5242486000061035, + "learning_rate": 1e-06, + "loss": 0.7017, + "mean_token_accuracy": 0.7635743618011475, + "num_tokens": 487257249.0, + "step": 18828 + }, + { + "epoch": 2.067757522512629, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4617726802825928, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7394222021102905, + "num_tokens": 487279719.0, + "step": 18829 + }, + { + "epoch": 2.0678673402152428, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.2319207191467285, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.728759229183197, + "num_tokens": 487305998.0, + "step": 18830 + }, + { + "epoch": 2.0679771579178565, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.167008638381958, + "learning_rate": 1e-06, + "loss": 0.8083, + "mean_token_accuracy": 0.7483965158462524, + "num_tokens": 487335168.0, + "step": 18831 + }, + { + "epoch": 2.06808697562047, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.361358642578125, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7359927296638489, + "num_tokens": 487359653.0, + "step": 18832 + }, + { + "epoch": 2.0681967933230836, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.617675304412842, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7514598965644836, + "num_tokens": 487379403.0, + "step": 18833 + }, + { + "epoch": 2.0683066110256974, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.432483196258545, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7322300672531128, + "num_tokens": 487403805.0, + "step": 18834 + }, + { + "epoch": 2.068416428728311, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.301503896713257, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7349761724472046, + "num_tokens": 487429302.0, + "step": 18835 + }, + { + "epoch": 2.068526246430925, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.262280225753784, + "learning_rate": 1e-06, + "loss": 0.7477, + "mean_token_accuracy": 0.7579748630523682, + "num_tokens": 487454107.0, + "step": 18836 + }, + { + "epoch": 2.068636064133538, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.2745399475097656, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7390874028205872, + "num_tokens": 487480299.0, + "step": 18837 + }, + { + "epoch": 2.068745881836152, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.321469306945801, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7235399484634399, + "num_tokens": 487507207.0, + "step": 18838 + }, + { + "epoch": 2.0688556995387657, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3892531394958496, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7463470101356506, + "num_tokens": 487532758.0, + "step": 18839 + }, + { + "epoch": 2.0689655172413794, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.6079354286193848, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7479680180549622, + "num_tokens": 487555940.0, + "step": 18840 + }, + { + "epoch": 2.0690753349439928, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.5099499225616455, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7231998443603516, + "num_tokens": 487581254.0, + "step": 18841 + }, + { + "epoch": 2.0691851526466065, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.759122371673584, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7541093826293945, + "num_tokens": 487600889.0, + "step": 18842 + }, + { + "epoch": 2.0692949703492203, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.669420003890991, + "learning_rate": 1e-06, + "loss": 0.7654, + "mean_token_accuracy": 0.7533097267150879, + "num_tokens": 487620106.0, + "step": 18843 + }, + { + "epoch": 2.069404788051834, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.026698350906372, + "learning_rate": 1e-06, + "loss": 0.7699, + "mean_token_accuracy": 0.7517002820968628, + "num_tokens": 487652019.0, + "step": 18844 + }, + { + "epoch": 2.069514605754448, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.099998950958252, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7339005470275879, + "num_tokens": 487682072.0, + "step": 18845 + }, + { + "epoch": 2.069624423457061, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.085139274597168, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7279807329177856, + "num_tokens": 487712303.0, + "step": 18846 + }, + { + "epoch": 2.069734241159675, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.5801615715026855, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7457133531570435, + "num_tokens": 487733598.0, + "step": 18847 + }, + { + "epoch": 2.0698440588622886, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.341402053833008, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7406752705574036, + "num_tokens": 487758116.0, + "step": 18848 + }, + { + "epoch": 2.0699538765649024, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.3380839824676514, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7168314456939697, + "num_tokens": 487784928.0, + "step": 18849 + }, + { + "epoch": 2.070063694267516, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4743382930755615, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7340818047523499, + "num_tokens": 487806999.0, + "step": 18850 + }, + { + "epoch": 2.0701735119701294, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4496541023254395, + "learning_rate": 1e-06, + "loss": 0.7487, + "mean_token_accuracy": 0.7523545622825623, + "num_tokens": 487829487.0, + "step": 18851 + }, + { + "epoch": 2.070283329672743, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.248528003692627, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7269484996795654, + "num_tokens": 487857701.0, + "step": 18852 + }, + { + "epoch": 2.070393147375357, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.229494333267212, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7251108288764954, + "num_tokens": 487886019.0, + "step": 18853 + }, + { + "epoch": 2.0705029650779707, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.6130638122558594, + "learning_rate": 1e-06, + "loss": 0.7728, + "mean_token_accuracy": 0.7553021907806396, + "num_tokens": 487906054.0, + "step": 18854 + }, + { + "epoch": 2.070612782780584, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.2112131118774414, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7431245446205139, + "num_tokens": 487932732.0, + "step": 18855 + }, + { + "epoch": 2.070722600483198, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.513561487197876, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7205824851989746, + "num_tokens": 487954950.0, + "step": 18856 + }, + { + "epoch": 2.0708324181858115, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 1.9333584308624268, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.727820634841919, + "num_tokens": 487989789.0, + "step": 18857 + }, + { + "epoch": 2.0709422358884253, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4220457077026367, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7324724197387695, + "num_tokens": 488013438.0, + "step": 18858 + }, + { + "epoch": 2.071052053591039, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.6607494354248047, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.742576003074646, + "num_tokens": 488033967.0, + "step": 18859 + }, + { + "epoch": 2.0711618712936524, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.658193349838257, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7349386215209961, + "num_tokens": 488055618.0, + "step": 18860 + }, + { + "epoch": 2.071271688996266, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4842612743377686, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7369129657745361, + "num_tokens": 488077824.0, + "step": 18861 + }, + { + "epoch": 2.07138150669888, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4361324310302734, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7233754396438599, + "num_tokens": 488103374.0, + "step": 18862 + }, + { + "epoch": 2.0714913244014936, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.2027475833892822, + "learning_rate": 1e-06, + "loss": 0.7776, + "mean_token_accuracy": 0.7468011379241943, + "num_tokens": 488129776.0, + "step": 18863 + }, + { + "epoch": 2.0716011421041074, + "ewc_loss": 1.1563301086425781e-05, + "grad_norm": 2.4043655395507812, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7234997153282166, + "num_tokens": 488154567.0, + "step": 18864 + }, + { + "epoch": 2.0717109598067207, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.6476151943206787, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7413061857223511, + "num_tokens": 488175092.0, + "step": 18865 + }, + { + "epoch": 2.0718207775093345, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.3670780658721924, + "learning_rate": 1e-06, + "loss": 0.9406, + "mean_token_accuracy": 0.7092341184616089, + "num_tokens": 488202329.0, + "step": 18866 + }, + { + "epoch": 2.0719305952119482, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.361738681793213, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7281557321548462, + "num_tokens": 488229448.0, + "step": 18867 + }, + { + "epoch": 2.072040412914562, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.8492045402526855, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7400148510932922, + "num_tokens": 488248341.0, + "step": 18868 + }, + { + "epoch": 2.0721502306171753, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.3005378246307373, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7160816192626953, + "num_tokens": 488276403.0, + "step": 18869 + }, + { + "epoch": 2.072260048319789, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4028825759887695, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7194270491600037, + "num_tokens": 488300829.0, + "step": 18870 + }, + { + "epoch": 2.072369866022403, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.1791064739227295, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7214866280555725, + "num_tokens": 488330090.0, + "step": 18871 + }, + { + "epoch": 2.0724796837250166, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.1844630241394043, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7348349094390869, + "num_tokens": 488357430.0, + "step": 18872 + }, + { + "epoch": 2.0725895014276303, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4964375495910645, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7338173985481262, + "num_tokens": 488382389.0, + "step": 18873 + }, + { + "epoch": 2.0726993191302436, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4183456897735596, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7450867891311646, + "num_tokens": 488405819.0, + "step": 18874 + }, + { + "epoch": 2.0728091368328574, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.438477039337158, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7285798192024231, + "num_tokens": 488428902.0, + "step": 18875 + }, + { + "epoch": 2.072918954535471, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.49228835105896, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7511066794395447, + "num_tokens": 488451267.0, + "step": 18876 + }, + { + "epoch": 2.073028772238085, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.2178761959075928, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7374498844146729, + "num_tokens": 488477786.0, + "step": 18877 + }, + { + "epoch": 2.073138589940698, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.417132616043091, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7474077939987183, + "num_tokens": 488501779.0, + "step": 18878 + }, + { + "epoch": 2.073248407643312, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.478041172027588, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7282896637916565, + "num_tokens": 488525927.0, + "step": 18879 + }, + { + "epoch": 2.0733582253459257, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.5608160495758057, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.734426736831665, + "num_tokens": 488549374.0, + "step": 18880 + }, + { + "epoch": 2.0734680430485395, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.091665267944336, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7171249985694885, + "num_tokens": 488581234.0, + "step": 18881 + }, + { + "epoch": 2.0735778607511532, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4618918895721436, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7384806871414185, + "num_tokens": 488604300.0, + "step": 18882 + }, + { + "epoch": 2.0736876784537666, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4504647254943848, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7267137169837952, + "num_tokens": 488628493.0, + "step": 18883 + }, + { + "epoch": 2.0737974961563803, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.593660354614258, + "learning_rate": 1e-06, + "loss": 0.7557, + "mean_token_accuracy": 0.7565034627914429, + "num_tokens": 488649579.0, + "step": 18884 + }, + { + "epoch": 2.073907313858994, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.219169855117798, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7179126739501953, + "num_tokens": 488679453.0, + "step": 18885 + }, + { + "epoch": 2.074017131561608, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.1820664405822754, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.738145649433136, + "num_tokens": 488706913.0, + "step": 18886 + }, + { + "epoch": 2.0741269492642216, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.24769926071167, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7379712462425232, + "num_tokens": 488733613.0, + "step": 18887 + }, + { + "epoch": 2.074236766966835, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.259105920791626, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7221980690956116, + "num_tokens": 488760801.0, + "step": 18888 + }, + { + "epoch": 2.0743465846694487, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.2880687713623047, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7225825786590576, + "num_tokens": 488788687.0, + "step": 18889 + }, + { + "epoch": 2.0744564023720624, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4055216312408447, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.735182523727417, + "num_tokens": 488814072.0, + "step": 18890 + }, + { + "epoch": 2.074566220074676, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.22796893119812, + "learning_rate": 1e-06, + "loss": 0.7776, + "mean_token_accuracy": 0.7578402757644653, + "num_tokens": 488841485.0, + "step": 18891 + }, + { + "epoch": 2.07467603777729, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.260347604751587, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7408128380775452, + "num_tokens": 488867353.0, + "step": 18892 + }, + { + "epoch": 2.0747858554799032, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.3574540615081787, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.702391505241394, + "num_tokens": 488896078.0, + "step": 18893 + }, + { + "epoch": 2.074895673182517, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.6968555450439453, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7334359884262085, + "num_tokens": 488917064.0, + "step": 18894 + }, + { + "epoch": 2.0750054908851308, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.209063768386841, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7238946557044983, + "num_tokens": 488945488.0, + "step": 18895 + }, + { + "epoch": 2.0751153085877445, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.336667060852051, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.738891065120697, + "num_tokens": 488968188.0, + "step": 18896 + }, + { + "epoch": 2.075225126290358, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.7974212169647217, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.742837131023407, + "num_tokens": 488989368.0, + "step": 18897 + }, + { + "epoch": 2.0753349439929716, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.2576067447662354, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.724291205406189, + "num_tokens": 489019061.0, + "step": 18898 + }, + { + "epoch": 2.0754447616955853, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.1918487548828125, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7179032564163208, + "num_tokens": 489049187.0, + "step": 18899 + }, + { + "epoch": 2.075554579398199, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4760243892669678, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7429633140563965, + "num_tokens": 489072262.0, + "step": 18900 + }, + { + "epoch": 2.075664397100813, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.316755771636963, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7362203598022461, + "num_tokens": 489098065.0, + "step": 18901 + }, + { + "epoch": 2.075774214803426, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.4554433822631836, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7335546016693115, + "num_tokens": 489122738.0, + "step": 18902 + }, + { + "epoch": 2.07588403250604, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.3752779960632324, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7193012237548828, + "num_tokens": 489149658.0, + "step": 18903 + }, + { + "epoch": 2.0759938502086537, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.602700710296631, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7466685771942139, + "num_tokens": 489170115.0, + "step": 18904 + }, + { + "epoch": 2.0761036679112674, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.274010181427002, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7088243365287781, + "num_tokens": 489198899.0, + "step": 18905 + }, + { + "epoch": 2.0762134856138807, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.606931686401367, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7326789498329163, + "num_tokens": 489220994.0, + "step": 18906 + }, + { + "epoch": 2.0763233033164945, + "ewc_loss": 1.1622905731201172e-05, + "grad_norm": 2.2026944160461426, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7266771197319031, + "num_tokens": 489247684.0, + "step": 18907 + }, + { + "epoch": 2.0764331210191083, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.286633253097534, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7425428628921509, + "num_tokens": 489274000.0, + "step": 18908 + }, + { + "epoch": 2.076542938721722, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.229372978210449, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.734426736831665, + "num_tokens": 489301138.0, + "step": 18909 + }, + { + "epoch": 2.0766527564243358, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.573248863220215, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7373631000518799, + "num_tokens": 489323949.0, + "step": 18910 + }, + { + "epoch": 2.076762574126949, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.340308904647827, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7311974763870239, + "num_tokens": 489348177.0, + "step": 18911 + }, + { + "epoch": 2.076872391829563, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.509359359741211, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7242376804351807, + "num_tokens": 489370653.0, + "step": 18912 + }, + { + "epoch": 2.0769822095321766, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.520071268081665, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7232716679573059, + "num_tokens": 489396076.0, + "step": 18913 + }, + { + "epoch": 2.0770920272347904, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.180492401123047, + "learning_rate": 1e-06, + "loss": 0.7507, + "mean_token_accuracy": 0.7640749216079712, + "num_tokens": 489422322.0, + "step": 18914 + }, + { + "epoch": 2.077201844937404, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2115230560302734, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7184785604476929, + "num_tokens": 489452241.0, + "step": 18915 + }, + { + "epoch": 2.0773116626400174, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.639117479324341, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7444831132888794, + "num_tokens": 489471717.0, + "step": 18916 + }, + { + "epoch": 2.077421480342631, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.5771586894989014, + "learning_rate": 1e-06, + "loss": 0.7874, + "mean_token_accuracy": 0.7540208101272583, + "num_tokens": 489493599.0, + "step": 18917 + }, + { + "epoch": 2.077531298045245, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.417745351791382, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7143824100494385, + "num_tokens": 489517991.0, + "step": 18918 + }, + { + "epoch": 2.0776411157478587, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.0648930072784424, + "learning_rate": 1e-06, + "loss": 0.9523, + "mean_token_accuracy": 0.7066969871520996, + "num_tokens": 489550331.0, + "step": 18919 + }, + { + "epoch": 2.077750933450472, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.212747097015381, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7221493124961853, + "num_tokens": 489576755.0, + "step": 18920 + }, + { + "epoch": 2.0778607511530858, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.274629831314087, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7415307760238647, + "num_tokens": 489603232.0, + "step": 18921 + }, + { + "epoch": 2.0779705688556995, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.135852098464966, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7105740308761597, + "num_tokens": 489633861.0, + "step": 18922 + }, + { + "epoch": 2.0780803865583133, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.5683400630950928, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.730513870716095, + "num_tokens": 489656715.0, + "step": 18923 + }, + { + "epoch": 2.078190204260927, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.695448398590088, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.736426830291748, + "num_tokens": 489678745.0, + "step": 18924 + }, + { + "epoch": 2.0783000219635404, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2811436653137207, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7274360656738281, + "num_tokens": 489705042.0, + "step": 18925 + }, + { + "epoch": 2.078409839666154, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.29826021194458, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7306180000305176, + "num_tokens": 489732389.0, + "step": 18926 + }, + { + "epoch": 2.078519657368768, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4188685417175293, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7046724557876587, + "num_tokens": 489756253.0, + "step": 18927 + }, + { + "epoch": 2.0786294750713816, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4916951656341553, + "learning_rate": 1e-06, + "loss": 0.7649, + "mean_token_accuracy": 0.7493669986724854, + "num_tokens": 489779374.0, + "step": 18928 + }, + { + "epoch": 2.078739292773995, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.5225348472595215, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.726679801940918, + "num_tokens": 489802145.0, + "step": 18929 + }, + { + "epoch": 2.0788491104766087, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1016244888305664, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7159351110458374, + "num_tokens": 489832481.0, + "step": 18930 + }, + { + "epoch": 2.0789589281792225, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4392764568328857, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7215321660041809, + "num_tokens": 489855325.0, + "step": 18931 + }, + { + "epoch": 2.079068745881836, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3377676010131836, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7304184436798096, + "num_tokens": 489880338.0, + "step": 18932 + }, + { + "epoch": 2.07917856358445, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4833879470825195, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7414132356643677, + "num_tokens": 489904091.0, + "step": 18933 + }, + { + "epoch": 2.0792883812870633, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4956212043762207, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7276085615158081, + "num_tokens": 489928676.0, + "step": 18934 + }, + { + "epoch": 2.079398198989677, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.213148832321167, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7269296646118164, + "num_tokens": 489956846.0, + "step": 18935 + }, + { + "epoch": 2.079508016692291, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.470125675201416, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7283521890640259, + "num_tokens": 489981568.0, + "step": 18936 + }, + { + "epoch": 2.0796178343949046, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.255143642425537, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7455469965934753, + "num_tokens": 490007835.0, + "step": 18937 + }, + { + "epoch": 2.0797276520975183, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.5276994705200195, + "learning_rate": 1e-06, + "loss": 0.6497, + "mean_token_accuracy": 0.7862871885299683, + "num_tokens": 490027937.0, + "step": 18938 + }, + { + "epoch": 2.0798374698001316, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.0538442134857178, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7175465226173401, + "num_tokens": 490063001.0, + "step": 18939 + }, + { + "epoch": 2.0799472875027454, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3429019451141357, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7423443794250488, + "num_tokens": 490088863.0, + "step": 18940 + }, + { + "epoch": 2.080057105205359, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.494884967803955, + "learning_rate": 1e-06, + "loss": 0.7344, + "mean_token_accuracy": 0.7612646222114563, + "num_tokens": 490110820.0, + "step": 18941 + }, + { + "epoch": 2.080166922907973, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1839089393615723, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7195398807525635, + "num_tokens": 490140847.0, + "step": 18942 + }, + { + "epoch": 2.0802767406105866, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.316868782043457, + "learning_rate": 1e-06, + "loss": 0.987, + "mean_token_accuracy": 0.7059743404388428, + "num_tokens": 490166679.0, + "step": 18943 + }, + { + "epoch": 2.0803865583132, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.6550750732421875, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.743940532207489, + "num_tokens": 490187894.0, + "step": 18944 + }, + { + "epoch": 2.0804963760158137, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 1.9081623554229736, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7278210520744324, + "num_tokens": 490221796.0, + "step": 18945 + }, + { + "epoch": 2.0806061937184275, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.541398286819458, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7317049503326416, + "num_tokens": 490243502.0, + "step": 18946 + }, + { + "epoch": 2.0807160114210412, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2078847885131836, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7474706172943115, + "num_tokens": 490271915.0, + "step": 18947 + }, + { + "epoch": 2.0808258291236545, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2546989917755127, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7354552745819092, + "num_tokens": 490300156.0, + "step": 18948 + }, + { + "epoch": 2.0809356468262683, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.249622344970703, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7032850980758667, + "num_tokens": 490330308.0, + "step": 18949 + }, + { + "epoch": 2.081045464528882, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2270889282226562, + "learning_rate": 1e-06, + "loss": 0.9794, + "mean_token_accuracy": 0.7137932777404785, + "num_tokens": 490358907.0, + "step": 18950 + }, + { + "epoch": 2.081155282231496, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4486842155456543, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.722808301448822, + "num_tokens": 490383541.0, + "step": 18951 + }, + { + "epoch": 2.0812650999341096, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.276247262954712, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7361716628074646, + "num_tokens": 490409125.0, + "step": 18952 + }, + { + "epoch": 2.081374917636723, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.421203136444092, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7404134273529053, + "num_tokens": 490432306.0, + "step": 18953 + }, + { + "epoch": 2.0814847353393366, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.585843801498413, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7245877385139465, + "num_tokens": 490454009.0, + "step": 18954 + }, + { + "epoch": 2.0815945530419504, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2639200687408447, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.717347264289856, + "num_tokens": 490482565.0, + "step": 18955 + }, + { + "epoch": 2.081704370744564, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 1.9495469331741333, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.7063741683959961, + "num_tokens": 490519834.0, + "step": 18956 + }, + { + "epoch": 2.0818141884471775, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.300727367401123, + "learning_rate": 1e-06, + "loss": 0.7503, + "mean_token_accuracy": 0.7556424140930176, + "num_tokens": 490546074.0, + "step": 18957 + }, + { + "epoch": 2.0819240061497912, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.515707015991211, + "learning_rate": 1e-06, + "loss": 0.7167, + "mean_token_accuracy": 0.7619524002075195, + "num_tokens": 490568911.0, + "step": 18958 + }, + { + "epoch": 2.082033823852405, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.305121421813965, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7176768183708191, + "num_tokens": 490594591.0, + "step": 18959 + }, + { + "epoch": 2.0821436415550187, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3270394802093506, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7354596257209778, + "num_tokens": 490619546.0, + "step": 18960 + }, + { + "epoch": 2.0822534592576325, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3003599643707275, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7334820032119751, + "num_tokens": 490648240.0, + "step": 18961 + }, + { + "epoch": 2.082363276960246, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.326274871826172, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.72918301820755, + "num_tokens": 490672451.0, + "step": 18962 + }, + { + "epoch": 2.0824730946628596, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4111733436584473, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7446551322937012, + "num_tokens": 490698015.0, + "step": 18963 + }, + { + "epoch": 2.0825829123654733, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3980541229248047, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7489296793937683, + "num_tokens": 490721407.0, + "step": 18964 + }, + { + "epoch": 2.082692730068087, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.470548629760742, + "learning_rate": 1e-06, + "loss": 0.7579, + "mean_token_accuracy": 0.7549464702606201, + "num_tokens": 490743600.0, + "step": 18965 + }, + { + "epoch": 2.082802547770701, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.447108030319214, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7362359166145325, + "num_tokens": 490766922.0, + "step": 18966 + }, + { + "epoch": 2.082912365473314, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.6918745040893555, + "learning_rate": 1e-06, + "loss": 0.7725, + "mean_token_accuracy": 0.7499616146087646, + "num_tokens": 490787071.0, + "step": 18967 + }, + { + "epoch": 2.083022183175928, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.0197529792785645, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7335455417633057, + "num_tokens": 490818928.0, + "step": 18968 + }, + { + "epoch": 2.0831320008785417, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.280998706817627, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7229324579238892, + "num_tokens": 490845146.0, + "step": 18969 + }, + { + "epoch": 2.0832418185811554, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.416938543319702, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7239044904708862, + "num_tokens": 490870347.0, + "step": 18970 + }, + { + "epoch": 2.0833516362837687, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4003405570983887, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7253957986831665, + "num_tokens": 490897708.0, + "step": 18971 + }, + { + "epoch": 2.0834614539863825, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.239727735519409, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7213395833969116, + "num_tokens": 490926444.0, + "step": 18972 + }, + { + "epoch": 2.0835712716889963, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3728885650634766, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.744805634021759, + "num_tokens": 490951897.0, + "step": 18973 + }, + { + "epoch": 2.08368108939161, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 1.9782358407974243, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7072256207466125, + "num_tokens": 490987745.0, + "step": 18974 + }, + { + "epoch": 2.0837909070942238, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1254642009735107, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.715217113494873, + "num_tokens": 491018623.0, + "step": 18975 + }, + { + "epoch": 2.083900724796837, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4434752464294434, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7351874113082886, + "num_tokens": 491042741.0, + "step": 18976 + }, + { + "epoch": 2.084010542499451, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.27681827545166, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7317358255386353, + "num_tokens": 491068750.0, + "step": 18977 + }, + { + "epoch": 2.0841203602020646, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.416917324066162, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7148633003234863, + "num_tokens": 491092239.0, + "step": 18978 + }, + { + "epoch": 2.0842301779046783, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.527461290359497, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7521851062774658, + "num_tokens": 491114475.0, + "step": 18979 + }, + { + "epoch": 2.084339995607292, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.678184747695923, + "learning_rate": 1e-06, + "loss": 0.7833, + "mean_token_accuracy": 0.7457537651062012, + "num_tokens": 491134611.0, + "step": 18980 + }, + { + "epoch": 2.0844498133099054, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.638620376586914, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7376258373260498, + "num_tokens": 491155132.0, + "step": 18981 + }, + { + "epoch": 2.084559631012519, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.539705276489258, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7444557547569275, + "num_tokens": 491178243.0, + "step": 18982 + }, + { + "epoch": 2.084669448715133, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.125166416168213, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.7091248035430908, + "num_tokens": 491209504.0, + "step": 18983 + }, + { + "epoch": 2.0847792664177467, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.569162607192993, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7287297248840332, + "num_tokens": 491234330.0, + "step": 18984 + }, + { + "epoch": 2.08488908412036, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1854498386383057, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7339392304420471, + "num_tokens": 491262831.0, + "step": 18985 + }, + { + "epoch": 2.0849989018229738, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3115310668945312, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7111833095550537, + "num_tokens": 491290913.0, + "step": 18986 + }, + { + "epoch": 2.0851087195255875, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.575955629348755, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7160376310348511, + "num_tokens": 491314614.0, + "step": 18987 + }, + { + "epoch": 2.0852185372282013, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.217625379562378, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7170313596725464, + "num_tokens": 491344133.0, + "step": 18988 + }, + { + "epoch": 2.085328354930815, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.381833076477051, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7311553955078125, + "num_tokens": 491370174.0, + "step": 18989 + }, + { + "epoch": 2.0854381726334283, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.363724708557129, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7391716837882996, + "num_tokens": 491396116.0, + "step": 18990 + }, + { + "epoch": 2.085547990336042, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1262919902801514, + "learning_rate": 1e-06, + "loss": 0.9735, + "mean_token_accuracy": 0.7036709785461426, + "num_tokens": 491430346.0, + "step": 18991 + }, + { + "epoch": 2.085657808038656, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3682398796081543, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7321580648422241, + "num_tokens": 491458979.0, + "step": 18992 + }, + { + "epoch": 2.0857676257412696, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.459843873977661, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7469086050987244, + "num_tokens": 491484176.0, + "step": 18993 + }, + { + "epoch": 2.0858774434438834, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3623499870300293, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7308215498924255, + "num_tokens": 491509964.0, + "step": 18994 + }, + { + "epoch": 2.0859872611464967, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.5926671028137207, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7305101156234741, + "num_tokens": 491531663.0, + "step": 18995 + }, + { + "epoch": 2.0860970788491104, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.761371374130249, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7220374345779419, + "num_tokens": 491554738.0, + "step": 18996 + }, + { + "epoch": 2.086206896551724, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4473912715911865, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.741872251033783, + "num_tokens": 491578738.0, + "step": 18997 + }, + { + "epoch": 2.086316714254338, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1995630264282227, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.7138798236846924, + "num_tokens": 491606169.0, + "step": 18998 + }, + { + "epoch": 2.0864265319569513, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.193227529525757, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7399390339851379, + "num_tokens": 491634519.0, + "step": 18999 + }, + { + "epoch": 2.086536349659565, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.2908246517181396, + "learning_rate": 1e-06, + "loss": 0.9379, + "mean_token_accuracy": 0.7077155113220215, + "num_tokens": 491664802.0, + "step": 19000 + }, + { + "epoch": 2.086646167362179, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 1.939644455909729, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7419496774673462, + "num_tokens": 491699769.0, + "step": 19001 + }, + { + "epoch": 2.0867559850647925, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.3412606716156006, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7315813302993774, + "num_tokens": 491726122.0, + "step": 19002 + }, + { + "epoch": 2.0868658027674063, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4296998977661133, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7424079179763794, + "num_tokens": 491751062.0, + "step": 19003 + }, + { + "epoch": 2.0869756204700196, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.341092824935913, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.736909806728363, + "num_tokens": 491777880.0, + "step": 19004 + }, + { + "epoch": 2.0870854381726334, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.302555561065674, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7289878726005554, + "num_tokens": 491803602.0, + "step": 19005 + }, + { + "epoch": 2.087195255875247, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.107475996017456, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7324750423431396, + "num_tokens": 491833549.0, + "step": 19006 + }, + { + "epoch": 2.087305073577861, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.580503225326538, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.709460437297821, + "num_tokens": 491855771.0, + "step": 19007 + }, + { + "epoch": 2.087414891280474, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.24332594871521, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.7312644720077515, + "num_tokens": 491883597.0, + "step": 19008 + }, + { + "epoch": 2.087524708983088, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.140105962753296, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7192318439483643, + "num_tokens": 491912746.0, + "step": 19009 + }, + { + "epoch": 2.0876345266857017, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.4199671745300293, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.737102210521698, + "num_tokens": 491937029.0, + "step": 19010 + }, + { + "epoch": 2.0877443443883155, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.504491090774536, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7427860498428345, + "num_tokens": 491959920.0, + "step": 19011 + }, + { + "epoch": 2.087854162090929, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.1550214290618896, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7454530000686646, + "num_tokens": 491989044.0, + "step": 19012 + }, + { + "epoch": 2.0879639797935425, + "ewc_loss": 1.1742115020751953e-05, + "grad_norm": 2.429959535598755, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7315157651901245, + "num_tokens": 492013024.0, + "step": 19013 + }, + { + "epoch": 2.0880737974961563, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.120478391647339, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7288083434104919, + "num_tokens": 492044906.0, + "step": 19014 + }, + { + "epoch": 2.08818361519877, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.310561418533325, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7354958653450012, + "num_tokens": 492070927.0, + "step": 19015 + }, + { + "epoch": 2.088293432901384, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4365975856781006, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7266427278518677, + "num_tokens": 492095050.0, + "step": 19016 + }, + { + "epoch": 2.0884032506039976, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.6366167068481445, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7331318855285645, + "num_tokens": 492115775.0, + "step": 19017 + }, + { + "epoch": 2.088513068306611, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4884541034698486, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7284338474273682, + "num_tokens": 492139040.0, + "step": 19018 + }, + { + "epoch": 2.0886228860092246, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3909239768981934, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.724555253982544, + "num_tokens": 492167280.0, + "step": 19019 + }, + { + "epoch": 2.0887327037118384, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4192497730255127, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7271339893341064, + "num_tokens": 492193315.0, + "step": 19020 + }, + { + "epoch": 2.088842521414452, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4964609146118164, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7281526327133179, + "num_tokens": 492217710.0, + "step": 19021 + }, + { + "epoch": 2.0889523391170655, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.7270216941833496, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7228744029998779, + "num_tokens": 492238561.0, + "step": 19022 + }, + { + "epoch": 2.089062156819679, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.543839931488037, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7305281162261963, + "num_tokens": 492261352.0, + "step": 19023 + }, + { + "epoch": 2.089171974522293, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.351895332336426, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7293233871459961, + "num_tokens": 492286699.0, + "step": 19024 + }, + { + "epoch": 2.0892817922249067, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.537168502807617, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7450168132781982, + "num_tokens": 492309374.0, + "step": 19025 + }, + { + "epoch": 2.0893916099275205, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.058194398880005, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7195504903793335, + "num_tokens": 492342546.0, + "step": 19026 + }, + { + "epoch": 2.089501427630134, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.293895959854126, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.75062096118927, + "num_tokens": 492369335.0, + "step": 19027 + }, + { + "epoch": 2.0896112453327476, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5214765071868896, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7251923680305481, + "num_tokens": 492393485.0, + "step": 19028 + }, + { + "epoch": 2.0897210630353613, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.6552772521972656, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7308846712112427, + "num_tokens": 492417059.0, + "step": 19029 + }, + { + "epoch": 2.089830880737975, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.7249767780303955, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7323737144470215, + "num_tokens": 492438264.0, + "step": 19030 + }, + { + "epoch": 2.089940698440589, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4080350399017334, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7163852453231812, + "num_tokens": 492463019.0, + "step": 19031 + }, + { + "epoch": 2.090050516143202, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.312894105911255, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7290181517601013, + "num_tokens": 492489882.0, + "step": 19032 + }, + { + "epoch": 2.090160333845816, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3680903911590576, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.7461493015289307, + "num_tokens": 492514387.0, + "step": 19033 + }, + { + "epoch": 2.0902701515484297, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.649312973022461, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7529993653297424, + "num_tokens": 492535709.0, + "step": 19034 + }, + { + "epoch": 2.0903799692510434, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4672441482543945, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7330877780914307, + "num_tokens": 492558357.0, + "step": 19035 + }, + { + "epoch": 2.0904897869536567, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.389248847961426, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7221094369888306, + "num_tokens": 492585376.0, + "step": 19036 + }, + { + "epoch": 2.0905996046562705, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.1521501541137695, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7129632830619812, + "num_tokens": 492614595.0, + "step": 19037 + }, + { + "epoch": 2.0907094223588842, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.7432899475097656, + "learning_rate": 1e-06, + "loss": 0.7843, + "mean_token_accuracy": 0.7471835613250732, + "num_tokens": 492632727.0, + "step": 19038 + }, + { + "epoch": 2.090819240061498, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.121290922164917, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7134643197059631, + "num_tokens": 492662994.0, + "step": 19039 + }, + { + "epoch": 2.0909290577641118, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3066797256469727, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7292499542236328, + "num_tokens": 492690017.0, + "step": 19040 + }, + { + "epoch": 2.091038875466725, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.239616870880127, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7336708903312683, + "num_tokens": 492718107.0, + "step": 19041 + }, + { + "epoch": 2.091148693169339, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.612588882446289, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7350767850875854, + "num_tokens": 492741305.0, + "step": 19042 + }, + { + "epoch": 2.0912585108719526, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.296557664871216, + "learning_rate": 1e-06, + "loss": 0.8384, + "mean_token_accuracy": 0.7305746078491211, + "num_tokens": 492766787.0, + "step": 19043 + }, + { + "epoch": 2.0913683285745663, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.297147274017334, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7331855893135071, + "num_tokens": 492793789.0, + "step": 19044 + }, + { + "epoch": 2.09147814627718, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2677011489868164, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7443410158157349, + "num_tokens": 492820917.0, + "step": 19045 + }, + { + "epoch": 2.0915879639797934, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.1661219596862793, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7123061418533325, + "num_tokens": 492849776.0, + "step": 19046 + }, + { + "epoch": 2.091697781682407, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.293956995010376, + "learning_rate": 1e-06, + "loss": 0.8154, + "mean_token_accuracy": 0.7415212988853455, + "num_tokens": 492877582.0, + "step": 19047 + }, + { + "epoch": 2.091807599385021, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.340271234512329, + "learning_rate": 1e-06, + "loss": 0.9811, + "mean_token_accuracy": 0.70499587059021, + "num_tokens": 492906048.0, + "step": 19048 + }, + { + "epoch": 2.0919174170876347, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5370678901672363, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7280522584915161, + "num_tokens": 492928817.0, + "step": 19049 + }, + { + "epoch": 2.092027234790248, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5745229721069336, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7383848428726196, + "num_tokens": 492951138.0, + "step": 19050 + }, + { + "epoch": 2.0921370524928617, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 1.97763192653656, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.740324854850769, + "num_tokens": 492985746.0, + "step": 19051 + }, + { + "epoch": 2.0922468701954755, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4144420623779297, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7383833527565002, + "num_tokens": 493008945.0, + "step": 19052 + }, + { + "epoch": 2.0923566878980893, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5036606788635254, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.737587571144104, + "num_tokens": 493032172.0, + "step": 19053 + }, + { + "epoch": 2.092466505600703, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.0115315914154053, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.72034752368927, + "num_tokens": 493064895.0, + "step": 19054 + }, + { + "epoch": 2.0925763233033163, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2678072452545166, + "learning_rate": 1e-06, + "loss": 0.9419, + "mean_token_accuracy": 0.7024219036102295, + "num_tokens": 493094111.0, + "step": 19055 + }, + { + "epoch": 2.09268614100593, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2374682426452637, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7354851961135864, + "num_tokens": 493120831.0, + "step": 19056 + }, + { + "epoch": 2.092795958708544, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2716126441955566, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7347055673599243, + "num_tokens": 493148749.0, + "step": 19057 + }, + { + "epoch": 2.0929057764111576, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.0514724254608154, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7223689556121826, + "num_tokens": 493187015.0, + "step": 19058 + }, + { + "epoch": 2.093015594113771, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.580096483230591, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.724536657333374, + "num_tokens": 493211688.0, + "step": 19059 + }, + { + "epoch": 2.0931254118163847, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2979986667633057, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7360159158706665, + "num_tokens": 493238035.0, + "step": 19060 + }, + { + "epoch": 2.0932352295189984, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2226765155792236, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7476203441619873, + "num_tokens": 493265714.0, + "step": 19061 + }, + { + "epoch": 2.093345047221612, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.208404064178467, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7509024143218994, + "num_tokens": 493294297.0, + "step": 19062 + }, + { + "epoch": 2.093454864924226, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.505204200744629, + "learning_rate": 1e-06, + "loss": 0.7817, + "mean_token_accuracy": 0.7548385858535767, + "num_tokens": 493315677.0, + "step": 19063 + }, + { + "epoch": 2.0935646826268393, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3833327293395996, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7158576250076294, + "num_tokens": 493342213.0, + "step": 19064 + }, + { + "epoch": 2.093674500329453, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4716525077819824, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7324793338775635, + "num_tokens": 493365307.0, + "step": 19065 + }, + { + "epoch": 2.0937843180320668, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.112536907196045, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7227522134780884, + "num_tokens": 493396239.0, + "step": 19066 + }, + { + "epoch": 2.0938941357346805, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.7284963130950928, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7339692115783691, + "num_tokens": 493416150.0, + "step": 19067 + }, + { + "epoch": 2.0940039534372943, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4236867427825928, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.71240234375, + "num_tokens": 493440906.0, + "step": 19068 + }, + { + "epoch": 2.0941137711399076, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.270279884338379, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7459190487861633, + "num_tokens": 493467869.0, + "step": 19069 + }, + { + "epoch": 2.0942235888425214, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.229076623916626, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7261934280395508, + "num_tokens": 493496899.0, + "step": 19070 + }, + { + "epoch": 2.094333406545135, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.690530300140381, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7447960376739502, + "num_tokens": 493518423.0, + "step": 19071 + }, + { + "epoch": 2.094443224247749, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.397130012512207, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7290183305740356, + "num_tokens": 493543772.0, + "step": 19072 + }, + { + "epoch": 2.0945530419503626, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.319462537765503, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7448509931564331, + "num_tokens": 493569302.0, + "step": 19073 + }, + { + "epoch": 2.094662859652976, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.7944796085357666, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7374149560928345, + "num_tokens": 493590965.0, + "step": 19074 + }, + { + "epoch": 2.0947726773555897, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.329084873199463, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7125763893127441, + "num_tokens": 493618059.0, + "step": 19075 + }, + { + "epoch": 2.0948824950582035, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.300647258758545, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7383978366851807, + "num_tokens": 493644737.0, + "step": 19076 + }, + { + "epoch": 2.094992312760817, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3732969760894775, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7188108563423157, + "num_tokens": 493670418.0, + "step": 19077 + }, + { + "epoch": 2.0951021304634305, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.0821444988250732, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7458117008209229, + "num_tokens": 493700392.0, + "step": 19078 + }, + { + "epoch": 2.0952119481660443, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.1871418952941895, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.746483564376831, + "num_tokens": 493727488.0, + "step": 19079 + }, + { + "epoch": 2.095321765868658, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4259886741638184, + "learning_rate": 1e-06, + "loss": 0.77, + "mean_token_accuracy": 0.7523813247680664, + "num_tokens": 493750634.0, + "step": 19080 + }, + { + "epoch": 2.095431583571272, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 3.2572543621063232, + "learning_rate": 1e-06, + "loss": 0.7354, + "mean_token_accuracy": 0.7551161646842957, + "num_tokens": 493766122.0, + "step": 19081 + }, + { + "epoch": 2.0955414012738856, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.275831699371338, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7066982984542847, + "num_tokens": 493792108.0, + "step": 19082 + }, + { + "epoch": 2.095651218976499, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2357630729675293, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7344956398010254, + "num_tokens": 493818761.0, + "step": 19083 + }, + { + "epoch": 2.0957610366791126, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.456730842590332, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.741546630859375, + "num_tokens": 493842579.0, + "step": 19084 + }, + { + "epoch": 2.0958708543817264, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5067315101623535, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7256588935852051, + "num_tokens": 493866585.0, + "step": 19085 + }, + { + "epoch": 2.09598067208434, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4455647468566895, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7323281764984131, + "num_tokens": 493891549.0, + "step": 19086 + }, + { + "epoch": 2.0960904897869534, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.364877462387085, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7112560272216797, + "num_tokens": 493917048.0, + "step": 19087 + }, + { + "epoch": 2.096200307489567, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.756101608276367, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.73356032371521, + "num_tokens": 493936071.0, + "step": 19088 + }, + { + "epoch": 2.096310125192181, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5564184188842773, + "learning_rate": 1e-06, + "loss": 0.8198, + "mean_token_accuracy": 0.7353887557983398, + "num_tokens": 493957816.0, + "step": 19089 + }, + { + "epoch": 2.0964199428947947, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.2945125102996826, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7160458564758301, + "num_tokens": 493983395.0, + "step": 19090 + }, + { + "epoch": 2.0965297605974085, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.311098098754883, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7118871212005615, + "num_tokens": 494012211.0, + "step": 19091 + }, + { + "epoch": 2.096639578300022, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3055050373077393, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7200126647949219, + "num_tokens": 494038658.0, + "step": 19092 + }, + { + "epoch": 2.0967493960026355, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.402348279953003, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7321309447288513, + "num_tokens": 494064052.0, + "step": 19093 + }, + { + "epoch": 2.0968592137052493, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.387399435043335, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7334156036376953, + "num_tokens": 494089375.0, + "step": 19094 + }, + { + "epoch": 2.096969031407863, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.219135046005249, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7289325594902039, + "num_tokens": 494118483.0, + "step": 19095 + }, + { + "epoch": 2.097078849110477, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.55305814743042, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7128238677978516, + "num_tokens": 494141958.0, + "step": 19096 + }, + { + "epoch": 2.09718866681309, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.1766273975372314, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7265782952308655, + "num_tokens": 494170703.0, + "step": 19097 + }, + { + "epoch": 2.097298484515704, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5346081256866455, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7209998369216919, + "num_tokens": 494194926.0, + "step": 19098 + }, + { + "epoch": 2.0974083022183176, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.376669406890869, + "learning_rate": 1e-06, + "loss": 0.7819, + "mean_token_accuracy": 0.7501352429389954, + "num_tokens": 494220067.0, + "step": 19099 + }, + { + "epoch": 2.0975181199209314, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.238255262374878, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7452383637428284, + "num_tokens": 494246743.0, + "step": 19100 + }, + { + "epoch": 2.0976279376235447, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.6238913536071777, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.72295081615448, + "num_tokens": 494267627.0, + "step": 19101 + }, + { + "epoch": 2.0977377553261585, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.0856857299804688, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7221555709838867, + "num_tokens": 494298287.0, + "step": 19102 + }, + { + "epoch": 2.0978475730287722, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.413424491882324, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7281782627105713, + "num_tokens": 494323848.0, + "step": 19103 + }, + { + "epoch": 2.097957390731386, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4648585319519043, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7391613721847534, + "num_tokens": 494348007.0, + "step": 19104 + }, + { + "epoch": 2.0980672084339997, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.4741463661193848, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7336146831512451, + "num_tokens": 494372682.0, + "step": 19105 + }, + { + "epoch": 2.098177026136613, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.3115618228912354, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7455388307571411, + "num_tokens": 494398639.0, + "step": 19106 + }, + { + "epoch": 2.098286843839227, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.6491029262542725, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7292277812957764, + "num_tokens": 494420579.0, + "step": 19107 + }, + { + "epoch": 2.0983966615418406, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.548375129699707, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7254326343536377, + "num_tokens": 494444564.0, + "step": 19108 + }, + { + "epoch": 2.0985064792444543, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.4638912677764893, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7234996557235718, + "num_tokens": 494468730.0, + "step": 19109 + }, + { + "epoch": 2.0986162969470676, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.450171709060669, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7285419702529907, + "num_tokens": 494494472.0, + "step": 19110 + }, + { + "epoch": 2.0987261146496814, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.212805986404419, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.718862771987915, + "num_tokens": 494523174.0, + "step": 19111 + }, + { + "epoch": 2.098835932352295, + "ewc_loss": 1.1801719665527344e-05, + "grad_norm": 2.5448415279388428, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7360987663269043, + "num_tokens": 494545274.0, + "step": 19112 + }, + { + "epoch": 2.098945750054909, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.4909005165100098, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7373517751693726, + "num_tokens": 494566831.0, + "step": 19113 + }, + { + "epoch": 2.0990555677575227, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.0552141666412354, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.72366863489151, + "num_tokens": 494598634.0, + "step": 19114 + }, + { + "epoch": 2.099165385460136, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.376821756362915, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7218536138534546, + "num_tokens": 494624358.0, + "step": 19115 + }, + { + "epoch": 2.0992752031627497, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2679026126861572, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7365396022796631, + "num_tokens": 494652414.0, + "step": 19116 + }, + { + "epoch": 2.0993850208653635, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.328634023666382, + "learning_rate": 1e-06, + "loss": 0.7749, + "mean_token_accuracy": 0.7506200075149536, + "num_tokens": 494677253.0, + "step": 19117 + }, + { + "epoch": 2.0994948385679773, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.771974563598633, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7496195435523987, + "num_tokens": 494697568.0, + "step": 19118 + }, + { + "epoch": 2.099604656270591, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2990663051605225, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7292407751083374, + "num_tokens": 494723570.0, + "step": 19119 + }, + { + "epoch": 2.0997144739732043, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.647205114364624, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7427924275398254, + "num_tokens": 494744675.0, + "step": 19120 + }, + { + "epoch": 2.099824291675818, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.1575238704681396, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.731977105140686, + "num_tokens": 494771771.0, + "step": 19121 + }, + { + "epoch": 2.099934109378432, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.521436929702759, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7232680916786194, + "num_tokens": 494796506.0, + "step": 19122 + }, + { + "epoch": 2.1000439270810456, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.0232672691345215, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7402150630950928, + "num_tokens": 494829970.0, + "step": 19123 + }, + { + "epoch": 2.1001537447836593, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.396400213241577, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7196797132492065, + "num_tokens": 494856312.0, + "step": 19124 + }, + { + "epoch": 2.1002635624862727, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.156501054763794, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7447667121887207, + "num_tokens": 494887483.0, + "step": 19125 + }, + { + "epoch": 2.1003733801888864, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.43353271484375, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7396602630615234, + "num_tokens": 494910484.0, + "step": 19126 + }, + { + "epoch": 2.1004831978915, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2117133140563965, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7502303123474121, + "num_tokens": 494942311.0, + "step": 19127 + }, + { + "epoch": 2.100593015594114, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.622976779937744, + "learning_rate": 1e-06, + "loss": 0.7396, + "mean_token_accuracy": 0.7591196894645691, + "num_tokens": 494963926.0, + "step": 19128 + }, + { + "epoch": 2.1007028332967272, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.143681526184082, + "learning_rate": 1e-06, + "loss": 0.7374, + "mean_token_accuracy": 0.764492392539978, + "num_tokens": 494992019.0, + "step": 19129 + }, + { + "epoch": 2.100812650999341, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 1.96087646484375, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7097622752189636, + "num_tokens": 495028204.0, + "step": 19130 + }, + { + "epoch": 2.1009224687019548, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.69594407081604, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.742883026599884, + "num_tokens": 495048069.0, + "step": 19131 + }, + { + "epoch": 2.1010322864045685, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.203449249267578, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7538951635360718, + "num_tokens": 495075347.0, + "step": 19132 + }, + { + "epoch": 2.1011421041071823, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.4835433959960938, + "learning_rate": 1e-06, + "loss": 0.7681, + "mean_token_accuracy": 0.7578155994415283, + "num_tokens": 495099759.0, + "step": 19133 + }, + { + "epoch": 2.1012519218097956, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2495481967926025, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7300081849098206, + "num_tokens": 495128353.0, + "step": 19134 + }, + { + "epoch": 2.1013617395124093, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.810195207595825, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.747901439666748, + "num_tokens": 495148321.0, + "step": 19135 + }, + { + "epoch": 2.101471557215023, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.404221534729004, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7275533676147461, + "num_tokens": 495173397.0, + "step": 19136 + }, + { + "epoch": 2.101581374917637, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.455113649368286, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7187650799751282, + "num_tokens": 495197877.0, + "step": 19137 + }, + { + "epoch": 2.10169119262025, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2027266025543213, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7249757051467896, + "num_tokens": 495227587.0, + "step": 19138 + }, + { + "epoch": 2.101801010322864, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.3070054054260254, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7240676879882812, + "num_tokens": 495253735.0, + "step": 19139 + }, + { + "epoch": 2.1019108280254777, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.527843713760376, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7354986667633057, + "num_tokens": 495275860.0, + "step": 19140 + }, + { + "epoch": 2.1020206457280914, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.674882411956787, + "learning_rate": 1e-06, + "loss": 0.7557, + "mean_token_accuracy": 0.7597071528434753, + "num_tokens": 495295552.0, + "step": 19141 + }, + { + "epoch": 2.102130463430705, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.552316665649414, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7374769449234009, + "num_tokens": 495318799.0, + "step": 19142 + }, + { + "epoch": 2.1022402811333185, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.657381057739258, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7422919869422913, + "num_tokens": 495340919.0, + "step": 19143 + }, + { + "epoch": 2.1023500988359323, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2691168785095215, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7302577495574951, + "num_tokens": 495366914.0, + "step": 19144 + }, + { + "epoch": 2.102459916538546, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.885915517807007, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7396306991577148, + "num_tokens": 495384879.0, + "step": 19145 + }, + { + "epoch": 2.10256973424116, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.4999349117279053, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7241642475128174, + "num_tokens": 495407837.0, + "step": 19146 + }, + { + "epoch": 2.1026795519437735, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.296844482421875, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7434543967247009, + "num_tokens": 495434520.0, + "step": 19147 + }, + { + "epoch": 2.102789369646387, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.9231390953063965, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7472578287124634, + "num_tokens": 495452239.0, + "step": 19148 + }, + { + "epoch": 2.1028991873490006, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.3137640953063965, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7265105247497559, + "num_tokens": 495478247.0, + "step": 19149 + }, + { + "epoch": 2.1030090050516144, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2266347408294678, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.72063148021698, + "num_tokens": 495507686.0, + "step": 19150 + }, + { + "epoch": 2.103118822754228, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.0693957805633545, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7231258749961853, + "num_tokens": 495542033.0, + "step": 19151 + }, + { + "epoch": 2.1032286404568414, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.3127529621124268, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7550196647644043, + "num_tokens": 495568759.0, + "step": 19152 + }, + { + "epoch": 2.103338458159455, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5021274089813232, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.731808066368103, + "num_tokens": 495591660.0, + "step": 19153 + }, + { + "epoch": 2.103448275862069, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2319366931915283, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7430093884468079, + "num_tokens": 495618271.0, + "step": 19154 + }, + { + "epoch": 2.1035580935646827, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5514187812805176, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7170372009277344, + "num_tokens": 495642256.0, + "step": 19155 + }, + { + "epoch": 2.1036679112672965, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.644378185272217, + "learning_rate": 1e-06, + "loss": 0.7495, + "mean_token_accuracy": 0.7575753331184387, + "num_tokens": 495662454.0, + "step": 19156 + }, + { + "epoch": 2.1037777289699098, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2551212310791016, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7209896445274353, + "num_tokens": 495690054.0, + "step": 19157 + }, + { + "epoch": 2.1038875466725235, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.4902443885803223, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7429253458976746, + "num_tokens": 495713834.0, + "step": 19158 + }, + { + "epoch": 2.1039973643751373, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.205059051513672, + "learning_rate": 1e-06, + "loss": 0.7485, + "mean_token_accuracy": 0.7578556537628174, + "num_tokens": 495740116.0, + "step": 19159 + }, + { + "epoch": 2.104107182077751, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5903635025024414, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7397720217704773, + "num_tokens": 495762199.0, + "step": 19160 + }, + { + "epoch": 2.104216999780365, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.274094820022583, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7451821565628052, + "num_tokens": 495787374.0, + "step": 19161 + }, + { + "epoch": 2.104326817482978, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.077903985977173, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7234891653060913, + "num_tokens": 495820269.0, + "step": 19162 + }, + { + "epoch": 2.104436635185592, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2678186893463135, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7273076772689819, + "num_tokens": 495848752.0, + "step": 19163 + }, + { + "epoch": 2.1045464528882056, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.560124397277832, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7295047044754028, + "num_tokens": 495871746.0, + "step": 19164 + }, + { + "epoch": 2.1046562705908194, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.6465885639190674, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7202959060668945, + "num_tokens": 495894868.0, + "step": 19165 + }, + { + "epoch": 2.1047660882934327, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.1850228309631348, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7150784730911255, + "num_tokens": 495923860.0, + "step": 19166 + }, + { + "epoch": 2.1048759059960465, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.0442047119140625, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7367767095565796, + "num_tokens": 495957197.0, + "step": 19167 + }, + { + "epoch": 2.10498572369866, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2869210243225098, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7199024558067322, + "num_tokens": 495985904.0, + "step": 19168 + }, + { + "epoch": 2.105095541401274, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.154179573059082, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.6943658590316772, + "num_tokens": 496015293.0, + "step": 19169 + }, + { + "epoch": 2.1052053591038877, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.304767370223999, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7287136912345886, + "num_tokens": 496042205.0, + "step": 19170 + }, + { + "epoch": 2.105315176806501, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.342681646347046, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7269233465194702, + "num_tokens": 496067972.0, + "step": 19171 + }, + { + "epoch": 2.105424994509115, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.420839786529541, + "learning_rate": 1e-06, + "loss": 0.793, + "mean_token_accuracy": 0.7491680979728699, + "num_tokens": 496092607.0, + "step": 19172 + }, + { + "epoch": 2.1055348122117286, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.178873062133789, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7284473776817322, + "num_tokens": 496122886.0, + "step": 19173 + }, + { + "epoch": 2.1056446299143423, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.541372537612915, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7169883251190186, + "num_tokens": 496146924.0, + "step": 19174 + }, + { + "epoch": 2.105754447616956, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5893118381500244, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7383509874343872, + "num_tokens": 496166899.0, + "step": 19175 + }, + { + "epoch": 2.1058642653195694, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.243350028991699, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7190970778465271, + "num_tokens": 496196377.0, + "step": 19176 + }, + { + "epoch": 2.105974083022183, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.305325984954834, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7264547944068909, + "num_tokens": 496223225.0, + "step": 19177 + }, + { + "epoch": 2.106083900724797, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2920780181884766, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7163351774215698, + "num_tokens": 496250112.0, + "step": 19178 + }, + { + "epoch": 2.1061937184274107, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.0722134113311768, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7413574457168579, + "num_tokens": 496279922.0, + "step": 19179 + }, + { + "epoch": 2.106303536130024, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.483489751815796, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7138835191726685, + "num_tokens": 496304436.0, + "step": 19180 + }, + { + "epoch": 2.1064133538326377, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.1699376106262207, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7217122912406921, + "num_tokens": 496334696.0, + "step": 19181 + }, + { + "epoch": 2.1065231715352515, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.5236384868621826, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7368491888046265, + "num_tokens": 496358059.0, + "step": 19182 + }, + { + "epoch": 2.1066329892378652, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.086251735687256, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7256600260734558, + "num_tokens": 496387740.0, + "step": 19183 + }, + { + "epoch": 2.106742806940479, + "ewc_loss": 1.1920928955078125e-05, + "grad_norm": 2.2188167572021484, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7203262448310852, + "num_tokens": 496415509.0, + "step": 19184 + }, + { + "epoch": 2.1068526246430923, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.737563371658325, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7339561581611633, + "num_tokens": 496437011.0, + "step": 19185 + }, + { + "epoch": 2.106962442345706, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.0357868671417236, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7241495847702026, + "num_tokens": 496469315.0, + "step": 19186 + }, + { + "epoch": 2.10707226004832, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.360874652862549, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7241833209991455, + "num_tokens": 496496872.0, + "step": 19187 + }, + { + "epoch": 2.1071820777509336, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.2096006870269775, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7435925602912903, + "num_tokens": 496526371.0, + "step": 19188 + }, + { + "epoch": 2.107291895453547, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.3436617851257324, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7291674613952637, + "num_tokens": 496552103.0, + "step": 19189 + }, + { + "epoch": 2.1074017131561606, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.24929141998291, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7197443842887878, + "num_tokens": 496579357.0, + "step": 19190 + }, + { + "epoch": 2.1075115308587744, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.4189977645874023, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7127247452735901, + "num_tokens": 496605039.0, + "step": 19191 + }, + { + "epoch": 2.107621348561388, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.9231417179107666, + "learning_rate": 1e-06, + "loss": 0.7366, + "mean_token_accuracy": 0.7571877837181091, + "num_tokens": 496621766.0, + "step": 19192 + }, + { + "epoch": 2.107731166264002, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.276277542114258, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7398692965507507, + "num_tokens": 496649137.0, + "step": 19193 + }, + { + "epoch": 2.1078409839666152, + "ewc_loss": 1.1980533599853516e-05, + "grad_norm": 2.3746752738952637, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7350751757621765, + "num_tokens": 496675600.0, + "step": 19194 + }, + { + "epoch": 2.107950801669229, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1802847385406494, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7317136526107788, + "num_tokens": 496702492.0, + "step": 19195 + }, + { + "epoch": 2.1080606193718427, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.3350236415863037, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7479764223098755, + "num_tokens": 496729085.0, + "step": 19196 + }, + { + "epoch": 2.1081704370744565, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4314024448394775, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.729555070400238, + "num_tokens": 496752094.0, + "step": 19197 + }, + { + "epoch": 2.1082802547770703, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.0694563388824463, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7025079727172852, + "num_tokens": 496785434.0, + "step": 19198 + }, + { + "epoch": 2.1083900724796836, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.159616470336914, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7199555039405823, + "num_tokens": 496814915.0, + "step": 19199 + }, + { + "epoch": 2.1084998901822973, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.0908925533294678, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7377138137817383, + "num_tokens": 496842686.0, + "step": 19200 + }, + { + "epoch": 2.108609707884911, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.6191556453704834, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.735831618309021, + "num_tokens": 496863892.0, + "step": 19201 + }, + { + "epoch": 2.108719525587525, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.39066481590271, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.733116090297699, + "num_tokens": 496886522.0, + "step": 19202 + }, + { + "epoch": 2.108829343290138, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5107691287994385, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7431713342666626, + "num_tokens": 496909244.0, + "step": 19203 + }, + { + "epoch": 2.108939160992752, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4650745391845703, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7217704057693481, + "num_tokens": 496935758.0, + "step": 19204 + }, + { + "epoch": 2.1090489786953657, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4096362590789795, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7386349439620972, + "num_tokens": 496960666.0, + "step": 19205 + }, + { + "epoch": 2.1091587963979794, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.22037672996521, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7271493673324585, + "num_tokens": 496987818.0, + "step": 19206 + }, + { + "epoch": 2.109268614100593, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4128568172454834, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.733437716960907, + "num_tokens": 497013415.0, + "step": 19207 + }, + { + "epoch": 2.1093784318032065, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.627156972885132, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7387799024581909, + "num_tokens": 497034957.0, + "step": 19208 + }, + { + "epoch": 2.1094882495058203, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5077743530273438, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.7411755323410034, + "num_tokens": 497057299.0, + "step": 19209 + }, + { + "epoch": 2.109598067208434, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5797841548919678, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7252731323242188, + "num_tokens": 497083160.0, + "step": 19210 + }, + { + "epoch": 2.1097078849110478, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.3846845626831055, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7264000773429871, + "num_tokens": 497110314.0, + "step": 19211 + }, + { + "epoch": 2.1098177026136615, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.425537586212158, + "learning_rate": 1e-06, + "loss": 0.7821, + "mean_token_accuracy": 0.7488416433334351, + "num_tokens": 497133958.0, + "step": 19212 + }, + { + "epoch": 2.109927520316275, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4519031047821045, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7371810674667358, + "num_tokens": 497156841.0, + "step": 19213 + }, + { + "epoch": 2.1100373380188886, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.416025161743164, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7513010501861572, + "num_tokens": 497181270.0, + "step": 19214 + }, + { + "epoch": 2.1101471557215024, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.379244804382324, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7347959280014038, + "num_tokens": 497206080.0, + "step": 19215 + }, + { + "epoch": 2.110256973424116, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.157763719558716, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7182853817939758, + "num_tokens": 497236514.0, + "step": 19216 + }, + { + "epoch": 2.1103667911267294, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.173349380493164, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7199394702911377, + "num_tokens": 497265054.0, + "step": 19217 + }, + { + "epoch": 2.110476608829343, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.7126712799072266, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7467293739318848, + "num_tokens": 497284782.0, + "step": 19218 + }, + { + "epoch": 2.110586426531957, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1010427474975586, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7329546213150024, + "num_tokens": 497315170.0, + "step": 19219 + }, + { + "epoch": 2.1106962442345707, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4022016525268555, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7159379720687866, + "num_tokens": 497342680.0, + "step": 19220 + }, + { + "epoch": 2.1108060619371845, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5959317684173584, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7198758721351624, + "num_tokens": 497365246.0, + "step": 19221 + }, + { + "epoch": 2.1109158796397978, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.619311571121216, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7247322797775269, + "num_tokens": 497386614.0, + "step": 19222 + }, + { + "epoch": 2.1110256973424115, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5705044269561768, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7353899478912354, + "num_tokens": 497408631.0, + "step": 19223 + }, + { + "epoch": 2.1111355150450253, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.2710516452789307, + "learning_rate": 1e-06, + "loss": 0.7642, + "mean_token_accuracy": 0.7534914016723633, + "num_tokens": 497435262.0, + "step": 19224 + }, + { + "epoch": 2.111245332747639, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.2456259727478027, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7361224889755249, + "num_tokens": 497462587.0, + "step": 19225 + }, + { + "epoch": 2.111355150450253, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4625625610351562, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7382729053497314, + "num_tokens": 497486300.0, + "step": 19226 + }, + { + "epoch": 2.111464968152866, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.5554466247558594, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7425217628479004, + "num_tokens": 497507945.0, + "step": 19227 + }, + { + "epoch": 2.11157478585548, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1214354038238525, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7073583602905273, + "num_tokens": 497540400.0, + "step": 19228 + }, + { + "epoch": 2.1116846035580936, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.672189712524414, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7368309497833252, + "num_tokens": 497562841.0, + "step": 19229 + }, + { + "epoch": 2.1117944212607074, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.455169677734375, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7199425101280212, + "num_tokens": 497586063.0, + "step": 19230 + }, + { + "epoch": 2.1119042389633207, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.3581254482269287, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7404223084449768, + "num_tokens": 497611524.0, + "step": 19231 + }, + { + "epoch": 2.1120140566659344, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.355931282043457, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7373790740966797, + "num_tokens": 497636834.0, + "step": 19232 + }, + { + "epoch": 2.112123874368548, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.4172956943511963, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7373455166816711, + "num_tokens": 497662050.0, + "step": 19233 + }, + { + "epoch": 2.112233692071162, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.1467108726501465, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7320167422294617, + "num_tokens": 497691458.0, + "step": 19234 + }, + { + "epoch": 2.1123435097737757, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.516433000564575, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7386760711669922, + "num_tokens": 497713683.0, + "step": 19235 + }, + { + "epoch": 2.112453327476389, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.374047040939331, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7314809560775757, + "num_tokens": 497740788.0, + "step": 19236 + }, + { + "epoch": 2.112563145179003, + "ewc_loss": 1.2040138244628906e-05, + "grad_norm": 2.283081293106079, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7244376540184021, + "num_tokens": 497767673.0, + "step": 19237 + }, + { + "epoch": 2.1126729628816165, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.737971544265747, + "learning_rate": 1e-06, + "loss": 0.7668, + "mean_token_accuracy": 0.7493878602981567, + "num_tokens": 497787761.0, + "step": 19238 + }, + { + "epoch": 2.1127827805842303, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.2473723888397217, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7258436679840088, + "num_tokens": 497818542.0, + "step": 19239 + }, + { + "epoch": 2.1128925982868436, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.378279685974121, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7197518348693848, + "num_tokens": 497845783.0, + "step": 19240 + }, + { + "epoch": 2.1130024159894574, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.1362216472625732, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7265048027038574, + "num_tokens": 497876838.0, + "step": 19241 + }, + { + "epoch": 2.113112233692071, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.1675665378570557, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7201601266860962, + "num_tokens": 497906561.0, + "step": 19242 + }, + { + "epoch": 2.113222051394685, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.225990056991577, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7075855731964111, + "num_tokens": 497936046.0, + "step": 19243 + }, + { + "epoch": 2.1133318690972986, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.5748825073242188, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7288783192634583, + "num_tokens": 497959831.0, + "step": 19244 + }, + { + "epoch": 2.113441686799912, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.40281343460083, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7317174673080444, + "num_tokens": 497986435.0, + "step": 19245 + }, + { + "epoch": 2.1135515045025257, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.212270498275757, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7239223122596741, + "num_tokens": 498016562.0, + "step": 19246 + }, + { + "epoch": 2.1136613222051395, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.2768731117248535, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7418531775474548, + "num_tokens": 498042669.0, + "step": 19247 + }, + { + "epoch": 2.1137711399077532, + "ewc_loss": 1.2099742889404297e-05, + "grad_norm": 2.787956714630127, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7268884181976318, + "num_tokens": 498062798.0, + "step": 19248 + }, + { + "epoch": 2.113880957610367, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.24654483795166, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.721078634262085, + "num_tokens": 498090075.0, + "step": 19249 + }, + { + "epoch": 2.1139907753129803, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1396751403808594, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7211338877677917, + "num_tokens": 498119591.0, + "step": 19250 + }, + { + "epoch": 2.114100593015594, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 1.957952857017517, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7270549535751343, + "num_tokens": 498154592.0, + "step": 19251 + }, + { + "epoch": 2.114210410718208, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.140127420425415, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7158179879188538, + "num_tokens": 498185943.0, + "step": 19252 + }, + { + "epoch": 2.1143202284208216, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.435678482055664, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7193872332572937, + "num_tokens": 498212773.0, + "step": 19253 + }, + { + "epoch": 2.1144300461234353, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.183685302734375, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7372178435325623, + "num_tokens": 498240260.0, + "step": 19254 + }, + { + "epoch": 2.1145398638260486, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.452268362045288, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7307490110397339, + "num_tokens": 498264336.0, + "step": 19255 + }, + { + "epoch": 2.1146496815286624, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2263247966766357, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7228175401687622, + "num_tokens": 498293781.0, + "step": 19256 + }, + { + "epoch": 2.114759499231276, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1832332611083984, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.723182201385498, + "num_tokens": 498322343.0, + "step": 19257 + }, + { + "epoch": 2.11486931693389, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3163576126098633, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.742742657661438, + "num_tokens": 498352780.0, + "step": 19258 + }, + { + "epoch": 2.114979134636503, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2309093475341797, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7421154975891113, + "num_tokens": 498377441.0, + "step": 19259 + }, + { + "epoch": 2.115088952339117, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.742347478866577, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7322460412979126, + "num_tokens": 498397880.0, + "step": 19260 + }, + { + "epoch": 2.1151987700417307, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2058279514312744, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7182872891426086, + "num_tokens": 498426227.0, + "step": 19261 + }, + { + "epoch": 2.1153085877443445, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.368218421936035, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7216455936431885, + "num_tokens": 498453888.0, + "step": 19262 + }, + { + "epoch": 2.1154184054469582, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3957583904266357, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7303860783576965, + "num_tokens": 498482431.0, + "step": 19263 + }, + { + "epoch": 2.1155282231495716, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1932828426361084, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7160676717758179, + "num_tokens": 498517225.0, + "step": 19264 + }, + { + "epoch": 2.1156380408521853, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.325623035430908, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7363049983978271, + "num_tokens": 498542827.0, + "step": 19265 + }, + { + "epoch": 2.115747858554799, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.106856346130371, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6998014450073242, + "num_tokens": 498574854.0, + "step": 19266 + }, + { + "epoch": 2.115857676257413, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4867098331451416, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.728345513343811, + "num_tokens": 498600075.0, + "step": 19267 + }, + { + "epoch": 2.115967493960026, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.271566390991211, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7227189540863037, + "num_tokens": 498627775.0, + "step": 19268 + }, + { + "epoch": 2.11607731166264, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.370238780975342, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7253901958465576, + "num_tokens": 498652204.0, + "step": 19269 + }, + { + "epoch": 2.1161871293652537, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.691310405731201, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.7450587749481201, + "num_tokens": 498672745.0, + "step": 19270 + }, + { + "epoch": 2.1162969470678674, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.494089126586914, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7371129989624023, + "num_tokens": 498695867.0, + "step": 19271 + }, + { + "epoch": 2.116406764770481, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.777996778488159, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7342304587364197, + "num_tokens": 498715297.0, + "step": 19272 + }, + { + "epoch": 2.1165165824730945, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4184863567352295, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7316912412643433, + "num_tokens": 498739354.0, + "step": 19273 + }, + { + "epoch": 2.1166264001757082, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4204752445220947, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.7452899217605591, + "num_tokens": 498762108.0, + "step": 19274 + }, + { + "epoch": 2.116736217878322, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3161673545837402, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.736286461353302, + "num_tokens": 498789026.0, + "step": 19275 + }, + { + "epoch": 2.1168460355809358, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2904481887817383, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7286978960037231, + "num_tokens": 498814879.0, + "step": 19276 + }, + { + "epoch": 2.1169558532835495, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5948214530944824, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7443169355392456, + "num_tokens": 498836829.0, + "step": 19277 + }, + { + "epoch": 2.117065670986163, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5174643993377686, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7267943620681763, + "num_tokens": 498861658.0, + "step": 19278 + }, + { + "epoch": 2.1171754886887766, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 3.0093235969543457, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7293508052825928, + "num_tokens": 498880637.0, + "step": 19279 + }, + { + "epoch": 2.1172853063913903, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.408005714416504, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7182372212409973, + "num_tokens": 498907284.0, + "step": 19280 + }, + { + "epoch": 2.117395124094004, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2961642742156982, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7311416864395142, + "num_tokens": 498936942.0, + "step": 19281 + }, + { + "epoch": 2.1175049417966174, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.094905138015747, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7519583702087402, + "num_tokens": 498966621.0, + "step": 19282 + }, + { + "epoch": 2.117614759499231, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.414267063140869, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7358566522598267, + "num_tokens": 498990140.0, + "step": 19283 + }, + { + "epoch": 2.117724577201845, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.622840166091919, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7313553094863892, + "num_tokens": 499011949.0, + "step": 19284 + }, + { + "epoch": 2.1178343949044587, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2595767974853516, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7348158359527588, + "num_tokens": 499038072.0, + "step": 19285 + }, + { + "epoch": 2.1179442126070724, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5494768619537354, + "learning_rate": 1e-06, + "loss": 0.7173, + "mean_token_accuracy": 0.7662353515625, + "num_tokens": 499058058.0, + "step": 19286 + }, + { + "epoch": 2.1180540303096858, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1472394466400146, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7188419699668884, + "num_tokens": 499090487.0, + "step": 19287 + }, + { + "epoch": 2.1181638480122995, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.343857526779175, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7187857627868652, + "num_tokens": 499117531.0, + "step": 19288 + }, + { + "epoch": 2.1182736657149133, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.260803699493408, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7442450523376465, + "num_tokens": 499146062.0, + "step": 19289 + }, + { + "epoch": 2.118383483417527, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1969637870788574, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7123186588287354, + "num_tokens": 499178273.0, + "step": 19290 + }, + { + "epoch": 2.118493301120141, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1462626457214355, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7327298521995544, + "num_tokens": 499209831.0, + "step": 19291 + }, + { + "epoch": 2.118603118822754, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2942686080932617, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7244905829429626, + "num_tokens": 499236931.0, + "step": 19292 + }, + { + "epoch": 2.118712936525368, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.0665407180786133, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7210265398025513, + "num_tokens": 499268058.0, + "step": 19293 + }, + { + "epoch": 2.1188227542279816, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.346564292907715, + "learning_rate": 1e-06, + "loss": 0.7843, + "mean_token_accuracy": 0.749506413936615, + "num_tokens": 499294168.0, + "step": 19294 + }, + { + "epoch": 2.1189325719305954, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.314539670944214, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7259137630462646, + "num_tokens": 499321024.0, + "step": 19295 + }, + { + "epoch": 2.1190423896332087, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.102872610092163, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7512849569320679, + "num_tokens": 499352640.0, + "step": 19296 + }, + { + "epoch": 2.1191522073358224, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.395501136779785, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7281291484832764, + "num_tokens": 499379641.0, + "step": 19297 + }, + { + "epoch": 2.119262025038436, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.133258104324341, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7213187217712402, + "num_tokens": 499411356.0, + "step": 19298 + }, + { + "epoch": 2.11937184274105, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.739665985107422, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.7445949912071228, + "num_tokens": 499431949.0, + "step": 19299 + }, + { + "epoch": 2.1194816604436637, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.916637420654297, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.737717866897583, + "num_tokens": 499448969.0, + "step": 19300 + }, + { + "epoch": 2.119591478146277, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.352847099304199, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7320228219032288, + "num_tokens": 499474514.0, + "step": 19301 + }, + { + "epoch": 2.1197012958488908, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5367085933685303, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7362345457077026, + "num_tokens": 499497833.0, + "step": 19302 + }, + { + "epoch": 2.1198111135515045, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1665329933166504, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7434920072555542, + "num_tokens": 499527147.0, + "step": 19303 + }, + { + "epoch": 2.1199209312541183, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3659324645996094, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7231802344322205, + "num_tokens": 499553289.0, + "step": 19304 + }, + { + "epoch": 2.120030748956732, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.7366843223571777, + "learning_rate": 1e-06, + "loss": 0.7382, + "mean_token_accuracy": 0.7576655149459839, + "num_tokens": 499575098.0, + "step": 19305 + }, + { + "epoch": 2.1201405666593454, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.16229510307312, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7154480218887329, + "num_tokens": 499605573.0, + "step": 19306 + }, + { + "epoch": 2.120250384361959, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3997957706451416, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7338111400604248, + "num_tokens": 499630419.0, + "step": 19307 + }, + { + "epoch": 2.120360202064573, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.345867156982422, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7396939992904663, + "num_tokens": 499655863.0, + "step": 19308 + }, + { + "epoch": 2.1204700197671866, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.0361883640289307, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7149361371994019, + "num_tokens": 499689025.0, + "step": 19309 + }, + { + "epoch": 2.1205798374698, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.159954309463501, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7367289066314697, + "num_tokens": 499717713.0, + "step": 19310 + }, + { + "epoch": 2.1206896551724137, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.586888551712036, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.753394365310669, + "num_tokens": 499739851.0, + "step": 19311 + }, + { + "epoch": 2.1207994728750275, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1024246215820312, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7356387972831726, + "num_tokens": 499770583.0, + "step": 19312 + }, + { + "epoch": 2.120909290577641, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.29085111618042, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7497226595878601, + "num_tokens": 499798119.0, + "step": 19313 + }, + { + "epoch": 2.121019108280255, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.202181339263916, + "learning_rate": 1e-06, + "loss": 0.956, + "mean_token_accuracy": 0.7033172249794006, + "num_tokens": 499828541.0, + "step": 19314 + }, + { + "epoch": 2.1211289259828683, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.577176094055176, + "learning_rate": 1e-06, + "loss": 0.7643, + "mean_token_accuracy": 0.7566031813621521, + "num_tokens": 499849585.0, + "step": 19315 + }, + { + "epoch": 2.121238743685482, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4145631790161133, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7410972714424133, + "num_tokens": 499871411.0, + "step": 19316 + }, + { + "epoch": 2.121348561388096, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5714476108551025, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7299636602401733, + "num_tokens": 499895292.0, + "step": 19317 + }, + { + "epoch": 2.1214583790907096, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2413547039031982, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7206097841262817, + "num_tokens": 499925480.0, + "step": 19318 + }, + { + "epoch": 2.121568196793323, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.6163148880004883, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.725570559501648, + "num_tokens": 499948452.0, + "step": 19319 + }, + { + "epoch": 2.1216780144959366, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3677186965942383, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7373611927032471, + "num_tokens": 499972771.0, + "step": 19320 + }, + { + "epoch": 2.1217878321985504, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.745293617248535, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.738318145275116, + "num_tokens": 499993166.0, + "step": 19321 + }, + { + "epoch": 2.121897649901164, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.186084032058716, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7429764866828918, + "num_tokens": 500024026.0, + "step": 19322 + }, + { + "epoch": 2.122007467603778, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3287394046783447, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7321479916572571, + "num_tokens": 500052901.0, + "step": 19323 + }, + { + "epoch": 2.122117285306391, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.474182605743408, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7270419597625732, + "num_tokens": 500076201.0, + "step": 19324 + }, + { + "epoch": 2.122227103009005, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2621097564697266, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7147622108459473, + "num_tokens": 500102261.0, + "step": 19325 + }, + { + "epoch": 2.1223369207116187, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.765127420425415, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7530758380889893, + "num_tokens": 500121653.0, + "step": 19326 + }, + { + "epoch": 2.1224467384142325, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.217941999435425, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7345930933952332, + "num_tokens": 500150513.0, + "step": 19327 + }, + { + "epoch": 2.1225565561168462, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3361525535583496, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7255361080169678, + "num_tokens": 500176494.0, + "step": 19328 + }, + { + "epoch": 2.1226663738194596, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.714892864227295, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7508912086486816, + "num_tokens": 500197664.0, + "step": 19329 + }, + { + "epoch": 2.1227761915220733, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.279252290725708, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7162075638771057, + "num_tokens": 500225618.0, + "step": 19330 + }, + { + "epoch": 2.122886009224687, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5053577423095703, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7345880270004272, + "num_tokens": 500250309.0, + "step": 19331 + }, + { + "epoch": 2.122995826927301, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4196150302886963, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7398495674133301, + "num_tokens": 500274450.0, + "step": 19332 + }, + { + "epoch": 2.123105644629914, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1074540615081787, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7233152389526367, + "num_tokens": 500305517.0, + "step": 19333 + }, + { + "epoch": 2.123215462332528, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4048354625701904, + "learning_rate": 1e-06, + "loss": 0.7235, + "mean_token_accuracy": 0.7667237520217896, + "num_tokens": 500329715.0, + "step": 19334 + }, + { + "epoch": 2.1233252800351416, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4283347129821777, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.7475485801696777, + "num_tokens": 500352933.0, + "step": 19335 + }, + { + "epoch": 2.1234350977377554, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4936962127685547, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.730461835861206, + "num_tokens": 500374961.0, + "step": 19336 + }, + { + "epoch": 2.123544915440369, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.214723825454712, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7389753460884094, + "num_tokens": 500404202.0, + "step": 19337 + }, + { + "epoch": 2.1236547331429825, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2521395683288574, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7332656383514404, + "num_tokens": 500431137.0, + "step": 19338 + }, + { + "epoch": 2.1237645508455962, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3192343711853027, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.745661735534668, + "num_tokens": 500456704.0, + "step": 19339 + }, + { + "epoch": 2.12387436854821, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5937113761901855, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7404206991195679, + "num_tokens": 500479314.0, + "step": 19340 + }, + { + "epoch": 2.1239841862508237, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.414646625518799, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.728807806968689, + "num_tokens": 500502601.0, + "step": 19341 + }, + { + "epoch": 2.1240940039534375, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.623439073562622, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7238396406173706, + "num_tokens": 500525637.0, + "step": 19342 + }, + { + "epoch": 2.124203821656051, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.279268980026245, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7257834672927856, + "num_tokens": 500551665.0, + "step": 19343 + }, + { + "epoch": 2.1243136393586646, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3371636867523193, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7215560674667358, + "num_tokens": 500580210.0, + "step": 19344 + }, + { + "epoch": 2.1244234570612783, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4456894397735596, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7198026180267334, + "num_tokens": 500605288.0, + "step": 19345 + }, + { + "epoch": 2.124533274763892, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.0784378051757812, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7197840213775635, + "num_tokens": 500634790.0, + "step": 19346 + }, + { + "epoch": 2.1246430924665054, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3340189456939697, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.710534930229187, + "num_tokens": 500661686.0, + "step": 19347 + }, + { + "epoch": 2.124752910169119, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.6902239322662354, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7465087175369263, + "num_tokens": 500682856.0, + "step": 19348 + }, + { + "epoch": 2.124862727871733, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1471874713897705, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7231883406639099, + "num_tokens": 500713087.0, + "step": 19349 + }, + { + "epoch": 2.1249725455743467, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.5181884765625, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7357220649719238, + "num_tokens": 500736800.0, + "step": 19350 + }, + { + "epoch": 2.1250823632769604, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4659502506256104, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7433646321296692, + "num_tokens": 500760036.0, + "step": 19351 + }, + { + "epoch": 2.1251921809795737, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4558193683624268, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7335155010223389, + "num_tokens": 500786040.0, + "step": 19352 + }, + { + "epoch": 2.1253019986821875, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2589337825775146, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7321114540100098, + "num_tokens": 500813124.0, + "step": 19353 + }, + { + "epoch": 2.1254118163848013, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4245431423187256, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7406147718429565, + "num_tokens": 500838213.0, + "step": 19354 + }, + { + "epoch": 2.125521634087415, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.427985429763794, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7377123236656189, + "num_tokens": 500863071.0, + "step": 19355 + }, + { + "epoch": 2.1256314517900288, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3639557361602783, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7285038232803345, + "num_tokens": 500889884.0, + "step": 19356 + }, + { + "epoch": 2.125741269492642, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2433862686157227, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7497619390487671, + "num_tokens": 500917979.0, + "step": 19357 + }, + { + "epoch": 2.125851087195256, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.3683998584747314, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7268002033233643, + "num_tokens": 500944795.0, + "step": 19358 + }, + { + "epoch": 2.1259609048978696, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.1990973949432373, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7260673642158508, + "num_tokens": 500973310.0, + "step": 19359 + }, + { + "epoch": 2.1260707226004834, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.233441114425659, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7342743873596191, + "num_tokens": 501001309.0, + "step": 19360 + }, + { + "epoch": 2.1261805403030967, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.194056272506714, + "learning_rate": 1e-06, + "loss": 0.921, + "mean_token_accuracy": 0.7134393453598022, + "num_tokens": 501031692.0, + "step": 19361 + }, + { + "epoch": 2.1262903580057104, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.532306671142578, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7322157621383667, + "num_tokens": 501055722.0, + "step": 19362 + }, + { + "epoch": 2.126400175708324, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.440429210662842, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7392179369926453, + "num_tokens": 501079954.0, + "step": 19363 + }, + { + "epoch": 2.126509993410938, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.490363836288452, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7407459020614624, + "num_tokens": 501102979.0, + "step": 19364 + }, + { + "epoch": 2.1266198111135517, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.2719836235046387, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7343071699142456, + "num_tokens": 501128962.0, + "step": 19365 + }, + { + "epoch": 2.126729628816165, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.4957339763641357, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7368154525756836, + "num_tokens": 501153358.0, + "step": 19366 + }, + { + "epoch": 2.1268394465187788, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.0757744312286377, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7325260639190674, + "num_tokens": 501183138.0, + "step": 19367 + }, + { + "epoch": 2.1269492642213925, + "ewc_loss": 1.2218952178955078e-05, + "grad_norm": 2.774623394012451, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7383607029914856, + "num_tokens": 501202590.0, + "step": 19368 + }, + { + "epoch": 2.1270590819240063, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5687952041625977, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7445301413536072, + "num_tokens": 501224546.0, + "step": 19369 + }, + { + "epoch": 2.1271688996266196, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.506290912628174, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7458921670913696, + "num_tokens": 501246890.0, + "step": 19370 + }, + { + "epoch": 2.1272787173292333, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5846478939056396, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7446765303611755, + "num_tokens": 501271218.0, + "step": 19371 + }, + { + "epoch": 2.127388535031847, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.268294334411621, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7219314575195312, + "num_tokens": 501298334.0, + "step": 19372 + }, + { + "epoch": 2.127498352734461, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.688328742980957, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7448967099189758, + "num_tokens": 501318225.0, + "step": 19373 + }, + { + "epoch": 2.1276081704370746, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.565502405166626, + "learning_rate": 1e-06, + "loss": 0.756, + "mean_token_accuracy": 0.7505910396575928, + "num_tokens": 501337679.0, + "step": 19374 + }, + { + "epoch": 2.127717988139688, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.6322078704833984, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7547509670257568, + "num_tokens": 501359800.0, + "step": 19375 + }, + { + "epoch": 2.1278278058423017, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.374537706375122, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7214818000793457, + "num_tokens": 501384640.0, + "step": 19376 + }, + { + "epoch": 2.1279376235449154, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4053921699523926, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7399678826332092, + "num_tokens": 501408346.0, + "step": 19377 + }, + { + "epoch": 2.128047441247529, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3955421447753906, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7397767305374146, + "num_tokens": 501434397.0, + "step": 19378 + }, + { + "epoch": 2.128157258950143, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5964646339416504, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7297459840774536, + "num_tokens": 501458928.0, + "step": 19379 + }, + { + "epoch": 2.1282670766527563, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.541529417037964, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7328765392303467, + "num_tokens": 501482911.0, + "step": 19380 + }, + { + "epoch": 2.12837689435537, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.276928424835205, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7294926643371582, + "num_tokens": 501511662.0, + "step": 19381 + }, + { + "epoch": 2.128486712057984, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1408631801605225, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7300385236740112, + "num_tokens": 501540540.0, + "step": 19382 + }, + { + "epoch": 2.1285965297605975, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1965525150299072, + "learning_rate": 1e-06, + "loss": 0.9469, + "mean_token_accuracy": 0.7088323831558228, + "num_tokens": 501570794.0, + "step": 19383 + }, + { + "epoch": 2.1287063474632113, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3040215969085693, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7151486277580261, + "num_tokens": 501598342.0, + "step": 19384 + }, + { + "epoch": 2.1288161651658246, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.661062240600586, + "learning_rate": 1e-06, + "loss": 0.7846, + "mean_token_accuracy": 0.7496286034584045, + "num_tokens": 501617992.0, + "step": 19385 + }, + { + "epoch": 2.1289259828684384, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5115561485290527, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7275356650352478, + "num_tokens": 501640359.0, + "step": 19386 + }, + { + "epoch": 2.129035800571052, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3526546955108643, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7372812032699585, + "num_tokens": 501666252.0, + "step": 19387 + }, + { + "epoch": 2.129145618273666, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.185335159301758, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7087763547897339, + "num_tokens": 501696625.0, + "step": 19388 + }, + { + "epoch": 2.129255435976279, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.412235975265503, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7098026871681213, + "num_tokens": 501721953.0, + "step": 19389 + }, + { + "epoch": 2.129365253678893, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.097036361694336, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7473754286766052, + "num_tokens": 501751758.0, + "step": 19390 + }, + { + "epoch": 2.1294750713815067, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3099875450134277, + "learning_rate": 1e-06, + "loss": 0.7614, + "mean_token_accuracy": 0.7565497159957886, + "num_tokens": 501777933.0, + "step": 19391 + }, + { + "epoch": 2.1295848890841205, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3029181957244873, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7051507234573364, + "num_tokens": 501805384.0, + "step": 19392 + }, + { + "epoch": 2.1296947067867342, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.8086798191070557, + "learning_rate": 1e-06, + "loss": 0.7618, + "mean_token_accuracy": 0.7539207339286804, + "num_tokens": 501825156.0, + "step": 19393 + }, + { + "epoch": 2.1298045244893475, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5066044330596924, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7542632222175598, + "num_tokens": 501847690.0, + "step": 19394 + }, + { + "epoch": 2.1299143421919613, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5516462326049805, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7374008297920227, + "num_tokens": 501871064.0, + "step": 19395 + }, + { + "epoch": 2.130024159894575, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2496519088745117, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7173914909362793, + "num_tokens": 501898573.0, + "step": 19396 + }, + { + "epoch": 2.130133977597189, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.262454032897949, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7300480604171753, + "num_tokens": 501926272.0, + "step": 19397 + }, + { + "epoch": 2.130243795299802, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2724781036376953, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7028589248657227, + "num_tokens": 501954207.0, + "step": 19398 + }, + { + "epoch": 2.130353613002416, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5774810314178467, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7331464290618896, + "num_tokens": 501977226.0, + "step": 19399 + }, + { + "epoch": 2.1304634307050296, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.315570592880249, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7480238080024719, + "num_tokens": 502001304.0, + "step": 19400 + }, + { + "epoch": 2.1305732484076434, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.6083927154541016, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7390611171722412, + "num_tokens": 502023302.0, + "step": 19401 + }, + { + "epoch": 2.130683066110257, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3305728435516357, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.734082043170929, + "num_tokens": 502048565.0, + "step": 19402 + }, + { + "epoch": 2.1307928838128705, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.523214340209961, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7228411436080933, + "num_tokens": 502071693.0, + "step": 19403 + }, + { + "epoch": 2.130902701515484, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1706736087799072, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7048888206481934, + "num_tokens": 502102292.0, + "step": 19404 + }, + { + "epoch": 2.131012519218098, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5796332359313965, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7463279962539673, + "num_tokens": 502124153.0, + "step": 19405 + }, + { + "epoch": 2.1311223369207117, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3668172359466553, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7471758127212524, + "num_tokens": 502149127.0, + "step": 19406 + }, + { + "epoch": 2.1312321546233255, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2546920776367188, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7416917085647583, + "num_tokens": 502175688.0, + "step": 19407 + }, + { + "epoch": 2.131341972325939, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 3.022935390472412, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7385127544403076, + "num_tokens": 502194729.0, + "step": 19408 + }, + { + "epoch": 2.1314517900285526, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.235896348953247, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7334566116333008, + "num_tokens": 502223616.0, + "step": 19409 + }, + { + "epoch": 2.1315616077311663, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5793509483337402, + "learning_rate": 1e-06, + "loss": 0.8443, + "mean_token_accuracy": 0.736553966999054, + "num_tokens": 502246377.0, + "step": 19410 + }, + { + "epoch": 2.13167142543378, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3426332473754883, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7153372764587402, + "num_tokens": 502274348.0, + "step": 19411 + }, + { + "epoch": 2.1317812431363934, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.143225908279419, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7391436696052551, + "num_tokens": 502304932.0, + "step": 19412 + }, + { + "epoch": 2.131891060839007, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1041815280914307, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7259626388549805, + "num_tokens": 502334295.0, + "step": 19413 + }, + { + "epoch": 2.132000878541621, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.199871063232422, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7317426204681396, + "num_tokens": 502361123.0, + "step": 19414 + }, + { + "epoch": 2.1321106962442347, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5639352798461914, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.741209864616394, + "num_tokens": 502385868.0, + "step": 19415 + }, + { + "epoch": 2.1322205139468484, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3288469314575195, + "learning_rate": 1e-06, + "loss": 0.7293, + "mean_token_accuracy": 0.7652533650398254, + "num_tokens": 502411193.0, + "step": 19416 + }, + { + "epoch": 2.1323303316494617, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4925074577331543, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7330489754676819, + "num_tokens": 502436624.0, + "step": 19417 + }, + { + "epoch": 2.1324401493520755, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3741214275360107, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7142196893692017, + "num_tokens": 502463703.0, + "step": 19418 + }, + { + "epoch": 2.1325499670546892, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3728113174438477, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7287806272506714, + "num_tokens": 502490220.0, + "step": 19419 + }, + { + "epoch": 2.132659784757303, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.00895619392395, + "learning_rate": 1e-06, + "loss": 0.7617, + "mean_token_accuracy": 0.7604421973228455, + "num_tokens": 502520829.0, + "step": 19420 + }, + { + "epoch": 2.1327696024599163, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.191650867462158, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7172790765762329, + "num_tokens": 502548809.0, + "step": 19421 + }, + { + "epoch": 2.13287942016253, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3059186935424805, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7363255620002747, + "num_tokens": 502572556.0, + "step": 19422 + }, + { + "epoch": 2.132989237865144, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.247462749481201, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7192524671554565, + "num_tokens": 502601085.0, + "step": 19423 + }, + { + "epoch": 2.1330990555677576, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.361389636993408, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7277195453643799, + "num_tokens": 502626949.0, + "step": 19424 + }, + { + "epoch": 2.1332088732703713, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.6439404487609863, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7541663646697998, + "num_tokens": 502648130.0, + "step": 19425 + }, + { + "epoch": 2.1333186909729847, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.423879861831665, + "learning_rate": 1e-06, + "loss": 0.9675, + "mean_token_accuracy": 0.704895555973053, + "num_tokens": 502672859.0, + "step": 19426 + }, + { + "epoch": 2.1334285086755984, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.139768362045288, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7110949754714966, + "num_tokens": 502704382.0, + "step": 19427 + }, + { + "epoch": 2.133538326378212, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1906864643096924, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7456739544868469, + "num_tokens": 502734070.0, + "step": 19428 + }, + { + "epoch": 2.133648144080826, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 3.0027339458465576, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7485904693603516, + "num_tokens": 502752470.0, + "step": 19429 + }, + { + "epoch": 2.1337579617834397, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4108524322509766, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7231053709983826, + "num_tokens": 502777691.0, + "step": 19430 + }, + { + "epoch": 2.133867779486053, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5100088119506836, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7257184982299805, + "num_tokens": 502802280.0, + "step": 19431 + }, + { + "epoch": 2.1339775971886668, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.0062243938446045, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7228753566741943, + "num_tokens": 502836765.0, + "step": 19432 + }, + { + "epoch": 2.1340874148912805, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1821775436401367, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7101732492446899, + "num_tokens": 502865359.0, + "step": 19433 + }, + { + "epoch": 2.1341972325938943, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3722126483917236, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7295799255371094, + "num_tokens": 502890557.0, + "step": 19434 + }, + { + "epoch": 2.134307050296508, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1808905601501465, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7337916493415833, + "num_tokens": 502919872.0, + "step": 19435 + }, + { + "epoch": 2.1344168679991213, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5861520767211914, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7095155119895935, + "num_tokens": 502945931.0, + "step": 19436 + }, + { + "epoch": 2.134526685701735, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2999253273010254, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.730154275894165, + "num_tokens": 502973651.0, + "step": 19437 + }, + { + "epoch": 2.134636503404349, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.393491744995117, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7235267758369446, + "num_tokens": 502999893.0, + "step": 19438 + }, + { + "epoch": 2.1347463211069626, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.8926503658294678, + "learning_rate": 1e-06, + "loss": 0.7758, + "mean_token_accuracy": 0.7491244077682495, + "num_tokens": 503019470.0, + "step": 19439 + }, + { + "epoch": 2.134856138809576, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5072929859161377, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.72933030128479, + "num_tokens": 503043003.0, + "step": 19440 + }, + { + "epoch": 2.1349659565121897, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1614181995391846, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7189065217971802, + "num_tokens": 503072935.0, + "step": 19441 + }, + { + "epoch": 2.1350757742148034, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3772637844085693, + "learning_rate": 1e-06, + "loss": 0.7322, + "mean_token_accuracy": 0.7612206339836121, + "num_tokens": 503097451.0, + "step": 19442 + }, + { + "epoch": 2.135185591917417, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4239213466644287, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7343007326126099, + "num_tokens": 503122261.0, + "step": 19443 + }, + { + "epoch": 2.135295409620031, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2247912883758545, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7159432172775269, + "num_tokens": 503152427.0, + "step": 19444 + }, + { + "epoch": 2.1354052273226443, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.434124708175659, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7136013507843018, + "num_tokens": 503177395.0, + "step": 19445 + }, + { + "epoch": 2.135515045025258, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5091216564178467, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7334616780281067, + "num_tokens": 503200906.0, + "step": 19446 + }, + { + "epoch": 2.1356248627278718, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2882492542266846, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7263786792755127, + "num_tokens": 503224808.0, + "step": 19447 + }, + { + "epoch": 2.1357346804304855, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3087995052337646, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7179297208786011, + "num_tokens": 503252145.0, + "step": 19448 + }, + { + "epoch": 2.135844498133099, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3469600677490234, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7485502362251282, + "num_tokens": 503280156.0, + "step": 19449 + }, + { + "epoch": 2.1359543158357126, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4126126766204834, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7524833679199219, + "num_tokens": 503305478.0, + "step": 19450 + }, + { + "epoch": 2.1360641335383264, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4398159980773926, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7238681316375732, + "num_tokens": 503328600.0, + "step": 19451 + }, + { + "epoch": 2.13617395124094, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.656442642211914, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7250473499298096, + "num_tokens": 503351610.0, + "step": 19452 + }, + { + "epoch": 2.136283768943554, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.75174617767334, + "learning_rate": 1e-06, + "loss": 0.7529, + "mean_token_accuracy": 0.7572019100189209, + "num_tokens": 503370742.0, + "step": 19453 + }, + { + "epoch": 2.136393586646167, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.302002191543579, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7384475469589233, + "num_tokens": 503397469.0, + "step": 19454 + }, + { + "epoch": 2.136503404348781, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 1.9957481622695923, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.7377854585647583, + "num_tokens": 503431983.0, + "step": 19455 + }, + { + "epoch": 2.1366132220513947, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2633168697357178, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.739587664604187, + "num_tokens": 503460751.0, + "step": 19456 + }, + { + "epoch": 2.1367230397540085, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3611435890197754, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7297546863555908, + "num_tokens": 503485362.0, + "step": 19457 + }, + { + "epoch": 2.136832857456622, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.436197519302368, + "learning_rate": 1e-06, + "loss": 0.7585, + "mean_token_accuracy": 0.7532508969306946, + "num_tokens": 503508116.0, + "step": 19458 + }, + { + "epoch": 2.1369426751592355, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.312711238861084, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7289241552352905, + "num_tokens": 503535310.0, + "step": 19459 + }, + { + "epoch": 2.1370524928618493, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.29215407371521, + "learning_rate": 1e-06, + "loss": 0.9743, + "mean_token_accuracy": 0.6975963115692139, + "num_tokens": 503565033.0, + "step": 19460 + }, + { + "epoch": 2.137162310564463, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.591592788696289, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7550184726715088, + "num_tokens": 503585580.0, + "step": 19461 + }, + { + "epoch": 2.137272128267077, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.397918701171875, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7388149499893188, + "num_tokens": 503610320.0, + "step": 19462 + }, + { + "epoch": 2.13738194596969, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.31667423248291, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7395844459533691, + "num_tokens": 503636171.0, + "step": 19463 + }, + { + "epoch": 2.137491763672304, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1217150688171387, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7232823371887207, + "num_tokens": 503667294.0, + "step": 19464 + }, + { + "epoch": 2.1376015813749176, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.320640802383423, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7467911243438721, + "num_tokens": 503693320.0, + "step": 19465 + }, + { + "epoch": 2.1377113990775314, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3936219215393066, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7424238324165344, + "num_tokens": 503718333.0, + "step": 19466 + }, + { + "epoch": 2.137821216780145, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.0534889698028564, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7366222143173218, + "num_tokens": 503750800.0, + "step": 19467 + }, + { + "epoch": 2.1379310344827585, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.247140645980835, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7491060495376587, + "num_tokens": 503776981.0, + "step": 19468 + }, + { + "epoch": 2.138040852185372, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.099484920501709, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7393697500228882, + "num_tokens": 503805743.0, + "step": 19469 + }, + { + "epoch": 2.138150669887986, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.466644287109375, + "learning_rate": 1e-06, + "loss": 0.7682, + "mean_token_accuracy": 0.7573096752166748, + "num_tokens": 503828450.0, + "step": 19470 + }, + { + "epoch": 2.1382604875905997, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5715415477752686, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7407066822052002, + "num_tokens": 503851585.0, + "step": 19471 + }, + { + "epoch": 2.138370305293213, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.32552433013916, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7412377595901489, + "num_tokens": 503877200.0, + "step": 19472 + }, + { + "epoch": 2.138480122995827, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.612659215927124, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7296426296234131, + "num_tokens": 503901070.0, + "step": 19473 + }, + { + "epoch": 2.1385899406984406, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.496140241622925, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7214281558990479, + "num_tokens": 503924246.0, + "step": 19474 + }, + { + "epoch": 2.1386997584010543, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1514854431152344, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7194088697433472, + "num_tokens": 503954041.0, + "step": 19475 + }, + { + "epoch": 2.138809576103668, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.219850778579712, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.733116865158081, + "num_tokens": 503981062.0, + "step": 19476 + }, + { + "epoch": 2.1389193938062814, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.728851795196533, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.736375093460083, + "num_tokens": 504002281.0, + "step": 19477 + }, + { + "epoch": 2.139029211508895, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.144894599914551, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.740935206413269, + "num_tokens": 504032604.0, + "step": 19478 + }, + { + "epoch": 2.139139029211509, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1545283794403076, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7207844257354736, + "num_tokens": 504061406.0, + "step": 19479 + }, + { + "epoch": 2.1392488469141226, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.579995632171631, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7398982048034668, + "num_tokens": 504082777.0, + "step": 19480 + }, + { + "epoch": 2.1393586646167364, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.412186622619629, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.725838303565979, + "num_tokens": 504107983.0, + "step": 19481 + }, + { + "epoch": 2.1394684823193497, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2109487056732178, + "learning_rate": 1e-06, + "loss": 0.9809, + "mean_token_accuracy": 0.7042564153671265, + "num_tokens": 504139751.0, + "step": 19482 + }, + { + "epoch": 2.1395783000219635, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.434506893157959, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7309750318527222, + "num_tokens": 504163631.0, + "step": 19483 + }, + { + "epoch": 2.1396881177245772, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4721827507019043, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7200648188591003, + "num_tokens": 504188728.0, + "step": 19484 + }, + { + "epoch": 2.139797935427191, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.414306879043579, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7273658514022827, + "num_tokens": 504214656.0, + "step": 19485 + }, + { + "epoch": 2.1399077531298047, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5573205947875977, + "learning_rate": 1e-06, + "loss": 0.7017, + "mean_token_accuracy": 0.7648686766624451, + "num_tokens": 504236576.0, + "step": 19486 + }, + { + "epoch": 2.140017570832418, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.304816484451294, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7375465631484985, + "num_tokens": 504263587.0, + "step": 19487 + }, + { + "epoch": 2.140127388535032, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4553239345550537, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7401166558265686, + "num_tokens": 504286875.0, + "step": 19488 + }, + { + "epoch": 2.1402372062376456, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.3897430896759033, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7467603087425232, + "num_tokens": 504310774.0, + "step": 19489 + }, + { + "epoch": 2.1403470239402593, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.6529769897460938, + "learning_rate": 1e-06, + "loss": 0.7598, + "mean_token_accuracy": 0.7535147666931152, + "num_tokens": 504332683.0, + "step": 19490 + }, + { + "epoch": 2.1404568416428726, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.379805088043213, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7025843858718872, + "num_tokens": 504359380.0, + "step": 19491 + }, + { + "epoch": 2.1405666593454864, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.213148593902588, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7304412722587585, + "num_tokens": 504389134.0, + "step": 19492 + }, + { + "epoch": 2.1406764770481, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.403240203857422, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.721806526184082, + "num_tokens": 504414439.0, + "step": 19493 + }, + { + "epoch": 2.140786294750714, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5828921794891357, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7215689420700073, + "num_tokens": 504437166.0, + "step": 19494 + }, + { + "epoch": 2.1408961124533277, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4233062267303467, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7325092554092407, + "num_tokens": 504460330.0, + "step": 19495 + }, + { + "epoch": 2.141005930155941, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.2906105518341064, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7127270102500916, + "num_tokens": 504487957.0, + "step": 19496 + }, + { + "epoch": 2.1411157478585547, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.183407783508301, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.732374906539917, + "num_tokens": 504516567.0, + "step": 19497 + }, + { + "epoch": 2.1412255655611685, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.281398296356201, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7274172306060791, + "num_tokens": 504545127.0, + "step": 19498 + }, + { + "epoch": 2.1413353832637823, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.908956527709961, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7470517158508301, + "num_tokens": 504563497.0, + "step": 19499 + }, + { + "epoch": 2.1414452009663956, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.4133498668670654, + "learning_rate": 1e-06, + "loss": 0.7601, + "mean_token_accuracy": 0.7488086223602295, + "num_tokens": 504587343.0, + "step": 19500 + }, + { + "epoch": 2.1415550186690093, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.075422525405884, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.728083610534668, + "num_tokens": 504620265.0, + "step": 19501 + }, + { + "epoch": 2.141664836371623, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.768566846847534, + "learning_rate": 1e-06, + "loss": 0.7032, + "mean_token_accuracy": 0.7686598896980286, + "num_tokens": 504638830.0, + "step": 19502 + }, + { + "epoch": 2.141774654074237, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5941030979156494, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7309368848800659, + "num_tokens": 504661914.0, + "step": 19503 + }, + { + "epoch": 2.1418844717768506, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.306128978729248, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.726669192314148, + "num_tokens": 504689335.0, + "step": 19504 + }, + { + "epoch": 2.141994289479464, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.1420493125915527, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.727375328540802, + "num_tokens": 504718675.0, + "step": 19505 + }, + { + "epoch": 2.1421041071820777, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.5045247077941895, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7420525550842285, + "num_tokens": 504741613.0, + "step": 19506 + }, + { + "epoch": 2.1422139248846914, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 3.4975249767303467, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7322288751602173, + "num_tokens": 504769859.0, + "step": 19507 + }, + { + "epoch": 2.142323742587305, + "ewc_loss": 1.2278556823730469e-05, + "grad_norm": 2.236431121826172, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7479788661003113, + "num_tokens": 504797627.0, + "step": 19508 + }, + { + "epoch": 2.142433560289919, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4434828758239746, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.7456227540969849, + "num_tokens": 504820414.0, + "step": 19509 + }, + { + "epoch": 2.1425433779925322, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.362656593322754, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7269373536109924, + "num_tokens": 504844428.0, + "step": 19510 + }, + { + "epoch": 2.142653195695146, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4266371726989746, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7015800476074219, + "num_tokens": 504871808.0, + "step": 19511 + }, + { + "epoch": 2.1427630133977598, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.407252073287964, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7285562753677368, + "num_tokens": 504897352.0, + "step": 19512 + }, + { + "epoch": 2.1428728311003735, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5362751483917236, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7407975196838379, + "num_tokens": 504920529.0, + "step": 19513 + }, + { + "epoch": 2.142982648802987, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5370326042175293, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7354822158813477, + "num_tokens": 504943383.0, + "step": 19514 + }, + { + "epoch": 2.1430924665056006, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2430691719055176, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7334519028663635, + "num_tokens": 504972413.0, + "step": 19515 + }, + { + "epoch": 2.1432022842082143, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.7309250831604004, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7384981513023376, + "num_tokens": 504994458.0, + "step": 19516 + }, + { + "epoch": 2.143312101910828, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3826937675476074, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7394425272941589, + "num_tokens": 505022774.0, + "step": 19517 + }, + { + "epoch": 2.143421919613442, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5221636295318604, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7308404445648193, + "num_tokens": 505046288.0, + "step": 19518 + }, + { + "epoch": 2.143531737316055, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.304338216781616, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7336171865463257, + "num_tokens": 505073053.0, + "step": 19519 + }, + { + "epoch": 2.143641555018669, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.248044967651367, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7292668223381042, + "num_tokens": 505101192.0, + "step": 19520 + }, + { + "epoch": 2.1437513727212827, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.409538507461548, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7267996668815613, + "num_tokens": 505127222.0, + "step": 19521 + }, + { + "epoch": 2.1438611904238964, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.261650323867798, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7024141550064087, + "num_tokens": 505155887.0, + "step": 19522 + }, + { + "epoch": 2.1439710081265098, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2878975868225098, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7249743938446045, + "num_tokens": 505184600.0, + "step": 19523 + }, + { + "epoch": 2.1440808258291235, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5552525520324707, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7216105461120605, + "num_tokens": 505208945.0, + "step": 19524 + }, + { + "epoch": 2.1441906435317373, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.589719772338867, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7573621273040771, + "num_tokens": 505232164.0, + "step": 19525 + }, + { + "epoch": 2.144300461234351, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2531847953796387, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7162142992019653, + "num_tokens": 505260252.0, + "step": 19526 + }, + { + "epoch": 2.144410278936965, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.519487142562866, + "learning_rate": 1e-06, + "loss": 0.7555, + "mean_token_accuracy": 0.7487311363220215, + "num_tokens": 505282177.0, + "step": 19527 + }, + { + "epoch": 2.144520096639578, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5495686531066895, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7350009679794312, + "num_tokens": 505305190.0, + "step": 19528 + }, + { + "epoch": 2.144629914342192, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.497799873352051, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7401052713394165, + "num_tokens": 505329180.0, + "step": 19529 + }, + { + "epoch": 2.1447397320448056, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1521239280700684, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7414836287498474, + "num_tokens": 505358431.0, + "step": 19530 + }, + { + "epoch": 2.1448495497474194, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1367850303649902, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7175610065460205, + "num_tokens": 505388397.0, + "step": 19531 + }, + { + "epoch": 2.144959367450033, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1908535957336426, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7321974039077759, + "num_tokens": 505420726.0, + "step": 19532 + }, + { + "epoch": 2.1450691851526464, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.367830991744995, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7500890493392944, + "num_tokens": 505447288.0, + "step": 19533 + }, + { + "epoch": 2.14517900285526, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.0258798599243164, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7251667976379395, + "num_tokens": 505481821.0, + "step": 19534 + }, + { + "epoch": 2.145288820557874, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.162381887435913, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7269566059112549, + "num_tokens": 505510728.0, + "step": 19535 + }, + { + "epoch": 2.1453986382604877, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.601740598678589, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7178679704666138, + "num_tokens": 505533522.0, + "step": 19536 + }, + { + "epoch": 2.1455084559631015, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2566733360290527, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7135694622993469, + "num_tokens": 505561078.0, + "step": 19537 + }, + { + "epoch": 2.145618273665715, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.0598273277282715, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7096047401428223, + "num_tokens": 505593858.0, + "step": 19538 + }, + { + "epoch": 2.1457280913683285, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.469573736190796, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7268690466880798, + "num_tokens": 505619141.0, + "step": 19539 + }, + { + "epoch": 2.1458379090709423, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.267723798751831, + "learning_rate": 1e-06, + "loss": 0.7809, + "mean_token_accuracy": 0.7533336281776428, + "num_tokens": 505645105.0, + "step": 19540 + }, + { + "epoch": 2.145947726773556, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.254169225692749, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7126233577728271, + "num_tokens": 505675618.0, + "step": 19541 + }, + { + "epoch": 2.1460575444761694, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.415343761444092, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7220776081085205, + "num_tokens": 505701316.0, + "step": 19542 + }, + { + "epoch": 2.146167362178783, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2783360481262207, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7184185981750488, + "num_tokens": 505728786.0, + "step": 19543 + }, + { + "epoch": 2.146277179881397, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3222126960754395, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7464507818222046, + "num_tokens": 505754729.0, + "step": 19544 + }, + { + "epoch": 2.1463869975840106, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.128354072570801, + "learning_rate": 1e-06, + "loss": 0.9427, + "mean_token_accuracy": 0.7103993892669678, + "num_tokens": 505786922.0, + "step": 19545 + }, + { + "epoch": 2.1464968152866244, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.24311900138855, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7292118668556213, + "num_tokens": 505816400.0, + "step": 19546 + }, + { + "epoch": 2.1466066329892377, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3499338626861572, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7640666961669922, + "num_tokens": 505839740.0, + "step": 19547 + }, + { + "epoch": 2.1467164506918515, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2165870666503906, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7385797500610352, + "num_tokens": 505868143.0, + "step": 19548 + }, + { + "epoch": 2.146826268394465, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3792130947113037, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7252118587493896, + "num_tokens": 505894277.0, + "step": 19549 + }, + { + "epoch": 2.146936086097079, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4475646018981934, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7363364696502686, + "num_tokens": 505919504.0, + "step": 19550 + }, + { + "epoch": 2.1470459037996923, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5674331188201904, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7336235642433167, + "num_tokens": 505942117.0, + "step": 19551 + }, + { + "epoch": 2.147155721502306, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.556389570236206, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7262346744537354, + "num_tokens": 505966337.0, + "step": 19552 + }, + { + "epoch": 2.14726553920492, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5119521617889404, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7205901741981506, + "num_tokens": 505991024.0, + "step": 19553 + }, + { + "epoch": 2.1473753569075336, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2035272121429443, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7272331714630127, + "num_tokens": 506018116.0, + "step": 19554 + }, + { + "epoch": 2.1474851746101473, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.366774320602417, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7261471152305603, + "num_tokens": 506042599.0, + "step": 19555 + }, + { + "epoch": 2.1475949923127606, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.224299430847168, + "learning_rate": 1e-06, + "loss": 0.9722, + "mean_token_accuracy": 0.697399914264679, + "num_tokens": 506071190.0, + "step": 19556 + }, + { + "epoch": 2.1477048100153744, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2015440464019775, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7368065118789673, + "num_tokens": 506101070.0, + "step": 19557 + }, + { + "epoch": 2.147814627717988, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4716386795043945, + "learning_rate": 1e-06, + "loss": 0.7462, + "mean_token_accuracy": 0.7555280923843384, + "num_tokens": 506123347.0, + "step": 19558 + }, + { + "epoch": 2.147924445420602, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.566746234893799, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7278062701225281, + "num_tokens": 506146118.0, + "step": 19559 + }, + { + "epoch": 2.1480342631232157, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3392839431762695, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.717984676361084, + "num_tokens": 506174054.0, + "step": 19560 + }, + { + "epoch": 2.148144080825829, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4617221355438232, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7455536127090454, + "num_tokens": 506197243.0, + "step": 19561 + }, + { + "epoch": 2.1482538985284427, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.06815767288208, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7154796719551086, + "num_tokens": 506229973.0, + "step": 19562 + }, + { + "epoch": 2.1483637162310565, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.6839067935943604, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7353165149688721, + "num_tokens": 506250745.0, + "step": 19563 + }, + { + "epoch": 2.1484735339336702, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.095851421356201, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7215301990509033, + "num_tokens": 506282315.0, + "step": 19564 + }, + { + "epoch": 2.148583351636284, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.0665102005004883, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.7197990417480469, + "num_tokens": 506316308.0, + "step": 19565 + }, + { + "epoch": 2.1486931693388973, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.190891742706299, + "learning_rate": 1e-06, + "loss": 0.8971, + "mean_token_accuracy": 0.7221383452415466, + "num_tokens": 506345180.0, + "step": 19566 + }, + { + "epoch": 2.148802987041511, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4332058429718018, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7392396926879883, + "num_tokens": 506369076.0, + "step": 19567 + }, + { + "epoch": 2.148912804744125, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4720969200134277, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7180861234664917, + "num_tokens": 506393479.0, + "step": 19568 + }, + { + "epoch": 2.1490226224467386, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.349644422531128, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7336069941520691, + "num_tokens": 506419932.0, + "step": 19569 + }, + { + "epoch": 2.149132440149352, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2758333683013916, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7244288325309753, + "num_tokens": 506448953.0, + "step": 19570 + }, + { + "epoch": 2.1492422578519657, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.16753888130188, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7309529185295105, + "num_tokens": 506478305.0, + "step": 19571 + }, + { + "epoch": 2.1493520755545794, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.226289749145508, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7333626747131348, + "num_tokens": 506503441.0, + "step": 19572 + }, + { + "epoch": 2.149461893257193, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4884016513824463, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7157412767410278, + "num_tokens": 506527930.0, + "step": 19573 + }, + { + "epoch": 2.149571710959807, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1983413696289062, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.723426103591919, + "num_tokens": 506557797.0, + "step": 19574 + }, + { + "epoch": 2.1496815286624202, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2803616523742676, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.744242787361145, + "num_tokens": 506584377.0, + "step": 19575 + }, + { + "epoch": 2.149791346365034, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 3.0022575855255127, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7357518672943115, + "num_tokens": 506603088.0, + "step": 19576 + }, + { + "epoch": 2.1499011640676478, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.357945680618286, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7144030332565308, + "num_tokens": 506630461.0, + "step": 19577 + }, + { + "epoch": 2.1500109817702615, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3332910537719727, + "learning_rate": 1e-06, + "loss": 0.725, + "mean_token_accuracy": 0.7647568583488464, + "num_tokens": 506654651.0, + "step": 19578 + }, + { + "epoch": 2.150120799472875, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1191837787628174, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7126789093017578, + "num_tokens": 506685387.0, + "step": 19579 + }, + { + "epoch": 2.1502306171754886, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2341880798339844, + "learning_rate": 1e-06, + "loss": 0.7286, + "mean_token_accuracy": 0.757247805595398, + "num_tokens": 506710343.0, + "step": 19580 + }, + { + "epoch": 2.1503404348781023, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2640395164489746, + "learning_rate": 1e-06, + "loss": 0.7535, + "mean_token_accuracy": 0.7550792694091797, + "num_tokens": 506736132.0, + "step": 19581 + }, + { + "epoch": 2.150450252580716, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.630990743637085, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7335590124130249, + "num_tokens": 506757653.0, + "step": 19582 + }, + { + "epoch": 2.15056007028333, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.709453821182251, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7411035895347595, + "num_tokens": 506777449.0, + "step": 19583 + }, + { + "epoch": 2.150669887985943, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.293762683868408, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7137339115142822, + "num_tokens": 506804917.0, + "step": 19584 + }, + { + "epoch": 2.150779705688557, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2658934593200684, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7437891960144043, + "num_tokens": 506831845.0, + "step": 19585 + }, + { + "epoch": 2.1508895233911707, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4673421382904053, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7272178530693054, + "num_tokens": 506858898.0, + "step": 19586 + }, + { + "epoch": 2.1509993410937844, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.650282382965088, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7363979816436768, + "num_tokens": 506881919.0, + "step": 19587 + }, + { + "epoch": 2.151109158796398, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.370161771774292, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7428774833679199, + "num_tokens": 506907384.0, + "step": 19588 + }, + { + "epoch": 2.1512189764990115, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4609713554382324, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7210421562194824, + "num_tokens": 506931758.0, + "step": 19589 + }, + { + "epoch": 2.1513287942016253, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.665531635284424, + "learning_rate": 1e-06, + "loss": 0.7576, + "mean_token_accuracy": 0.7566298246383667, + "num_tokens": 506950804.0, + "step": 19590 + }, + { + "epoch": 2.151438611904239, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5903449058532715, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7376613616943359, + "num_tokens": 506973831.0, + "step": 19591 + }, + { + "epoch": 2.1515484296068528, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2941083908081055, + "learning_rate": 1e-06, + "loss": 0.9411, + "mean_token_accuracy": 0.7098148465156555, + "num_tokens": 507002361.0, + "step": 19592 + }, + { + "epoch": 2.151658247309466, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.187528371810913, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7355628609657288, + "num_tokens": 507031774.0, + "step": 19593 + }, + { + "epoch": 2.15176806501208, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3446872234344482, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7076176404953003, + "num_tokens": 507056790.0, + "step": 19594 + }, + { + "epoch": 2.1518778827146936, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4264445304870605, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7207036018371582, + "num_tokens": 507082694.0, + "step": 19595 + }, + { + "epoch": 2.1519877004173074, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.770716667175293, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.7445003390312195, + "num_tokens": 507102802.0, + "step": 19596 + }, + { + "epoch": 2.152097518119921, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.373762845993042, + "learning_rate": 1e-06, + "loss": 0.7692, + "mean_token_accuracy": 0.7539258003234863, + "num_tokens": 507126400.0, + "step": 19597 + }, + { + "epoch": 2.1522073358225344, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2721104621887207, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7382059097290039, + "num_tokens": 507153282.0, + "step": 19598 + }, + { + "epoch": 2.152317153525148, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.7215592861175537, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7513871192932129, + "num_tokens": 507174702.0, + "step": 19599 + }, + { + "epoch": 2.152426971227762, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2523577213287354, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7195801734924316, + "num_tokens": 507202298.0, + "step": 19600 + }, + { + "epoch": 2.1525367889303757, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.395949602127075, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7166476845741272, + "num_tokens": 507227837.0, + "step": 19601 + }, + { + "epoch": 2.152646606632989, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2296411991119385, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7240525484085083, + "num_tokens": 507255427.0, + "step": 19602 + }, + { + "epoch": 2.1527564243356028, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2313873767852783, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7306318283081055, + "num_tokens": 507282890.0, + "step": 19603 + }, + { + "epoch": 2.1528662420382165, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5435781478881836, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7488481402397156, + "num_tokens": 507306909.0, + "step": 19604 + }, + { + "epoch": 2.1529760597408303, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.009469509124756, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7478511929512024, + "num_tokens": 507338753.0, + "step": 19605 + }, + { + "epoch": 2.153085877443444, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.6040968894958496, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7206152081489563, + "num_tokens": 507360979.0, + "step": 19606 + }, + { + "epoch": 2.1531956951460574, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.1244916915893555, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7342578768730164, + "num_tokens": 507390687.0, + "step": 19607 + }, + { + "epoch": 2.153305512848671, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3494997024536133, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7046390771865845, + "num_tokens": 507415851.0, + "step": 19608 + }, + { + "epoch": 2.153415330551285, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.786700963973999, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7351256608963013, + "num_tokens": 507436547.0, + "step": 19609 + }, + { + "epoch": 2.1535251482538986, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5776145458221436, + "learning_rate": 1e-06, + "loss": 0.7178, + "mean_token_accuracy": 0.7701015472412109, + "num_tokens": 507456399.0, + "step": 19610 + }, + { + "epoch": 2.1536349659565124, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3405489921569824, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7090305685997009, + "num_tokens": 507483644.0, + "step": 19611 + }, + { + "epoch": 2.1537447836591257, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.67618727684021, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7420662641525269, + "num_tokens": 507507416.0, + "step": 19612 + }, + { + "epoch": 2.1538546013617395, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2943801879882812, + "learning_rate": 1e-06, + "loss": 0.9583, + "mean_token_accuracy": 0.6989566683769226, + "num_tokens": 507534168.0, + "step": 19613 + }, + { + "epoch": 2.153964419064353, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2708959579467773, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7364118099212646, + "num_tokens": 507560904.0, + "step": 19614 + }, + { + "epoch": 2.154074236766967, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.6232681274414062, + "learning_rate": 1e-06, + "loss": 0.7673, + "mean_token_accuracy": 0.7541909217834473, + "num_tokens": 507583246.0, + "step": 19615 + }, + { + "epoch": 2.1541840544695807, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.364595651626587, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7440352439880371, + "num_tokens": 507608366.0, + "step": 19616 + }, + { + "epoch": 2.154293872172194, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2599565982818604, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7210732102394104, + "num_tokens": 507637581.0, + "step": 19617 + }, + { + "epoch": 2.154403689874808, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.715481996536255, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7387402653694153, + "num_tokens": 507657453.0, + "step": 19618 + }, + { + "epoch": 2.1545135075774215, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.297870635986328, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7126324772834778, + "num_tokens": 507685092.0, + "step": 19619 + }, + { + "epoch": 2.1546233252800353, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3698511123657227, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7377250790596008, + "num_tokens": 507710248.0, + "step": 19620 + }, + { + "epoch": 2.1547331429826486, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.469041347503662, + "learning_rate": 1e-06, + "loss": 0.7497, + "mean_token_accuracy": 0.7602705955505371, + "num_tokens": 507735881.0, + "step": 19621 + }, + { + "epoch": 2.1548429606852624, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3882954120635986, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7228624820709229, + "num_tokens": 507762092.0, + "step": 19622 + }, + { + "epoch": 2.154952778387876, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2175486087799072, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7163247466087341, + "num_tokens": 507789685.0, + "step": 19623 + }, + { + "epoch": 2.15506259609049, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.4517414569854736, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7142547965049744, + "num_tokens": 507815097.0, + "step": 19624 + }, + { + "epoch": 2.1551724137931036, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3760640621185303, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7372623682022095, + "num_tokens": 507839109.0, + "step": 19625 + }, + { + "epoch": 2.155282231495717, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.43725323677063, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7060756087303162, + "num_tokens": 507864532.0, + "step": 19626 + }, + { + "epoch": 2.1553920491983307, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.732130765914917, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7326637506484985, + "num_tokens": 507886816.0, + "step": 19627 + }, + { + "epoch": 2.1555018669009445, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.384882688522339, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.746606171131134, + "num_tokens": 507910570.0, + "step": 19628 + }, + { + "epoch": 2.1556116846035582, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.904634952545166, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7517702579498291, + "num_tokens": 507930748.0, + "step": 19629 + }, + { + "epoch": 2.1557215023061715, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.534519672393799, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7216304540634155, + "num_tokens": 507954800.0, + "step": 19630 + }, + { + "epoch": 2.1558313200087853, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.8076488971710205, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7406308054924011, + "num_tokens": 507974285.0, + "step": 19631 + }, + { + "epoch": 2.155941137711399, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.248206615447998, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7419684529304504, + "num_tokens": 508001138.0, + "step": 19632 + }, + { + "epoch": 2.156050955414013, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2984678745269775, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.7366135120391846, + "num_tokens": 508027055.0, + "step": 19633 + }, + { + "epoch": 2.1561607731166266, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2385642528533936, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7144260406494141, + "num_tokens": 508055242.0, + "step": 19634 + }, + { + "epoch": 2.15627059081924, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.261570930480957, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7117820978164673, + "num_tokens": 508084001.0, + "step": 19635 + }, + { + "epoch": 2.1563804085218536, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.250002861022949, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7345152497291565, + "num_tokens": 508111763.0, + "step": 19636 + }, + { + "epoch": 2.1564902262244674, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3404016494750977, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7433006763458252, + "num_tokens": 508137475.0, + "step": 19637 + }, + { + "epoch": 2.156600043927081, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2169110774993896, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7281692028045654, + "num_tokens": 508165164.0, + "step": 19638 + }, + { + "epoch": 2.156709861629695, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.893294095993042, + "learning_rate": 1e-06, + "loss": 0.7472, + "mean_token_accuracy": 0.7571507692337036, + "num_tokens": 508183366.0, + "step": 19639 + }, + { + "epoch": 2.1568196793323082, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3412668704986572, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7167437076568604, + "num_tokens": 508210922.0, + "step": 19640 + }, + { + "epoch": 2.156929497034922, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.460418224334717, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7277069091796875, + "num_tokens": 508233612.0, + "step": 19641 + }, + { + "epoch": 2.1570393147375357, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.2059524059295654, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7167258262634277, + "num_tokens": 508262678.0, + "step": 19642 + }, + { + "epoch": 2.1571491324401495, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.292910099029541, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7200007438659668, + "num_tokens": 508290696.0, + "step": 19643 + }, + { + "epoch": 2.157258950142763, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.134871006011963, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7432206869125366, + "num_tokens": 508321748.0, + "step": 19644 + }, + { + "epoch": 2.1573687678453766, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.407686471939087, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7336845397949219, + "num_tokens": 508345438.0, + "step": 19645 + }, + { + "epoch": 2.1574785855479903, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.314924955368042, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.716098427772522, + "num_tokens": 508372099.0, + "step": 19646 + }, + { + "epoch": 2.157588403250604, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.346097707748413, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.734978973865509, + "num_tokens": 508398974.0, + "step": 19647 + }, + { + "epoch": 2.157698220953218, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.250065326690674, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7283544540405273, + "num_tokens": 508428562.0, + "step": 19648 + }, + { + "epoch": 2.157808038655831, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.399547815322876, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7126423120498657, + "num_tokens": 508456487.0, + "step": 19649 + }, + { + "epoch": 2.157917856358445, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.522684335708618, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7405427694320679, + "num_tokens": 508481014.0, + "step": 19650 + }, + { + "epoch": 2.1580276740610587, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.407517433166504, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7349849939346313, + "num_tokens": 508506398.0, + "step": 19651 + }, + { + "epoch": 2.1581374917636724, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.5877814292907715, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7410608530044556, + "num_tokens": 508530440.0, + "step": 19652 + }, + { + "epoch": 2.1582473094662857, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.556199073791504, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7235097289085388, + "num_tokens": 508553552.0, + "step": 19653 + }, + { + "epoch": 2.1583571271688995, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3180336952209473, + "learning_rate": 1e-06, + "loss": 0.7823, + "mean_token_accuracy": 0.747954785823822, + "num_tokens": 508579793.0, + "step": 19654 + }, + { + "epoch": 2.1584669448715132, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.013385534286499, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.7158595323562622, + "num_tokens": 508614825.0, + "step": 19655 + }, + { + "epoch": 2.158576762574127, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.646965265274048, + "learning_rate": 1e-06, + "loss": 0.7816, + "mean_token_accuracy": 0.7478635311126709, + "num_tokens": 508636198.0, + "step": 19656 + }, + { + "epoch": 2.1586865802767408, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.162154197692871, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.724295973777771, + "num_tokens": 508665107.0, + "step": 19657 + }, + { + "epoch": 2.158796397979354, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.1822848320007324, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7327408194541931, + "num_tokens": 508692624.0, + "step": 19658 + }, + { + "epoch": 2.158906215681968, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.429159164428711, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7361552119255066, + "num_tokens": 508716756.0, + "step": 19659 + }, + { + "epoch": 2.1590160333845816, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2854537963867188, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.722836971282959, + "num_tokens": 508745627.0, + "step": 19660 + }, + { + "epoch": 2.1591258510871953, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.519231081008911, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7180262804031372, + "num_tokens": 508770712.0, + "step": 19661 + }, + { + "epoch": 2.159235668789809, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.155700445175171, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7017683386802673, + "num_tokens": 508800724.0, + "step": 19662 + }, + { + "epoch": 2.1593454864924224, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.3084306716918945, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7354006171226501, + "num_tokens": 508826714.0, + "step": 19663 + }, + { + "epoch": 2.159455304195036, + "ewc_loss": 1.233816146850586e-05, + "grad_norm": 2.359827756881714, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7277029752731323, + "num_tokens": 508852337.0, + "step": 19664 + }, + { + "epoch": 2.15956512189765, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.7738921642303467, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7284786701202393, + "num_tokens": 508872900.0, + "step": 19665 + }, + { + "epoch": 2.1596749396002637, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.415276527404785, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.740513801574707, + "num_tokens": 508898461.0, + "step": 19666 + }, + { + "epoch": 2.1597847573028774, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2692794799804688, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.733850359916687, + "num_tokens": 508925970.0, + "step": 19667 + }, + { + "epoch": 2.1598945750054908, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.1728672981262207, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7217211127281189, + "num_tokens": 508954639.0, + "step": 19668 + }, + { + "epoch": 2.1600043927081045, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.0814578533172607, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7164443731307983, + "num_tokens": 508987944.0, + "step": 19669 + }, + { + "epoch": 2.1601142104107183, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.9841182231903076, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7400910258293152, + "num_tokens": 509006625.0, + "step": 19670 + }, + { + "epoch": 2.160224028113332, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.126603364944458, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7239258289337158, + "num_tokens": 509036342.0, + "step": 19671 + }, + { + "epoch": 2.1603338458159453, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2207977771759033, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7173481583595276, + "num_tokens": 509064156.0, + "step": 19672 + }, + { + "epoch": 2.160443663518559, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.420644760131836, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7325277328491211, + "num_tokens": 509088825.0, + "step": 19673 + }, + { + "epoch": 2.160553481221173, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.145820379257202, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7229065895080566, + "num_tokens": 509120868.0, + "step": 19674 + }, + { + "epoch": 2.1606632989237866, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1497042179107666, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7092952728271484, + "num_tokens": 509150995.0, + "step": 19675 + }, + { + "epoch": 2.1607731166264004, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.256995677947998, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7465931177139282, + "num_tokens": 509177385.0, + "step": 19676 + }, + { + "epoch": 2.1608829343290137, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5375709533691406, + "learning_rate": 1e-06, + "loss": 0.9872, + "mean_token_accuracy": 0.6969422101974487, + "num_tokens": 509203424.0, + "step": 19677 + }, + { + "epoch": 2.1609927520316274, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3510401248931885, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7172577381134033, + "num_tokens": 509228899.0, + "step": 19678 + }, + { + "epoch": 2.161102569734241, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3486881256103516, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7399269938468933, + "num_tokens": 509256290.0, + "step": 19679 + }, + { + "epoch": 2.161212387436855, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.249593496322632, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7323139309883118, + "num_tokens": 509283170.0, + "step": 19680 + }, + { + "epoch": 2.1613222051394683, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2095630168914795, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7185097336769104, + "num_tokens": 509312279.0, + "step": 19681 + }, + { + "epoch": 2.161432022842082, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.736788749694824, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7422654628753662, + "num_tokens": 509338350.0, + "step": 19682 + }, + { + "epoch": 2.161541840544696, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.190307140350342, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7391489744186401, + "num_tokens": 509366366.0, + "step": 19683 + }, + { + "epoch": 2.1616516582473095, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.768467426300049, + "learning_rate": 1e-06, + "loss": 0.6795, + "mean_token_accuracy": 0.7787637710571289, + "num_tokens": 509383927.0, + "step": 19684 + }, + { + "epoch": 2.1617614759499233, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.412367582321167, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7171227931976318, + "num_tokens": 509411164.0, + "step": 19685 + }, + { + "epoch": 2.1618712936525366, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.458420515060425, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7199670076370239, + "num_tokens": 509436736.0, + "step": 19686 + }, + { + "epoch": 2.1619811113551504, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4972283840179443, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.708388090133667, + "num_tokens": 509462765.0, + "step": 19687 + }, + { + "epoch": 2.162090929057764, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.555903196334839, + "learning_rate": 1e-06, + "loss": 0.7823, + "mean_token_accuracy": 0.7467337846755981, + "num_tokens": 509485486.0, + "step": 19688 + }, + { + "epoch": 2.162200746760378, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3667099475860596, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7552801370620728, + "num_tokens": 509511127.0, + "step": 19689 + }, + { + "epoch": 2.1623105644629916, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.27905535697937, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7423611283302307, + "num_tokens": 509536626.0, + "step": 19690 + }, + { + "epoch": 2.162420382165605, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3712503910064697, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7528329491615295, + "num_tokens": 509560467.0, + "step": 19691 + }, + { + "epoch": 2.1625301998682187, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4733471870422363, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7144680023193359, + "num_tokens": 509587425.0, + "step": 19692 + }, + { + "epoch": 2.1626400175708325, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.557831287384033, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.7449499368667603, + "num_tokens": 509610781.0, + "step": 19693 + }, + { + "epoch": 2.162749835273446, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.506108522415161, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7299397587776184, + "num_tokens": 509635032.0, + "step": 19694 + }, + { + "epoch": 2.16285965297606, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.23157000541687, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.727409303188324, + "num_tokens": 509663336.0, + "step": 19695 + }, + { + "epoch": 2.1629694706786733, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.688615560531616, + "learning_rate": 1e-06, + "loss": 0.9497, + "mean_token_accuracy": 0.711699366569519, + "num_tokens": 509687814.0, + "step": 19696 + }, + { + "epoch": 2.163079288381287, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.355668783187866, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.71886146068573, + "num_tokens": 509713456.0, + "step": 19697 + }, + { + "epoch": 2.163189106083901, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.456421375274658, + "learning_rate": 1e-06, + "loss": 0.7521, + "mean_token_accuracy": 0.7561302185058594, + "num_tokens": 509735769.0, + "step": 19698 + }, + { + "epoch": 2.1632989237865146, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4856181144714355, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7256031036376953, + "num_tokens": 509761707.0, + "step": 19699 + }, + { + "epoch": 2.163408741489128, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.064237117767334, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7338576316833496, + "num_tokens": 509793233.0, + "step": 19700 + }, + { + "epoch": 2.1635185591917416, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1668193340301514, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7266606092453003, + "num_tokens": 509820851.0, + "step": 19701 + }, + { + "epoch": 2.1636283768943554, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1830708980560303, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7281891107559204, + "num_tokens": 509848138.0, + "step": 19702 + }, + { + "epoch": 2.163738194596969, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3497769832611084, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7243435382843018, + "num_tokens": 509875119.0, + "step": 19703 + }, + { + "epoch": 2.1638480122995825, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.386823892593384, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7318991422653198, + "num_tokens": 509902039.0, + "step": 19704 + }, + { + "epoch": 2.163957830002196, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.342358112335205, + "learning_rate": 1e-06, + "loss": 0.8146, + "mean_token_accuracy": 0.737684965133667, + "num_tokens": 509926147.0, + "step": 19705 + }, + { + "epoch": 2.16406764770481, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.7279083728790283, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7233725786209106, + "num_tokens": 509949592.0, + "step": 19706 + }, + { + "epoch": 2.1641774654074237, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3264272212982178, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7355489730834961, + "num_tokens": 509975177.0, + "step": 19707 + }, + { + "epoch": 2.1642872831100375, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.437177896499634, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7388994693756104, + "num_tokens": 509998908.0, + "step": 19708 + }, + { + "epoch": 2.164397100812651, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4385263919830322, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7385355830192566, + "num_tokens": 510024977.0, + "step": 19709 + }, + { + "epoch": 2.1645069185152646, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2001709938049316, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.737945556640625, + "num_tokens": 510054957.0, + "step": 19710 + }, + { + "epoch": 2.1646167362178783, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3744356632232666, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7364062666893005, + "num_tokens": 510081631.0, + "step": 19711 + }, + { + "epoch": 2.164726553920492, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.377950668334961, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7424989938735962, + "num_tokens": 510107994.0, + "step": 19712 + }, + { + "epoch": 2.164836371623106, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2522456645965576, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.707476794719696, + "num_tokens": 510136020.0, + "step": 19713 + }, + { + "epoch": 2.164946189325719, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.448803663253784, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7398021817207336, + "num_tokens": 510161525.0, + "step": 19714 + }, + { + "epoch": 2.165056007028333, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2633872032165527, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.714104950428009, + "num_tokens": 510188589.0, + "step": 19715 + }, + { + "epoch": 2.1651658247309467, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.1657135486602783, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7347303628921509, + "num_tokens": 510216497.0, + "step": 19716 + }, + { + "epoch": 2.1652756424335604, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.3345844745635986, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.738113284111023, + "num_tokens": 510240815.0, + "step": 19717 + }, + { + "epoch": 2.165385460136174, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4971628189086914, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7330325841903687, + "num_tokens": 510263707.0, + "step": 19718 + }, + { + "epoch": 2.1654952778387875, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.3505990505218506, + "learning_rate": 1e-06, + "loss": 0.7308, + "mean_token_accuracy": 0.761380672454834, + "num_tokens": 510287514.0, + "step": 19719 + }, + { + "epoch": 2.1656050955414012, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.193056583404541, + "learning_rate": 1e-06, + "loss": 0.753, + "mean_token_accuracy": 0.7505729794502258, + "num_tokens": 510315369.0, + "step": 19720 + }, + { + "epoch": 2.165714913244015, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.5052502155303955, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.744003415107727, + "num_tokens": 510339585.0, + "step": 19721 + }, + { + "epoch": 2.1658247309466288, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.323866844177246, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7147699594497681, + "num_tokens": 510367976.0, + "step": 19722 + }, + { + "epoch": 2.165934548649242, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.5844552516937256, + "learning_rate": 1e-06, + "loss": 0.7739, + "mean_token_accuracy": 0.7524571418762207, + "num_tokens": 510389909.0, + "step": 19723 + }, + { + "epoch": 2.166044366351856, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.5376245975494385, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7231262922286987, + "num_tokens": 510412666.0, + "step": 19724 + }, + { + "epoch": 2.1661541840544696, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.814509630203247, + "learning_rate": 1e-06, + "loss": 0.7451, + "mean_token_accuracy": 0.7559886574745178, + "num_tokens": 510431793.0, + "step": 19725 + }, + { + "epoch": 2.1662640017570833, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4467482566833496, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7392054796218872, + "num_tokens": 510457084.0, + "step": 19726 + }, + { + "epoch": 2.166373819459697, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.36279559135437, + "learning_rate": 1e-06, + "loss": 0.8419, + "mean_token_accuracy": 0.7297372817993164, + "num_tokens": 510483651.0, + "step": 19727 + }, + { + "epoch": 2.1664836371623104, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.493892192840576, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7419147491455078, + "num_tokens": 510508228.0, + "step": 19728 + }, + { + "epoch": 2.166593454864924, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2734110355377197, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7184872031211853, + "num_tokens": 510535082.0, + "step": 19729 + }, + { + "epoch": 2.166703272567538, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.285626173019409, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7494732141494751, + "num_tokens": 510561521.0, + "step": 19730 + }, + { + "epoch": 2.1668130902701517, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4627468585968018, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7389521598815918, + "num_tokens": 510586160.0, + "step": 19731 + }, + { + "epoch": 2.166922907972765, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4729418754577637, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7362859845161438, + "num_tokens": 510611053.0, + "step": 19732 + }, + { + "epoch": 2.1670327256753787, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.321492910385132, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7212972640991211, + "num_tokens": 510639136.0, + "step": 19733 + }, + { + "epoch": 2.1671425433779925, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.266777276992798, + "learning_rate": 1e-06, + "loss": 0.9804, + "mean_token_accuracy": 0.6993241310119629, + "num_tokens": 510667618.0, + "step": 19734 + }, + { + "epoch": 2.1672523610806063, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.5593388080596924, + "learning_rate": 1e-06, + "loss": 0.7901, + "mean_token_accuracy": 0.7544536590576172, + "num_tokens": 510690341.0, + "step": 19735 + }, + { + "epoch": 2.16736217878322, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.205186605453491, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7344216108322144, + "num_tokens": 510717531.0, + "step": 19736 + }, + { + "epoch": 2.1674719964858333, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.0294151306152344, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7017594575881958, + "num_tokens": 510750234.0, + "step": 19737 + }, + { + "epoch": 2.167581814188447, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2126731872558594, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7293096780776978, + "num_tokens": 510779786.0, + "step": 19738 + }, + { + "epoch": 2.167691631891061, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.5445444583892822, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7394282817840576, + "num_tokens": 510802666.0, + "step": 19739 + }, + { + "epoch": 2.1678014495936746, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4528377056121826, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7334554195404053, + "num_tokens": 510825954.0, + "step": 19740 + }, + { + "epoch": 2.1679112672962884, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.6471378803253174, + "learning_rate": 1e-06, + "loss": 0.75, + "mean_token_accuracy": 0.7601286172866821, + "num_tokens": 510847338.0, + "step": 19741 + }, + { + "epoch": 2.1680210849989017, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4235923290252686, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7193751335144043, + "num_tokens": 510873730.0, + "step": 19742 + }, + { + "epoch": 2.1681309027015154, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.2534892559051514, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7159127593040466, + "num_tokens": 510903789.0, + "step": 19743 + }, + { + "epoch": 2.168240720404129, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.283092498779297, + "learning_rate": 1e-06, + "loss": 0.6957, + "mean_token_accuracy": 0.7762197256088257, + "num_tokens": 510929730.0, + "step": 19744 + }, + { + "epoch": 2.168350538106743, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.4027175903320312, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.74518883228302, + "num_tokens": 510955077.0, + "step": 19745 + }, + { + "epoch": 2.1684603558093567, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.39544939994812, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7472183704376221, + "num_tokens": 510979271.0, + "step": 19746 + }, + { + "epoch": 2.16857017351197, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.02482271194458, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7137342095375061, + "num_tokens": 511012254.0, + "step": 19747 + }, + { + "epoch": 2.1686799912145838, + "ewc_loss": 1.239776611328125e-05, + "grad_norm": 2.689145803451538, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7432141304016113, + "num_tokens": 511033036.0, + "step": 19748 + }, + { + "epoch": 2.1687898089171975, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3840837478637695, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7487595081329346, + "num_tokens": 511058387.0, + "step": 19749 + }, + { + "epoch": 2.1688996266198113, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.228912115097046, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7332241535186768, + "num_tokens": 511084828.0, + "step": 19750 + }, + { + "epoch": 2.1690094443224246, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.306689739227295, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7316832542419434, + "num_tokens": 511112041.0, + "step": 19751 + }, + { + "epoch": 2.1691192620250384, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1067891120910645, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.719127893447876, + "num_tokens": 511143885.0, + "step": 19752 + }, + { + "epoch": 2.169229079727652, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4722445011138916, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7253811955451965, + "num_tokens": 511168625.0, + "step": 19753 + }, + { + "epoch": 2.169338897430266, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.318751096725464, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7314685583114624, + "num_tokens": 511194843.0, + "step": 19754 + }, + { + "epoch": 2.1694487151328796, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.6281232833862305, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7253192067146301, + "num_tokens": 511216030.0, + "step": 19755 + }, + { + "epoch": 2.169558532835493, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.421933650970459, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7193615436553955, + "num_tokens": 511241673.0, + "step": 19756 + }, + { + "epoch": 2.1696683505381067, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.416901111602783, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7273628115653992, + "num_tokens": 511265301.0, + "step": 19757 + }, + { + "epoch": 2.1697781682407205, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.7396280765533447, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7397988438606262, + "num_tokens": 511284891.0, + "step": 19758 + }, + { + "epoch": 2.169887985943334, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3688080310821533, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7190201282501221, + "num_tokens": 511310615.0, + "step": 19759 + }, + { + "epoch": 2.1699978036459475, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.113145589828491, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7401366829872131, + "num_tokens": 511339053.0, + "step": 19760 + }, + { + "epoch": 2.1701076213485613, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.0271949768066406, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7119330167770386, + "num_tokens": 511371657.0, + "step": 19761 + }, + { + "epoch": 2.170217439051175, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2569143772125244, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.747912585735321, + "num_tokens": 511400108.0, + "step": 19762 + }, + { + "epoch": 2.170327256753789, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4453632831573486, + "learning_rate": 1e-06, + "loss": 0.7311, + "mean_token_accuracy": 0.7595863938331604, + "num_tokens": 511422134.0, + "step": 19763 + }, + { + "epoch": 2.1704370744564025, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.135958194732666, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.733029305934906, + "num_tokens": 511452620.0, + "step": 19764 + }, + { + "epoch": 2.170546892159016, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.481882095336914, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7320704460144043, + "num_tokens": 511477119.0, + "step": 19765 + }, + { + "epoch": 2.1706567098616296, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1748557090759277, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7258957028388977, + "num_tokens": 511506474.0, + "step": 19766 + }, + { + "epoch": 2.1707665275642434, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4951372146606445, + "learning_rate": 1e-06, + "loss": 0.7633, + "mean_token_accuracy": 0.7566565275192261, + "num_tokens": 511529501.0, + "step": 19767 + }, + { + "epoch": 2.170876345266857, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2729525566101074, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7404628396034241, + "num_tokens": 511558257.0, + "step": 19768 + }, + { + "epoch": 2.170986162969471, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3941595554351807, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7287055850028992, + "num_tokens": 511583088.0, + "step": 19769 + }, + { + "epoch": 2.171095980672084, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.171455144882202, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.725658655166626, + "num_tokens": 511612614.0, + "step": 19770 + }, + { + "epoch": 2.171205798374698, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.6084446907043457, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7429854869842529, + "num_tokens": 511633909.0, + "step": 19771 + }, + { + "epoch": 2.1713156160773117, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2995386123657227, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.725825846195221, + "num_tokens": 511659968.0, + "step": 19772 + }, + { + "epoch": 2.1714254337799255, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5513734817504883, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7429977655410767, + "num_tokens": 511682095.0, + "step": 19773 + }, + { + "epoch": 2.171535251482539, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2405786514282227, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.721182107925415, + "num_tokens": 511710990.0, + "step": 19774 + }, + { + "epoch": 2.1716450691851525, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.31139874458313, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7472017407417297, + "num_tokens": 511739624.0, + "step": 19775 + }, + { + "epoch": 2.1717548868877663, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4007794857025146, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7132720947265625, + "num_tokens": 511764140.0, + "step": 19776 + }, + { + "epoch": 2.17186470459038, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.291825771331787, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7395379543304443, + "num_tokens": 511791977.0, + "step": 19777 + }, + { + "epoch": 2.171974522292994, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4924874305725098, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7354380488395691, + "num_tokens": 511815727.0, + "step": 19778 + }, + { + "epoch": 2.172084339995607, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2653439044952393, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7163255214691162, + "num_tokens": 511844926.0, + "step": 19779 + }, + { + "epoch": 2.172194157698221, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.083287477493286, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7444616556167603, + "num_tokens": 511877010.0, + "step": 19780 + }, + { + "epoch": 2.1723039754008346, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5913193225860596, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7352012395858765, + "num_tokens": 511897821.0, + "step": 19781 + }, + { + "epoch": 2.1724137931034484, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3337864875793457, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7245171070098877, + "num_tokens": 511925977.0, + "step": 19782 + }, + { + "epoch": 2.1725236108060617, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.411942720413208, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7382224798202515, + "num_tokens": 511949715.0, + "step": 19783 + }, + { + "epoch": 2.1726334285086755, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3293895721435547, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7534216642379761, + "num_tokens": 511977205.0, + "step": 19784 + }, + { + "epoch": 2.1727432462112892, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.366659164428711, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7354608178138733, + "num_tokens": 512002980.0, + "step": 19785 + }, + { + "epoch": 2.172853063913903, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5748531818389893, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7483679056167603, + "num_tokens": 512023265.0, + "step": 19786 + }, + { + "epoch": 2.1729628816165167, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2161083221435547, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7317078113555908, + "num_tokens": 512049915.0, + "step": 19787 + }, + { + "epoch": 2.17307269931913, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.155515432357788, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.732564389705658, + "num_tokens": 512078755.0, + "step": 19788 + }, + { + "epoch": 2.173182517021744, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.070260524749756, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7423496246337891, + "num_tokens": 512109347.0, + "step": 19789 + }, + { + "epoch": 2.1732923347243576, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2020180225372314, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.737912654876709, + "num_tokens": 512137702.0, + "step": 19790 + }, + { + "epoch": 2.1734021524269713, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3320722579956055, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7342313528060913, + "num_tokens": 512163873.0, + "step": 19791 + }, + { + "epoch": 2.173511970129585, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.505610466003418, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7354862689971924, + "num_tokens": 512187004.0, + "step": 19792 + }, + { + "epoch": 2.1736217878321984, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.854640245437622, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.748166561126709, + "num_tokens": 512206001.0, + "step": 19793 + }, + { + "epoch": 2.173731605534812, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.071594715118408, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.716304361820221, + "num_tokens": 512239300.0, + "step": 19794 + }, + { + "epoch": 2.173841423237426, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.182904005050659, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7395429015159607, + "num_tokens": 512267649.0, + "step": 19795 + }, + { + "epoch": 2.1739512409400397, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.580531358718872, + "learning_rate": 1e-06, + "loss": 0.7208, + "mean_token_accuracy": 0.767496645450592, + "num_tokens": 512287428.0, + "step": 19796 + }, + { + "epoch": 2.1740610586426534, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.382199764251709, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7382762432098389, + "num_tokens": 512312435.0, + "step": 19797 + }, + { + "epoch": 2.1741708763452667, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3655669689178467, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7461597919464111, + "num_tokens": 512336657.0, + "step": 19798 + }, + { + "epoch": 2.1742806940478805, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.386594295501709, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.71596759557724, + "num_tokens": 512363372.0, + "step": 19799 + }, + { + "epoch": 2.1743905117504942, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.380807876586914, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.7290085554122925, + "num_tokens": 512391054.0, + "step": 19800 + }, + { + "epoch": 2.174500329453108, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.6363682746887207, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.74190753698349, + "num_tokens": 512413378.0, + "step": 19801 + }, + { + "epoch": 2.1746101471557213, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.523364782333374, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7377781271934509, + "num_tokens": 512436309.0, + "step": 19802 + }, + { + "epoch": 2.174719964858335, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5546069145202637, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7382016181945801, + "num_tokens": 512458612.0, + "step": 19803 + }, + { + "epoch": 2.174829782560949, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.476017475128174, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7177103757858276, + "num_tokens": 512483455.0, + "step": 19804 + }, + { + "epoch": 2.1749396002635626, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3395776748657227, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7310603857040405, + "num_tokens": 512509908.0, + "step": 19805 + }, + { + "epoch": 2.1750494179661763, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4837303161621094, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7248358726501465, + "num_tokens": 512534717.0, + "step": 19806 + }, + { + "epoch": 2.1751592356687897, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5394861698150635, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7444626092910767, + "num_tokens": 512557093.0, + "step": 19807 + }, + { + "epoch": 2.1752690533714034, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.467984437942505, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7152663469314575, + "num_tokens": 512584806.0, + "step": 19808 + }, + { + "epoch": 2.175378871074017, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3348605632781982, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7124615907669067, + "num_tokens": 512610521.0, + "step": 19809 + }, + { + "epoch": 2.175488688776631, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.129119396209717, + "learning_rate": 1e-06, + "loss": 0.7586, + "mean_token_accuracy": 0.7543218731880188, + "num_tokens": 512638067.0, + "step": 19810 + }, + { + "epoch": 2.1755985064792442, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.604985237121582, + "learning_rate": 1e-06, + "loss": 0.7133, + "mean_token_accuracy": 0.764077365398407, + "num_tokens": 512657400.0, + "step": 19811 + }, + { + "epoch": 2.175708324181858, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3779399394989014, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7416220903396606, + "num_tokens": 512683831.0, + "step": 19812 + }, + { + "epoch": 2.1758181418844718, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3198347091674805, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7308143377304077, + "num_tokens": 512711437.0, + "step": 19813 + }, + { + "epoch": 2.1759279595870855, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.207308292388916, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7205380201339722, + "num_tokens": 512739807.0, + "step": 19814 + }, + { + "epoch": 2.1760377772896993, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.044118642807007, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7424443960189819, + "num_tokens": 512772090.0, + "step": 19815 + }, + { + "epoch": 2.1761475949923126, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5160202980041504, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7246306538581848, + "num_tokens": 512796315.0, + "step": 19816 + }, + { + "epoch": 2.1762574126949263, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2595276832580566, + "learning_rate": 1e-06, + "loss": 0.9273, + "mean_token_accuracy": 0.7158262729644775, + "num_tokens": 512823980.0, + "step": 19817 + }, + { + "epoch": 2.17636723039754, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.115396499633789, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7449964284896851, + "num_tokens": 512853809.0, + "step": 19818 + }, + { + "epoch": 2.176477048100154, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4929568767547607, + "learning_rate": 1e-06, + "loss": 0.8062, + "mean_token_accuracy": 0.744356632232666, + "num_tokens": 512877992.0, + "step": 19819 + }, + { + "epoch": 2.1765868658027676, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5903241634368896, + "learning_rate": 1e-06, + "loss": 0.7824, + "mean_token_accuracy": 0.746641218662262, + "num_tokens": 512900783.0, + "step": 19820 + }, + { + "epoch": 2.176696683505381, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.5612902641296387, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7284919023513794, + "num_tokens": 512926616.0, + "step": 19821 + }, + { + "epoch": 2.1768065012079947, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4421865940093994, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.754702627658844, + "num_tokens": 512950669.0, + "step": 19822 + }, + { + "epoch": 2.1769163189106084, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4209110736846924, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7221884727478027, + "num_tokens": 512975107.0, + "step": 19823 + }, + { + "epoch": 2.177026136613222, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2863776683807373, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7472855448722839, + "num_tokens": 513001713.0, + "step": 19824 + }, + { + "epoch": 2.1771359543158355, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.608997106552124, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7151215672492981, + "num_tokens": 513026129.0, + "step": 19825 + }, + { + "epoch": 2.1772457720184493, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4434027671813965, + "learning_rate": 1e-06, + "loss": 0.7441, + "mean_token_accuracy": 0.7569121718406677, + "num_tokens": 513049616.0, + "step": 19826 + }, + { + "epoch": 2.177355589721063, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.423536539077759, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7244340181350708, + "num_tokens": 513076361.0, + "step": 19827 + }, + { + "epoch": 2.177465407423677, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3555233478546143, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7280932664871216, + "num_tokens": 513103893.0, + "step": 19828 + }, + { + "epoch": 2.1775752251262905, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.6300277709960938, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7308073043823242, + "num_tokens": 513125256.0, + "step": 19829 + }, + { + "epoch": 2.177685042828904, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2823121547698975, + "learning_rate": 1e-06, + "loss": 0.9441, + "mean_token_accuracy": 0.7050861120223999, + "num_tokens": 513152505.0, + "step": 19830 + }, + { + "epoch": 2.1777948605315176, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.036952018737793, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7415262460708618, + "num_tokens": 513182326.0, + "step": 19831 + }, + { + "epoch": 2.1779046782341314, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.451096534729004, + "learning_rate": 1e-06, + "loss": 0.7706, + "mean_token_accuracy": 0.7616285085678101, + "num_tokens": 513204562.0, + "step": 19832 + }, + { + "epoch": 2.178014495936745, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.155040979385376, + "learning_rate": 1e-06, + "loss": 0.7521, + "mean_token_accuracy": 0.7593729496002197, + "num_tokens": 513231469.0, + "step": 19833 + }, + { + "epoch": 2.1781243136393584, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1323187351226807, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.73891282081604, + "num_tokens": 513260543.0, + "step": 19834 + }, + { + "epoch": 2.178234131341972, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.542215347290039, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7540874481201172, + "num_tokens": 513283046.0, + "step": 19835 + }, + { + "epoch": 2.178343949044586, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.412637233734131, + "learning_rate": 1e-06, + "loss": 0.9851, + "mean_token_accuracy": 0.6954197287559509, + "num_tokens": 513309369.0, + "step": 19836 + }, + { + "epoch": 2.1784537667471997, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3507792949676514, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.6986674070358276, + "num_tokens": 513337836.0, + "step": 19837 + }, + { + "epoch": 2.1785635844498135, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.0597105026245117, + "learning_rate": 1e-06, + "loss": 0.9934, + "mean_token_accuracy": 0.6922898888587952, + "num_tokens": 513370115.0, + "step": 19838 + }, + { + "epoch": 2.1786734021524268, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2969627380371094, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.722064733505249, + "num_tokens": 513398324.0, + "step": 19839 + }, + { + "epoch": 2.1787832198550405, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.268857479095459, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7322604656219482, + "num_tokens": 513427101.0, + "step": 19840 + }, + { + "epoch": 2.1788930375576543, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3775510787963867, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7514834403991699, + "num_tokens": 513450935.0, + "step": 19841 + }, + { + "epoch": 2.179002855260268, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.3899948596954346, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7283210754394531, + "num_tokens": 513476918.0, + "step": 19842 + }, + { + "epoch": 2.179112672962882, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.1870205402374268, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7303334474563599, + "num_tokens": 513508803.0, + "step": 19843 + }, + { + "epoch": 2.179222490665495, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4734690189361572, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7299654483795166, + "num_tokens": 513533374.0, + "step": 19844 + }, + { + "epoch": 2.179332308368109, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.486135959625244, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7446470260620117, + "num_tokens": 513557068.0, + "step": 19845 + }, + { + "epoch": 2.1794421260707226, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.4981753826141357, + "learning_rate": 1e-06, + "loss": 0.7952, + "mean_token_accuracy": 0.7412198781967163, + "num_tokens": 513579110.0, + "step": 19846 + }, + { + "epoch": 2.1795519437733364, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.316512107849121, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7275242209434509, + "num_tokens": 513604626.0, + "step": 19847 + }, + { + "epoch": 2.17966176147595, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.303826332092285, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.718335747718811, + "num_tokens": 513630768.0, + "step": 19848 + }, + { + "epoch": 2.1797715791785635, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.26446795463562, + "learning_rate": 1e-06, + "loss": 0.789, + "mean_token_accuracy": 0.7457725405693054, + "num_tokens": 513656694.0, + "step": 19849 + }, + { + "epoch": 2.179881396881177, + "ewc_loss": 1.245737075805664e-05, + "grad_norm": 2.2760984897613525, + "learning_rate": 1e-06, + "loss": 0.9226, + "mean_token_accuracy": 0.7189116477966309, + "num_tokens": 513685769.0, + "step": 19850 + }, + { + "epoch": 2.179991214583791, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.055900812149048, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7064113616943359, + "num_tokens": 513719014.0, + "step": 19851 + }, + { + "epoch": 2.1801010322864047, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.5073297023773193, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7233863472938538, + "num_tokens": 513745490.0, + "step": 19852 + }, + { + "epoch": 2.180210849989018, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.238762140274048, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7211079597473145, + "num_tokens": 513774431.0, + "step": 19853 + }, + { + "epoch": 2.180320667691632, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3424715995788574, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7323073148727417, + "num_tokens": 513799223.0, + "step": 19854 + }, + { + "epoch": 2.1804304853942456, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2502591609954834, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7239160537719727, + "num_tokens": 513826121.0, + "step": 19855 + }, + { + "epoch": 2.1805403030968593, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.063551425933838, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7211794853210449, + "num_tokens": 513857479.0, + "step": 19856 + }, + { + "epoch": 2.180650120799473, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.6603610515594482, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7220005989074707, + "num_tokens": 513878498.0, + "step": 19857 + }, + { + "epoch": 2.1807599385020864, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.758180856704712, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.735750138759613, + "num_tokens": 513899898.0, + "step": 19858 + }, + { + "epoch": 2.1808697562047, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1486809253692627, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7210039496421814, + "num_tokens": 513929478.0, + "step": 19859 + }, + { + "epoch": 2.180979573907314, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2330775260925293, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7265907526016235, + "num_tokens": 513956905.0, + "step": 19860 + }, + { + "epoch": 2.1810893916099277, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.186678409576416, + "learning_rate": 1e-06, + "loss": 0.7742, + "mean_token_accuracy": 0.7448464632034302, + "num_tokens": 513984325.0, + "step": 19861 + }, + { + "epoch": 2.181199209312541, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1855480670928955, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7091546058654785, + "num_tokens": 514013885.0, + "step": 19862 + }, + { + "epoch": 2.1813090270151547, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.054783582687378, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7235945463180542, + "num_tokens": 514045712.0, + "step": 19863 + }, + { + "epoch": 2.1814188447177685, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.6714975833892822, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7372466325759888, + "num_tokens": 514067777.0, + "step": 19864 + }, + { + "epoch": 2.1815286624203822, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 3.0561723709106445, + "learning_rate": 1e-06, + "loss": 0.735, + "mean_token_accuracy": 0.7579784989356995, + "num_tokens": 514084239.0, + "step": 19865 + }, + { + "epoch": 2.181638480122996, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.351463556289673, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7142013311386108, + "num_tokens": 514111071.0, + "step": 19866 + }, + { + "epoch": 2.1817482978256093, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.263558864593506, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7465872764587402, + "num_tokens": 514139515.0, + "step": 19867 + }, + { + "epoch": 2.181858115528223, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3361778259277344, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7175511121749878, + "num_tokens": 514166181.0, + "step": 19868 + }, + { + "epoch": 2.181967933230837, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.18803334236145, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7439451217651367, + "num_tokens": 514194972.0, + "step": 19869 + }, + { + "epoch": 2.1820777509334506, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.227041006088257, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7205073833465576, + "num_tokens": 514224534.0, + "step": 19870 + }, + { + "epoch": 2.1821875686360643, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.077864408493042, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7432906031608582, + "num_tokens": 514256824.0, + "step": 19871 + }, + { + "epoch": 2.1822973863386776, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1787517070770264, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7382469177246094, + "num_tokens": 514284891.0, + "step": 19872 + }, + { + "epoch": 2.1824072040412914, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.6195387840270996, + "learning_rate": 1e-06, + "loss": 0.7829, + "mean_token_accuracy": 0.7471281290054321, + "num_tokens": 514306297.0, + "step": 19873 + }, + { + "epoch": 2.182517021743905, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3734352588653564, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7408546209335327, + "num_tokens": 514331575.0, + "step": 19874 + }, + { + "epoch": 2.182626839446519, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.4685823917388916, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7314510345458984, + "num_tokens": 514355046.0, + "step": 19875 + }, + { + "epoch": 2.1827366571491327, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3386213779449463, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7314095497131348, + "num_tokens": 514381567.0, + "step": 19876 + }, + { + "epoch": 2.182846474851746, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.443654775619507, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7418623566627502, + "num_tokens": 514404830.0, + "step": 19877 + }, + { + "epoch": 2.1829562925543597, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1759681701660156, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7376400232315063, + "num_tokens": 514434722.0, + "step": 19878 + }, + { + "epoch": 2.1830661102569735, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.569822311401367, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7227604389190674, + "num_tokens": 514457611.0, + "step": 19879 + }, + { + "epoch": 2.1831759279595873, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.8435258865356445, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7479300498962402, + "num_tokens": 514477830.0, + "step": 19880 + }, + { + "epoch": 2.1832857456622006, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2589924335479736, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7427152991294861, + "num_tokens": 514503871.0, + "step": 19881 + }, + { + "epoch": 2.1833955633648143, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.641892671585083, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7318611145019531, + "num_tokens": 514525892.0, + "step": 19882 + }, + { + "epoch": 2.183505381067428, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.1047658920288086, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.720786452293396, + "num_tokens": 514557113.0, + "step": 19883 + }, + { + "epoch": 2.183615198770042, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.126498222351074, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.736469566822052, + "num_tokens": 514588219.0, + "step": 19884 + }, + { + "epoch": 2.1837250164726556, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.226419687271118, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.740821123123169, + "num_tokens": 514615084.0, + "step": 19885 + }, + { + "epoch": 2.183834834175269, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.331798553466797, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7358057498931885, + "num_tokens": 514639856.0, + "step": 19886 + }, + { + "epoch": 2.1839446518778827, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.9080495834350586, + "learning_rate": 1e-06, + "loss": 0.7936, + "mean_token_accuracy": 0.7472293972969055, + "num_tokens": 514659269.0, + "step": 19887 + }, + { + "epoch": 2.1840544695804964, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.5030908584594727, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7312668561935425, + "num_tokens": 514682620.0, + "step": 19888 + }, + { + "epoch": 2.18416428728311, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.6719870567321777, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7214788794517517, + "num_tokens": 514705976.0, + "step": 19889 + }, + { + "epoch": 2.1842741049857235, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.503138303756714, + "learning_rate": 1e-06, + "loss": 0.7479, + "mean_token_accuracy": 0.7558602094650269, + "num_tokens": 514727483.0, + "step": 19890 + }, + { + "epoch": 2.1843839226883373, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.525365114212036, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7291964292526245, + "num_tokens": 514750353.0, + "step": 19891 + }, + { + "epoch": 2.184493740390951, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2309112548828125, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7218596935272217, + "num_tokens": 514776854.0, + "step": 19892 + }, + { + "epoch": 2.1846035580935648, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.0603315830230713, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7327207326889038, + "num_tokens": 514809145.0, + "step": 19893 + }, + { + "epoch": 2.1847133757961785, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.19512677192688, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7302426099777222, + "num_tokens": 514837614.0, + "step": 19894 + }, + { + "epoch": 2.184823193498792, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.4123947620391846, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7027996778488159, + "num_tokens": 514863122.0, + "step": 19895 + }, + { + "epoch": 2.1849330112014056, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.543466329574585, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7466821074485779, + "num_tokens": 514884835.0, + "step": 19896 + }, + { + "epoch": 2.1850428289040194, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.366316318511963, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7380080223083496, + "num_tokens": 514910027.0, + "step": 19897 + }, + { + "epoch": 2.185152646606633, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.350299596786499, + "learning_rate": 1e-06, + "loss": 0.7858, + "mean_token_accuracy": 0.7586883306503296, + "num_tokens": 514934175.0, + "step": 19898 + }, + { + "epoch": 2.185262464309247, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.612093687057495, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7273338437080383, + "num_tokens": 514956012.0, + "step": 19899 + }, + { + "epoch": 2.18537228201186, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.303272247314453, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7349156141281128, + "num_tokens": 514982460.0, + "step": 19900 + }, + { + "epoch": 2.185482099714474, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.547257900238037, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7498534321784973, + "num_tokens": 515005156.0, + "step": 19901 + }, + { + "epoch": 2.1855919174170877, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.4455854892730713, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.725433349609375, + "num_tokens": 515030699.0, + "step": 19902 + }, + { + "epoch": 2.1857017351197014, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3393373489379883, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7313783168792725, + "num_tokens": 515055876.0, + "step": 19903 + }, + { + "epoch": 2.1858115528223148, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.1688990592956543, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7154239416122437, + "num_tokens": 515086888.0, + "step": 19904 + }, + { + "epoch": 2.1859213705249285, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.2560083866119385, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7393883466720581, + "num_tokens": 515115200.0, + "step": 19905 + }, + { + "epoch": 2.1860311882275423, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.3218274116516113, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7323592305183411, + "num_tokens": 515141586.0, + "step": 19906 + }, + { + "epoch": 2.186141005930156, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3344764709472656, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7336575984954834, + "num_tokens": 515167298.0, + "step": 19907 + }, + { + "epoch": 2.18625082363277, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.5297322273254395, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7364134788513184, + "num_tokens": 515190325.0, + "step": 19908 + }, + { + "epoch": 2.186360641335383, + "ewc_loss": 1.2576580047607422e-05, + "grad_norm": 2.4021248817443848, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.7458711862564087, + "num_tokens": 515215158.0, + "step": 19909 + }, + { + "epoch": 2.186470459037997, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.1641998291015625, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7442373633384705, + "num_tokens": 515244882.0, + "step": 19910 + }, + { + "epoch": 2.1865802767406106, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.6624724864959717, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7393125295639038, + "num_tokens": 515265951.0, + "step": 19911 + }, + { + "epoch": 2.1866900944432244, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.621129274368286, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7264835834503174, + "num_tokens": 515287808.0, + "step": 19912 + }, + { + "epoch": 2.1867999121458377, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.488868474960327, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7376312613487244, + "num_tokens": 515312532.0, + "step": 19913 + }, + { + "epoch": 2.1869097298484514, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 3.0080652236938477, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.749543309211731, + "num_tokens": 515330522.0, + "step": 19914 + }, + { + "epoch": 2.187019547551065, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.6141774654388428, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7089241147041321, + "num_tokens": 515353629.0, + "step": 19915 + }, + { + "epoch": 2.187129365253679, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.549215078353882, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7396510243415833, + "num_tokens": 515377622.0, + "step": 19916 + }, + { + "epoch": 2.1872391829562927, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.484840154647827, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7269906401634216, + "num_tokens": 515402491.0, + "step": 19917 + }, + { + "epoch": 2.187349000658906, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.474381923675537, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7536128759384155, + "num_tokens": 515427669.0, + "step": 19918 + }, + { + "epoch": 2.18745881836152, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.280848741531372, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7047505378723145, + "num_tokens": 515455641.0, + "step": 19919 + }, + { + "epoch": 2.1875686360641335, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.442554473876953, + "learning_rate": 1e-06, + "loss": 0.7906, + "mean_token_accuracy": 0.7451807260513306, + "num_tokens": 515480433.0, + "step": 19920 + }, + { + "epoch": 2.1876784537667473, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.314608097076416, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7323395609855652, + "num_tokens": 515506270.0, + "step": 19921 + }, + { + "epoch": 2.187788271469361, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4297640323638916, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7341271638870239, + "num_tokens": 515532306.0, + "step": 19922 + }, + { + "epoch": 2.1878980891719744, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.363905429840088, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7193139791488647, + "num_tokens": 515556984.0, + "step": 19923 + }, + { + "epoch": 2.188007906874588, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.474168062210083, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7316430807113647, + "num_tokens": 515581961.0, + "step": 19924 + }, + { + "epoch": 2.188117724577202, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.8936703205108643, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7399625778198242, + "num_tokens": 515599473.0, + "step": 19925 + }, + { + "epoch": 2.1882275422798156, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3643975257873535, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7159678936004639, + "num_tokens": 515623550.0, + "step": 19926 + }, + { + "epoch": 2.1883373599824294, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.408079147338867, + "learning_rate": 1e-06, + "loss": 0.7103, + "mean_token_accuracy": 0.7629085779190063, + "num_tokens": 515647512.0, + "step": 19927 + }, + { + "epoch": 2.1884471776850427, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.774435043334961, + "learning_rate": 1e-06, + "loss": 0.753, + "mean_token_accuracy": 0.7613952159881592, + "num_tokens": 515666805.0, + "step": 19928 + }, + { + "epoch": 2.1885569953876565, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5325419902801514, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7434681057929993, + "num_tokens": 515691097.0, + "step": 19929 + }, + { + "epoch": 2.1886668130902702, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.152381658554077, + "learning_rate": 1e-06, + "loss": 0.9828, + "mean_token_accuracy": 0.7026693820953369, + "num_tokens": 515721818.0, + "step": 19930 + }, + { + "epoch": 2.188776630792884, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.207947015762329, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7109636664390564, + "num_tokens": 515754030.0, + "step": 19931 + }, + { + "epoch": 2.1888864484954973, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.334534168243408, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7337642312049866, + "num_tokens": 515780601.0, + "step": 19932 + }, + { + "epoch": 2.188996266198111, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.417358160018921, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7355833649635315, + "num_tokens": 515803712.0, + "step": 19933 + }, + { + "epoch": 2.189106083900725, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.436518430709839, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.753423810005188, + "num_tokens": 515826421.0, + "step": 19934 + }, + { + "epoch": 2.1892159016033386, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.400269031524658, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7144193053245544, + "num_tokens": 515852185.0, + "step": 19935 + }, + { + "epoch": 2.1893257193059523, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3179609775543213, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7219638824462891, + "num_tokens": 515880345.0, + "step": 19936 + }, + { + "epoch": 2.1894355370085656, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.196049213409424, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7282676100730896, + "num_tokens": 515909406.0, + "step": 19937 + }, + { + "epoch": 2.1895453547111794, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.159445285797119, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7357612252235413, + "num_tokens": 515938440.0, + "step": 19938 + }, + { + "epoch": 2.189655172413793, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5722949504852295, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7325780391693115, + "num_tokens": 515960058.0, + "step": 19939 + }, + { + "epoch": 2.189764990116407, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.721881628036499, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7374438047409058, + "num_tokens": 515980139.0, + "step": 19940 + }, + { + "epoch": 2.18987480781902, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.6079986095428467, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.73570317029953, + "num_tokens": 516000716.0, + "step": 19941 + }, + { + "epoch": 2.189984625521634, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.98452091217041, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.74395352602005, + "num_tokens": 516019694.0, + "step": 19942 + }, + { + "epoch": 2.1900944432242477, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2213194370269775, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7152647972106934, + "num_tokens": 516049134.0, + "step": 19943 + }, + { + "epoch": 2.1902042609268615, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2836947441101074, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7223738431930542, + "num_tokens": 516076668.0, + "step": 19944 + }, + { + "epoch": 2.1903140786294752, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4297304153442383, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7367098927497864, + "num_tokens": 516101344.0, + "step": 19945 + }, + { + "epoch": 2.1904238963320886, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3917882442474365, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7169589400291443, + "num_tokens": 516127699.0, + "step": 19946 + }, + { + "epoch": 2.1905337140347023, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.308737277984619, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.74106365442276, + "num_tokens": 516153496.0, + "step": 19947 + }, + { + "epoch": 2.190643531737316, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.485860824584961, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7114638090133667, + "num_tokens": 516179007.0, + "step": 19948 + }, + { + "epoch": 2.19075334943993, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.356111526489258, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7312789559364319, + "num_tokens": 516203928.0, + "step": 19949 + }, + { + "epoch": 2.1908631671425436, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4736671447753906, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7415621280670166, + "num_tokens": 516227878.0, + "step": 19950 + }, + { + "epoch": 2.190972984845157, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3485052585601807, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7266067266464233, + "num_tokens": 516254967.0, + "step": 19951 + }, + { + "epoch": 2.1910828025477707, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5105698108673096, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.755100667476654, + "num_tokens": 516277432.0, + "step": 19952 + }, + { + "epoch": 2.1911926202503844, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.663196325302124, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7230000495910645, + "num_tokens": 516300329.0, + "step": 19953 + }, + { + "epoch": 2.191302437952998, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2825186252593994, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7135103940963745, + "num_tokens": 516328649.0, + "step": 19954 + }, + { + "epoch": 2.1914122556556115, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4203383922576904, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7246588468551636, + "num_tokens": 516353113.0, + "step": 19955 + }, + { + "epoch": 2.1915220733582252, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.117403984069824, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7417305707931519, + "num_tokens": 516383463.0, + "step": 19956 + }, + { + "epoch": 2.191631891060839, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4394524097442627, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7506399154663086, + "num_tokens": 516407754.0, + "step": 19957 + }, + { + "epoch": 2.1917417087634528, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.622690439224243, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7508234977722168, + "num_tokens": 516428975.0, + "step": 19958 + }, + { + "epoch": 2.1918515264660665, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.8780994415283203, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7475095987319946, + "num_tokens": 516448000.0, + "step": 19959 + }, + { + "epoch": 2.19196134416868, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.199477195739746, + "learning_rate": 1e-06, + "loss": 0.9147, + "mean_token_accuracy": 0.7133335471153259, + "num_tokens": 516475023.0, + "step": 19960 + }, + { + "epoch": 2.1920711618712936, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.313462495803833, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7404449582099915, + "num_tokens": 516501229.0, + "step": 19961 + }, + { + "epoch": 2.1921809795739073, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2352354526519775, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7339428067207336, + "num_tokens": 516532095.0, + "step": 19962 + }, + { + "epoch": 2.192290797276521, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.423673152923584, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7312558889389038, + "num_tokens": 516555590.0, + "step": 19963 + }, + { + "epoch": 2.1924006149791344, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.1819510459899902, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7249475717544556, + "num_tokens": 516586018.0, + "step": 19964 + }, + { + "epoch": 2.192510432681748, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4867613315582275, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7199602127075195, + "num_tokens": 516610101.0, + "step": 19965 + }, + { + "epoch": 2.192620250384362, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4405012130737305, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7331824898719788, + "num_tokens": 516634371.0, + "step": 19966 + }, + { + "epoch": 2.1927300680869757, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.459709644317627, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7475938200950623, + "num_tokens": 516659537.0, + "step": 19967 + }, + { + "epoch": 2.1928398857895894, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.510899782180786, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7370623350143433, + "num_tokens": 516684554.0, + "step": 19968 + }, + { + "epoch": 2.1929497034922028, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5426409244537354, + "learning_rate": 1e-06, + "loss": 0.7606, + "mean_token_accuracy": 0.7540789842605591, + "num_tokens": 516706127.0, + "step": 19969 + }, + { + "epoch": 2.1930595211948165, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.440661668777466, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7358941435813904, + "num_tokens": 516729918.0, + "step": 19970 + }, + { + "epoch": 2.1931693388974303, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3113903999328613, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7462549805641174, + "num_tokens": 516754609.0, + "step": 19971 + }, + { + "epoch": 2.193279156600044, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5718131065368652, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7170776724815369, + "num_tokens": 516779618.0, + "step": 19972 + }, + { + "epoch": 2.193388974302658, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3481369018554688, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7401861548423767, + "num_tokens": 516803725.0, + "step": 19973 + }, + { + "epoch": 2.193498792005271, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.299466848373413, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7524908185005188, + "num_tokens": 516829951.0, + "step": 19974 + }, + { + "epoch": 2.193608609707885, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.112971544265747, + "learning_rate": 1e-06, + "loss": 0.9499, + "mean_token_accuracy": 0.7048261165618896, + "num_tokens": 516861823.0, + "step": 19975 + }, + { + "epoch": 2.1937184274104986, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.341118812561035, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7228167057037354, + "num_tokens": 516888694.0, + "step": 19976 + }, + { + "epoch": 2.1938282451131124, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2953906059265137, + "learning_rate": 1e-06, + "loss": 0.764, + "mean_token_accuracy": 0.7536365985870361, + "num_tokens": 516913859.0, + "step": 19977 + }, + { + "epoch": 2.193938062815726, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.7337191104888916, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7365379929542542, + "num_tokens": 516934106.0, + "step": 19978 + }, + { + "epoch": 2.1940478805183394, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2431247234344482, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7397282123565674, + "num_tokens": 516962220.0, + "step": 19979 + }, + { + "epoch": 2.194157698220953, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.219752788543701, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7275339365005493, + "num_tokens": 516992097.0, + "step": 19980 + }, + { + "epoch": 2.194267515923567, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.0898537635803223, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7310127019882202, + "num_tokens": 517024985.0, + "step": 19981 + }, + { + "epoch": 2.1943773336261807, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.435133218765259, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7316092252731323, + "num_tokens": 517049686.0, + "step": 19982 + }, + { + "epoch": 2.194487151328794, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.581505537033081, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7330489158630371, + "num_tokens": 517073345.0, + "step": 19983 + }, + { + "epoch": 2.1945969690314078, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.31882381439209, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7349357604980469, + "num_tokens": 517099659.0, + "step": 19984 + }, + { + "epoch": 2.1947067867340215, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.1522204875946045, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7232768535614014, + "num_tokens": 517127778.0, + "step": 19985 + }, + { + "epoch": 2.1948166044366353, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.688244581222534, + "learning_rate": 1e-06, + "loss": 0.7013, + "mean_token_accuracy": 0.7652751207351685, + "num_tokens": 517147665.0, + "step": 19986 + }, + { + "epoch": 2.194926422139249, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2642550468444824, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.719490647315979, + "num_tokens": 517173900.0, + "step": 19987 + }, + { + "epoch": 2.1950362398418624, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3073642253875732, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7406023144721985, + "num_tokens": 517201170.0, + "step": 19988 + }, + { + "epoch": 2.195146057544476, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2091457843780518, + "learning_rate": 1e-06, + "loss": 0.9673, + "mean_token_accuracy": 0.6917482614517212, + "num_tokens": 517230427.0, + "step": 19989 + }, + { + "epoch": 2.19525587524709, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.0943455696105957, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7150763273239136, + "num_tokens": 517265792.0, + "step": 19990 + }, + { + "epoch": 2.1953656929497036, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.626110076904297, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.737857460975647, + "num_tokens": 517287592.0, + "step": 19991 + }, + { + "epoch": 2.195475510652317, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5203282833099365, + "learning_rate": 1e-06, + "loss": 0.7728, + "mean_token_accuracy": 0.7530618906021118, + "num_tokens": 517310886.0, + "step": 19992 + }, + { + "epoch": 2.1955853283549307, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.6848511695861816, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7440031170845032, + "num_tokens": 517332572.0, + "step": 19993 + }, + { + "epoch": 2.1956951460575445, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.601161003112793, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7358613014221191, + "num_tokens": 517355634.0, + "step": 19994 + }, + { + "epoch": 2.195804963760158, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.122257947921753, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7027269601821899, + "num_tokens": 517385468.0, + "step": 19995 + }, + { + "epoch": 2.195914781462772, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.0168869495391846, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7229005098342896, + "num_tokens": 517417204.0, + "step": 19996 + }, + { + "epoch": 2.1960245991653853, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.4517221450805664, + "learning_rate": 1e-06, + "loss": 0.7464, + "mean_token_accuracy": 0.7585890889167786, + "num_tokens": 517441033.0, + "step": 19997 + }, + { + "epoch": 2.196134416867999, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3015687465667725, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7122939229011536, + "num_tokens": 517469258.0, + "step": 19998 + }, + { + "epoch": 2.196244234570613, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.349961042404175, + "learning_rate": 1e-06, + "loss": 0.9568, + "mean_token_accuracy": 0.7097072601318359, + "num_tokens": 517495697.0, + "step": 19999 + }, + { + "epoch": 2.1963540522732266, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.117992877960205, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7315869331359863, + "num_tokens": 517526328.0, + "step": 20000 + }, + { + "epoch": 2.1964638699758403, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.245379686355591, + "learning_rate": 1e-06, + "loss": 0.949, + "mean_token_accuracy": 0.7053927183151245, + "num_tokens": 517556477.0, + "step": 20001 + }, + { + "epoch": 2.1965736876784536, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.2768924236297607, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7295393347740173, + "num_tokens": 517583923.0, + "step": 20002 + }, + { + "epoch": 2.1966835053810674, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.0972564220428467, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7125523090362549, + "num_tokens": 517618071.0, + "step": 20003 + }, + { + "epoch": 2.196793323083681, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.5243847370147705, + "learning_rate": 1e-06, + "loss": 0.766, + "mean_token_accuracy": 0.7643997669219971, + "num_tokens": 517640507.0, + "step": 20004 + }, + { + "epoch": 2.196903140786295, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.3583240509033203, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7387356162071228, + "num_tokens": 517665619.0, + "step": 20005 + }, + { + "epoch": 2.197012958488908, + "ewc_loss": 1.2636184692382812e-05, + "grad_norm": 2.17183256149292, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7439788579940796, + "num_tokens": 517693362.0, + "step": 20006 + }, + { + "epoch": 2.197122776191522, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.467597007751465, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7309024333953857, + "num_tokens": 517716288.0, + "step": 20007 + }, + { + "epoch": 2.1972325938941357, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.348986864089966, + "learning_rate": 1e-06, + "loss": 0.9256, + "mean_token_accuracy": 0.7129626274108887, + "num_tokens": 517741774.0, + "step": 20008 + }, + { + "epoch": 2.1973424115967495, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1760191917419434, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7170419692993164, + "num_tokens": 517773664.0, + "step": 20009 + }, + { + "epoch": 2.1974522292993632, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5767557621002197, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7433487176895142, + "num_tokens": 517793926.0, + "step": 20010 + }, + { + "epoch": 2.1975620470019765, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1486756801605225, + "learning_rate": 1e-06, + "loss": 0.9104, + "mean_token_accuracy": 0.721469521522522, + "num_tokens": 517823769.0, + "step": 20011 + }, + { + "epoch": 2.1976718647045903, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5007903575897217, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7355819344520569, + "num_tokens": 517849202.0, + "step": 20012 + }, + { + "epoch": 2.197781682407204, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.389216661453247, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7224817276000977, + "num_tokens": 517877265.0, + "step": 20013 + }, + { + "epoch": 2.197891500109818, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4272701740264893, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7380098104476929, + "num_tokens": 517900377.0, + "step": 20014 + }, + { + "epoch": 2.198001317812431, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.135516881942749, + "learning_rate": 1e-06, + "loss": 0.7432, + "mean_token_accuracy": 0.7575721740722656, + "num_tokens": 517927383.0, + "step": 20015 + }, + { + "epoch": 2.198111135515045, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1524534225463867, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7102341651916504, + "num_tokens": 517957042.0, + "step": 20016 + }, + { + "epoch": 2.1982209532176586, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3792264461517334, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7198337912559509, + "num_tokens": 517982093.0, + "step": 20017 + }, + { + "epoch": 2.1983307709202724, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3147854804992676, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7508927583694458, + "num_tokens": 518006726.0, + "step": 20018 + }, + { + "epoch": 2.198440588622886, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.171304941177368, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7271051406860352, + "num_tokens": 518038182.0, + "step": 20019 + }, + { + "epoch": 2.1985504063254995, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4232821464538574, + "learning_rate": 1e-06, + "loss": 0.7468, + "mean_token_accuracy": 0.7583955526351929, + "num_tokens": 518061637.0, + "step": 20020 + }, + { + "epoch": 2.1986602240281132, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2336928844451904, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7195228338241577, + "num_tokens": 518091519.0, + "step": 20021 + }, + { + "epoch": 2.198770041730727, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.415344476699829, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.71108078956604, + "num_tokens": 518116665.0, + "step": 20022 + }, + { + "epoch": 2.1988798594333407, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1100480556488037, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7182812690734863, + "num_tokens": 518151433.0, + "step": 20023 + }, + { + "epoch": 2.1989896771359545, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.7753398418426514, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7144657969474792, + "num_tokens": 518172320.0, + "step": 20024 + }, + { + "epoch": 2.199099494838568, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.6529934406280518, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7644591331481934, + "num_tokens": 518192686.0, + "step": 20025 + }, + { + "epoch": 2.1992093125411816, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.365306854248047, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7177627682685852, + "num_tokens": 518219581.0, + "step": 20026 + }, + { + "epoch": 2.1993191302437953, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.25553560256958, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7303850650787354, + "num_tokens": 518247489.0, + "step": 20027 + }, + { + "epoch": 2.199428947946409, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.596606492996216, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7387862205505371, + "num_tokens": 518269276.0, + "step": 20028 + }, + { + "epoch": 2.199538765649023, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3269450664520264, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7497797012329102, + "num_tokens": 518297160.0, + "step": 20029 + }, + { + "epoch": 2.199648583351636, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.386939287185669, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7121330499649048, + "num_tokens": 518324758.0, + "step": 20030 + }, + { + "epoch": 2.19975840105425, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.44645619392395, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.7461647987365723, + "num_tokens": 518351349.0, + "step": 20031 + }, + { + "epoch": 2.1998682187568637, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2281503677368164, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7421202659606934, + "num_tokens": 518379043.0, + "step": 20032 + }, + { + "epoch": 2.1999780364594774, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.213362216949463, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7268105745315552, + "num_tokens": 518408717.0, + "step": 20033 + }, + { + "epoch": 2.2000878541620907, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5673067569732666, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7265040874481201, + "num_tokens": 518433853.0, + "step": 20034 + }, + { + "epoch": 2.2001976718647045, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.6175239086151123, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7172138690948486, + "num_tokens": 518457724.0, + "step": 20035 + }, + { + "epoch": 2.2003074895673183, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2011802196502686, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7539800405502319, + "num_tokens": 518485575.0, + "step": 20036 + }, + { + "epoch": 2.200417307269932, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.264657497406006, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7125952243804932, + "num_tokens": 518517005.0, + "step": 20037 + }, + { + "epoch": 2.2005271249725458, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4900569915771484, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7146457433700562, + "num_tokens": 518541236.0, + "step": 20038 + }, + { + "epoch": 2.200636942675159, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.6269474029541016, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7406795620918274, + "num_tokens": 518565424.0, + "step": 20039 + }, + { + "epoch": 2.200746760377773, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.338155508041382, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7492145299911499, + "num_tokens": 518589389.0, + "step": 20040 + }, + { + "epoch": 2.2008565780803866, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4148006439208984, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7299239635467529, + "num_tokens": 518614095.0, + "step": 20041 + }, + { + "epoch": 2.2009663957830004, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4966464042663574, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7266783714294434, + "num_tokens": 518638191.0, + "step": 20042 + }, + { + "epoch": 2.2010762134856137, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.6854846477508545, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7349629402160645, + "num_tokens": 518660484.0, + "step": 20043 + }, + { + "epoch": 2.2011860311882274, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2651240825653076, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7157652974128723, + "num_tokens": 518687926.0, + "step": 20044 + }, + { + "epoch": 2.201295848890841, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1460580825805664, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7165054082870483, + "num_tokens": 518719279.0, + "step": 20045 + }, + { + "epoch": 2.201405666593455, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.0488221645355225, + "learning_rate": 1e-06, + "loss": 0.9215, + "mean_token_accuracy": 0.7203314304351807, + "num_tokens": 518753491.0, + "step": 20046 + }, + { + "epoch": 2.2015154842960687, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2253053188323975, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7221391797065735, + "num_tokens": 518780926.0, + "step": 20047 + }, + { + "epoch": 2.201625301998682, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5103955268859863, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7079306244850159, + "num_tokens": 518805478.0, + "step": 20048 + }, + { + "epoch": 2.2017351197012958, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.36763858795166, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7326001524925232, + "num_tokens": 518830675.0, + "step": 20049 + }, + { + "epoch": 2.2018449374039095, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.507042407989502, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7510253190994263, + "num_tokens": 518854521.0, + "step": 20050 + }, + { + "epoch": 2.2019547551065233, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.313094139099121, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7223007678985596, + "num_tokens": 518882182.0, + "step": 20051 + }, + { + "epoch": 2.202064572809137, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2345399856567383, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7294324040412903, + "num_tokens": 518911153.0, + "step": 20052 + }, + { + "epoch": 2.2021743905117503, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2777953147888184, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7220167517662048, + "num_tokens": 518938259.0, + "step": 20053 + }, + { + "epoch": 2.202284208214364, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4278323650360107, + "learning_rate": 1e-06, + "loss": 0.7166, + "mean_token_accuracy": 0.7670197486877441, + "num_tokens": 518960702.0, + "step": 20054 + }, + { + "epoch": 2.202394025916978, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.260744094848633, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7154011726379395, + "num_tokens": 518989298.0, + "step": 20055 + }, + { + "epoch": 2.2025038436195916, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.390756845474243, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7123122811317444, + "num_tokens": 519014972.0, + "step": 20056 + }, + { + "epoch": 2.2026136613222054, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.603898763656616, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.740277886390686, + "num_tokens": 519038925.0, + "step": 20057 + }, + { + "epoch": 2.2027234790248187, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3367366790771484, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7305731773376465, + "num_tokens": 519064807.0, + "step": 20058 + }, + { + "epoch": 2.2028332967274324, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.5899832248687744, + "learning_rate": 1e-06, + "loss": 0.7421, + "mean_token_accuracy": 0.7589433193206787, + "num_tokens": 519086927.0, + "step": 20059 + }, + { + "epoch": 2.202943114430046, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.349914789199829, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.7570511102676392, + "num_tokens": 519111766.0, + "step": 20060 + }, + { + "epoch": 2.20305293213266, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1980366706848145, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.715829074382782, + "num_tokens": 519140981.0, + "step": 20061 + }, + { + "epoch": 2.2031627498352733, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4243812561035156, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7184041738510132, + "num_tokens": 519165772.0, + "step": 20062 + }, + { + "epoch": 2.203272567537887, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3040122985839844, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7497768998146057, + "num_tokens": 519191829.0, + "step": 20063 + }, + { + "epoch": 2.203382385240501, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4909205436706543, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.713788628578186, + "num_tokens": 519216179.0, + "step": 20064 + }, + { + "epoch": 2.2034922029431145, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.220597982406616, + "learning_rate": 1e-06, + "loss": 0.7745, + "mean_token_accuracy": 0.7509949207305908, + "num_tokens": 519245075.0, + "step": 20065 + }, + { + "epoch": 2.2036020206457283, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.281054735183716, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7355437278747559, + "num_tokens": 519272624.0, + "step": 20066 + }, + { + "epoch": 2.2037118383483416, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4152512550354004, + "learning_rate": 1e-06, + "loss": 0.879, + "mean_token_accuracy": 0.7290465831756592, + "num_tokens": 519297407.0, + "step": 20067 + }, + { + "epoch": 2.2038216560509554, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.0049567222595215, + "learning_rate": 1e-06, + "loss": 0.9312, + "mean_token_accuracy": 0.707792341709137, + "num_tokens": 519335419.0, + "step": 20068 + }, + { + "epoch": 2.203931473753569, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.649746894836426, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7548539042472839, + "num_tokens": 519355629.0, + "step": 20069 + }, + { + "epoch": 2.204041291456183, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.365511894226074, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7334425449371338, + "num_tokens": 519381439.0, + "step": 20070 + }, + { + "epoch": 2.204151109158796, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.435466766357422, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7205647826194763, + "num_tokens": 519404641.0, + "step": 20071 + }, + { + "epoch": 2.20426092686141, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1902365684509277, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7237187027931213, + "num_tokens": 519432333.0, + "step": 20072 + }, + { + "epoch": 2.2043707445640237, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2543487548828125, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7204180955886841, + "num_tokens": 519462964.0, + "step": 20073 + }, + { + "epoch": 2.2044805622666375, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.358893632888794, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7364458441734314, + "num_tokens": 519487752.0, + "step": 20074 + }, + { + "epoch": 2.2045903799692512, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2215359210968018, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7251434326171875, + "num_tokens": 519516482.0, + "step": 20075 + }, + { + "epoch": 2.2047001976718645, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4452767372131348, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7516617774963379, + "num_tokens": 519540401.0, + "step": 20076 + }, + { + "epoch": 2.2048100153744783, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.500744104385376, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7186315655708313, + "num_tokens": 519564891.0, + "step": 20077 + }, + { + "epoch": 2.204919833077092, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1075212955474854, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7199370265007019, + "num_tokens": 519595185.0, + "step": 20078 + }, + { + "epoch": 2.205029650779706, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.3631019592285156, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7285543084144592, + "num_tokens": 519620293.0, + "step": 20079 + }, + { + "epoch": 2.2051394684823196, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.6319169998168945, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7419060468673706, + "num_tokens": 519643636.0, + "step": 20080 + }, + { + "epoch": 2.205249286184933, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.163062572479248, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7192962765693665, + "num_tokens": 519671611.0, + "step": 20081 + }, + { + "epoch": 2.2053591038875466, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.0346100330352783, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7167654037475586, + "num_tokens": 519706511.0, + "step": 20082 + }, + { + "epoch": 2.2054689215901604, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4944281578063965, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7427868843078613, + "num_tokens": 519729225.0, + "step": 20083 + }, + { + "epoch": 2.205578739292774, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1797561645507812, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7284677028656006, + "num_tokens": 519757241.0, + "step": 20084 + }, + { + "epoch": 2.2056885569953875, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2439849376678467, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7365860342979431, + "num_tokens": 519783692.0, + "step": 20085 + }, + { + "epoch": 2.205798374698001, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.591247081756592, + "learning_rate": 1e-06, + "loss": 0.7775, + "mean_token_accuracy": 0.7541993856430054, + "num_tokens": 519807339.0, + "step": 20086 + }, + { + "epoch": 2.205908192400615, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.492748498916626, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7440720796585083, + "num_tokens": 519830355.0, + "step": 20087 + }, + { + "epoch": 2.2060180101032287, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1748154163360596, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7309294939041138, + "num_tokens": 519859643.0, + "step": 20088 + }, + { + "epoch": 2.2061278278058425, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.4955785274505615, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7436232566833496, + "num_tokens": 519883210.0, + "step": 20089 + }, + { + "epoch": 2.206237645508456, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.69288969039917, + "learning_rate": 1e-06, + "loss": 0.759, + "mean_token_accuracy": 0.7566799521446228, + "num_tokens": 519903342.0, + "step": 20090 + }, + { + "epoch": 2.2063474632110696, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2029190063476562, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7382246851921082, + "num_tokens": 519930300.0, + "step": 20091 + }, + { + "epoch": 2.2064572809136833, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.1748430728912354, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7226608395576477, + "num_tokens": 519960474.0, + "step": 20092 + }, + { + "epoch": 2.206567098616297, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.2563107013702393, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7395612001419067, + "num_tokens": 519986070.0, + "step": 20093 + }, + { + "epoch": 2.2066769163189104, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.431046724319458, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7422045469284058, + "num_tokens": 520011594.0, + "step": 20094 + }, + { + "epoch": 2.206786734021524, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.44394850730896, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7157202959060669, + "num_tokens": 520037705.0, + "step": 20095 + }, + { + "epoch": 2.206896551724138, + "ewc_loss": 1.2695789337158203e-05, + "grad_norm": 2.722540855407715, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7427308559417725, + "num_tokens": 520056695.0, + "step": 20096 + }, + { + "epoch": 2.2070063694267517, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5232179164886475, + "learning_rate": 1e-06, + "loss": 0.7826, + "mean_token_accuracy": 0.7433587312698364, + "num_tokens": 520079533.0, + "step": 20097 + }, + { + "epoch": 2.2071161871293654, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.519993543624878, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7292918562889099, + "num_tokens": 520104959.0, + "step": 20098 + }, + { + "epoch": 2.2072260048319787, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5860207080841064, + "learning_rate": 1e-06, + "loss": 0.7797, + "mean_token_accuracy": 0.7491213083267212, + "num_tokens": 520126584.0, + "step": 20099 + }, + { + "epoch": 2.2073358225345925, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4319515228271484, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7350499629974365, + "num_tokens": 520149976.0, + "step": 20100 + }, + { + "epoch": 2.2074456402372062, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.427204132080078, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7344872355461121, + "num_tokens": 520177916.0, + "step": 20101 + }, + { + "epoch": 2.20755545793982, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.246143341064453, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7251521348953247, + "num_tokens": 520205584.0, + "step": 20102 + }, + { + "epoch": 2.2076652756424338, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.295788288116455, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7245970964431763, + "num_tokens": 520234122.0, + "step": 20103 + }, + { + "epoch": 2.207775093345047, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5326881408691406, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7091825008392334, + "num_tokens": 520258717.0, + "step": 20104 + }, + { + "epoch": 2.207884911047661, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.431363821029663, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7433537244796753, + "num_tokens": 520284535.0, + "step": 20105 + }, + { + "epoch": 2.2079947287502746, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.305957078933716, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7544563412666321, + "num_tokens": 520310647.0, + "step": 20106 + }, + { + "epoch": 2.2081045464528883, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.035216808319092, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7113662362098694, + "num_tokens": 520344625.0, + "step": 20107 + }, + { + "epoch": 2.208214364155502, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3551249504089355, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7385263442993164, + "num_tokens": 520368818.0, + "step": 20108 + }, + { + "epoch": 2.2083241818581154, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.149123191833496, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7335679531097412, + "num_tokens": 520396062.0, + "step": 20109 + }, + { + "epoch": 2.208433999560729, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2497403621673584, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7297830581665039, + "num_tokens": 520423499.0, + "step": 20110 + }, + { + "epoch": 2.208543817263343, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 1.920880913734436, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7168848514556885, + "num_tokens": 520460909.0, + "step": 20111 + }, + { + "epoch": 2.2086536349659567, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2463326454162598, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7377868890762329, + "num_tokens": 520487116.0, + "step": 20112 + }, + { + "epoch": 2.20876345266857, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.204094409942627, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7128175497055054, + "num_tokens": 520516152.0, + "step": 20113 + }, + { + "epoch": 2.2088732703711838, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4631497859954834, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7468146085739136, + "num_tokens": 520539944.0, + "step": 20114 + }, + { + "epoch": 2.2089830880737975, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2049546241760254, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7369183897972107, + "num_tokens": 520567663.0, + "step": 20115 + }, + { + "epoch": 2.2090929057764113, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.6260883808135986, + "learning_rate": 1e-06, + "loss": 0.7593, + "mean_token_accuracy": 0.7542118430137634, + "num_tokens": 520590056.0, + "step": 20116 + }, + { + "epoch": 2.209202723479025, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5436296463012695, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7420805096626282, + "num_tokens": 520612078.0, + "step": 20117 + }, + { + "epoch": 2.2093125411816383, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.7899868488311768, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7379326224327087, + "num_tokens": 520631927.0, + "step": 20118 + }, + { + "epoch": 2.209422358884252, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.7141644954681396, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7282313108444214, + "num_tokens": 520658703.0, + "step": 20119 + }, + { + "epoch": 2.209532176586866, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.7205348014831543, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7418941855430603, + "num_tokens": 520677388.0, + "step": 20120 + }, + { + "epoch": 2.2096419942894796, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2401554584503174, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.715566098690033, + "num_tokens": 520706453.0, + "step": 20121 + }, + { + "epoch": 2.209751811992093, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.232368230819702, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7334433794021606, + "num_tokens": 520736876.0, + "step": 20122 + }, + { + "epoch": 2.2098616296947067, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5195627212524414, + "learning_rate": 1e-06, + "loss": 0.9723, + "mean_token_accuracy": 0.6975370645523071, + "num_tokens": 520762241.0, + "step": 20123 + }, + { + "epoch": 2.2099714473973204, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.510547637939453, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7346084117889404, + "num_tokens": 520785181.0, + "step": 20124 + }, + { + "epoch": 2.210081265099934, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.244868040084839, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7268766164779663, + "num_tokens": 520812759.0, + "step": 20125 + }, + { + "epoch": 2.210191082802548, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3345699310302734, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7433187961578369, + "num_tokens": 520839324.0, + "step": 20126 + }, + { + "epoch": 2.2103009005051613, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.298053026199341, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7345218062400818, + "num_tokens": 520867844.0, + "step": 20127 + }, + { + "epoch": 2.210410718207775, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2339813709259033, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.712287187576294, + "num_tokens": 520897337.0, + "step": 20128 + }, + { + "epoch": 2.2105205359103888, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3755712509155273, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7557902336120605, + "num_tokens": 520922710.0, + "step": 20129 + }, + { + "epoch": 2.2106303536130025, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.427661657333374, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7204879522323608, + "num_tokens": 520948579.0, + "step": 20130 + }, + { + "epoch": 2.2107401713156163, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3138225078582764, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7208940982818604, + "num_tokens": 520976304.0, + "step": 20131 + }, + { + "epoch": 2.2108499890182296, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2102391719818115, + "learning_rate": 1e-06, + "loss": 0.9665, + "mean_token_accuracy": 0.6954731941223145, + "num_tokens": 521005547.0, + "step": 20132 + }, + { + "epoch": 2.2109598067208434, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3466956615448, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7314164638519287, + "num_tokens": 521032555.0, + "step": 20133 + }, + { + "epoch": 2.211069624423457, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.432457685470581, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7108293175697327, + "num_tokens": 521059832.0, + "step": 20134 + }, + { + "epoch": 2.211179442126071, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2660439014434814, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7199560403823853, + "num_tokens": 521088278.0, + "step": 20135 + }, + { + "epoch": 2.211289259828684, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5379669666290283, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7196313142776489, + "num_tokens": 521111519.0, + "step": 20136 + }, + { + "epoch": 2.211399077531298, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2186577320098877, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7381224632263184, + "num_tokens": 521139870.0, + "step": 20137 + }, + { + "epoch": 2.2115088952339117, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4057960510253906, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7339128851890564, + "num_tokens": 521167763.0, + "step": 20138 + }, + { + "epoch": 2.2116187129365255, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.631342649459839, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7312588095664978, + "num_tokens": 521193950.0, + "step": 20139 + }, + { + "epoch": 2.211728530639139, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2265570163726807, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7243192791938782, + "num_tokens": 521227519.0, + "step": 20140 + }, + { + "epoch": 2.2118383483417525, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.8311703205108643, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.739352285861969, + "num_tokens": 521247804.0, + "step": 20141 + }, + { + "epoch": 2.2119481660443663, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.660853624343872, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7232458591461182, + "num_tokens": 521269010.0, + "step": 20142 + }, + { + "epoch": 2.21205798374698, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.442108392715454, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.72441565990448, + "num_tokens": 521293769.0, + "step": 20143 + }, + { + "epoch": 2.212167801449594, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4481418132781982, + "learning_rate": 1e-06, + "loss": 0.8404, + "mean_token_accuracy": 0.7406362295150757, + "num_tokens": 521318842.0, + "step": 20144 + }, + { + "epoch": 2.212277619152207, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.2427520751953125, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7295593619346619, + "num_tokens": 521346698.0, + "step": 20145 + }, + { + "epoch": 2.212387436854821, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.4426026344299316, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7433046102523804, + "num_tokens": 521371578.0, + "step": 20146 + }, + { + "epoch": 2.2124972545574346, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.3453428745269775, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7451808452606201, + "num_tokens": 521397721.0, + "step": 20147 + }, + { + "epoch": 2.2126070722600484, + "ewc_loss": 1.2755393981933594e-05, + "grad_norm": 2.5648157596588135, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7304655313491821, + "num_tokens": 521418959.0, + "step": 20148 + }, + { + "epoch": 2.212716889962662, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2696845531463623, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7233326435089111, + "num_tokens": 521447973.0, + "step": 20149 + }, + { + "epoch": 2.2128267076652754, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2717158794403076, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7282933592796326, + "num_tokens": 521475234.0, + "step": 20150 + }, + { + "epoch": 2.212936525367889, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4327118396759033, + "learning_rate": 1e-06, + "loss": 0.7883, + "mean_token_accuracy": 0.7429554462432861, + "num_tokens": 521499190.0, + "step": 20151 + }, + { + "epoch": 2.213046343070503, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4412200450897217, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7304719090461731, + "num_tokens": 521525491.0, + "step": 20152 + }, + { + "epoch": 2.2131561607731167, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5296285152435303, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7244043946266174, + "num_tokens": 521548624.0, + "step": 20153 + }, + { + "epoch": 2.2132659784757305, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.112236261367798, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.7147783041000366, + "num_tokens": 521579307.0, + "step": 20154 + }, + { + "epoch": 2.213375796178344, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.475399971008301, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7540773153305054, + "num_tokens": 521602114.0, + "step": 20155 + }, + { + "epoch": 2.2134856138809575, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4218671321868896, + "learning_rate": 1e-06, + "loss": 0.7535, + "mean_token_accuracy": 0.7529650926589966, + "num_tokens": 521625068.0, + "step": 20156 + }, + { + "epoch": 2.2135954315835713, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3503544330596924, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7383233308792114, + "num_tokens": 521649794.0, + "step": 20157 + }, + { + "epoch": 2.213705249286185, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.766113758087158, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7309131026268005, + "num_tokens": 521669433.0, + "step": 20158 + }, + { + "epoch": 2.213815066988799, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.551525354385376, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7241483926773071, + "num_tokens": 521691755.0, + "step": 20159 + }, + { + "epoch": 2.213924884691412, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3152763843536377, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7331136465072632, + "num_tokens": 521717300.0, + "step": 20160 + }, + { + "epoch": 2.214034702394026, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.247936725616455, + "learning_rate": 1e-06, + "loss": 0.9667, + "mean_token_accuracy": 0.7029567956924438, + "num_tokens": 521746826.0, + "step": 20161 + }, + { + "epoch": 2.2141445200966396, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.6627795696258545, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7446096539497375, + "num_tokens": 521767377.0, + "step": 20162 + }, + { + "epoch": 2.2142543377992534, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3462278842926025, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7329858541488647, + "num_tokens": 521792448.0, + "step": 20163 + }, + { + "epoch": 2.2143641555018667, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3914053440093994, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7343237400054932, + "num_tokens": 521817314.0, + "step": 20164 + }, + { + "epoch": 2.2144739732044805, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.480665445327759, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7233902215957642, + "num_tokens": 521844677.0, + "step": 20165 + }, + { + "epoch": 2.2145837909070942, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.6390693187713623, + "learning_rate": 1e-06, + "loss": 0.7719, + "mean_token_accuracy": 0.7615443468093872, + "num_tokens": 521865563.0, + "step": 20166 + }, + { + "epoch": 2.214693608609708, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.8043527603149414, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7409859299659729, + "num_tokens": 521885758.0, + "step": 20167 + }, + { + "epoch": 2.2148034263123217, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.579944610595703, + "learning_rate": 1e-06, + "loss": 0.7404, + "mean_token_accuracy": 0.7652132511138916, + "num_tokens": 521907417.0, + "step": 20168 + }, + { + "epoch": 2.214913244014935, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2932817935943604, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7287092208862305, + "num_tokens": 521934415.0, + "step": 20169 + }, + { + "epoch": 2.215023061717549, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.638336181640625, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.723754346370697, + "num_tokens": 521956542.0, + "step": 20170 + }, + { + "epoch": 2.2151328794201626, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2895853519439697, + "learning_rate": 1e-06, + "loss": 0.7737, + "mean_token_accuracy": 0.7490315437316895, + "num_tokens": 521981406.0, + "step": 20171 + }, + { + "epoch": 2.2152426971227763, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3771743774414062, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7257459163665771, + "num_tokens": 522006913.0, + "step": 20172 + }, + { + "epoch": 2.2153525148253896, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3703064918518066, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7202298641204834, + "num_tokens": 522032649.0, + "step": 20173 + }, + { + "epoch": 2.2154623325280034, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.688627243041992, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7394617795944214, + "num_tokens": 522054575.0, + "step": 20174 + }, + { + "epoch": 2.215572150230617, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.0051655769348145, + "learning_rate": 1e-06, + "loss": 0.9473, + "mean_token_accuracy": 0.7035919427871704, + "num_tokens": 522088974.0, + "step": 20175 + }, + { + "epoch": 2.215681967933231, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5459866523742676, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7297916412353516, + "num_tokens": 522112488.0, + "step": 20176 + }, + { + "epoch": 2.2157917856358447, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.665069818496704, + "learning_rate": 1e-06, + "loss": 0.7477, + "mean_token_accuracy": 0.7560831308364868, + "num_tokens": 522133063.0, + "step": 20177 + }, + { + "epoch": 2.215901603338458, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2618167400360107, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7201125621795654, + "num_tokens": 522161195.0, + "step": 20178 + }, + { + "epoch": 2.2160114210410717, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.273344039916992, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7337835431098938, + "num_tokens": 522188462.0, + "step": 20179 + }, + { + "epoch": 2.2161212387436855, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.1979787349700928, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7167525291442871, + "num_tokens": 522218385.0, + "step": 20180 + }, + { + "epoch": 2.2162310564462993, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.10221266746521, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7393785715103149, + "num_tokens": 522250015.0, + "step": 20181 + }, + { + "epoch": 2.216340874148913, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3130741119384766, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7283290028572083, + "num_tokens": 522276403.0, + "step": 20182 + }, + { + "epoch": 2.2164506918515263, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.1219937801361084, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7158291339874268, + "num_tokens": 522309096.0, + "step": 20183 + }, + { + "epoch": 2.21656050955414, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2391724586486816, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7105600237846375, + "num_tokens": 522337840.0, + "step": 20184 + }, + { + "epoch": 2.216670327256754, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.381051540374756, + "learning_rate": 1e-06, + "loss": 0.7192, + "mean_token_accuracy": 0.7649644017219543, + "num_tokens": 522361322.0, + "step": 20185 + }, + { + "epoch": 2.2167801449593676, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.597548484802246, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7515703439712524, + "num_tokens": 522382718.0, + "step": 20186 + }, + { + "epoch": 2.216889962661981, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.350924253463745, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7523214221000671, + "num_tokens": 522407048.0, + "step": 20187 + }, + { + "epoch": 2.2169997803645947, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2654154300689697, + "learning_rate": 1e-06, + "loss": 0.7136, + "mean_token_accuracy": 0.7729096412658691, + "num_tokens": 522432975.0, + "step": 20188 + }, + { + "epoch": 2.2171095980672084, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.303243398666382, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7374430894851685, + "num_tokens": 522460947.0, + "step": 20189 + }, + { + "epoch": 2.217219415769822, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.1288394927978516, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7359787225723267, + "num_tokens": 522491659.0, + "step": 20190 + }, + { + "epoch": 2.217329233472436, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5731396675109863, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7363850474357605, + "num_tokens": 522513786.0, + "step": 20191 + }, + { + "epoch": 2.2174390511750492, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.319805383682251, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7394136190414429, + "num_tokens": 522540420.0, + "step": 20192 + }, + { + "epoch": 2.217548868877663, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.7440361976623535, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7572208642959595, + "num_tokens": 522560474.0, + "step": 20193 + }, + { + "epoch": 2.2176586865802768, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4774892330169678, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7448955774307251, + "num_tokens": 522583465.0, + "step": 20194 + }, + { + "epoch": 2.2177685042828905, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3261027336120605, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7364013195037842, + "num_tokens": 522609422.0, + "step": 20195 + }, + { + "epoch": 2.217878321985504, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.354576349258423, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7271837592124939, + "num_tokens": 522634318.0, + "step": 20196 + }, + { + "epoch": 2.2179881396881176, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3988046646118164, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7470943927764893, + "num_tokens": 522658245.0, + "step": 20197 + }, + { + "epoch": 2.2180979573907313, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.7389109134674072, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7326807379722595, + "num_tokens": 522682109.0, + "step": 20198 + }, + { + "epoch": 2.218207775093345, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.119886875152588, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7273191809654236, + "num_tokens": 522713748.0, + "step": 20199 + }, + { + "epoch": 2.218317592795959, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3231332302093506, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7054270505905151, + "num_tokens": 522742450.0, + "step": 20200 + }, + { + "epoch": 2.218427410498572, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.249983072280884, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7229693531990051, + "num_tokens": 522771817.0, + "step": 20201 + }, + { + "epoch": 2.218537228201186, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2090234756469727, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7156198024749756, + "num_tokens": 522799701.0, + "step": 20202 + }, + { + "epoch": 2.2186470459037997, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4273681640625, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7337921261787415, + "num_tokens": 522824493.0, + "step": 20203 + }, + { + "epoch": 2.2187568636064134, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.598487138748169, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.723886251449585, + "num_tokens": 522846289.0, + "step": 20204 + }, + { + "epoch": 2.218866681309027, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.659566879272461, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.748811662197113, + "num_tokens": 522867454.0, + "step": 20205 + }, + { + "epoch": 2.2189764990116405, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.504814863204956, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7337809801101685, + "num_tokens": 522889945.0, + "step": 20206 + }, + { + "epoch": 2.2190863167142543, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 1.9682563543319702, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7364197373390198, + "num_tokens": 522923535.0, + "step": 20207 + }, + { + "epoch": 2.219196134416868, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.6087253093719482, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7346116900444031, + "num_tokens": 522945158.0, + "step": 20208 + }, + { + "epoch": 2.219305952119482, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.111846685409546, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7303875684738159, + "num_tokens": 522975291.0, + "step": 20209 + }, + { + "epoch": 2.2194157698220955, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.175706386566162, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.738030195236206, + "num_tokens": 523002590.0, + "step": 20210 + }, + { + "epoch": 2.219525587524709, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.7101407051086426, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7211389541625977, + "num_tokens": 523024099.0, + "step": 20211 + }, + { + "epoch": 2.2196354052273226, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.551316976547241, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7305352687835693, + "num_tokens": 523046434.0, + "step": 20212 + }, + { + "epoch": 2.2197452229299364, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5653069019317627, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7414082288742065, + "num_tokens": 523069020.0, + "step": 20213 + }, + { + "epoch": 2.21985504063255, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5602476596832275, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7402458190917969, + "num_tokens": 523092677.0, + "step": 20214 + }, + { + "epoch": 2.2199648583351634, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.1172902584075928, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7265206575393677, + "num_tokens": 523125967.0, + "step": 20215 + }, + { + "epoch": 2.220074676037777, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.34840726852417, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.735657811164856, + "num_tokens": 523152192.0, + "step": 20216 + }, + { + "epoch": 2.220184493740391, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5374536514282227, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7258517742156982, + "num_tokens": 523174814.0, + "step": 20217 + }, + { + "epoch": 2.2202943114430047, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.465373992919922, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7202142477035522, + "num_tokens": 523199753.0, + "step": 20218 + }, + { + "epoch": 2.2204041291456185, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2411949634552, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7226037383079529, + "num_tokens": 523228109.0, + "step": 20219 + }, + { + "epoch": 2.220513946848232, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2295846939086914, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7235069274902344, + "num_tokens": 523257988.0, + "step": 20220 + }, + { + "epoch": 2.2206237645508455, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.734069347381592, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7451502084732056, + "num_tokens": 523277505.0, + "step": 20221 + }, + { + "epoch": 2.2207335822534593, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2173163890838623, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7260974645614624, + "num_tokens": 523304940.0, + "step": 20222 + }, + { + "epoch": 2.220843399956073, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2501087188720703, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7176761031150818, + "num_tokens": 523332632.0, + "step": 20223 + }, + { + "epoch": 2.2209532176586864, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4050450325012207, + "learning_rate": 1e-06, + "loss": 0.9468, + "mean_token_accuracy": 0.7042988538742065, + "num_tokens": 523361444.0, + "step": 20224 + }, + { + "epoch": 2.2210630353613, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5901403427124023, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7370766997337341, + "num_tokens": 523383465.0, + "step": 20225 + }, + { + "epoch": 2.221172853063914, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4654924869537354, + "learning_rate": 1e-06, + "loss": 0.758, + "mean_token_accuracy": 0.751116156578064, + "num_tokens": 523406351.0, + "step": 20226 + }, + { + "epoch": 2.2212826707665276, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4004673957824707, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.7485336065292358, + "num_tokens": 523432018.0, + "step": 20227 + }, + { + "epoch": 2.2213924884691414, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2340590953826904, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.713379979133606, + "num_tokens": 523463555.0, + "step": 20228 + }, + { + "epoch": 2.2215023061717547, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.6622731685638428, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7376685738563538, + "num_tokens": 523485185.0, + "step": 20229 + }, + { + "epoch": 2.2216121238743685, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4853334426879883, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7129271626472473, + "num_tokens": 523509089.0, + "step": 20230 + }, + { + "epoch": 2.221721941576982, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.422576904296875, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7323622703552246, + "num_tokens": 523534084.0, + "step": 20231 + }, + { + "epoch": 2.221831759279596, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.174213171005249, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7188619375228882, + "num_tokens": 523564309.0, + "step": 20232 + }, + { + "epoch": 2.2219415769822097, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3349735736846924, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7434999942779541, + "num_tokens": 523588865.0, + "step": 20233 + }, + { + "epoch": 2.222051394684823, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.136307716369629, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7092417478561401, + "num_tokens": 523620490.0, + "step": 20234 + }, + { + "epoch": 2.222161212387437, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.405745267868042, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7380333542823792, + "num_tokens": 523645323.0, + "step": 20235 + }, + { + "epoch": 2.2222710300900506, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5315101146698, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7211809158325195, + "num_tokens": 523668337.0, + "step": 20236 + }, + { + "epoch": 2.2223808477926643, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3542563915252686, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7382261157035828, + "num_tokens": 523692874.0, + "step": 20237 + }, + { + "epoch": 2.222490665495278, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5532853603363037, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7540841102600098, + "num_tokens": 523714459.0, + "step": 20238 + }, + { + "epoch": 2.2226004831978914, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5880532264709473, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7366573214530945, + "num_tokens": 523736494.0, + "step": 20239 + }, + { + "epoch": 2.222710300900505, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2201669216156006, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7182735204696655, + "num_tokens": 523766414.0, + "step": 20240 + }, + { + "epoch": 2.222820118603119, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2319962978363037, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7238993048667908, + "num_tokens": 523794278.0, + "step": 20241 + }, + { + "epoch": 2.2229299363057327, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 1.9555081129074097, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7462125420570374, + "num_tokens": 523829092.0, + "step": 20242 + }, + { + "epoch": 2.223039754008346, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.906254291534424, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.7449992895126343, + "num_tokens": 523847786.0, + "step": 20243 + }, + { + "epoch": 2.2231495717109597, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.6380743980407715, + "learning_rate": 1e-06, + "loss": 0.9398, + "mean_token_accuracy": 0.7149054408073425, + "num_tokens": 523873224.0, + "step": 20244 + }, + { + "epoch": 2.2232593894135735, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3914124965667725, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7427780032157898, + "num_tokens": 523899560.0, + "step": 20245 + }, + { + "epoch": 2.2233692071161872, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2474403381347656, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7252060174942017, + "num_tokens": 523926135.0, + "step": 20246 + }, + { + "epoch": 2.223479024818801, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.718120813369751, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7387411594390869, + "num_tokens": 523947372.0, + "step": 20247 + }, + { + "epoch": 2.2235888425214143, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.448803663253784, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7281469702720642, + "num_tokens": 523972720.0, + "step": 20248 + }, + { + "epoch": 2.223698660224028, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.914689064025879, + "learning_rate": 1e-06, + "loss": 0.7235, + "mean_token_accuracy": 0.7638672590255737, + "num_tokens": 523990408.0, + "step": 20249 + }, + { + "epoch": 2.223808477926642, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4370779991149902, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7354236245155334, + "num_tokens": 524018298.0, + "step": 20250 + }, + { + "epoch": 2.2239182956292556, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2443246841430664, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.723950207233429, + "num_tokens": 524047897.0, + "step": 20251 + }, + { + "epoch": 2.224028113331869, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.498105525970459, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7119684219360352, + "num_tokens": 524072670.0, + "step": 20252 + }, + { + "epoch": 2.2241379310344827, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5981926918029785, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.753046989440918, + "num_tokens": 524094237.0, + "step": 20253 + }, + { + "epoch": 2.2242477487370964, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.534213066101074, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7142727375030518, + "num_tokens": 524119778.0, + "step": 20254 + }, + { + "epoch": 2.22435756643971, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.21401309967041, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7323216199874878, + "num_tokens": 524147937.0, + "step": 20255 + }, + { + "epoch": 2.224467384142324, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 1.9594676494598389, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7329448461532593, + "num_tokens": 524182165.0, + "step": 20256 + }, + { + "epoch": 2.2245772018449372, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.727710485458374, + "learning_rate": 1e-06, + "loss": 0.767, + "mean_token_accuracy": 0.7580441236495972, + "num_tokens": 524201340.0, + "step": 20257 + }, + { + "epoch": 2.224687019547551, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4256060123443604, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.729488730430603, + "num_tokens": 524226267.0, + "step": 20258 + }, + { + "epoch": 2.2247968372501647, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.4071736335754395, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7408066987991333, + "num_tokens": 524249983.0, + "step": 20259 + }, + { + "epoch": 2.2249066549527785, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.245687246322632, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7095794081687927, + "num_tokens": 524277426.0, + "step": 20260 + }, + { + "epoch": 2.2250164726553923, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.463571310043335, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.718400776386261, + "num_tokens": 524302820.0, + "step": 20261 + }, + { + "epoch": 2.2251262903580056, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2566428184509277, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7311534285545349, + "num_tokens": 524331358.0, + "step": 20262 + }, + { + "epoch": 2.2252361080606193, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2811949253082275, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7263141870498657, + "num_tokens": 524358418.0, + "step": 20263 + }, + { + "epoch": 2.225345925763233, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5730490684509277, + "learning_rate": 1e-06, + "loss": 0.7506, + "mean_token_accuracy": 0.7638565301895142, + "num_tokens": 524379393.0, + "step": 20264 + }, + { + "epoch": 2.225455743465847, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2700421810150146, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7435346841812134, + "num_tokens": 524406388.0, + "step": 20265 + }, + { + "epoch": 2.22556556116846, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.1900854110717773, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7156102061271667, + "num_tokens": 524434241.0, + "step": 20266 + }, + { + "epoch": 2.225675378871074, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.2459089756011963, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.717755138874054, + "num_tokens": 524461854.0, + "step": 20267 + }, + { + "epoch": 2.2257851965736877, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.0613605976104736, + "learning_rate": 1e-06, + "loss": 0.9888, + "mean_token_accuracy": 0.6941695213317871, + "num_tokens": 524496859.0, + "step": 20268 + }, + { + "epoch": 2.2258950142763014, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.495044231414795, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7199535369873047, + "num_tokens": 524523319.0, + "step": 20269 + }, + { + "epoch": 2.226004831978915, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.7337098121643066, + "learning_rate": 1e-06, + "loss": 0.7748, + "mean_token_accuracy": 0.755107045173645, + "num_tokens": 524543772.0, + "step": 20270 + }, + { + "epoch": 2.2261146496815285, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3157284259796143, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.715802788734436, + "num_tokens": 524570526.0, + "step": 20271 + }, + { + "epoch": 2.2262244673841423, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.845505952835083, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7424271702766418, + "num_tokens": 524591893.0, + "step": 20272 + }, + { + "epoch": 2.226334285086756, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.454061508178711, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7323499917984009, + "num_tokens": 524617959.0, + "step": 20273 + }, + { + "epoch": 2.2264441027893698, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.110290765762329, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7401477098464966, + "num_tokens": 524650018.0, + "step": 20274 + }, + { + "epoch": 2.226553920491983, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3217649459838867, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7292763590812683, + "num_tokens": 524677642.0, + "step": 20275 + }, + { + "epoch": 2.226663738194597, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.5881102085113525, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7288151979446411, + "num_tokens": 524700131.0, + "step": 20276 + }, + { + "epoch": 2.2267735558972106, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.486259698867798, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.749448299407959, + "num_tokens": 524723625.0, + "step": 20277 + }, + { + "epoch": 2.2268833735998244, + "ewc_loss": 1.2814998626708984e-05, + "grad_norm": 2.3505756855010986, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7232242822647095, + "num_tokens": 524749654.0, + "step": 20278 + }, + { + "epoch": 2.226993191302438, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.625225782394409, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7442972660064697, + "num_tokens": 524770999.0, + "step": 20279 + }, + { + "epoch": 2.2271030090050514, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.2321488857269287, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7494754791259766, + "num_tokens": 524798476.0, + "step": 20280 + }, + { + "epoch": 2.227212826707665, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.193075656890869, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7196958661079407, + "num_tokens": 524830257.0, + "step": 20281 + }, + { + "epoch": 2.227322644410279, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.181979179382324, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7396174669265747, + "num_tokens": 524860125.0, + "step": 20282 + }, + { + "epoch": 2.2274324621128927, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.2150814533233643, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7342430353164673, + "num_tokens": 524886906.0, + "step": 20283 + }, + { + "epoch": 2.2275422798155065, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.363780975341797, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.730002760887146, + "num_tokens": 524913202.0, + "step": 20284 + }, + { + "epoch": 2.2276520975181198, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.807265520095825, + "learning_rate": 1e-06, + "loss": 0.7899, + "mean_token_accuracy": 0.7549936771392822, + "num_tokens": 524934388.0, + "step": 20285 + }, + { + "epoch": 2.2277619152207335, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.7556018829345703, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7458491325378418, + "num_tokens": 524954701.0, + "step": 20286 + }, + { + "epoch": 2.2278717329233473, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.256344795227051, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7442233562469482, + "num_tokens": 524982373.0, + "step": 20287 + }, + { + "epoch": 2.227981550625961, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4597766399383545, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7437477707862854, + "num_tokens": 525006604.0, + "step": 20288 + }, + { + "epoch": 2.228091368328575, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.5507683753967285, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7490956783294678, + "num_tokens": 525030729.0, + "step": 20289 + }, + { + "epoch": 2.228201186031188, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4345791339874268, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.721354067325592, + "num_tokens": 525057137.0, + "step": 20290 + }, + { + "epoch": 2.228311003733802, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.510697841644287, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7280619144439697, + "num_tokens": 525083368.0, + "step": 20291 + }, + { + "epoch": 2.2284208214364156, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.49808931350708, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7258681654930115, + "num_tokens": 525106564.0, + "step": 20292 + }, + { + "epoch": 2.2285306391390294, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.3892698287963867, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7369629144668579, + "num_tokens": 525131346.0, + "step": 20293 + }, + { + "epoch": 2.2286404568416427, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4415054321289062, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.728013277053833, + "num_tokens": 525153578.0, + "step": 20294 + }, + { + "epoch": 2.2287502745442564, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.2635860443115234, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7367871403694153, + "num_tokens": 525180802.0, + "step": 20295 + }, + { + "epoch": 2.22886009224687, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.318091630935669, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7152537107467651, + "num_tokens": 525207531.0, + "step": 20296 + }, + { + "epoch": 2.228969909949484, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.180217981338501, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7420681715011597, + "num_tokens": 525235532.0, + "step": 20297 + }, + { + "epoch": 2.2290797276520977, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.5334644317626953, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7157970666885376, + "num_tokens": 525260748.0, + "step": 20298 + }, + { + "epoch": 2.229189545354711, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.241398572921753, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7264933586120605, + "num_tokens": 525290637.0, + "step": 20299 + }, + { + "epoch": 2.229299363057325, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.6873724460601807, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7310544848442078, + "num_tokens": 525311476.0, + "step": 20300 + }, + { + "epoch": 2.2294091807599385, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.359469175338745, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.742745041847229, + "num_tokens": 525336307.0, + "step": 20301 + }, + { + "epoch": 2.2295189984625523, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.4522383213043213, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7313556671142578, + "num_tokens": 525360801.0, + "step": 20302 + }, + { + "epoch": 2.2296288161651656, + "ewc_loss": 1.2874603271484375e-05, + "grad_norm": 2.7758095264434814, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7300083637237549, + "num_tokens": 525381272.0, + "step": 20303 + }, + { + "epoch": 2.2297386338677794, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.44054913520813, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7134615182876587, + "num_tokens": 525405143.0, + "step": 20304 + }, + { + "epoch": 2.229848451570393, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.3010449409484863, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7233920693397522, + "num_tokens": 525432417.0, + "step": 20305 + }, + { + "epoch": 2.229958269273007, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2778260707855225, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7272863984107971, + "num_tokens": 525458740.0, + "step": 20306 + }, + { + "epoch": 2.2300680869756206, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.437344789505005, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.728279709815979, + "num_tokens": 525483970.0, + "step": 20307 + }, + { + "epoch": 2.230177904678234, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.3168842792510986, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7316231727600098, + "num_tokens": 525511566.0, + "step": 20308 + }, + { + "epoch": 2.2302877223808477, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.497873067855835, + "learning_rate": 1e-06, + "loss": 0.791, + "mean_token_accuracy": 0.7506693601608276, + "num_tokens": 525534404.0, + "step": 20309 + }, + { + "epoch": 2.2303975400834615, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4362099170684814, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7256431579589844, + "num_tokens": 525559670.0, + "step": 20310 + }, + { + "epoch": 2.2305073577860752, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.5310769081115723, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7469052672386169, + "num_tokens": 525582828.0, + "step": 20311 + }, + { + "epoch": 2.230617175488689, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.57108211517334, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7353727221488953, + "num_tokens": 525605437.0, + "step": 20312 + }, + { + "epoch": 2.2307269931913023, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4877960681915283, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7470300793647766, + "num_tokens": 525628750.0, + "step": 20313 + }, + { + "epoch": 2.230836810893916, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4071500301361084, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7322795391082764, + "num_tokens": 525652681.0, + "step": 20314 + }, + { + "epoch": 2.23094662859653, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.511352777481079, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7459172010421753, + "num_tokens": 525675705.0, + "step": 20315 + }, + { + "epoch": 2.2310564462991436, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2681965827941895, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7357032895088196, + "num_tokens": 525702444.0, + "step": 20316 + }, + { + "epoch": 2.231166264001757, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4958250522613525, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7489976286888123, + "num_tokens": 525725438.0, + "step": 20317 + }, + { + "epoch": 2.2312760817043706, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.0683255195617676, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7324258089065552, + "num_tokens": 525755652.0, + "step": 20318 + }, + { + "epoch": 2.2313858994069844, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2949254512786865, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7388615012168884, + "num_tokens": 525781826.0, + "step": 20319 + }, + { + "epoch": 2.231495717109598, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 1.9601337909698486, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.7079867124557495, + "num_tokens": 525820208.0, + "step": 20320 + }, + { + "epoch": 2.231605534812212, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4476757049560547, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7392151355743408, + "num_tokens": 525841822.0, + "step": 20321 + }, + { + "epoch": 2.2317153525148252, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.145320177078247, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.7065029740333557, + "num_tokens": 525873699.0, + "step": 20322 + }, + { + "epoch": 2.231825170217439, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.6712207794189453, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7187241911888123, + "num_tokens": 525896504.0, + "step": 20323 + }, + { + "epoch": 2.2319349879200527, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.287933826446533, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7470771670341492, + "num_tokens": 525923052.0, + "step": 20324 + }, + { + "epoch": 2.2320448056226665, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.35611629486084, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7229306697845459, + "num_tokens": 525950202.0, + "step": 20325 + }, + { + "epoch": 2.23215462332528, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.025657892227173, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.7020542025566101, + "num_tokens": 525983447.0, + "step": 20326 + }, + { + "epoch": 2.2322644410278936, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.3365440368652344, + "learning_rate": 1e-06, + "loss": 0.9465, + "mean_token_accuracy": 0.7031251192092896, + "num_tokens": 526010645.0, + "step": 20327 + }, + { + "epoch": 2.2323742587305073, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.6302220821380615, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7214884757995605, + "num_tokens": 526033189.0, + "step": 20328 + }, + { + "epoch": 2.232484076433121, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.481560468673706, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7253603339195251, + "num_tokens": 526057076.0, + "step": 20329 + }, + { + "epoch": 2.232593894135735, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.18955135345459, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7118908166885376, + "num_tokens": 526088194.0, + "step": 20330 + }, + { + "epoch": 2.232703711838348, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.423395872116089, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7112064361572266, + "num_tokens": 526112889.0, + "step": 20331 + }, + { + "epoch": 2.232813529540962, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.233577251434326, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7329446077346802, + "num_tokens": 526140702.0, + "step": 20332 + }, + { + "epoch": 2.2329233472435757, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.077200412750244, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.6994885206222534, + "num_tokens": 526172674.0, + "step": 20333 + }, + { + "epoch": 2.2330331649461894, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.618929624557495, + "learning_rate": 1e-06, + "loss": 0.78, + "mean_token_accuracy": 0.750643253326416, + "num_tokens": 526194778.0, + "step": 20334 + }, + { + "epoch": 2.233142982648803, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.402249336242676, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7269541025161743, + "num_tokens": 526221856.0, + "step": 20335 + }, + { + "epoch": 2.2332528003514165, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.5008490085601807, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7414206862449646, + "num_tokens": 526245796.0, + "step": 20336 + }, + { + "epoch": 2.2333626180540302, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.38651442527771, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7420488595962524, + "num_tokens": 526269914.0, + "step": 20337 + }, + { + "epoch": 2.233472435756644, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4624252319335938, + "learning_rate": 1e-06, + "loss": 0.7561, + "mean_token_accuracy": 0.7638213634490967, + "num_tokens": 526292531.0, + "step": 20338 + }, + { + "epoch": 2.2335822534592578, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.744361400604248, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.744454026222229, + "num_tokens": 526311733.0, + "step": 20339 + }, + { + "epoch": 2.2336920711618715, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.169558048248291, + "learning_rate": 1e-06, + "loss": 0.7579, + "mean_token_accuracy": 0.7577123045921326, + "num_tokens": 526338515.0, + "step": 20340 + }, + { + "epoch": 2.233801888864485, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1042773723602295, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7473092079162598, + "num_tokens": 526371474.0, + "step": 20341 + }, + { + "epoch": 2.2339117065670986, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.895808219909668, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.743828535079956, + "num_tokens": 526390129.0, + "step": 20342 + }, + { + "epoch": 2.2340215242697123, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.286836624145508, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7373017072677612, + "num_tokens": 526417402.0, + "step": 20343 + }, + { + "epoch": 2.234131341972326, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.4217188358306885, + "learning_rate": 1e-06, + "loss": 0.7763, + "mean_token_accuracy": 0.7594995498657227, + "num_tokens": 526440645.0, + "step": 20344 + }, + { + "epoch": 2.2342411596749394, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2790169715881348, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7589155435562134, + "num_tokens": 526466874.0, + "step": 20345 + }, + { + "epoch": 2.234350977377553, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.501188278198242, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7354859709739685, + "num_tokens": 526490122.0, + "step": 20346 + }, + { + "epoch": 2.234460795080167, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1749165058135986, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7100749611854553, + "num_tokens": 526517745.0, + "step": 20347 + }, + { + "epoch": 2.2345706127827807, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.586355209350586, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7465612888336182, + "num_tokens": 526542473.0, + "step": 20348 + }, + { + "epoch": 2.2346804304853944, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.392601728439331, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7078037261962891, + "num_tokens": 526567896.0, + "step": 20349 + }, + { + "epoch": 2.2347902481880078, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.7187399864196777, + "learning_rate": 1e-06, + "loss": 0.7712, + "mean_token_accuracy": 0.7494506239891052, + "num_tokens": 526588527.0, + "step": 20350 + }, + { + "epoch": 2.2349000658906215, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.580911874771118, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7507656216621399, + "num_tokens": 526610904.0, + "step": 20351 + }, + { + "epoch": 2.2350098835932353, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2473316192626953, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7356307506561279, + "num_tokens": 526638307.0, + "step": 20352 + }, + { + "epoch": 2.235119701295849, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.6520426273345947, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.7513802647590637, + "num_tokens": 526657833.0, + "step": 20353 + }, + { + "epoch": 2.2352295189984623, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1277618408203125, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7158262729644775, + "num_tokens": 526688820.0, + "step": 20354 + }, + { + "epoch": 2.235339336701076, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.828709840774536, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7326897978782654, + "num_tokens": 526712067.0, + "step": 20355 + }, + { + "epoch": 2.23544915440369, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.544931650161743, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7429811954498291, + "num_tokens": 526734605.0, + "step": 20356 + }, + { + "epoch": 2.2355589721063036, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.236204147338867, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7180105447769165, + "num_tokens": 526763488.0, + "step": 20357 + }, + { + "epoch": 2.2356687898089174, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.3724122047424316, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7328008413314819, + "num_tokens": 526788452.0, + "step": 20358 + }, + { + "epoch": 2.2357786075115307, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.451158046722412, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7318801283836365, + "num_tokens": 526812108.0, + "step": 20359 + }, + { + "epoch": 2.2358884252141444, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.478079319000244, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7393940091133118, + "num_tokens": 526836705.0, + "step": 20360 + }, + { + "epoch": 2.235998242916758, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.3965530395507812, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7521306276321411, + "num_tokens": 526860268.0, + "step": 20361 + }, + { + "epoch": 2.236108060619372, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.2126669883728027, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7293441295623779, + "num_tokens": 526888612.0, + "step": 20362 + }, + { + "epoch": 2.2362178783219857, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1213765144348145, + "learning_rate": 1e-06, + "loss": 0.7546, + "mean_token_accuracy": 0.7588565349578857, + "num_tokens": 526918780.0, + "step": 20363 + }, + { + "epoch": 2.236327696024599, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.1560757160186768, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7417120337486267, + "num_tokens": 526947039.0, + "step": 20364 + }, + { + "epoch": 2.236437513727213, + "ewc_loss": 1.2993812561035156e-05, + "grad_norm": 2.6540238857269287, + "learning_rate": 1e-06, + "loss": 0.7225, + "mean_token_accuracy": 0.7660319805145264, + "num_tokens": 526967088.0, + "step": 20365 + }, + { + "epoch": 2.2365473314298265, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.66507887840271, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7310658693313599, + "num_tokens": 526988400.0, + "step": 20366 + }, + { + "epoch": 2.2366571491324403, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3777530193328857, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7186496257781982, + "num_tokens": 527015572.0, + "step": 20367 + }, + { + "epoch": 2.2367669668350536, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2808609008789062, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7443015575408936, + "num_tokens": 527041790.0, + "step": 20368 + }, + { + "epoch": 2.2368767845376674, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4662275314331055, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7357006072998047, + "num_tokens": 527066764.0, + "step": 20369 + }, + { + "epoch": 2.236986602240281, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4226434230804443, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7338319420814514, + "num_tokens": 527092131.0, + "step": 20370 + }, + { + "epoch": 2.237096419942895, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4871721267700195, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7182180881500244, + "num_tokens": 527115857.0, + "step": 20371 + }, + { + "epoch": 2.2372062376455086, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2070119380950928, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7190288305282593, + "num_tokens": 527146876.0, + "step": 20372 + }, + { + "epoch": 2.237316055348122, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.650247573852539, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7480432391166687, + "num_tokens": 527167614.0, + "step": 20373 + }, + { + "epoch": 2.2374258730507357, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4967052936553955, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7331392168998718, + "num_tokens": 527191952.0, + "step": 20374 + }, + { + "epoch": 2.2375356907533495, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.612769842147827, + "learning_rate": 1e-06, + "loss": 0.7535, + "mean_token_accuracy": 0.7594200372695923, + "num_tokens": 527211500.0, + "step": 20375 + }, + { + "epoch": 2.237645508455963, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4434657096862793, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7417163848876953, + "num_tokens": 527236686.0, + "step": 20376 + }, + { + "epoch": 2.2377553261585765, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4812419414520264, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7399176955223083, + "num_tokens": 527260395.0, + "step": 20377 + }, + { + "epoch": 2.2378651438611903, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.6041126251220703, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7313103079795837, + "num_tokens": 527282151.0, + "step": 20378 + }, + { + "epoch": 2.237974961563804, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5960047245025635, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7305238246917725, + "num_tokens": 527306096.0, + "step": 20379 + }, + { + "epoch": 2.238084779266418, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.796607494354248, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7325130701065063, + "num_tokens": 527325235.0, + "step": 20380 + }, + { + "epoch": 2.2381945969690316, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.6064453125, + "learning_rate": 1e-06, + "loss": 0.7732, + "mean_token_accuracy": 0.7520481944084167, + "num_tokens": 527347389.0, + "step": 20381 + }, + { + "epoch": 2.238304414671645, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4257094860076904, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.715597927570343, + "num_tokens": 527373511.0, + "step": 20382 + }, + { + "epoch": 2.2384142323742586, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.1006274223327637, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7312145233154297, + "num_tokens": 527402166.0, + "step": 20383 + }, + { + "epoch": 2.2385240500768724, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.7645418643951416, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7408269047737122, + "num_tokens": 527422959.0, + "step": 20384 + }, + { + "epoch": 2.238633867779486, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.337733268737793, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7244845628738403, + "num_tokens": 527450587.0, + "step": 20385 + }, + { + "epoch": 2.2387436854821, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.328749418258667, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7455697059631348, + "num_tokens": 527476126.0, + "step": 20386 + }, + { + "epoch": 2.238853503184713, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.6797726154327393, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7330299615859985, + "num_tokens": 527496509.0, + "step": 20387 + }, + { + "epoch": 2.238963320887327, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5954205989837646, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7426233291625977, + "num_tokens": 527518868.0, + "step": 20388 + }, + { + "epoch": 2.2390731385899407, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3379697799682617, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7164225578308105, + "num_tokens": 527547509.0, + "step": 20389 + }, + { + "epoch": 2.2391829562925545, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.227407693862915, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7269539833068848, + "num_tokens": 527579173.0, + "step": 20390 + }, + { + "epoch": 2.2392927739951682, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.768174648284912, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7513918280601501, + "num_tokens": 527599196.0, + "step": 20391 + }, + { + "epoch": 2.2394025916977816, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.0912253856658936, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.737013578414917, + "num_tokens": 527629669.0, + "step": 20392 + }, + { + "epoch": 2.2395124094003953, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.1150529384613037, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.731425404548645, + "num_tokens": 527660922.0, + "step": 20393 + }, + { + "epoch": 2.239622227103009, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.724698543548584, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7486990690231323, + "num_tokens": 527680675.0, + "step": 20394 + }, + { + "epoch": 2.239732044805623, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.334808111190796, + "learning_rate": 1e-06, + "loss": 0.7708, + "mean_token_accuracy": 0.7473810911178589, + "num_tokens": 527703669.0, + "step": 20395 + }, + { + "epoch": 2.239841862508236, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2967681884765625, + "learning_rate": 1e-06, + "loss": 0.9714, + "mean_token_accuracy": 0.7008810043334961, + "num_tokens": 527735082.0, + "step": 20396 + }, + { + "epoch": 2.23995168021085, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.393763780593872, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.7428282499313354, + "num_tokens": 527759437.0, + "step": 20397 + }, + { + "epoch": 2.2400614979134637, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.1974265575408936, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7028298377990723, + "num_tokens": 527789906.0, + "step": 20398 + }, + { + "epoch": 2.2401713156160774, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5185680389404297, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7433561682701111, + "num_tokens": 527812118.0, + "step": 20399 + }, + { + "epoch": 2.240281133318691, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4717931747436523, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7214851379394531, + "num_tokens": 527837353.0, + "step": 20400 + }, + { + "epoch": 2.2403909510213045, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.0970373153686523, + "learning_rate": 1e-06, + "loss": 0.9387, + "mean_token_accuracy": 0.7118227481842041, + "num_tokens": 527868683.0, + "step": 20401 + }, + { + "epoch": 2.2405007687239182, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3783984184265137, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7431268692016602, + "num_tokens": 527892039.0, + "step": 20402 + }, + { + "epoch": 2.240610586426532, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3893542289733887, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7237122058868408, + "num_tokens": 527918095.0, + "step": 20403 + }, + { + "epoch": 2.2407204041291457, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.700310707092285, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7216855883598328, + "num_tokens": 527938217.0, + "step": 20404 + }, + { + "epoch": 2.240830221831759, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2100436687469482, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.722884476184845, + "num_tokens": 527965025.0, + "step": 20405 + }, + { + "epoch": 2.240940039534373, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4478609561920166, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7134467363357544, + "num_tokens": 527989612.0, + "step": 20406 + }, + { + "epoch": 2.2410498572369866, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.6460349559783936, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7560523152351379, + "num_tokens": 528009851.0, + "step": 20407 + }, + { + "epoch": 2.2411596749396003, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 1.9841605424880981, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7153047323226929, + "num_tokens": 528043087.0, + "step": 20408 + }, + { + "epoch": 2.241269492642214, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.209974527359009, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7092400789260864, + "num_tokens": 528072194.0, + "step": 20409 + }, + { + "epoch": 2.2413793103448274, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3240127563476562, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7166748642921448, + "num_tokens": 528098834.0, + "step": 20410 + }, + { + "epoch": 2.241489128047441, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5209124088287354, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7399523258209229, + "num_tokens": 528122187.0, + "step": 20411 + }, + { + "epoch": 2.241598945750055, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.0905566215515137, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7071975469589233, + "num_tokens": 528154044.0, + "step": 20412 + }, + { + "epoch": 2.2417087634526687, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5071208477020264, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7290011048316956, + "num_tokens": 528175938.0, + "step": 20413 + }, + { + "epoch": 2.2418185811552824, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2263309955596924, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7079825401306152, + "num_tokens": 528204258.0, + "step": 20414 + }, + { + "epoch": 2.2419283988578957, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.075906276702881, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7204378843307495, + "num_tokens": 528237244.0, + "step": 20415 + }, + { + "epoch": 2.2420382165605095, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.322439432144165, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7184750437736511, + "num_tokens": 528264249.0, + "step": 20416 + }, + { + "epoch": 2.2421480342631233, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.698552131652832, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7388078570365906, + "num_tokens": 528284839.0, + "step": 20417 + }, + { + "epoch": 2.242257851965737, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.38832688331604, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7199796438217163, + "num_tokens": 528310191.0, + "step": 20418 + }, + { + "epoch": 2.2423676696683508, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.306086540222168, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.728725254535675, + "num_tokens": 528338473.0, + "step": 20419 + }, + { + "epoch": 2.242477487370964, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.1100094318389893, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7246189713478088, + "num_tokens": 528368942.0, + "step": 20420 + }, + { + "epoch": 2.242587305073578, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.451559543609619, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.7222514152526855, + "num_tokens": 528394289.0, + "step": 20421 + }, + { + "epoch": 2.2426971227761916, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.1998720169067383, + "learning_rate": 1e-06, + "loss": 0.9977, + "mean_token_accuracy": 0.6964170336723328, + "num_tokens": 528424723.0, + "step": 20422 + }, + { + "epoch": 2.2428069404788054, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3405563831329346, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7065289616584778, + "num_tokens": 528451787.0, + "step": 20423 + }, + { + "epoch": 2.2429167581814187, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.120079755783081, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7129042148590088, + "num_tokens": 528484557.0, + "step": 20424 + }, + { + "epoch": 2.2430265758840324, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.379258394241333, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7187669277191162, + "num_tokens": 528511132.0, + "step": 20425 + }, + { + "epoch": 2.243136393586646, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2371914386749268, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.7298340201377869, + "num_tokens": 528539163.0, + "step": 20426 + }, + { + "epoch": 2.24324621128926, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2423553466796875, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7394065260887146, + "num_tokens": 528567797.0, + "step": 20427 + }, + { + "epoch": 2.2433560289918737, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.268561601638794, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7151645421981812, + "num_tokens": 528595182.0, + "step": 20428 + }, + { + "epoch": 2.243465846694487, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.331681728363037, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7424302101135254, + "num_tokens": 528620815.0, + "step": 20429 + }, + { + "epoch": 2.2435756643971008, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.0961570739746094, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7190068960189819, + "num_tokens": 528653451.0, + "step": 20430 + }, + { + "epoch": 2.2436854820997145, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.127176284790039, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7218955755233765, + "num_tokens": 528682954.0, + "step": 20431 + }, + { + "epoch": 2.2437952998023283, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.326442241668701, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7122470140457153, + "num_tokens": 528713588.0, + "step": 20432 + }, + { + "epoch": 2.2439051175049416, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.369858980178833, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7433093190193176, + "num_tokens": 528738256.0, + "step": 20433 + }, + { + "epoch": 2.2440149352075554, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3105907440185547, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7342541217803955, + "num_tokens": 528764360.0, + "step": 20434 + }, + { + "epoch": 2.244124752910169, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.319080352783203, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7265080213546753, + "num_tokens": 528791968.0, + "step": 20435 + }, + { + "epoch": 2.244234570612783, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2453534603118896, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7162388563156128, + "num_tokens": 528818306.0, + "step": 20436 + }, + { + "epoch": 2.2443443883153966, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4069554805755615, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7441372871398926, + "num_tokens": 528844044.0, + "step": 20437 + }, + { + "epoch": 2.24445420601801, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.375396728515625, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7500620484352112, + "num_tokens": 528869218.0, + "step": 20438 + }, + { + "epoch": 2.2445640237206237, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2759487628936768, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7412862777709961, + "num_tokens": 528895694.0, + "step": 20439 + }, + { + "epoch": 2.2446738414232374, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4328088760375977, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7145897150039673, + "num_tokens": 528924203.0, + "step": 20440 + }, + { + "epoch": 2.244783659125851, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.393749713897705, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.731417179107666, + "num_tokens": 528948436.0, + "step": 20441 + }, + { + "epoch": 2.244893476828465, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.602836847305298, + "learning_rate": 1e-06, + "loss": 0.8036, + "mean_token_accuracy": 0.7461986541748047, + "num_tokens": 528969233.0, + "step": 20442 + }, + { + "epoch": 2.2450032945310783, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.330531597137451, + "learning_rate": 1e-06, + "loss": 0.7439, + "mean_token_accuracy": 0.759445071220398, + "num_tokens": 528993458.0, + "step": 20443 + }, + { + "epoch": 2.245113112233692, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4411227703094482, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7355935573577881, + "num_tokens": 529019352.0, + "step": 20444 + }, + { + "epoch": 2.245222929936306, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.507375717163086, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.723310112953186, + "num_tokens": 529043241.0, + "step": 20445 + }, + { + "epoch": 2.2453327476389195, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3048596382141113, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7217853665351868, + "num_tokens": 529071459.0, + "step": 20446 + }, + { + "epoch": 2.245442565341533, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2723772525787354, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7283939719200134, + "num_tokens": 529097966.0, + "step": 20447 + }, + { + "epoch": 2.2455523830441466, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4036505222320557, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7194199562072754, + "num_tokens": 529123413.0, + "step": 20448 + }, + { + "epoch": 2.2456622007467604, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3445487022399902, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7088310718536377, + "num_tokens": 529151037.0, + "step": 20449 + }, + { + "epoch": 2.245772018449374, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5574862957000732, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7355411052703857, + "num_tokens": 529173773.0, + "step": 20450 + }, + { + "epoch": 2.245881836151988, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.5935566425323486, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7502926588058472, + "num_tokens": 529195203.0, + "step": 20451 + }, + { + "epoch": 2.245991653854601, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4720678329467773, + "learning_rate": 1e-06, + "loss": 0.7644, + "mean_token_accuracy": 0.747684895992279, + "num_tokens": 529218317.0, + "step": 20452 + }, + { + "epoch": 2.246101471557215, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.406916618347168, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7441989779472351, + "num_tokens": 529242615.0, + "step": 20453 + }, + { + "epoch": 2.2462112892598287, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3381967544555664, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7297101020812988, + "num_tokens": 529268984.0, + "step": 20454 + }, + { + "epoch": 2.2463211069624425, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.35842227935791, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7367771863937378, + "num_tokens": 529293638.0, + "step": 20455 + }, + { + "epoch": 2.246430924665056, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4976003170013428, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7352521419525146, + "num_tokens": 529317290.0, + "step": 20456 + }, + { + "epoch": 2.2465407423676695, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.7322635650634766, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.746272087097168, + "num_tokens": 529336431.0, + "step": 20457 + }, + { + "epoch": 2.2466505600702833, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.174339771270752, + "learning_rate": 1e-06, + "loss": 0.9407, + "mean_token_accuracy": 0.7043224573135376, + "num_tokens": 529366266.0, + "step": 20458 + }, + { + "epoch": 2.246760377772897, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.859020709991455, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.751482367515564, + "num_tokens": 529383851.0, + "step": 20459 + }, + { + "epoch": 2.246870195475511, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.3791184425354004, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.734329104423523, + "num_tokens": 529409177.0, + "step": 20460 + }, + { + "epoch": 2.246980013178124, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.093064785003662, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7269597053527832, + "num_tokens": 529439888.0, + "step": 20461 + }, + { + "epoch": 2.247089830880738, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.533139228820801, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7501238584518433, + "num_tokens": 529462820.0, + "step": 20462 + }, + { + "epoch": 2.2471996485833516, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.4733662605285645, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.736780047416687, + "num_tokens": 529486913.0, + "step": 20463 + }, + { + "epoch": 2.2473094662859654, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2323203086853027, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7381499409675598, + "num_tokens": 529513992.0, + "step": 20464 + }, + { + "epoch": 2.247419283988579, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.381152868270874, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.738167941570282, + "num_tokens": 529540114.0, + "step": 20465 + }, + { + "epoch": 2.2475291016911925, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.2557148933410645, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7348005771636963, + "num_tokens": 529566298.0, + "step": 20466 + }, + { + "epoch": 2.2476389193938062, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.36566162109375, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7283727526664734, + "num_tokens": 529591700.0, + "step": 20467 + }, + { + "epoch": 2.24774873709642, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.696200370788574, + "learning_rate": 1e-06, + "loss": 0.709, + "mean_token_accuracy": 0.7678871750831604, + "num_tokens": 529612780.0, + "step": 20468 + }, + { + "epoch": 2.2478585547990337, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.201759099960327, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7163035273551941, + "num_tokens": 529641782.0, + "step": 20469 + }, + { + "epoch": 2.2479683725016475, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.249596357345581, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7289979457855225, + "num_tokens": 529668437.0, + "step": 20470 + }, + { + "epoch": 2.248078190204261, + "ewc_loss": 1.3053417205810547e-05, + "grad_norm": 2.52103590965271, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7162636518478394, + "num_tokens": 529692229.0, + "step": 20471 + }, + { + "epoch": 2.2481880079068746, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1229748725891113, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.73690265417099, + "num_tokens": 529722955.0, + "step": 20472 + }, + { + "epoch": 2.2482978256094883, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2629058361053467, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7400025725364685, + "num_tokens": 529750229.0, + "step": 20473 + }, + { + "epoch": 2.248407643312102, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2855186462402344, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7225303053855896, + "num_tokens": 529777962.0, + "step": 20474 + }, + { + "epoch": 2.2485174610147154, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2358829975128174, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7314472198486328, + "num_tokens": 529805968.0, + "step": 20475 + }, + { + "epoch": 2.248627278717329, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.495496988296509, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7526285648345947, + "num_tokens": 529828306.0, + "step": 20476 + }, + { + "epoch": 2.248737096419943, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3332347869873047, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7236288785934448, + "num_tokens": 529855477.0, + "step": 20477 + }, + { + "epoch": 2.2488469141225567, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4807968139648438, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7482240796089172, + "num_tokens": 529879914.0, + "step": 20478 + }, + { + "epoch": 2.2489567318251704, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.504241943359375, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7404436469078064, + "num_tokens": 529902911.0, + "step": 20479 + }, + { + "epoch": 2.2490665495277837, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.054831027984619, + "learning_rate": 1e-06, + "loss": 0.9486, + "mean_token_accuracy": 0.7074806690216064, + "num_tokens": 529936411.0, + "step": 20480 + }, + { + "epoch": 2.2491763672303975, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.6062965393066406, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7363314628601074, + "num_tokens": 529957626.0, + "step": 20481 + }, + { + "epoch": 2.2492861849330112, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5794904232025146, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7280986309051514, + "num_tokens": 529981233.0, + "step": 20482 + }, + { + "epoch": 2.249396002635625, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.346914768218994, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7357175946235657, + "num_tokens": 530008794.0, + "step": 20483 + }, + { + "epoch": 2.2495058203382383, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2370965480804443, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7241066098213196, + "num_tokens": 530036435.0, + "step": 20484 + }, + { + "epoch": 2.249615638040852, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3460471630096436, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7366782426834106, + "num_tokens": 530060470.0, + "step": 20485 + }, + { + "epoch": 2.249725455743466, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.29166316986084, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.718063235282898, + "num_tokens": 530090529.0, + "step": 20486 + }, + { + "epoch": 2.2498352734460796, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.315614938735962, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7373634576797485, + "num_tokens": 530117324.0, + "step": 20487 + }, + { + "epoch": 2.2499450911486933, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4430928230285645, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7363565564155579, + "num_tokens": 530139881.0, + "step": 20488 + }, + { + "epoch": 2.2500549088513067, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.7351789474487305, + "learning_rate": 1e-06, + "loss": 0.7106, + "mean_token_accuracy": 0.7593945264816284, + "num_tokens": 530159955.0, + "step": 20489 + }, + { + "epoch": 2.2501647265539204, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4171323776245117, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7274206280708313, + "num_tokens": 530183815.0, + "step": 20490 + }, + { + "epoch": 2.250274544256534, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4002022743225098, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7324459552764893, + "num_tokens": 530209864.0, + "step": 20491 + }, + { + "epoch": 2.250384361959148, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.404667854309082, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.739872932434082, + "num_tokens": 530236326.0, + "step": 20492 + }, + { + "epoch": 2.2504941796617617, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.6553795337677, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7359918355941772, + "num_tokens": 530256498.0, + "step": 20493 + }, + { + "epoch": 2.250603997364375, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.409959554672241, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7310256361961365, + "num_tokens": 530280928.0, + "step": 20494 + }, + { + "epoch": 2.2507138150669888, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.593698501586914, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7337135076522827, + "num_tokens": 530301881.0, + "step": 20495 + }, + { + "epoch": 2.2508236327696025, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2111704349517822, + "learning_rate": 1e-06, + "loss": 0.9505, + "mean_token_accuracy": 0.7059445977210999, + "num_tokens": 530333726.0, + "step": 20496 + }, + { + "epoch": 2.2509334504722163, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.353262186050415, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7455372214317322, + "num_tokens": 530360414.0, + "step": 20497 + }, + { + "epoch": 2.25104326817483, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.273705005645752, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7123098969459534, + "num_tokens": 530387789.0, + "step": 20498 + }, + { + "epoch": 2.2511530858774433, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5864956378936768, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7261806726455688, + "num_tokens": 530410661.0, + "step": 20499 + }, + { + "epoch": 2.251262903580057, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2293548583984375, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7246664762496948, + "num_tokens": 530438675.0, + "step": 20500 + }, + { + "epoch": 2.251372721282671, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.246203660964966, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7283533215522766, + "num_tokens": 530466771.0, + "step": 20501 + }, + { + "epoch": 2.2514825389852846, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4081544876098633, + "learning_rate": 1e-06, + "loss": 0.7192, + "mean_token_accuracy": 0.7652245163917542, + "num_tokens": 530489353.0, + "step": 20502 + }, + { + "epoch": 2.251592356687898, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4949944019317627, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7282658219337463, + "num_tokens": 530514382.0, + "step": 20503 + }, + { + "epoch": 2.2517021743905117, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.37876558303833, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7202286720275879, + "num_tokens": 530539874.0, + "step": 20504 + }, + { + "epoch": 2.2518119920931254, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4312405586242676, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7394112348556519, + "num_tokens": 530564468.0, + "step": 20505 + }, + { + "epoch": 2.251921809795739, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1463422775268555, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7431110739707947, + "num_tokens": 530592630.0, + "step": 20506 + }, + { + "epoch": 2.2520316274983525, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.280778408050537, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7198258638381958, + "num_tokens": 530621981.0, + "step": 20507 + }, + { + "epoch": 2.2521414452009663, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5635409355163574, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7388322353363037, + "num_tokens": 530646013.0, + "step": 20508 + }, + { + "epoch": 2.25225126290358, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4802141189575195, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7432224750518799, + "num_tokens": 530669801.0, + "step": 20509 + }, + { + "epoch": 2.2523610806061938, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2926747798919678, + "learning_rate": 1e-06, + "loss": 0.7495, + "mean_token_accuracy": 0.7653086185455322, + "num_tokens": 530695794.0, + "step": 20510 + }, + { + "epoch": 2.2524708983088075, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2420759201049805, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7276507616043091, + "num_tokens": 530725142.0, + "step": 20511 + }, + { + "epoch": 2.252580716011421, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2057788372039795, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7399835586547852, + "num_tokens": 530754745.0, + "step": 20512 + }, + { + "epoch": 2.2526905337140346, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5136728286743164, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7303581237792969, + "num_tokens": 530778615.0, + "step": 20513 + }, + { + "epoch": 2.2528003514166484, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3021256923675537, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.742110013961792, + "num_tokens": 530806303.0, + "step": 20514 + }, + { + "epoch": 2.252910169119262, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1238157749176025, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7326409220695496, + "num_tokens": 530836596.0, + "step": 20515 + }, + { + "epoch": 2.253019986821876, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3473987579345703, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7279213070869446, + "num_tokens": 530862211.0, + "step": 20516 + }, + { + "epoch": 2.253129804524489, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2006964683532715, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7083475589752197, + "num_tokens": 530892342.0, + "step": 20517 + }, + { + "epoch": 2.253239622227103, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.129211902618408, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7014427185058594, + "num_tokens": 530924070.0, + "step": 20518 + }, + { + "epoch": 2.2533494399297167, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.243316888809204, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7048545479774475, + "num_tokens": 530953116.0, + "step": 20519 + }, + { + "epoch": 2.2534592576323305, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.219756841659546, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7311967611312866, + "num_tokens": 530981887.0, + "step": 20520 + }, + { + "epoch": 2.253569075334944, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.219249725341797, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7336721420288086, + "num_tokens": 531009953.0, + "step": 20521 + }, + { + "epoch": 2.2536788930375575, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1188974380493164, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7282137870788574, + "num_tokens": 531042200.0, + "step": 20522 + }, + { + "epoch": 2.2537887107401713, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.21536922454834, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.750379204750061, + "num_tokens": 531071165.0, + "step": 20523 + }, + { + "epoch": 2.253898528442785, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4724888801574707, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7469016909599304, + "num_tokens": 531095433.0, + "step": 20524 + }, + { + "epoch": 2.254008346145399, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3466854095458984, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7283602952957153, + "num_tokens": 531120481.0, + "step": 20525 + }, + { + "epoch": 2.254118163848012, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.673837184906006, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7449996471405029, + "num_tokens": 531143677.0, + "step": 20526 + }, + { + "epoch": 2.254227981550626, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2333109378814697, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7159442901611328, + "num_tokens": 531173405.0, + "step": 20527 + }, + { + "epoch": 2.2543377992532396, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.418023109436035, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7105404734611511, + "num_tokens": 531200292.0, + "step": 20528 + }, + { + "epoch": 2.2544476169558534, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.524376392364502, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7348265647888184, + "num_tokens": 531223427.0, + "step": 20529 + }, + { + "epoch": 2.2545574346584667, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4087045192718506, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7171664237976074, + "num_tokens": 531248597.0, + "step": 20530 + }, + { + "epoch": 2.2546672523610805, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.548002004623413, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7450127601623535, + "num_tokens": 531269512.0, + "step": 20531 + }, + { + "epoch": 2.254777070063694, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.257052183151245, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7221611738204956, + "num_tokens": 531298178.0, + "step": 20532 + }, + { + "epoch": 2.254886887766308, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2557315826416016, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.738994836807251, + "num_tokens": 531325600.0, + "step": 20533 + }, + { + "epoch": 2.2549967054689217, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.374055862426758, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7288992404937744, + "num_tokens": 531351729.0, + "step": 20534 + }, + { + "epoch": 2.255106523171535, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.492812156677246, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7345077991485596, + "num_tokens": 531376139.0, + "step": 20535 + }, + { + "epoch": 2.255216340874149, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4845097064971924, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.722136378288269, + "num_tokens": 531402112.0, + "step": 20536 + }, + { + "epoch": 2.2553261585767626, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.165818214416504, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7206124663352966, + "num_tokens": 531431035.0, + "step": 20537 + }, + { + "epoch": 2.2554359762793763, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2174158096313477, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7231847047805786, + "num_tokens": 531459968.0, + "step": 20538 + }, + { + "epoch": 2.25554579398199, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5317702293395996, + "learning_rate": 1e-06, + "loss": 0.8694, + "mean_token_accuracy": 0.7195979952812195, + "num_tokens": 531482442.0, + "step": 20539 + }, + { + "epoch": 2.2556556116846034, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.356409788131714, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7474004030227661, + "num_tokens": 531506980.0, + "step": 20540 + }, + { + "epoch": 2.255765429387217, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1956787109375, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7348096966743469, + "num_tokens": 531536863.0, + "step": 20541 + }, + { + "epoch": 2.255875247089831, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4214797019958496, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7419452667236328, + "num_tokens": 531561662.0, + "step": 20542 + }, + { + "epoch": 2.2559850647924446, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.7587430477142334, + "learning_rate": 1e-06, + "loss": 0.7743, + "mean_token_accuracy": 0.7507436871528625, + "num_tokens": 531583732.0, + "step": 20543 + }, + { + "epoch": 2.2560948824950584, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.192397117614746, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7550647258758545, + "num_tokens": 531611244.0, + "step": 20544 + }, + { + "epoch": 2.2562047001976717, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.9894843101501465, + "learning_rate": 1e-06, + "loss": 0.7214, + "mean_token_accuracy": 0.7643305063247681, + "num_tokens": 531629118.0, + "step": 20545 + }, + { + "epoch": 2.2563145179002855, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1147727966308594, + "learning_rate": 1e-06, + "loss": 0.9901, + "mean_token_accuracy": 0.6947252750396729, + "num_tokens": 531660232.0, + "step": 20546 + }, + { + "epoch": 2.2564243356028992, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1740691661834717, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7399173974990845, + "num_tokens": 531691246.0, + "step": 20547 + }, + { + "epoch": 2.256534153305513, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 1.9244344234466553, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7106106281280518, + "num_tokens": 531727836.0, + "step": 20548 + }, + { + "epoch": 2.2566439710081267, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.686777114868164, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7454502582550049, + "num_tokens": 531748360.0, + "step": 20549 + }, + { + "epoch": 2.25675378871074, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.127972364425659, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.6980181932449341, + "num_tokens": 531782369.0, + "step": 20550 + }, + { + "epoch": 2.256863606413354, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.694303512573242, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7183864712715149, + "num_tokens": 531803350.0, + "step": 20551 + }, + { + "epoch": 2.2569734241159676, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.358128547668457, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7268069982528687, + "num_tokens": 531830245.0, + "step": 20552 + }, + { + "epoch": 2.2570832418185813, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2574431896209717, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7380240559577942, + "num_tokens": 531857075.0, + "step": 20553 + }, + { + "epoch": 2.2571930595211946, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.098151206970215, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7215597629547119, + "num_tokens": 531887055.0, + "step": 20554 + }, + { + "epoch": 2.2573028772238084, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3328022956848145, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7218097448348999, + "num_tokens": 531914390.0, + "step": 20555 + }, + { + "epoch": 2.257412694926422, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1607272624969482, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7294309139251709, + "num_tokens": 531944376.0, + "step": 20556 + }, + { + "epoch": 2.257522512629036, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3879990577697754, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7323638796806335, + "num_tokens": 531970531.0, + "step": 20557 + }, + { + "epoch": 2.2576323303316492, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.350567579269409, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7403826713562012, + "num_tokens": 531997450.0, + "step": 20558 + }, + { + "epoch": 2.257742148034263, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2548751831054688, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7391108274459839, + "num_tokens": 532023100.0, + "step": 20559 + }, + { + "epoch": 2.2578519657368767, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.261312246322632, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7299274206161499, + "num_tokens": 532049239.0, + "step": 20560 + }, + { + "epoch": 2.2579617834394905, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.040844202041626, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7265504598617554, + "num_tokens": 532083272.0, + "step": 20561 + }, + { + "epoch": 2.2580716011421043, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.171900749206543, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7266521453857422, + "num_tokens": 532112550.0, + "step": 20562 + }, + { + "epoch": 2.2581814188447176, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1334750652313232, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7254236936569214, + "num_tokens": 532141991.0, + "step": 20563 + }, + { + "epoch": 2.2582912365473313, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1734912395477295, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7389949560165405, + "num_tokens": 532169669.0, + "step": 20564 + }, + { + "epoch": 2.258401054249945, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5761899948120117, + "learning_rate": 1e-06, + "loss": 0.6583, + "mean_token_accuracy": 0.7843859195709229, + "num_tokens": 532189423.0, + "step": 20565 + }, + { + "epoch": 2.258510871952559, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.539721727371216, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7331123352050781, + "num_tokens": 532213736.0, + "step": 20566 + }, + { + "epoch": 2.2586206896551726, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.201655864715576, + "learning_rate": 1e-06, + "loss": 0.9271, + "mean_token_accuracy": 0.7071398496627808, + "num_tokens": 532242845.0, + "step": 20567 + }, + { + "epoch": 2.258730507357786, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2591936588287354, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7439560890197754, + "num_tokens": 532269830.0, + "step": 20568 + }, + { + "epoch": 2.2588403250603997, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4045379161834717, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.744705080986023, + "num_tokens": 532294985.0, + "step": 20569 + }, + { + "epoch": 2.2589501427630134, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.247110605239868, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7484174966812134, + "num_tokens": 532322489.0, + "step": 20570 + }, + { + "epoch": 2.259059960465627, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4667530059814453, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7300366163253784, + "num_tokens": 532346418.0, + "step": 20571 + }, + { + "epoch": 2.259169778168241, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.5580174922943115, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7494891881942749, + "num_tokens": 532368992.0, + "step": 20572 + }, + { + "epoch": 2.2592795958708543, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.476210117340088, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7195025682449341, + "num_tokens": 532394273.0, + "step": 20573 + }, + { + "epoch": 2.259389413573468, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.7291100025177, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7283973693847656, + "num_tokens": 532416996.0, + "step": 20574 + }, + { + "epoch": 2.2594992312760818, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4955499172210693, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7325087785720825, + "num_tokens": 532442497.0, + "step": 20575 + }, + { + "epoch": 2.2596090489786955, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.428382158279419, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7220814228057861, + "num_tokens": 532469051.0, + "step": 20576 + }, + { + "epoch": 2.259718866681309, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.40558123588562, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7318708896636963, + "num_tokens": 532495620.0, + "step": 20577 + }, + { + "epoch": 2.2598286843839226, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.522043228149414, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7415922284126282, + "num_tokens": 532518777.0, + "step": 20578 + }, + { + "epoch": 2.2599385020865363, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4405131340026855, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7475389242172241, + "num_tokens": 532543082.0, + "step": 20579 + }, + { + "epoch": 2.26004831978915, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.198289155960083, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7237381935119629, + "num_tokens": 532571941.0, + "step": 20580 + }, + { + "epoch": 2.260158137491764, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3631021976470947, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7367995977401733, + "num_tokens": 532596864.0, + "step": 20581 + }, + { + "epoch": 2.260267955194377, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.273052215576172, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7158030271530151, + "num_tokens": 532624412.0, + "step": 20582 + }, + { + "epoch": 2.260377772896991, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.6338305473327637, + "learning_rate": 1e-06, + "loss": 0.7439, + "mean_token_accuracy": 0.7546219229698181, + "num_tokens": 532645362.0, + "step": 20583 + }, + { + "epoch": 2.2604875905996047, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.83048415184021, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7346938252449036, + "num_tokens": 532665259.0, + "step": 20584 + }, + { + "epoch": 2.2605974083022184, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4528865814208984, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7184149026870728, + "num_tokens": 532688492.0, + "step": 20585 + }, + { + "epoch": 2.2607072260048318, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.676309823989868, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7422347068786621, + "num_tokens": 532712808.0, + "step": 20586 + }, + { + "epoch": 2.2608170437074455, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.190110206604004, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7411797046661377, + "num_tokens": 532741960.0, + "step": 20587 + }, + { + "epoch": 2.2609268614100593, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.0491254329681396, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7270137071609497, + "num_tokens": 532775010.0, + "step": 20588 + }, + { + "epoch": 2.261036679112673, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3781850337982178, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7306065559387207, + "num_tokens": 532799610.0, + "step": 20589 + }, + { + "epoch": 2.261146496815287, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.534060478210449, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7415416240692139, + "num_tokens": 532822318.0, + "step": 20590 + }, + { + "epoch": 2.2612563145179, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3546667098999023, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7439309358596802, + "num_tokens": 532848644.0, + "step": 20591 + }, + { + "epoch": 2.261366132220514, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2973790168762207, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7193988561630249, + "num_tokens": 532874857.0, + "step": 20592 + }, + { + "epoch": 2.2614759499231276, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.310088872909546, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.7363406419754028, + "num_tokens": 532901140.0, + "step": 20593 + }, + { + "epoch": 2.2615857676257414, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.119842052459717, + "learning_rate": 1e-06, + "loss": 0.9814, + "mean_token_accuracy": 0.6985786557197571, + "num_tokens": 532933633.0, + "step": 20594 + }, + { + "epoch": 2.261695585328355, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.383007764816284, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.720183789730072, + "num_tokens": 532960356.0, + "step": 20595 + }, + { + "epoch": 2.2618054030309684, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.681297540664673, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7231051921844482, + "num_tokens": 532982102.0, + "step": 20596 + }, + { + "epoch": 2.261915220733582, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3854382038116455, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7383612394332886, + "num_tokens": 533008657.0, + "step": 20597 + }, + { + "epoch": 2.262025038436196, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.0115315914154053, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.716377854347229, + "num_tokens": 533040050.0, + "step": 20598 + }, + { + "epoch": 2.2621348561388097, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.378021717071533, + "learning_rate": 1e-06, + "loss": 0.8965, + "mean_token_accuracy": 0.7191262245178223, + "num_tokens": 533065654.0, + "step": 20599 + }, + { + "epoch": 2.2622446738414235, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4508116245269775, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7487961053848267, + "num_tokens": 533089382.0, + "step": 20600 + }, + { + "epoch": 2.262354491544037, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1892175674438477, + "learning_rate": 1e-06, + "loss": 1.062, + "mean_token_accuracy": 0.6980253458023071, + "num_tokens": 533119669.0, + "step": 20601 + }, + { + "epoch": 2.2624643092466505, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.7566263675689697, + "learning_rate": 1e-06, + "loss": 0.701, + "mean_token_accuracy": 0.7761943340301514, + "num_tokens": 533138406.0, + "step": 20602 + }, + { + "epoch": 2.2625741269492643, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2606940269470215, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7403969168663025, + "num_tokens": 533165108.0, + "step": 20603 + }, + { + "epoch": 2.262683944651878, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.160407066345215, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7310814261436462, + "num_tokens": 533196382.0, + "step": 20604 + }, + { + "epoch": 2.2627937623544914, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3936126232147217, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7304364442825317, + "num_tokens": 533221981.0, + "step": 20605 + }, + { + "epoch": 2.262903580057105, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.427285671234131, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7174279689788818, + "num_tokens": 533248532.0, + "step": 20606 + }, + { + "epoch": 2.263013397759719, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1853814125061035, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7211695909500122, + "num_tokens": 533276848.0, + "step": 20607 + }, + { + "epoch": 2.2631232154623326, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.109628438949585, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7057148218154907, + "num_tokens": 533309991.0, + "step": 20608 + }, + { + "epoch": 2.263233033164946, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.313223123550415, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7419670224189758, + "num_tokens": 533337717.0, + "step": 20609 + }, + { + "epoch": 2.2633428508675597, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4555509090423584, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7259804010391235, + "num_tokens": 533361471.0, + "step": 20610 + }, + { + "epoch": 2.2634526685701735, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.271649122238159, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7349891662597656, + "num_tokens": 533390149.0, + "step": 20611 + }, + { + "epoch": 2.263562486272787, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4543299674987793, + "learning_rate": 1e-06, + "loss": 0.7333, + "mean_token_accuracy": 0.7618555426597595, + "num_tokens": 533412367.0, + "step": 20612 + }, + { + "epoch": 2.263672303975401, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3861424922943115, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7342543005943298, + "num_tokens": 533439118.0, + "step": 20613 + }, + { + "epoch": 2.2637821216780143, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4678380489349365, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7188948392868042, + "num_tokens": 533465411.0, + "step": 20614 + }, + { + "epoch": 2.263891939380628, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4931230545043945, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7268330454826355, + "num_tokens": 533491447.0, + "step": 20615 + }, + { + "epoch": 2.264001757083242, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.6082041263580322, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7431308627128601, + "num_tokens": 533513943.0, + "step": 20616 + }, + { + "epoch": 2.2641115747858556, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4145607948303223, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7114508748054504, + "num_tokens": 533541238.0, + "step": 20617 + }, + { + "epoch": 2.2642213924884693, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3288769721984863, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7431521415710449, + "num_tokens": 533565011.0, + "step": 20618 + }, + { + "epoch": 2.2643312101910826, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.0351626873016357, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7395320534706116, + "num_tokens": 533598006.0, + "step": 20619 + }, + { + "epoch": 2.2644410278936964, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.4630637168884277, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.709968090057373, + "num_tokens": 533623268.0, + "step": 20620 + }, + { + "epoch": 2.26455084559631, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.187086582183838, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7304717302322388, + "num_tokens": 533652105.0, + "step": 20621 + }, + { + "epoch": 2.264660663298924, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.84159779548645, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7375480532646179, + "num_tokens": 533670980.0, + "step": 20622 + }, + { + "epoch": 2.2647704810015377, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.446791410446167, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7258760333061218, + "num_tokens": 533697461.0, + "step": 20623 + }, + { + "epoch": 2.264880298704151, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.268009662628174, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7287575006484985, + "num_tokens": 533723770.0, + "step": 20624 + }, + { + "epoch": 2.2649901164067647, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1134097576141357, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7417845726013184, + "num_tokens": 533752403.0, + "step": 20625 + }, + { + "epoch": 2.2650999341093785, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3931703567504883, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7364879846572876, + "num_tokens": 533777072.0, + "step": 20626 + }, + { + "epoch": 2.2652097518119922, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.6013474464416504, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7441205978393555, + "num_tokens": 533800909.0, + "step": 20627 + }, + { + "epoch": 2.265319569514606, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.439253807067871, + "learning_rate": 1e-06, + "loss": 0.757, + "mean_token_accuracy": 0.7564716935157776, + "num_tokens": 533825942.0, + "step": 20628 + }, + { + "epoch": 2.2654293872172193, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.295814037322998, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7471518516540527, + "num_tokens": 533851246.0, + "step": 20629 + }, + { + "epoch": 2.265539204919833, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.7196285724639893, + "learning_rate": 1e-06, + "loss": 0.7631, + "mean_token_accuracy": 0.748113751411438, + "num_tokens": 533871288.0, + "step": 20630 + }, + { + "epoch": 2.265649022622447, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.364929676055908, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7168474197387695, + "num_tokens": 533897718.0, + "step": 20631 + }, + { + "epoch": 2.2657588403250606, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.48176908493042, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7438122034072876, + "num_tokens": 533919924.0, + "step": 20632 + }, + { + "epoch": 2.265868658027674, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.413437843322754, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7240656614303589, + "num_tokens": 533945249.0, + "step": 20633 + }, + { + "epoch": 2.2659784757302877, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2554585933685303, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.709202229976654, + "num_tokens": 533973516.0, + "step": 20634 + }, + { + "epoch": 2.2660882934329014, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.48712158203125, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.7484198212623596, + "num_tokens": 533997573.0, + "step": 20635 + }, + { + "epoch": 2.266198111135515, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.8430588245391846, + "learning_rate": 1e-06, + "loss": 0.7317, + "mean_token_accuracy": 0.7591492533683777, + "num_tokens": 534016343.0, + "step": 20636 + }, + { + "epoch": 2.2663079288381285, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.1507699489593506, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7245662212371826, + "num_tokens": 534046554.0, + "step": 20637 + }, + { + "epoch": 2.2664177465407422, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.525970458984375, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7387993335723877, + "num_tokens": 534070506.0, + "step": 20638 + }, + { + "epoch": 2.266527564243356, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2715554237365723, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7156948447227478, + "num_tokens": 534098458.0, + "step": 20639 + }, + { + "epoch": 2.2666373819459698, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 3.2501373291015625, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7478902339935303, + "num_tokens": 534119676.0, + "step": 20640 + }, + { + "epoch": 2.2667471996485835, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3089165687561035, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7575472593307495, + "num_tokens": 534147125.0, + "step": 20641 + }, + { + "epoch": 2.266857017351197, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.2667698860168457, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.6949765682220459, + "num_tokens": 534175831.0, + "step": 20642 + }, + { + "epoch": 2.2669668350538106, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.3138062953948975, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7474941611289978, + "num_tokens": 534202981.0, + "step": 20643 + }, + { + "epoch": 2.2670766527564243, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.325610637664795, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7149089574813843, + "num_tokens": 534232202.0, + "step": 20644 + }, + { + "epoch": 2.267186470459038, + "ewc_loss": 1.3113021850585938e-05, + "grad_norm": 2.498318672180176, + "learning_rate": 1e-06, + "loss": 0.772, + "mean_token_accuracy": 0.7543131113052368, + "num_tokens": 534256887.0, + "step": 20645 + }, + { + "epoch": 2.267296288161652, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.278202772140503, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7462384104728699, + "num_tokens": 534283370.0, + "step": 20646 + }, + { + "epoch": 2.267406105864265, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.564985752105713, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7512709498405457, + "num_tokens": 534305855.0, + "step": 20647 + }, + { + "epoch": 2.267515923566879, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.9197254180908203, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7319681644439697, + "num_tokens": 534324343.0, + "step": 20648 + }, + { + "epoch": 2.2676257412694927, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.4103469848632812, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7455868124961853, + "num_tokens": 534349733.0, + "step": 20649 + }, + { + "epoch": 2.2677355589721064, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.431957721710205, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7212164998054504, + "num_tokens": 534376612.0, + "step": 20650 + }, + { + "epoch": 2.26784537667472, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.1813135147094727, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7371121644973755, + "num_tokens": 534405067.0, + "step": 20651 + }, + { + "epoch": 2.2679551943773335, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.1558687686920166, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7036862373352051, + "num_tokens": 534434465.0, + "step": 20652 + }, + { + "epoch": 2.2680650120799473, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.575549364089966, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7281438112258911, + "num_tokens": 534457313.0, + "step": 20653 + }, + { + "epoch": 2.268174829782561, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.3381597995758057, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7405462265014648, + "num_tokens": 534483692.0, + "step": 20654 + }, + { + "epoch": 2.2682846474851748, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.449338674545288, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.72791588306427, + "num_tokens": 534506119.0, + "step": 20655 + }, + { + "epoch": 2.268394465187788, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.193862199783325, + "learning_rate": 1e-06, + "loss": 0.9554, + "mean_token_accuracy": 0.7073630094528198, + "num_tokens": 534539490.0, + "step": 20656 + }, + { + "epoch": 2.268504282890402, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.307230234146118, + "learning_rate": 1e-06, + "loss": 0.7729, + "mean_token_accuracy": 0.7596237659454346, + "num_tokens": 534566135.0, + "step": 20657 + }, + { + "epoch": 2.2686141005930156, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.4510862827301025, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7460503578186035, + "num_tokens": 534591151.0, + "step": 20658 + }, + { + "epoch": 2.2687239182956294, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.234572410583496, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7279590368270874, + "num_tokens": 534620416.0, + "step": 20659 + }, + { + "epoch": 2.2688337359982427, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.523822546005249, + "learning_rate": 1e-06, + "loss": 0.7946, + "mean_token_accuracy": 0.7449020147323608, + "num_tokens": 534643272.0, + "step": 20660 + }, + { + "epoch": 2.2689435537008564, + "ewc_loss": 1.3172626495361328e-05, + "grad_norm": 2.197662830352783, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.717147707939148, + "num_tokens": 534674244.0, + "step": 20661 + }, + { + "epoch": 2.26905337140347, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1585493087768555, + "learning_rate": 1e-06, + "loss": 0.7875, + "mean_token_accuracy": 0.7458425760269165, + "num_tokens": 534702621.0, + "step": 20662 + }, + { + "epoch": 2.269163189106084, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.617696762084961, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7412038445472717, + "num_tokens": 534724554.0, + "step": 20663 + }, + { + "epoch": 2.2692730068086977, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3124921321868896, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7149007320404053, + "num_tokens": 534754776.0, + "step": 20664 + }, + { + "epoch": 2.269382824511311, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4096333980560303, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.732347846031189, + "num_tokens": 534781638.0, + "step": 20665 + }, + { + "epoch": 2.2694926422139248, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.7530245780944824, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7475844621658325, + "num_tokens": 534800711.0, + "step": 20666 + }, + { + "epoch": 2.2696024599165385, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.183696985244751, + "learning_rate": 1e-06, + "loss": 0.9484, + "mean_token_accuracy": 0.7056626081466675, + "num_tokens": 534833402.0, + "step": 20667 + }, + { + "epoch": 2.2697122776191523, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3776345252990723, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7463263273239136, + "num_tokens": 534858057.0, + "step": 20668 + }, + { + "epoch": 2.269822095321766, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2805073261260986, + "learning_rate": 1e-06, + "loss": 0.9035, + "mean_token_accuracy": 0.7106801271438599, + "num_tokens": 534884852.0, + "step": 20669 + }, + { + "epoch": 2.2699319130243794, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.60613751411438, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7388870716094971, + "num_tokens": 534906785.0, + "step": 20670 + }, + { + "epoch": 2.270041730726993, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1818153858184814, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7286128997802734, + "num_tokens": 534939871.0, + "step": 20671 + }, + { + "epoch": 2.270151548429607, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.8926258087158203, + "learning_rate": 1e-06, + "loss": 0.7585, + "mean_token_accuracy": 0.7583330869674683, + "num_tokens": 534957889.0, + "step": 20672 + }, + { + "epoch": 2.2702613661322206, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.262563467025757, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7133930325508118, + "num_tokens": 534988918.0, + "step": 20673 + }, + { + "epoch": 2.2703711838348344, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4165585041046143, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7213001847267151, + "num_tokens": 535016145.0, + "step": 20674 + }, + { + "epoch": 2.2704810015374477, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.489173650741577, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.746335506439209, + "num_tokens": 535040159.0, + "step": 20675 + }, + { + "epoch": 2.2705908192400615, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1474621295928955, + "learning_rate": 1e-06, + "loss": 0.8019, + "mean_token_accuracy": 0.7485805749893188, + "num_tokens": 535068420.0, + "step": 20676 + }, + { + "epoch": 2.270700636942675, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.8452742099761963, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7324295043945312, + "num_tokens": 535086914.0, + "step": 20677 + }, + { + "epoch": 2.270810454645289, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5576860904693604, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7466107606887817, + "num_tokens": 535107860.0, + "step": 20678 + }, + { + "epoch": 2.2709202723479027, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3466267585754395, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7394408583641052, + "num_tokens": 535134240.0, + "step": 20679 + }, + { + "epoch": 2.271030090050516, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.650594472885132, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.716646671295166, + "num_tokens": 535159061.0, + "step": 20680 + }, + { + "epoch": 2.27113990775313, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.360293388366699, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7397801280021667, + "num_tokens": 535183824.0, + "step": 20681 + }, + { + "epoch": 2.2712497254557436, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.481316089630127, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7271991968154907, + "num_tokens": 535209253.0, + "step": 20682 + }, + { + "epoch": 2.2713595431583573, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.573430061340332, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7225635647773743, + "num_tokens": 535232310.0, + "step": 20683 + }, + { + "epoch": 2.2714693608609706, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.477752208709717, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7398309111595154, + "num_tokens": 535254604.0, + "step": 20684 + }, + { + "epoch": 2.2715791785635844, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.096787929534912, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7183037996292114, + "num_tokens": 535288578.0, + "step": 20685 + }, + { + "epoch": 2.271688996266198, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4671788215637207, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7237173318862915, + "num_tokens": 535313776.0, + "step": 20686 + }, + { + "epoch": 2.271798813968812, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.113717794418335, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.729324221611023, + "num_tokens": 535343650.0, + "step": 20687 + }, + { + "epoch": 2.271908631671425, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.754669666290283, + "learning_rate": 1e-06, + "loss": 0.7737, + "mean_token_accuracy": 0.7475043535232544, + "num_tokens": 535363387.0, + "step": 20688 + }, + { + "epoch": 2.272018449374039, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.655674695968628, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7561557292938232, + "num_tokens": 535385358.0, + "step": 20689 + }, + { + "epoch": 2.2721282670766527, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2924816608428955, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.6986408233642578, + "num_tokens": 535416079.0, + "step": 20690 + }, + { + "epoch": 2.2722380847792665, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4083147048950195, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7404594421386719, + "num_tokens": 535440675.0, + "step": 20691 + }, + { + "epoch": 2.2723479024818802, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.231733560562134, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7204447388648987, + "num_tokens": 535469567.0, + "step": 20692 + }, + { + "epoch": 2.2724577201844935, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.0786900520324707, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7068072557449341, + "num_tokens": 535503782.0, + "step": 20693 + }, + { + "epoch": 2.2725675378871073, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3044323921203613, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.7479897141456604, + "num_tokens": 535529052.0, + "step": 20694 + }, + { + "epoch": 2.272677355589721, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.762037992477417, + "learning_rate": 1e-06, + "loss": 0.7646, + "mean_token_accuracy": 0.7511364221572876, + "num_tokens": 535550344.0, + "step": 20695 + }, + { + "epoch": 2.272787173292335, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.49393367767334, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7296627759933472, + "num_tokens": 535574011.0, + "step": 20696 + }, + { + "epoch": 2.2728969909949486, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1779160499572754, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7235357761383057, + "num_tokens": 535604376.0, + "step": 20697 + }, + { + "epoch": 2.273006808697562, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2646901607513428, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7261977195739746, + "num_tokens": 535633623.0, + "step": 20698 + }, + { + "epoch": 2.2731166264001756, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.466322898864746, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.723945140838623, + "num_tokens": 535658667.0, + "step": 20699 + }, + { + "epoch": 2.2732264441027894, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.415783405303955, + "learning_rate": 1e-06, + "loss": 0.7529, + "mean_token_accuracy": 0.7624698877334595, + "num_tokens": 535682680.0, + "step": 20700 + }, + { + "epoch": 2.273336261805403, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5466151237487793, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7535839676856995, + "num_tokens": 535704776.0, + "step": 20701 + }, + { + "epoch": 2.273446079508017, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.503458023071289, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7284849882125854, + "num_tokens": 535731321.0, + "step": 20702 + }, + { + "epoch": 2.2735558972106302, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.843806505203247, + "learning_rate": 1e-06, + "loss": 0.7362, + "mean_token_accuracy": 0.7549760937690735, + "num_tokens": 535749357.0, + "step": 20703 + }, + { + "epoch": 2.273665714913244, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2178680896759033, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7461028099060059, + "num_tokens": 535776414.0, + "step": 20704 + }, + { + "epoch": 2.2737755326158577, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5325229167938232, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7110908627510071, + "num_tokens": 535800583.0, + "step": 20705 + }, + { + "epoch": 2.2738853503184715, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2604403495788574, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7279062867164612, + "num_tokens": 535827948.0, + "step": 20706 + }, + { + "epoch": 2.273995168021085, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.271289110183716, + "learning_rate": 1e-06, + "loss": 0.7672, + "mean_token_accuracy": 0.7496682405471802, + "num_tokens": 535855953.0, + "step": 20707 + }, + { + "epoch": 2.2741049857236986, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5425288677215576, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7529481649398804, + "num_tokens": 535878268.0, + "step": 20708 + }, + { + "epoch": 2.2742148034263123, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2481982707977295, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.73112553358078, + "num_tokens": 535906332.0, + "step": 20709 + }, + { + "epoch": 2.274324621128926, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.617947578430176, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7298319339752197, + "num_tokens": 535930170.0, + "step": 20710 + }, + { + "epoch": 2.2744344388315394, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4292423725128174, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7331052422523499, + "num_tokens": 535956811.0, + "step": 20711 + }, + { + "epoch": 2.274544256534153, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1949222087860107, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7318296432495117, + "num_tokens": 535987062.0, + "step": 20712 + }, + { + "epoch": 2.274654074236767, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.230520009994507, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.746256947517395, + "num_tokens": 536016727.0, + "step": 20713 + }, + { + "epoch": 2.2747638919393807, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.513298988342285, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7275961637496948, + "num_tokens": 536040280.0, + "step": 20714 + }, + { + "epoch": 2.2748737096419944, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3867809772491455, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7209049463272095, + "num_tokens": 536067056.0, + "step": 20715 + }, + { + "epoch": 2.2749835273446077, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5984208583831787, + "learning_rate": 1e-06, + "loss": 0.9603, + "mean_token_accuracy": 0.7142292261123657, + "num_tokens": 536090216.0, + "step": 20716 + }, + { + "epoch": 2.2750933450472215, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.495551824569702, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7431224584579468, + "num_tokens": 536113676.0, + "step": 20717 + }, + { + "epoch": 2.2752031627498353, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3358163833618164, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7449432015419006, + "num_tokens": 536140983.0, + "step": 20718 + }, + { + "epoch": 2.275312980452449, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5565686225891113, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7102394700050354, + "num_tokens": 536164566.0, + "step": 20719 + }, + { + "epoch": 2.2754227981550628, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3660624027252197, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.7451152801513672, + "num_tokens": 536189222.0, + "step": 20720 + }, + { + "epoch": 2.275532615857676, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.773409366607666, + "learning_rate": 1e-06, + "loss": 0.7647, + "mean_token_accuracy": 0.7461211681365967, + "num_tokens": 536207842.0, + "step": 20721 + }, + { + "epoch": 2.27564243356029, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.250767946243286, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7357414364814758, + "num_tokens": 536238776.0, + "step": 20722 + }, + { + "epoch": 2.2757522512629036, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.6963634490966797, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7371439933776855, + "num_tokens": 536258384.0, + "step": 20723 + }, + { + "epoch": 2.2758620689655173, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.1889753341674805, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7230445146560669, + "num_tokens": 536284965.0, + "step": 20724 + }, + { + "epoch": 2.275971886668131, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3111205101013184, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7348623871803284, + "num_tokens": 536313428.0, + "step": 20725 + }, + { + "epoch": 2.2760817043707444, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.535109519958496, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7189053893089294, + "num_tokens": 536339517.0, + "step": 20726 + }, + { + "epoch": 2.276191522073358, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.254460573196411, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.734405517578125, + "num_tokens": 536365847.0, + "step": 20727 + }, + { + "epoch": 2.276301339775972, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2638957500457764, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7089087963104248, + "num_tokens": 536393986.0, + "step": 20728 + }, + { + "epoch": 2.2764111574785857, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2701199054718018, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7076395750045776, + "num_tokens": 536422355.0, + "step": 20729 + }, + { + "epoch": 2.2765209751811994, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.500802993774414, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7271297574043274, + "num_tokens": 536447075.0, + "step": 20730 + }, + { + "epoch": 2.2766307928838128, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2033793926239014, + "learning_rate": 1e-06, + "loss": 0.9744, + "mean_token_accuracy": 0.6969101428985596, + "num_tokens": 536478409.0, + "step": 20731 + }, + { + "epoch": 2.2767406105864265, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.392461061477661, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7194249629974365, + "num_tokens": 536505080.0, + "step": 20732 + }, + { + "epoch": 2.2768504282890403, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.355686902999878, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7525402307510376, + "num_tokens": 536529626.0, + "step": 20733 + }, + { + "epoch": 2.276960245991654, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2219979763031006, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7458809614181519, + "num_tokens": 536556253.0, + "step": 20734 + }, + { + "epoch": 2.2770700636942673, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.6269569396972656, + "learning_rate": 1e-06, + "loss": 0.7328, + "mean_token_accuracy": 0.7598652839660645, + "num_tokens": 536578185.0, + "step": 20735 + }, + { + "epoch": 2.277179881396881, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.6135356426239014, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7274566888809204, + "num_tokens": 536601047.0, + "step": 20736 + }, + { + "epoch": 2.277289699099495, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.175969123840332, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7400013208389282, + "num_tokens": 536631078.0, + "step": 20737 + }, + { + "epoch": 2.2773995168021086, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.8559327125549316, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7306472063064575, + "num_tokens": 536649282.0, + "step": 20738 + }, + { + "epoch": 2.277509334504722, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.342681646347046, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7193191051483154, + "num_tokens": 536676114.0, + "step": 20739 + }, + { + "epoch": 2.2776191522073357, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.51754093170166, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7431215643882751, + "num_tokens": 536701464.0, + "step": 20740 + }, + { + "epoch": 2.2777289699099494, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.035782814025879, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7188010811805725, + "num_tokens": 536733933.0, + "step": 20741 + }, + { + "epoch": 2.277838787612563, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.066660165786743, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.7427911758422852, + "num_tokens": 536764702.0, + "step": 20742 + }, + { + "epoch": 2.277948605315177, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.372234344482422, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7171447277069092, + "num_tokens": 536789400.0, + "step": 20743 + }, + { + "epoch": 2.2780584230177903, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.614046335220337, + "learning_rate": 1e-06, + "loss": 0.777, + "mean_token_accuracy": 0.7573810815811157, + "num_tokens": 536810565.0, + "step": 20744 + }, + { + "epoch": 2.278168240720404, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.386089563369751, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7328001856803894, + "num_tokens": 536835174.0, + "step": 20745 + }, + { + "epoch": 2.278278058423018, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4155895709991455, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.728634238243103, + "num_tokens": 536860085.0, + "step": 20746 + }, + { + "epoch": 2.2783878761256315, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5474679470062256, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7343641519546509, + "num_tokens": 536883060.0, + "step": 20747 + }, + { + "epoch": 2.2784976938282453, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.176959991455078, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7245804071426392, + "num_tokens": 536912050.0, + "step": 20748 + }, + { + "epoch": 2.2786075115308586, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5062873363494873, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7383679151535034, + "num_tokens": 536934709.0, + "step": 20749 + }, + { + "epoch": 2.2787173292334724, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2011256217956543, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7402957081794739, + "num_tokens": 536963123.0, + "step": 20750 + }, + { + "epoch": 2.278827146936086, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.12898588180542, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7191617488861084, + "num_tokens": 536994318.0, + "step": 20751 + }, + { + "epoch": 2.2789369646387, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.917160749435425, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7407785654067993, + "num_tokens": 537013036.0, + "step": 20752 + }, + { + "epoch": 2.2790467823413136, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.27750563621521, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7067101001739502, + "num_tokens": 537042769.0, + "step": 20753 + }, + { + "epoch": 2.279156600043927, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2796669006347656, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7376102209091187, + "num_tokens": 537070224.0, + "step": 20754 + }, + { + "epoch": 2.2792664177465407, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.24422025680542, + "learning_rate": 1e-06, + "loss": 0.9798, + "mean_token_accuracy": 0.7032618522644043, + "num_tokens": 537100398.0, + "step": 20755 + }, + { + "epoch": 2.2793762354491545, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3192996978759766, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7413305640220642, + "num_tokens": 537124562.0, + "step": 20756 + }, + { + "epoch": 2.279486053151768, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5508248805999756, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7208138704299927, + "num_tokens": 537148606.0, + "step": 20757 + }, + { + "epoch": 2.2795958708543815, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2144412994384766, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7203162908554077, + "num_tokens": 537177375.0, + "step": 20758 + }, + { + "epoch": 2.2797056885569953, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4925312995910645, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7523436546325684, + "num_tokens": 537198936.0, + "step": 20759 + }, + { + "epoch": 2.279815506259609, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4989733695983887, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7520236372947693, + "num_tokens": 537222002.0, + "step": 20760 + }, + { + "epoch": 2.279925323962223, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2239811420440674, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7246878743171692, + "num_tokens": 537252011.0, + "step": 20761 + }, + { + "epoch": 2.2800351416648366, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3897006511688232, + "learning_rate": 1e-06, + "loss": 0.792, + "mean_token_accuracy": 0.7446506023406982, + "num_tokens": 537275960.0, + "step": 20762 + }, + { + "epoch": 2.28014495936745, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.0843873023986816, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7343748807907104, + "num_tokens": 537306553.0, + "step": 20763 + }, + { + "epoch": 2.2802547770700636, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.5413806438446045, + "learning_rate": 1e-06, + "loss": 0.7962, + "mean_token_accuracy": 0.743560791015625, + "num_tokens": 537329446.0, + "step": 20764 + }, + { + "epoch": 2.2803645947726774, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.7098748683929443, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7389201521873474, + "num_tokens": 537352940.0, + "step": 20765 + }, + { + "epoch": 2.280474412475291, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.2917892932891846, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7412787675857544, + "num_tokens": 537381267.0, + "step": 20766 + }, + { + "epoch": 2.2805842301779045, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3504092693328857, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7333601117134094, + "num_tokens": 537407946.0, + "step": 20767 + }, + { + "epoch": 2.280694047880518, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.7318077087402344, + "learning_rate": 1e-06, + "loss": 0.7809, + "mean_token_accuracy": 0.750687837600708, + "num_tokens": 537426873.0, + "step": 20768 + }, + { + "epoch": 2.280803865583132, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3506884574890137, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7319617867469788, + "num_tokens": 537453441.0, + "step": 20769 + }, + { + "epoch": 2.2809136832857457, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.4498326778411865, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7306311726570129, + "num_tokens": 537477762.0, + "step": 20770 + }, + { + "epoch": 2.2810235009883595, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3286051750183105, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7238383293151855, + "num_tokens": 537504972.0, + "step": 20771 + }, + { + "epoch": 2.281133318690973, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.067514181137085, + "learning_rate": 1e-06, + "loss": 0.9726, + "mean_token_accuracy": 0.7044186592102051, + "num_tokens": 537538606.0, + "step": 20772 + }, + { + "epoch": 2.2812431363935866, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1083898544311523, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7197517156600952, + "num_tokens": 537570654.0, + "step": 20773 + }, + { + "epoch": 2.2813529540962003, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3450756072998047, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7259199023246765, + "num_tokens": 537597100.0, + "step": 20774 + }, + { + "epoch": 2.281462771798814, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4034619331359863, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7380377650260925, + "num_tokens": 537621795.0, + "step": 20775 + }, + { + "epoch": 2.281572589501428, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.506258964538574, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7355778813362122, + "num_tokens": 537645966.0, + "step": 20776 + }, + { + "epoch": 2.281682407204041, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.248716115951538, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7163601517677307, + "num_tokens": 537675897.0, + "step": 20777 + }, + { + "epoch": 2.281792224906655, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3940813541412354, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7340720295906067, + "num_tokens": 537703630.0, + "step": 20778 + }, + { + "epoch": 2.2819020426092687, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.269730806350708, + "learning_rate": 1e-06, + "loss": 0.95, + "mean_token_accuracy": 0.7031241655349731, + "num_tokens": 537731596.0, + "step": 20779 + }, + { + "epoch": 2.2820118603118824, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2419917583465576, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7274547219276428, + "num_tokens": 537759309.0, + "step": 20780 + }, + { + "epoch": 2.282121678014496, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.609257221221924, + "learning_rate": 1e-06, + "loss": 0.8168, + "mean_token_accuracy": 0.7364470958709717, + "num_tokens": 537779453.0, + "step": 20781 + }, + { + "epoch": 2.2822314957171095, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4473466873168945, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7458076477050781, + "num_tokens": 537804432.0, + "step": 20782 + }, + { + "epoch": 2.2823413134197232, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1215999126434326, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7371150255203247, + "num_tokens": 537833522.0, + "step": 20783 + }, + { + "epoch": 2.282451131122337, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2685205936431885, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7283907532691956, + "num_tokens": 537861995.0, + "step": 20784 + }, + { + "epoch": 2.2825609488249508, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.6923835277557373, + "learning_rate": 1e-06, + "loss": 0.7702, + "mean_token_accuracy": 0.7488746643066406, + "num_tokens": 537883858.0, + "step": 20785 + }, + { + "epoch": 2.282670766527564, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3058738708496094, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7123075127601624, + "num_tokens": 537910691.0, + "step": 20786 + }, + { + "epoch": 2.282780584230178, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.7840747833251953, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7307420372962952, + "num_tokens": 537932131.0, + "step": 20787 + }, + { + "epoch": 2.2828904019327916, + "ewc_loss": 1.3232231140136719e-05, + "grad_norm": 2.3946168422698975, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7124353647232056, + "num_tokens": 537959397.0, + "step": 20788 + }, + { + "epoch": 2.2830002196354053, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.7670505046844482, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7228681445121765, + "num_tokens": 537980465.0, + "step": 20789 + }, + { + "epoch": 2.2831100373380186, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1187777519226074, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7187285423278809, + "num_tokens": 538012996.0, + "step": 20790 + }, + { + "epoch": 2.2832198550406324, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1121840476989746, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.7107205390930176, + "num_tokens": 538042730.0, + "step": 20791 + }, + { + "epoch": 2.283329672743246, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.7626571655273438, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7363475561141968, + "num_tokens": 538064188.0, + "step": 20792 + }, + { + "epoch": 2.28343949044586, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4490151405334473, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7486851215362549, + "num_tokens": 538089270.0, + "step": 20793 + }, + { + "epoch": 2.2835493081484737, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2300772666931152, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7266296148300171, + "num_tokens": 538117966.0, + "step": 20794 + }, + { + "epoch": 2.283659125851087, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2452852725982666, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7449519634246826, + "num_tokens": 538146195.0, + "step": 20795 + }, + { + "epoch": 2.2837689435537007, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3207294940948486, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7335165739059448, + "num_tokens": 538171123.0, + "step": 20796 + }, + { + "epoch": 2.2838787612563145, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.5173275470733643, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7142581939697266, + "num_tokens": 538195696.0, + "step": 20797 + }, + { + "epoch": 2.2839885789589283, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.1885640621185303, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7284233570098877, + "num_tokens": 538224843.0, + "step": 20798 + }, + { + "epoch": 2.284098396661542, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.8363406658172607, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7441776990890503, + "num_tokens": 538243669.0, + "step": 20799 + }, + { + "epoch": 2.2842082143641553, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2390174865722656, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7371440529823303, + "num_tokens": 538271782.0, + "step": 20800 + }, + { + "epoch": 2.284318032066769, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4104843139648438, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.728600025177002, + "num_tokens": 538295796.0, + "step": 20801 + }, + { + "epoch": 2.284427849769383, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.5524182319641113, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7208649516105652, + "num_tokens": 538319690.0, + "step": 20802 + }, + { + "epoch": 2.2845376674719966, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4079489707946777, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7258181571960449, + "num_tokens": 538345376.0, + "step": 20803 + }, + { + "epoch": 2.2846474851746104, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.380760431289673, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.7410800457000732, + "num_tokens": 538368600.0, + "step": 20804 + }, + { + "epoch": 2.2847573028772237, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.269339084625244, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7337753176689148, + "num_tokens": 538396941.0, + "step": 20805 + }, + { + "epoch": 2.2848671205798374, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4764928817749023, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7067780494689941, + "num_tokens": 538423386.0, + "step": 20806 + }, + { + "epoch": 2.284976938282451, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.413297653198242, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7238616347312927, + "num_tokens": 538449073.0, + "step": 20807 + }, + { + "epoch": 2.285086755985065, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1359689235687256, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7286453247070312, + "num_tokens": 538481914.0, + "step": 20808 + }, + { + "epoch": 2.2851965736876787, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6209425926208496, + "learning_rate": 1e-06, + "loss": 0.7651, + "mean_token_accuracy": 0.7518770098686218, + "num_tokens": 538502236.0, + "step": 20809 + }, + { + "epoch": 2.285306391390292, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.707348108291626, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7329182624816895, + "num_tokens": 538522773.0, + "step": 20810 + }, + { + "epoch": 2.2854162090929058, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.51267671585083, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.740209698677063, + "num_tokens": 538547524.0, + "step": 20811 + }, + { + "epoch": 2.2855260267955195, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.607999563217163, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.747541069984436, + "num_tokens": 538569996.0, + "step": 20812 + }, + { + "epoch": 2.2856358444981333, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.234086513519287, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7267732620239258, + "num_tokens": 538597151.0, + "step": 20813 + }, + { + "epoch": 2.2857456622007466, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.16582989692688, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7089163661003113, + "num_tokens": 538627382.0, + "step": 20814 + }, + { + "epoch": 2.2858554799033604, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4715378284454346, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7328084111213684, + "num_tokens": 538652426.0, + "step": 20815 + }, + { + "epoch": 2.285965297605974, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2694311141967773, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.690801739692688, + "num_tokens": 538681280.0, + "step": 20816 + }, + { + "epoch": 2.286075115308588, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.2197265625, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7219865918159485, + "num_tokens": 538709258.0, + "step": 20817 + }, + { + "epoch": 2.286184933011201, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5148355960845947, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7336918711662292, + "num_tokens": 538734645.0, + "step": 20818 + }, + { + "epoch": 2.286294750713815, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4068470001220703, + "learning_rate": 1e-06, + "loss": 0.791, + "mean_token_accuracy": 0.7436275482177734, + "num_tokens": 538759285.0, + "step": 20819 + }, + { + "epoch": 2.2864045684164287, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1243741512298584, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.703067421913147, + "num_tokens": 538789988.0, + "step": 20820 + }, + { + "epoch": 2.2865143861190425, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.496722936630249, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7398879528045654, + "num_tokens": 538813693.0, + "step": 20821 + }, + { + "epoch": 2.286624203821656, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.289851188659668, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7447500228881836, + "num_tokens": 538839601.0, + "step": 20822 + }, + { + "epoch": 2.2867340215242695, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.0892515182495117, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7288789749145508, + "num_tokens": 538869111.0, + "step": 20823 + }, + { + "epoch": 2.2868438392268833, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4285333156585693, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7283299565315247, + "num_tokens": 538895327.0, + "step": 20824 + }, + { + "epoch": 2.286953656929497, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.613356351852417, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7140140533447266, + "num_tokens": 538917551.0, + "step": 20825 + }, + { + "epoch": 2.287063474632111, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.14638352394104, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7167304754257202, + "num_tokens": 538948810.0, + "step": 20826 + }, + { + "epoch": 2.2871732923347246, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.3595869541168213, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7055021524429321, + "num_tokens": 538977071.0, + "step": 20827 + }, + { + "epoch": 2.287283110037338, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.255945920944214, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7434438467025757, + "num_tokens": 539005221.0, + "step": 20828 + }, + { + "epoch": 2.2873929277399516, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.0954477787017822, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7212345600128174, + "num_tokens": 539037391.0, + "step": 20829 + }, + { + "epoch": 2.2875027454425654, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 3.079375743865967, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7411324977874756, + "num_tokens": 539054853.0, + "step": 20830 + }, + { + "epoch": 2.287612563145179, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.5642664432525635, + "learning_rate": 1e-06, + "loss": 0.7915, + "mean_token_accuracy": 0.7527534365653992, + "num_tokens": 539076794.0, + "step": 20831 + }, + { + "epoch": 2.287722380847793, + "ewc_loss": 1.329183578491211e-05, + "grad_norm": 2.4524261951446533, + "learning_rate": 1e-06, + "loss": 0.7557, + "mean_token_accuracy": 0.7640564441680908, + "num_tokens": 539099218.0, + "step": 20832 + }, + { + "epoch": 2.287832198550406, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3271336555480957, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7284018397331238, + "num_tokens": 539124379.0, + "step": 20833 + }, + { + "epoch": 2.28794201625302, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.313756227493286, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7275556921958923, + "num_tokens": 539150985.0, + "step": 20834 + }, + { + "epoch": 2.2880518339556337, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6473474502563477, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7464671730995178, + "num_tokens": 539171635.0, + "step": 20835 + }, + { + "epoch": 2.2881616516582475, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.485391855239868, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.741853654384613, + "num_tokens": 539195741.0, + "step": 20836 + }, + { + "epoch": 2.288271469360861, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.20593523979187, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7361642122268677, + "num_tokens": 539224786.0, + "step": 20837 + }, + { + "epoch": 2.2883812870634745, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.501344680786133, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7154048681259155, + "num_tokens": 539249631.0, + "step": 20838 + }, + { + "epoch": 2.2884911047660883, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2268667221069336, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7292085886001587, + "num_tokens": 539277615.0, + "step": 20839 + }, + { + "epoch": 2.288600922468702, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5012059211730957, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7328245639801025, + "num_tokens": 539300415.0, + "step": 20840 + }, + { + "epoch": 2.2887107401713154, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3893892765045166, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.744184672832489, + "num_tokens": 539324690.0, + "step": 20841 + }, + { + "epoch": 2.288820557873929, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2485568523406982, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.74165278673172, + "num_tokens": 539353522.0, + "step": 20842 + }, + { + "epoch": 2.288930375576543, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2936527729034424, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7336044311523438, + "num_tokens": 539380717.0, + "step": 20843 + }, + { + "epoch": 2.2890401932791566, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.48861026763916, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7061060667037964, + "num_tokens": 539406095.0, + "step": 20844 + }, + { + "epoch": 2.2891500109817704, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.694403886795044, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7425822615623474, + "num_tokens": 539428055.0, + "step": 20845 + }, + { + "epoch": 2.2892598286843837, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3733391761779785, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.745998740196228, + "num_tokens": 539453392.0, + "step": 20846 + }, + { + "epoch": 2.2893696463869975, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3367936611175537, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7366293668746948, + "num_tokens": 539478657.0, + "step": 20847 + }, + { + "epoch": 2.2894794640896112, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.311255693435669, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7370625734329224, + "num_tokens": 539503239.0, + "step": 20848 + }, + { + "epoch": 2.289589281792225, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2798359394073486, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.751693844795227, + "num_tokens": 539529013.0, + "step": 20849 + }, + { + "epoch": 2.2896990994948387, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.404951572418213, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.746924877166748, + "num_tokens": 539552767.0, + "step": 20850 + }, + { + "epoch": 2.289808917197452, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4349448680877686, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7556657791137695, + "num_tokens": 539579488.0, + "step": 20851 + }, + { + "epoch": 2.289918734900066, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2944185733795166, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7220116853713989, + "num_tokens": 539606583.0, + "step": 20852 + }, + { + "epoch": 2.2900285526026796, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4770429134368896, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7185033559799194, + "num_tokens": 539631989.0, + "step": 20853 + }, + { + "epoch": 2.2901383703052933, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5139360427856445, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7289954423904419, + "num_tokens": 539656172.0, + "step": 20854 + }, + { + "epoch": 2.290248188007907, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2867324352264404, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7249166369438171, + "num_tokens": 539684394.0, + "step": 20855 + }, + { + "epoch": 2.2903580057105204, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2813704013824463, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7080103754997253, + "num_tokens": 539712695.0, + "step": 20856 + }, + { + "epoch": 2.290467823413134, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2408127784729004, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7150620222091675, + "num_tokens": 539741458.0, + "step": 20857 + }, + { + "epoch": 2.290577641115748, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2722811698913574, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7262212038040161, + "num_tokens": 539770286.0, + "step": 20858 + }, + { + "epoch": 2.2906874588183617, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.391636848449707, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7180496454238892, + "num_tokens": 539797632.0, + "step": 20859 + }, + { + "epoch": 2.2907972765209754, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.032212018966675, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7167584300041199, + "num_tokens": 539831806.0, + "step": 20860 + }, + { + "epoch": 2.2909070942235887, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.156153440475464, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7266213893890381, + "num_tokens": 539859437.0, + "step": 20861 + }, + { + "epoch": 2.2910169119262025, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6587114334106445, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7213908433914185, + "num_tokens": 539880540.0, + "step": 20862 + }, + { + "epoch": 2.2911267296288162, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.8045499324798584, + "learning_rate": 1e-06, + "loss": 0.7641, + "mean_token_accuracy": 0.7519423961639404, + "num_tokens": 539899973.0, + "step": 20863 + }, + { + "epoch": 2.29123654733143, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 1.9771454334259033, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7206896543502808, + "num_tokens": 539934351.0, + "step": 20864 + }, + { + "epoch": 2.2913463650340433, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.585035800933838, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7431582808494568, + "num_tokens": 539956011.0, + "step": 20865 + }, + { + "epoch": 2.291456182736657, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.635598659515381, + "learning_rate": 1e-06, + "loss": 0.8075, + "mean_token_accuracy": 0.744778037071228, + "num_tokens": 539976916.0, + "step": 20866 + }, + { + "epoch": 2.291566000439271, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.323374032974243, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7185888290405273, + "num_tokens": 540002841.0, + "step": 20867 + }, + { + "epoch": 2.2916758181418846, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3092381954193115, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7289330959320068, + "num_tokens": 540028328.0, + "step": 20868 + }, + { + "epoch": 2.291785635844498, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.275865316390991, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7267888784408569, + "num_tokens": 540054231.0, + "step": 20869 + }, + { + "epoch": 2.2918954535471117, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.73500919342041, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7370823621749878, + "num_tokens": 540072985.0, + "step": 20870 + }, + { + "epoch": 2.2920052712497254, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.575491428375244, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7450915575027466, + "num_tokens": 540093628.0, + "step": 20871 + }, + { + "epoch": 2.292115088952339, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5960474014282227, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7352330684661865, + "num_tokens": 540115163.0, + "step": 20872 + }, + { + "epoch": 2.292224906654953, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5698294639587402, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.750176191329956, + "num_tokens": 540137017.0, + "step": 20873 + }, + { + "epoch": 2.2923347243575662, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.643984317779541, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7332494258880615, + "num_tokens": 540158223.0, + "step": 20874 + }, + { + "epoch": 2.29244454206018, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3641037940979004, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7147946357727051, + "num_tokens": 540184447.0, + "step": 20875 + }, + { + "epoch": 2.2925543597627938, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1963748931884766, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.719667911529541, + "num_tokens": 540214098.0, + "step": 20876 + }, + { + "epoch": 2.2926641774654075, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.498305082321167, + "learning_rate": 1e-06, + "loss": 0.7157, + "mean_token_accuracy": 0.7684699296951294, + "num_tokens": 540236665.0, + "step": 20877 + }, + { + "epoch": 2.2927739951680213, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3795955181121826, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7475503087043762, + "num_tokens": 540262127.0, + "step": 20878 + }, + { + "epoch": 2.2928838128706346, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.31141996383667, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7351194620132446, + "num_tokens": 540286459.0, + "step": 20879 + }, + { + "epoch": 2.2929936305732483, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6550209522247314, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7231650352478027, + "num_tokens": 540308637.0, + "step": 20880 + }, + { + "epoch": 2.293103448275862, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.465149164199829, + "learning_rate": 1e-06, + "loss": 0.7501, + "mean_token_accuracy": 0.7546144127845764, + "num_tokens": 540330801.0, + "step": 20881 + }, + { + "epoch": 2.293213265978476, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.7492101192474365, + "learning_rate": 1e-06, + "loss": 0.7305, + "mean_token_accuracy": 0.7601534128189087, + "num_tokens": 540350290.0, + "step": 20882 + }, + { + "epoch": 2.2933230836810896, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.352259397506714, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7398780584335327, + "num_tokens": 540375822.0, + "step": 20883 + }, + { + "epoch": 2.293432901383703, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5643537044525146, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.727888286113739, + "num_tokens": 540400247.0, + "step": 20884 + }, + { + "epoch": 2.2935427190863167, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.763312816619873, + "learning_rate": 1e-06, + "loss": 0.704, + "mean_token_accuracy": 0.7643887400627136, + "num_tokens": 540419036.0, + "step": 20885 + }, + { + "epoch": 2.2936525367889304, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 1.9892438650131226, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7000186443328857, + "num_tokens": 540454269.0, + "step": 20886 + }, + { + "epoch": 2.293762354491544, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.384737253189087, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7414050102233887, + "num_tokens": 540478961.0, + "step": 20887 + }, + { + "epoch": 2.2938721721941575, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.305513381958008, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.7456550598144531, + "num_tokens": 540505132.0, + "step": 20888 + }, + { + "epoch": 2.2939819898967713, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.8878395557403564, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7372957468032837, + "num_tokens": 540524365.0, + "step": 20889 + }, + { + "epoch": 2.294091807599385, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4083731174468994, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7235117554664612, + "num_tokens": 540551399.0, + "step": 20890 + }, + { + "epoch": 2.294201625301999, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.448946952819824, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7466696500778198, + "num_tokens": 540575806.0, + "step": 20891 + }, + { + "epoch": 2.294311443004612, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4047763347625732, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.730316162109375, + "num_tokens": 540599400.0, + "step": 20892 + }, + { + "epoch": 2.294421260707226, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1850695610046387, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7331206798553467, + "num_tokens": 540629135.0, + "step": 20893 + }, + { + "epoch": 2.2945310784098396, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.1909167766571045, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.7013075351715088, + "num_tokens": 540659440.0, + "step": 20894 + }, + { + "epoch": 2.2946408961124534, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.685272455215454, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7477818131446838, + "num_tokens": 540680501.0, + "step": 20895 + }, + { + "epoch": 2.294750713815067, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.474769115447998, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.735049843788147, + "num_tokens": 540703039.0, + "step": 20896 + }, + { + "epoch": 2.2948605315176804, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.251180410385132, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7195241451263428, + "num_tokens": 540732665.0, + "step": 20897 + }, + { + "epoch": 2.294970349220294, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.364238739013672, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7302195429801941, + "num_tokens": 540759246.0, + "step": 20898 + }, + { + "epoch": 2.295080166922908, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.178109645843506, + "learning_rate": 1e-06, + "loss": 0.9326, + "mean_token_accuracy": 0.7139996290206909, + "num_tokens": 540788127.0, + "step": 20899 + }, + { + "epoch": 2.2951899846255217, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.8964993953704834, + "learning_rate": 1e-06, + "loss": 0.7829, + "mean_token_accuracy": 0.7492408752441406, + "num_tokens": 540806931.0, + "step": 20900 + }, + { + "epoch": 2.2952998023281355, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.856170654296875, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7366506457328796, + "num_tokens": 540826294.0, + "step": 20901 + }, + { + "epoch": 2.2954096200307488, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3855409622192383, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.726806640625, + "num_tokens": 540851938.0, + "step": 20902 + }, + { + "epoch": 2.2955194377333625, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.25022029876709, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7134878635406494, + "num_tokens": 540879803.0, + "step": 20903 + }, + { + "epoch": 2.2956292554359763, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.321026563644409, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7115979790687561, + "num_tokens": 540908534.0, + "step": 20904 + }, + { + "epoch": 2.29573907313859, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.6373445987701416, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7229231595993042, + "num_tokens": 540930373.0, + "step": 20905 + }, + { + "epoch": 2.295848890841204, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3972227573394775, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7403112649917603, + "num_tokens": 540954331.0, + "step": 20906 + }, + { + "epoch": 2.295958708543817, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.58944034576416, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7458518743515015, + "num_tokens": 540975078.0, + "step": 20907 + }, + { + "epoch": 2.296068526246431, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.410069227218628, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.72021484375, + "num_tokens": 541000601.0, + "step": 20908 + }, + { + "epoch": 2.2961783439490446, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.3479318618774414, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7184813022613525, + "num_tokens": 541028698.0, + "step": 20909 + }, + { + "epoch": 2.2962881616516584, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2087619304656982, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7084404826164246, + "num_tokens": 541060964.0, + "step": 20910 + }, + { + "epoch": 2.296397979354272, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.4808340072631836, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7279071807861328, + "num_tokens": 541083273.0, + "step": 20911 + }, + { + "epoch": 2.2965077970568855, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.5006415843963623, + "learning_rate": 1e-06, + "loss": 0.9653, + "mean_token_accuracy": 0.7064329981803894, + "num_tokens": 541107607.0, + "step": 20912 + }, + { + "epoch": 2.296617614759499, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.409982442855835, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7324353456497192, + "num_tokens": 541131056.0, + "step": 20913 + }, + { + "epoch": 2.296727432462113, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.2342591285705566, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7211321592330933, + "num_tokens": 541158773.0, + "step": 20914 + }, + { + "epoch": 2.2968372501647267, + "ewc_loss": 1.33514404296875e-05, + "grad_norm": 2.468982696533203, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7373496294021606, + "num_tokens": 541181050.0, + "step": 20915 + }, + { + "epoch": 2.29694706786734, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3157811164855957, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7301796078681946, + "num_tokens": 541210007.0, + "step": 20916 + }, + { + "epoch": 2.297056885569954, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2898154258728027, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7215441465377808, + "num_tokens": 541235748.0, + "step": 20917 + }, + { + "epoch": 2.2971667032725676, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.158756732940674, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7260838747024536, + "num_tokens": 541267118.0, + "step": 20918 + }, + { + "epoch": 2.2972765209751813, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.445256233215332, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7367548942565918, + "num_tokens": 541290251.0, + "step": 20919 + }, + { + "epoch": 2.2973863386777946, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0784709453582764, + "learning_rate": 1e-06, + "loss": 0.9362, + "mean_token_accuracy": 0.7115087509155273, + "num_tokens": 541321159.0, + "step": 20920 + }, + { + "epoch": 2.2974961563804084, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.219620704650879, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7210900783538818, + "num_tokens": 541348450.0, + "step": 20921 + }, + { + "epoch": 2.297605974083022, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.392932176589966, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7439802885055542, + "num_tokens": 541372135.0, + "step": 20922 + }, + { + "epoch": 2.297715791785636, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.209221124649048, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7236193418502808, + "num_tokens": 541401150.0, + "step": 20923 + }, + { + "epoch": 2.2978256094882497, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0205748081207275, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7165224552154541, + "num_tokens": 541435158.0, + "step": 20924 + }, + { + "epoch": 2.297935427190863, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.9708549976348877, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7366535067558289, + "num_tokens": 541453374.0, + "step": 20925 + }, + { + "epoch": 2.2980452448934767, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.465386390686035, + "learning_rate": 1e-06, + "loss": 0.6747, + "mean_token_accuracy": 0.7852188348770142, + "num_tokens": 541475660.0, + "step": 20926 + }, + { + "epoch": 2.2981550625960905, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2455484867095947, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7529431581497192, + "num_tokens": 541503522.0, + "step": 20927 + }, + { + "epoch": 2.2982648802987042, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4212582111358643, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7379913926124573, + "num_tokens": 541528360.0, + "step": 20928 + }, + { + "epoch": 2.298374698001318, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5190176963806152, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7207781672477722, + "num_tokens": 541554598.0, + "step": 20929 + }, + { + "epoch": 2.2984845157039313, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4224257469177246, + "learning_rate": 1e-06, + "loss": 0.7524, + "mean_token_accuracy": 0.7568091154098511, + "num_tokens": 541577840.0, + "step": 20930 + }, + { + "epoch": 2.298594333406545, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.24409818649292, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7097885012626648, + "num_tokens": 541608864.0, + "step": 20931 + }, + { + "epoch": 2.298704151109159, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2981414794921875, + "learning_rate": 1e-06, + "loss": 0.9275, + "mean_token_accuracy": 0.7038449645042419, + "num_tokens": 541636803.0, + "step": 20932 + }, + { + "epoch": 2.2988139688117726, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2540297508239746, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7292395830154419, + "num_tokens": 541664820.0, + "step": 20933 + }, + { + "epoch": 2.2989237865143863, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4402570724487305, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.7239561080932617, + "num_tokens": 541690577.0, + "step": 20934 + }, + { + "epoch": 2.2990336042169996, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.299166202545166, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7300011515617371, + "num_tokens": 541716375.0, + "step": 20935 + }, + { + "epoch": 2.2991434219196134, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4022719860076904, + "learning_rate": 1e-06, + "loss": 0.7922, + "mean_token_accuracy": 0.7453113794326782, + "num_tokens": 541742331.0, + "step": 20936 + }, + { + "epoch": 2.299253239622227, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3488264083862305, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7376277446746826, + "num_tokens": 541767319.0, + "step": 20937 + }, + { + "epoch": 2.299363057324841, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2984328269958496, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7077248096466064, + "num_tokens": 541795125.0, + "step": 20938 + }, + { + "epoch": 2.2994728750274542, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.047628402709961, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7142496109008789, + "num_tokens": 541830697.0, + "step": 20939 + }, + { + "epoch": 2.299582692730068, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3683063983917236, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7412316203117371, + "num_tokens": 541856540.0, + "step": 20940 + }, + { + "epoch": 2.2996925104326817, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.7321698665618896, + "learning_rate": 1e-06, + "loss": 0.7575, + "mean_token_accuracy": 0.7530852556228638, + "num_tokens": 541876615.0, + "step": 20941 + }, + { + "epoch": 2.2998023281352955, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.166534662246704, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7243596315383911, + "num_tokens": 541903735.0, + "step": 20942 + }, + { + "epoch": 2.2999121458379093, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0952231884002686, + "learning_rate": 1e-06, + "loss": 0.9292, + "mean_token_accuracy": 0.7126862406730652, + "num_tokens": 541935807.0, + "step": 20943 + }, + { + "epoch": 2.3000219635405226, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.329806327819824, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7401145100593567, + "num_tokens": 541962688.0, + "step": 20944 + }, + { + "epoch": 2.3001317812431363, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3144052028656006, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7207130193710327, + "num_tokens": 541988938.0, + "step": 20945 + }, + { + "epoch": 2.30024159894575, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6527042388916016, + "learning_rate": 1e-06, + "loss": 0.7354, + "mean_token_accuracy": 0.7615343928337097, + "num_tokens": 542010466.0, + "step": 20946 + }, + { + "epoch": 2.300351416648364, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.379074811935425, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7367805242538452, + "num_tokens": 542034293.0, + "step": 20947 + }, + { + "epoch": 2.300461234350977, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.465999126434326, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.712982177734375, + "num_tokens": 542059149.0, + "step": 20948 + }, + { + "epoch": 2.300571052053591, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0653295516967773, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.729322075843811, + "num_tokens": 542089035.0, + "step": 20949 + }, + { + "epoch": 2.3006808697562047, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3302807807922363, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7006427645683289, + "num_tokens": 542116054.0, + "step": 20950 + }, + { + "epoch": 2.3007906874588184, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.245638370513916, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7209533452987671, + "num_tokens": 542145231.0, + "step": 20951 + }, + { + "epoch": 2.300900505161432, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.21529483795166, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7257635593414307, + "num_tokens": 542173544.0, + "step": 20952 + }, + { + "epoch": 2.3010103228640455, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.201538324356079, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7111158967018127, + "num_tokens": 542202205.0, + "step": 20953 + }, + { + "epoch": 2.3011201405666593, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3745131492614746, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7241520881652832, + "num_tokens": 542227842.0, + "step": 20954 + }, + { + "epoch": 2.301229958269273, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1270928382873535, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7250102162361145, + "num_tokens": 542255069.0, + "step": 20955 + }, + { + "epoch": 2.3013397759718868, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.296961784362793, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7313158512115479, + "num_tokens": 542282401.0, + "step": 20956 + }, + { + "epoch": 2.3014495936745005, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2929940223693848, + "learning_rate": 1e-06, + "loss": 0.7489, + "mean_token_accuracy": 0.7556535601615906, + "num_tokens": 542308418.0, + "step": 20957 + }, + { + "epoch": 2.301559411377114, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3991878032684326, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7277822494506836, + "num_tokens": 542333647.0, + "step": 20958 + }, + { + "epoch": 2.3016692290797276, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.285264730453491, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.732149600982666, + "num_tokens": 542359218.0, + "step": 20959 + }, + { + "epoch": 2.3017790467823414, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2003746032714844, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7396787405014038, + "num_tokens": 542388516.0, + "step": 20960 + }, + { + "epoch": 2.301888864484955, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.243119478225708, + "learning_rate": 1e-06, + "loss": 0.9148, + "mean_token_accuracy": 0.7180994749069214, + "num_tokens": 542417076.0, + "step": 20961 + }, + { + "epoch": 2.301998682187569, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.568206787109375, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7314461469650269, + "num_tokens": 542439792.0, + "step": 20962 + }, + { + "epoch": 2.302108499890182, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.189931631088257, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7253157496452332, + "num_tokens": 542469589.0, + "step": 20963 + }, + { + "epoch": 2.302218317592796, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4040629863739014, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7237739562988281, + "num_tokens": 542494322.0, + "step": 20964 + }, + { + "epoch": 2.3023281352954097, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3531107902526855, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7500649094581604, + "num_tokens": 542519882.0, + "step": 20965 + }, + { + "epoch": 2.3024379529980235, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2699856758117676, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7227676510810852, + "num_tokens": 542548456.0, + "step": 20966 + }, + { + "epoch": 2.3025477707006368, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5063440799713135, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7302882671356201, + "num_tokens": 542573067.0, + "step": 20967 + }, + { + "epoch": 2.3026575884032505, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.227670431137085, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7212465405464172, + "num_tokens": 542602039.0, + "step": 20968 + }, + { + "epoch": 2.3027674061058643, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.8302793502807617, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7506964206695557, + "num_tokens": 542622463.0, + "step": 20969 + }, + { + "epoch": 2.302877223808478, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2786288261413574, + "learning_rate": 1e-06, + "loss": 0.9314, + "mean_token_accuracy": 0.7089731693267822, + "num_tokens": 542649997.0, + "step": 20970 + }, + { + "epoch": 2.3029870415110913, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.190441131591797, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7403619289398193, + "num_tokens": 542679946.0, + "step": 20971 + }, + { + "epoch": 2.303096859213705, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4887471199035645, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7371823787689209, + "num_tokens": 542704940.0, + "step": 20972 + }, + { + "epoch": 2.303206676916319, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1364591121673584, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7261347770690918, + "num_tokens": 542735386.0, + "step": 20973 + }, + { + "epoch": 2.3033164946189326, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3965156078338623, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7400570511817932, + "num_tokens": 542760679.0, + "step": 20974 + }, + { + "epoch": 2.3034263123215464, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0834970474243164, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7427283525466919, + "num_tokens": 542789594.0, + "step": 20975 + }, + { + "epoch": 2.3035361300241597, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.20021915435791, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7216348648071289, + "num_tokens": 542817666.0, + "step": 20976 + }, + { + "epoch": 2.3036459477267734, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.366856813430786, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7505549788475037, + "num_tokens": 542841983.0, + "step": 20977 + }, + { + "epoch": 2.303755765429387, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.295971393585205, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7349319458007812, + "num_tokens": 542869857.0, + "step": 20978 + }, + { + "epoch": 2.303865583132001, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1837575435638428, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7252271771430969, + "num_tokens": 542900501.0, + "step": 20979 + }, + { + "epoch": 2.3039754008346147, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.7453248500823975, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7349807620048523, + "num_tokens": 542922771.0, + "step": 20980 + }, + { + "epoch": 2.304085218537228, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4739441871643066, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7529529929161072, + "num_tokens": 542945980.0, + "step": 20981 + }, + { + "epoch": 2.304195036239842, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5621190071105957, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7197832465171814, + "num_tokens": 542969733.0, + "step": 20982 + }, + { + "epoch": 2.3043048539424555, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2727346420288086, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7444066405296326, + "num_tokens": 542997008.0, + "step": 20983 + }, + { + "epoch": 2.3044146716450693, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.403540849685669, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7367693185806274, + "num_tokens": 543021143.0, + "step": 20984 + }, + { + "epoch": 2.304524489347683, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3702151775360107, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7438245415687561, + "num_tokens": 543045313.0, + "step": 20985 + }, + { + "epoch": 2.3046343070502964, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4829657077789307, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.742910623550415, + "num_tokens": 543068448.0, + "step": 20986 + }, + { + "epoch": 2.30474412475291, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.116344928741455, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7230780124664307, + "num_tokens": 543099282.0, + "step": 20987 + }, + { + "epoch": 2.304853942455524, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6596436500549316, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7345352172851562, + "num_tokens": 543120078.0, + "step": 20988 + }, + { + "epoch": 2.3049637601581376, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.476179599761963, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7348026037216187, + "num_tokens": 543143251.0, + "step": 20989 + }, + { + "epoch": 2.3050735778607514, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3872172832489014, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7210943698883057, + "num_tokens": 543167052.0, + "step": 20990 + }, + { + "epoch": 2.3051833955633647, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.651216506958008, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7296470403671265, + "num_tokens": 543188857.0, + "step": 20991 + }, + { + "epoch": 2.3052932132659785, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.428227424621582, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.734825611114502, + "num_tokens": 543215243.0, + "step": 20992 + }, + { + "epoch": 2.3054030309685922, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5365705490112305, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.725310206413269, + "num_tokens": 543239720.0, + "step": 20993 + }, + { + "epoch": 2.305512848671206, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6703715324401855, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7384009957313538, + "num_tokens": 543261966.0, + "step": 20994 + }, + { + "epoch": 2.3056226663738193, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6853086948394775, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7493253946304321, + "num_tokens": 543281693.0, + "step": 20995 + }, + { + "epoch": 2.305732484076433, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5560994148254395, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7439888715744019, + "num_tokens": 543303138.0, + "step": 20996 + }, + { + "epoch": 2.305842301779047, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.471575975418091, + "learning_rate": 1e-06, + "loss": 0.9365, + "mean_token_accuracy": 0.7076716423034668, + "num_tokens": 543330442.0, + "step": 20997 + }, + { + "epoch": 2.3059521194816606, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1645820140838623, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7171261310577393, + "num_tokens": 543362534.0, + "step": 20998 + }, + { + "epoch": 2.306061937184274, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6870765686035156, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.740638017654419, + "num_tokens": 543382404.0, + "step": 20999 + }, + { + "epoch": 2.3061717548868876, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4604992866516113, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7201635837554932, + "num_tokens": 543407543.0, + "step": 21000 + }, + { + "epoch": 2.3062815725895014, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5195813179016113, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7447007298469543, + "num_tokens": 543430168.0, + "step": 21001 + }, + { + "epoch": 2.306391390292115, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2785322666168213, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7237957715988159, + "num_tokens": 543458488.0, + "step": 21002 + }, + { + "epoch": 2.306501207994729, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6292810440063477, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7414913773536682, + "num_tokens": 543481233.0, + "step": 21003 + }, + { + "epoch": 2.306611025697342, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.472444772720337, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7287611961364746, + "num_tokens": 543504692.0, + "step": 21004 + }, + { + "epoch": 2.306720843399956, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6056175231933594, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7403291463851929, + "num_tokens": 543526621.0, + "step": 21005 + }, + { + "epoch": 2.3068306611025697, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.162721872329712, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7273303866386414, + "num_tokens": 543554414.0, + "step": 21006 + }, + { + "epoch": 2.3069404788051835, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0930161476135254, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.7069034576416016, + "num_tokens": 543587486.0, + "step": 21007 + }, + { + "epoch": 2.3070502965077972, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2845101356506348, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7188718914985657, + "num_tokens": 543615678.0, + "step": 21008 + }, + { + "epoch": 2.3071601142104106, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4808530807495117, + "learning_rate": 1e-06, + "loss": 0.7585, + "mean_token_accuracy": 0.7544287443161011, + "num_tokens": 543639927.0, + "step": 21009 + }, + { + "epoch": 2.3072699319130243, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.304058313369751, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7146387696266174, + "num_tokens": 543668013.0, + "step": 21010 + }, + { + "epoch": 2.307379749615638, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5168769359588623, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7502959966659546, + "num_tokens": 543692129.0, + "step": 21011 + }, + { + "epoch": 2.307489567318252, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6984055042266846, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7440890073776245, + "num_tokens": 543714848.0, + "step": 21012 + }, + { + "epoch": 2.3075993850208656, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4544479846954346, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7285810112953186, + "num_tokens": 543739756.0, + "step": 21013 + }, + { + "epoch": 2.307709202723479, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.558187246322632, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7288392782211304, + "num_tokens": 543761497.0, + "step": 21014 + }, + { + "epoch": 2.3078190204260927, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.729919195175171, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7337601184844971, + "num_tokens": 543783793.0, + "step": 21015 + }, + { + "epoch": 2.3079288381287064, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.407726764678955, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7227439880371094, + "num_tokens": 543809473.0, + "step": 21016 + }, + { + "epoch": 2.30803865583132, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4838554859161377, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7380383014678955, + "num_tokens": 543832119.0, + "step": 21017 + }, + { + "epoch": 2.3081484735339335, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.408550262451172, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7287591695785522, + "num_tokens": 543858050.0, + "step": 21018 + }, + { + "epoch": 2.3082582912365472, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4104373455047607, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7293801307678223, + "num_tokens": 543883102.0, + "step": 21019 + }, + { + "epoch": 2.308368108939161, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6300997734069824, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.7533479928970337, + "num_tokens": 543904191.0, + "step": 21020 + }, + { + "epoch": 2.3084779266417748, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6142287254333496, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7375795841217041, + "num_tokens": 543925282.0, + "step": 21021 + }, + { + "epoch": 2.308587744344388, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.597079277038574, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7287247180938721, + "num_tokens": 543948199.0, + "step": 21022 + }, + { + "epoch": 2.308697562047002, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.048827886581421, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7516964673995972, + "num_tokens": 543978822.0, + "step": 21023 + }, + { + "epoch": 2.3088073797496156, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1922974586486816, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7098748087882996, + "num_tokens": 544009634.0, + "step": 21024 + }, + { + "epoch": 2.3089171974522293, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 1.9821199178695679, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7159017324447632, + "num_tokens": 544045825.0, + "step": 21025 + }, + { + "epoch": 2.309027015154843, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.599276065826416, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7391800284385681, + "num_tokens": 544066655.0, + "step": 21026 + }, + { + "epoch": 2.3091368328574564, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.557349443435669, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7232962846755981, + "num_tokens": 544088602.0, + "step": 21027 + }, + { + "epoch": 2.30924665056007, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.415540933609009, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7442630529403687, + "num_tokens": 544111969.0, + "step": 21028 + }, + { + "epoch": 2.309356468262684, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.36568284034729, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7508052587509155, + "num_tokens": 544136419.0, + "step": 21029 + }, + { + "epoch": 2.3094662859652977, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.0973920822143555, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7236564755439758, + "num_tokens": 544167652.0, + "step": 21030 + }, + { + "epoch": 2.3095761036679114, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.1610000133514404, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7260148525238037, + "num_tokens": 544196676.0, + "step": 21031 + }, + { + "epoch": 2.3096859213705248, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.399730682373047, + "learning_rate": 1e-06, + "loss": 0.7978, + "mean_token_accuracy": 0.7396544218063354, + "num_tokens": 544222160.0, + "step": 21032 + }, + { + "epoch": 2.3097957390731385, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3219923973083496, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7176563739776611, + "num_tokens": 544250547.0, + "step": 21033 + }, + { + "epoch": 2.3099055567757523, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.511599540710449, + "learning_rate": 1e-06, + "loss": 0.912, + "mean_token_accuracy": 0.7166304588317871, + "num_tokens": 544275216.0, + "step": 21034 + }, + { + "epoch": 2.310015374478366, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2500393390655518, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7300698757171631, + "num_tokens": 544302957.0, + "step": 21035 + }, + { + "epoch": 2.31012519218098, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.429527997970581, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7220466732978821, + "num_tokens": 544330432.0, + "step": 21036 + }, + { + "epoch": 2.310235009883593, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.437654972076416, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7300113439559937, + "num_tokens": 544358067.0, + "step": 21037 + }, + { + "epoch": 2.310344827586207, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6764161586761475, + "learning_rate": 1e-06, + "loss": 0.6439, + "mean_token_accuracy": 0.7887054681777954, + "num_tokens": 544378540.0, + "step": 21038 + }, + { + "epoch": 2.3104546452888206, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.450479507446289, + "learning_rate": 1e-06, + "loss": 0.7588, + "mean_token_accuracy": 0.7616499662399292, + "num_tokens": 544402832.0, + "step": 21039 + }, + { + "epoch": 2.3105644629914344, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5141193866729736, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7576903104782104, + "num_tokens": 544424797.0, + "step": 21040 + }, + { + "epoch": 2.310674280694048, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.14884090423584, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7233860492706299, + "num_tokens": 544457150.0, + "step": 21041 + }, + { + "epoch": 2.3107840983966614, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.079080104827881, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7298954725265503, + "num_tokens": 544488923.0, + "step": 21042 + }, + { + "epoch": 2.310893916099275, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4389941692352295, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7368484139442444, + "num_tokens": 544518724.0, + "step": 21043 + }, + { + "epoch": 2.311003733801889, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 1.9955112934112549, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7356926798820496, + "num_tokens": 544551721.0, + "step": 21044 + }, + { + "epoch": 2.3111135515045027, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.575392246246338, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.717876672744751, + "num_tokens": 544575492.0, + "step": 21045 + }, + { + "epoch": 2.311223369207116, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2313122749328613, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7252002954483032, + "num_tokens": 544602952.0, + "step": 21046 + }, + { + "epoch": 2.3113331869097298, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.435643196105957, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7446591854095459, + "num_tokens": 544628679.0, + "step": 21047 + }, + { + "epoch": 2.3114430046123435, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.3891820907592773, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7334986925125122, + "num_tokens": 544655406.0, + "step": 21048 + }, + { + "epoch": 2.3115528223149573, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.5543694496154785, + "learning_rate": 1e-06, + "loss": 0.9171, + "mean_token_accuracy": 0.7212207317352295, + "num_tokens": 544676038.0, + "step": 21049 + }, + { + "epoch": 2.3116626400175706, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6183881759643555, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.7517153024673462, + "num_tokens": 544696723.0, + "step": 21050 + }, + { + "epoch": 2.3117724577201844, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4053335189819336, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.740563154220581, + "num_tokens": 544720954.0, + "step": 21051 + }, + { + "epoch": 2.311882275422798, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2775681018829346, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7291151285171509, + "num_tokens": 544748396.0, + "step": 21052 + }, + { + "epoch": 2.311992093125412, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4404001235961914, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7383739948272705, + "num_tokens": 544771867.0, + "step": 21053 + }, + { + "epoch": 2.3121019108280256, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.4452929496765137, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7506800293922424, + "num_tokens": 544794970.0, + "step": 21054 + }, + { + "epoch": 2.312211728530639, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.278578519821167, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.7455188035964966, + "num_tokens": 544821841.0, + "step": 21055 + }, + { + "epoch": 2.3123215462332527, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.331604480743408, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7272167205810547, + "num_tokens": 544849110.0, + "step": 21056 + }, + { + "epoch": 2.3124313639358665, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.393688917160034, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7251914739608765, + "num_tokens": 544876829.0, + "step": 21057 + }, + { + "epoch": 2.31254118163848, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.211700677871704, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7476640939712524, + "num_tokens": 544904983.0, + "step": 21058 + }, + { + "epoch": 2.312650999341094, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.6807825565338135, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7172398567199707, + "num_tokens": 544926799.0, + "step": 21059 + }, + { + "epoch": 2.3127608170437073, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.231790781021118, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.7040888071060181, + "num_tokens": 544956576.0, + "step": 21060 + }, + { + "epoch": 2.312870634746321, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.635185480117798, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.7501325011253357, + "num_tokens": 544977466.0, + "step": 21061 + }, + { + "epoch": 2.312980452448935, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.2437198162078857, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7152570486068726, + "num_tokens": 545006051.0, + "step": 21062 + }, + { + "epoch": 2.3130902701515486, + "ewc_loss": 1.341104507446289e-05, + "grad_norm": 2.331609010696411, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7173258066177368, + "num_tokens": 545032919.0, + "step": 21063 + }, + { + "epoch": 2.3132000878541623, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2909610271453857, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.729200005531311, + "num_tokens": 545058743.0, + "step": 21064 + }, + { + "epoch": 2.3133099055567756, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5301783084869385, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7498950958251953, + "num_tokens": 545080941.0, + "step": 21065 + }, + { + "epoch": 2.3134197232593894, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.241565704345703, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.7195866703987122, + "num_tokens": 545111743.0, + "step": 21066 + }, + { + "epoch": 2.313529540962003, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.430760622024536, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7204749584197998, + "num_tokens": 545137372.0, + "step": 21067 + }, + { + "epoch": 2.313639358664617, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5164480209350586, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7252280712127686, + "num_tokens": 545161184.0, + "step": 21068 + }, + { + "epoch": 2.31374917636723, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4233922958374023, + "learning_rate": 1e-06, + "loss": 0.7288, + "mean_token_accuracy": 0.7641885280609131, + "num_tokens": 545184304.0, + "step": 21069 + }, + { + "epoch": 2.313858994069844, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3144819736480713, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7235520482063293, + "num_tokens": 545210656.0, + "step": 21070 + }, + { + "epoch": 2.3139688117724577, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4231016635894775, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7573232650756836, + "num_tokens": 545235013.0, + "step": 21071 + }, + { + "epoch": 2.3140786294750715, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3462276458740234, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7298458814620972, + "num_tokens": 545261614.0, + "step": 21072 + }, + { + "epoch": 2.314188447177685, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2722573280334473, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7346967458724976, + "num_tokens": 545288530.0, + "step": 21073 + }, + { + "epoch": 2.3142982648802986, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.27951717376709, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7176201939582825, + "num_tokens": 545316653.0, + "step": 21074 + }, + { + "epoch": 2.3144080825829123, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.327712059020996, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7337590456008911, + "num_tokens": 545344921.0, + "step": 21075 + }, + { + "epoch": 2.314517900285526, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.121863603591919, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7080439925193787, + "num_tokens": 545375523.0, + "step": 21076 + }, + { + "epoch": 2.31462771798814, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.0375375747680664, + "learning_rate": 1e-06, + "loss": 0.8721, + "mean_token_accuracy": 0.7320967316627502, + "num_tokens": 545407587.0, + "step": 21077 + }, + { + "epoch": 2.314737535690753, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.426873207092285, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7451478838920593, + "num_tokens": 545432350.0, + "step": 21078 + }, + { + "epoch": 2.314847353393367, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2772376537323, + "learning_rate": 1e-06, + "loss": 0.8036, + "mean_token_accuracy": 0.7484068870544434, + "num_tokens": 545460827.0, + "step": 21079 + }, + { + "epoch": 2.3149571710959806, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.46083402633667, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7379258871078491, + "num_tokens": 545484607.0, + "step": 21080 + }, + { + "epoch": 2.3150669887985944, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3741540908813477, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7268068790435791, + "num_tokens": 545508854.0, + "step": 21081 + }, + { + "epoch": 2.315176806501208, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3167953491210938, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7025027275085449, + "num_tokens": 545536439.0, + "step": 21082 + }, + { + "epoch": 2.3152866242038215, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3306567668914795, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7298810482025146, + "num_tokens": 545562043.0, + "step": 21083 + }, + { + "epoch": 2.3153964419064352, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.23417592048645, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7340415716171265, + "num_tokens": 545589760.0, + "step": 21084 + }, + { + "epoch": 2.315506259609049, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.322641134262085, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7417421340942383, + "num_tokens": 545614440.0, + "step": 21085 + }, + { + "epoch": 2.3156160773116627, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.303887367248535, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7367323637008667, + "num_tokens": 545639962.0, + "step": 21086 + }, + { + "epoch": 2.3157258950142765, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.6369361877441406, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7185076475143433, + "num_tokens": 545663324.0, + "step": 21087 + }, + { + "epoch": 2.31583571271689, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.621678352355957, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7528063654899597, + "num_tokens": 545685668.0, + "step": 21088 + }, + { + "epoch": 2.3159455304195036, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.454389810562134, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7284115552902222, + "num_tokens": 545709631.0, + "step": 21089 + }, + { + "epoch": 2.3160553481221173, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.512063980102539, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7468242645263672, + "num_tokens": 545733222.0, + "step": 21090 + }, + { + "epoch": 2.316165165824731, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.308026075363159, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7196703553199768, + "num_tokens": 545759942.0, + "step": 21091 + }, + { + "epoch": 2.316274983527345, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.314283847808838, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7479308843612671, + "num_tokens": 545785642.0, + "step": 21092 + }, + { + "epoch": 2.316384801229958, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2399895191192627, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7379138469696045, + "num_tokens": 545814865.0, + "step": 21093 + }, + { + "epoch": 2.316494618932572, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.654695510864258, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.724605917930603, + "num_tokens": 545838748.0, + "step": 21094 + }, + { + "epoch": 2.3166044366351857, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 1.9335386753082275, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7166093587875366, + "num_tokens": 545876189.0, + "step": 21095 + }, + { + "epoch": 2.3167142543377994, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.214766502380371, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.714724063873291, + "num_tokens": 545906308.0, + "step": 21096 + }, + { + "epoch": 2.3168240720404127, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3559043407440186, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7369568347930908, + "num_tokens": 545931183.0, + "step": 21097 + }, + { + "epoch": 2.3169338897430265, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.244917869567871, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7501610517501831, + "num_tokens": 545958608.0, + "step": 21098 + }, + { + "epoch": 2.3170437074456403, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.153040647506714, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7236189842224121, + "num_tokens": 545988236.0, + "step": 21099 + }, + { + "epoch": 2.317153525148254, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.265411376953125, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7270588278770447, + "num_tokens": 546016863.0, + "step": 21100 + }, + { + "epoch": 2.3172633428508673, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.603146553039551, + "learning_rate": 1e-06, + "loss": 0.7401, + "mean_token_accuracy": 0.7586833834648132, + "num_tokens": 546038858.0, + "step": 21101 + }, + { + "epoch": 2.317373160553481, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.377392053604126, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7431102991104126, + "num_tokens": 546064374.0, + "step": 21102 + }, + { + "epoch": 2.317482978256095, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.309626817703247, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7391371726989746, + "num_tokens": 546091560.0, + "step": 21103 + }, + { + "epoch": 2.3175927959587086, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.217413902282715, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7360103130340576, + "num_tokens": 546121413.0, + "step": 21104 + }, + { + "epoch": 2.3177026136613224, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3446342945098877, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7073347568511963, + "num_tokens": 546146179.0, + "step": 21105 + }, + { + "epoch": 2.3178124313639357, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4700255393981934, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.749536395072937, + "num_tokens": 546169669.0, + "step": 21106 + }, + { + "epoch": 2.3179222490665494, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2629597187042236, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.709361732006073, + "num_tokens": 546196400.0, + "step": 21107 + }, + { + "epoch": 2.318032066769163, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.47637939453125, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7374119758605957, + "num_tokens": 546219794.0, + "step": 21108 + }, + { + "epoch": 2.318141884471777, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.082131862640381, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7122890949249268, + "num_tokens": 546252721.0, + "step": 21109 + }, + { + "epoch": 2.3182517021743907, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4505696296691895, + "learning_rate": 1e-06, + "loss": 0.7528, + "mean_token_accuracy": 0.7615009546279907, + "num_tokens": 546275903.0, + "step": 21110 + }, + { + "epoch": 2.318361519877004, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.8594748973846436, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7434122562408447, + "num_tokens": 546296258.0, + "step": 21111 + }, + { + "epoch": 2.3184713375796178, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.459789514541626, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7251589298248291, + "num_tokens": 546320414.0, + "step": 21112 + }, + { + "epoch": 2.3185811552822315, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4787228107452393, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7540680170059204, + "num_tokens": 546341684.0, + "step": 21113 + }, + { + "epoch": 2.3186909729848453, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5429508686065674, + "learning_rate": 1e-06, + "loss": 0.7309, + "mean_token_accuracy": 0.7591621279716492, + "num_tokens": 546362874.0, + "step": 21114 + }, + { + "epoch": 2.318800790687459, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.322341203689575, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7194540500640869, + "num_tokens": 546390052.0, + "step": 21115 + }, + { + "epoch": 2.3189106083900723, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.1591274738311768, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.722256600856781, + "num_tokens": 546419092.0, + "step": 21116 + }, + { + "epoch": 2.319020426092686, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5591189861297607, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.733757734298706, + "num_tokens": 546441330.0, + "step": 21117 + }, + { + "epoch": 2.3191302437953, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.6237473487854004, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7361196279525757, + "num_tokens": 546462439.0, + "step": 21118 + }, + { + "epoch": 2.3192400614979136, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.31138014793396, + "learning_rate": 1e-06, + "loss": 0.7289, + "mean_token_accuracy": 0.7577328681945801, + "num_tokens": 546487977.0, + "step": 21119 + }, + { + "epoch": 2.319349879200527, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.994328022003174, + "learning_rate": 1e-06, + "loss": 0.7652, + "mean_token_accuracy": 0.7524856925010681, + "num_tokens": 546505219.0, + "step": 21120 + }, + { + "epoch": 2.3194596969031407, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.1810269355773926, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7333245873451233, + "num_tokens": 546533972.0, + "step": 21121 + }, + { + "epoch": 2.3195695146057544, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4394683837890625, + "learning_rate": 1e-06, + "loss": 0.8886, + "mean_token_accuracy": 0.7181743383407593, + "num_tokens": 546559537.0, + "step": 21122 + }, + { + "epoch": 2.319679332308368, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5056657791137695, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7131418585777283, + "num_tokens": 546583319.0, + "step": 21123 + }, + { + "epoch": 2.319789150010982, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2473607063293457, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7312840819358826, + "num_tokens": 546609774.0, + "step": 21124 + }, + { + "epoch": 2.3198989677135953, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2068305015563965, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7293534278869629, + "num_tokens": 546637470.0, + "step": 21125 + }, + { + "epoch": 2.320008785416209, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.350721597671509, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7467093467712402, + "num_tokens": 546663033.0, + "step": 21126 + }, + { + "epoch": 2.320118603118823, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.630506992340088, + "learning_rate": 1e-06, + "loss": 0.9865, + "mean_token_accuracy": 0.6935315132141113, + "num_tokens": 546686081.0, + "step": 21127 + }, + { + "epoch": 2.3202284208214365, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4076178073883057, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7454013824462891, + "num_tokens": 546711171.0, + "step": 21128 + }, + { + "epoch": 2.32033823852405, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5803885459899902, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7132937908172607, + "num_tokens": 546733502.0, + "step": 21129 + }, + { + "epoch": 2.3204480562266636, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.249375820159912, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7190216183662415, + "num_tokens": 546762358.0, + "step": 21130 + }, + { + "epoch": 2.3205578739292774, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.442129373550415, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.730563759803772, + "num_tokens": 546787678.0, + "step": 21131 + }, + { + "epoch": 2.320667691631891, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.1413025856018066, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7239418625831604, + "num_tokens": 546818176.0, + "step": 21132 + }, + { + "epoch": 2.320777509334505, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.445728063583374, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7354849576950073, + "num_tokens": 546842222.0, + "step": 21133 + }, + { + "epoch": 2.320887327037118, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4151461124420166, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7333718538284302, + "num_tokens": 546866707.0, + "step": 21134 + }, + { + "epoch": 2.320997144739732, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.404115676879883, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7435941696166992, + "num_tokens": 546891770.0, + "step": 21135 + }, + { + "epoch": 2.3211069624423457, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.6517982482910156, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7268255949020386, + "num_tokens": 546915119.0, + "step": 21136 + }, + { + "epoch": 2.3212167801449595, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.215024948120117, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.7161902189254761, + "num_tokens": 546946013.0, + "step": 21137 + }, + { + "epoch": 2.3213265978475732, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.441842555999756, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.7456125020980835, + "num_tokens": 546971462.0, + "step": 21138 + }, + { + "epoch": 2.3214364155501865, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.247324228286743, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7268392443656921, + "num_tokens": 547000981.0, + "step": 21139 + }, + { + "epoch": 2.3215462332528003, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.522798538208008, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7458516359329224, + "num_tokens": 547025222.0, + "step": 21140 + }, + { + "epoch": 2.321656050955414, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5510008335113525, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7320065498352051, + "num_tokens": 547049261.0, + "step": 21141 + }, + { + "epoch": 2.321765868658028, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4473769664764404, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7488703727722168, + "num_tokens": 547071238.0, + "step": 21142 + }, + { + "epoch": 2.3218756863606416, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.1857640743255615, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7180057764053345, + "num_tokens": 547102348.0, + "step": 21143 + }, + { + "epoch": 2.321985504063255, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.294891119003296, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.722224771976471, + "num_tokens": 547130826.0, + "step": 21144 + }, + { + "epoch": 2.3220953217658686, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.716243267059326, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7400485873222351, + "num_tokens": 547152390.0, + "step": 21145 + }, + { + "epoch": 2.3222051394684824, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3824222087860107, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7264299392700195, + "num_tokens": 547176962.0, + "step": 21146 + }, + { + "epoch": 2.322314957171096, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.333033561706543, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7320706844329834, + "num_tokens": 547201818.0, + "step": 21147 + }, + { + "epoch": 2.3224247748737095, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.354475259780884, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7407147884368896, + "num_tokens": 547227966.0, + "step": 21148 + }, + { + "epoch": 2.322534592576323, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.2984278202056885, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7456614971160889, + "num_tokens": 547254343.0, + "step": 21149 + }, + { + "epoch": 2.322644410278937, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.36582088470459, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7380920052528381, + "num_tokens": 547281730.0, + "step": 21150 + }, + { + "epoch": 2.3227542279815507, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.143603563308716, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7241648435592651, + "num_tokens": 547310001.0, + "step": 21151 + }, + { + "epoch": 2.322864045684164, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.714108943939209, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7481563091278076, + "num_tokens": 547331369.0, + "step": 21152 + }, + { + "epoch": 2.322973863386778, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3372762203216553, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7308865785598755, + "num_tokens": 547357263.0, + "step": 21153 + }, + { + "epoch": 2.3230836810893916, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.241058588027954, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7119197249412537, + "num_tokens": 547386568.0, + "step": 21154 + }, + { + "epoch": 2.3231934987920053, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 1.9714230298995972, + "learning_rate": 1e-06, + "loss": 1.017, + "mean_token_accuracy": 0.6864427328109741, + "num_tokens": 547422711.0, + "step": 21155 + }, + { + "epoch": 2.323303316494619, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3412599563598633, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7165120840072632, + "num_tokens": 547448999.0, + "step": 21156 + }, + { + "epoch": 2.3234131341972324, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.3984122276306152, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7350162267684937, + "num_tokens": 547473272.0, + "step": 21157 + }, + { + "epoch": 2.323522951899846, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.796497344970703, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7379648089408875, + "num_tokens": 547492875.0, + "step": 21158 + }, + { + "epoch": 2.32363276960246, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.104360580444336, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7189657688140869, + "num_tokens": 547526202.0, + "step": 21159 + }, + { + "epoch": 2.3237425873050737, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4477782249450684, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7321478128433228, + "num_tokens": 547551740.0, + "step": 21160 + }, + { + "epoch": 2.3238524050076874, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5332889556884766, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7213870286941528, + "num_tokens": 547576042.0, + "step": 21161 + }, + { + "epoch": 2.3239622227103007, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.202129364013672, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7188624143600464, + "num_tokens": 547605542.0, + "step": 21162 + }, + { + "epoch": 2.3240720404129145, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.0645289421081543, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7049821615219116, + "num_tokens": 547636849.0, + "step": 21163 + }, + { + "epoch": 2.3241818581155282, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.5436854362487793, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7193357348442078, + "num_tokens": 547661953.0, + "step": 21164 + }, + { + "epoch": 2.324291675818142, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.4813225269317627, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7437615394592285, + "num_tokens": 547685654.0, + "step": 21165 + }, + { + "epoch": 2.3244014935207558, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.353530168533325, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7152904272079468, + "num_tokens": 547712398.0, + "step": 21166 + }, + { + "epoch": 2.324511311223369, + "ewc_loss": 1.3530254364013672e-05, + "grad_norm": 2.358680486679077, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.734542191028595, + "num_tokens": 547737811.0, + "step": 21167 + }, + { + "epoch": 2.324621128925983, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.493905544281006, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7312747240066528, + "num_tokens": 547762265.0, + "step": 21168 + }, + { + "epoch": 2.3247309466285966, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.617466926574707, + "learning_rate": 1e-06, + "loss": 0.7966, + "mean_token_accuracy": 0.7440476417541504, + "num_tokens": 547783120.0, + "step": 21169 + }, + { + "epoch": 2.3248407643312103, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.1084609031677246, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7133973836898804, + "num_tokens": 547814700.0, + "step": 21170 + }, + { + "epoch": 2.324950582033824, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.3428637981414795, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7132483720779419, + "num_tokens": 547842383.0, + "step": 21171 + }, + { + "epoch": 2.3250603997364374, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.4154086112976074, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7211906313896179, + "num_tokens": 547868389.0, + "step": 21172 + }, + { + "epoch": 2.325170217439051, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.2530322074890137, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7212477326393127, + "num_tokens": 547898639.0, + "step": 21173 + }, + { + "epoch": 2.325280035141665, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.339564085006714, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7355461120605469, + "num_tokens": 547925825.0, + "step": 21174 + }, + { + "epoch": 2.3253898528442787, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.528934955596924, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7333174347877502, + "num_tokens": 547949645.0, + "step": 21175 + }, + { + "epoch": 2.325499670546892, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.608962297439575, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7198190689086914, + "num_tokens": 547974075.0, + "step": 21176 + }, + { + "epoch": 2.3256094882495058, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.3826510906219482, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7311030626296997, + "num_tokens": 548000066.0, + "step": 21177 + }, + { + "epoch": 2.3257193059521195, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.2342374324798584, + "learning_rate": 1e-06, + "loss": 0.926, + "mean_token_accuracy": 0.7115790843963623, + "num_tokens": 548028726.0, + "step": 21178 + }, + { + "epoch": 2.3258291236547333, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.4944801330566406, + "learning_rate": 1e-06, + "loss": 0.7058, + "mean_token_accuracy": 0.7603170275688171, + "num_tokens": 548050967.0, + "step": 21179 + }, + { + "epoch": 2.3259389413573466, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.4704511165618896, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7291589975357056, + "num_tokens": 548075738.0, + "step": 21180 + }, + { + "epoch": 2.3260487590599603, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.5702707767486572, + "learning_rate": 1e-06, + "loss": 0.7173, + "mean_token_accuracy": 0.767199695110321, + "num_tokens": 548097259.0, + "step": 21181 + }, + { + "epoch": 2.326158576762574, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.176938772201538, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7152765989303589, + "num_tokens": 548128313.0, + "step": 21182 + }, + { + "epoch": 2.326268394465188, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.403902053833008, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7241848707199097, + "num_tokens": 548155143.0, + "step": 21183 + }, + { + "epoch": 2.3263782121678016, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.7032978534698486, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7389750480651855, + "num_tokens": 548176848.0, + "step": 21184 + }, + { + "epoch": 2.326488029870415, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.8057069778442383, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7438747882843018, + "num_tokens": 548196503.0, + "step": 21185 + }, + { + "epoch": 2.3265978475730287, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.821051597595215, + "learning_rate": 1e-06, + "loss": 0.7943, + "mean_token_accuracy": 0.7418611645698547, + "num_tokens": 548214433.0, + "step": 21186 + }, + { + "epoch": 2.3267076652756424, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.414112091064453, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7139348387718201, + "num_tokens": 548242208.0, + "step": 21187 + }, + { + "epoch": 2.326817482978256, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.4096055030822754, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7321619987487793, + "num_tokens": 548267871.0, + "step": 21188 + }, + { + "epoch": 2.32692730068087, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.904792547225952, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7374613881111145, + "num_tokens": 548285796.0, + "step": 21189 + }, + { + "epoch": 2.3270371183834833, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.460103988647461, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7256463170051575, + "num_tokens": 548314029.0, + "step": 21190 + }, + { + "epoch": 2.327146936086097, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.414815664291382, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7590720653533936, + "num_tokens": 548339195.0, + "step": 21191 + }, + { + "epoch": 2.3272567537887108, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.450731039047241, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7284242510795593, + "num_tokens": 548363915.0, + "step": 21192 + }, + { + "epoch": 2.3273665714913245, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.260704278945923, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7259423732757568, + "num_tokens": 548392975.0, + "step": 21193 + }, + { + "epoch": 2.3274763891939383, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.1094610691070557, + "learning_rate": 1e-06, + "loss": 0.9374, + "mean_token_accuracy": 0.7056372761726379, + "num_tokens": 548423278.0, + "step": 21194 + }, + { + "epoch": 2.3275862068965516, + "ewc_loss": 1.3589859008789062e-05, + "grad_norm": 2.358879566192627, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7365065217018127, + "num_tokens": 548451510.0, + "step": 21195 + }, + { + "epoch": 2.3276960245991654, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3622257709503174, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7389642596244812, + "num_tokens": 548477218.0, + "step": 21196 + }, + { + "epoch": 2.327805842301779, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3246068954467773, + "learning_rate": 1e-06, + "loss": 0.8057, + "mean_token_accuracy": 0.7495732307434082, + "num_tokens": 548503026.0, + "step": 21197 + }, + { + "epoch": 2.327915660004393, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.6259212493896484, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7399467825889587, + "num_tokens": 548525537.0, + "step": 21198 + }, + { + "epoch": 2.328025477707006, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.2872655391693115, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7161758542060852, + "num_tokens": 548554097.0, + "step": 21199 + }, + { + "epoch": 2.32813529540962, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4421536922454834, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7346646785736084, + "num_tokens": 548578518.0, + "step": 21200 + }, + { + "epoch": 2.3282451131122337, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3810575008392334, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7350128293037415, + "num_tokens": 548603344.0, + "step": 21201 + }, + { + "epoch": 2.3283549308148475, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.2266318798065186, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.7480179667472839, + "num_tokens": 548627826.0, + "step": 21202 + }, + { + "epoch": 2.3284647485174608, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.375241994857788, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7417446374893188, + "num_tokens": 548652117.0, + "step": 21203 + }, + { + "epoch": 2.3285745662200745, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4369752407073975, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7488899230957031, + "num_tokens": 548677237.0, + "step": 21204 + }, + { + "epoch": 2.3286843839226883, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.102390766143799, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7268047332763672, + "num_tokens": 548709854.0, + "step": 21205 + }, + { + "epoch": 2.328794201625302, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3560898303985596, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7216627597808838, + "num_tokens": 548736894.0, + "step": 21206 + }, + { + "epoch": 2.328904019327916, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5247116088867188, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7401870489120483, + "num_tokens": 548758761.0, + "step": 21207 + }, + { + "epoch": 2.329013837030529, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.428184747695923, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7247231006622314, + "num_tokens": 548783443.0, + "step": 21208 + }, + { + "epoch": 2.329123654733143, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3939368724823, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7194607257843018, + "num_tokens": 548809750.0, + "step": 21209 + }, + { + "epoch": 2.3292334724357566, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1176064014434814, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7277861833572388, + "num_tokens": 548839854.0, + "step": 21210 + }, + { + "epoch": 2.3293432901383704, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.386646270751953, + "learning_rate": 1e-06, + "loss": 0.7649, + "mean_token_accuracy": 0.7557470798492432, + "num_tokens": 548864134.0, + "step": 21211 + }, + { + "epoch": 2.329453107840984, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3345954418182373, + "learning_rate": 1e-06, + "loss": 0.9443, + "mean_token_accuracy": 0.7050338387489319, + "num_tokens": 548893718.0, + "step": 21212 + }, + { + "epoch": 2.3295629255435975, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3533596992492676, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7186712026596069, + "num_tokens": 548920479.0, + "step": 21213 + }, + { + "epoch": 2.329672743246211, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.6108179092407227, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7256301045417786, + "num_tokens": 548944177.0, + "step": 21214 + }, + { + "epoch": 2.329782560948825, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.535536527633667, + "learning_rate": 1e-06, + "loss": 0.9013, + "mean_token_accuracy": 0.7167830467224121, + "num_tokens": 548969981.0, + "step": 21215 + }, + { + "epoch": 2.3298923786514387, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.6479711532592773, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7196589112281799, + "num_tokens": 548991299.0, + "step": 21216 + }, + { + "epoch": 2.3300021963540525, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.593169927597046, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7352360486984253, + "num_tokens": 549013991.0, + "step": 21217 + }, + { + "epoch": 2.330112014056666, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3402035236358643, + "learning_rate": 1e-06, + "loss": 0.7417, + "mean_token_accuracy": 0.7612709403038025, + "num_tokens": 549038032.0, + "step": 21218 + }, + { + "epoch": 2.3302218317592795, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3605434894561768, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7084908485412598, + "num_tokens": 549064922.0, + "step": 21219 + }, + { + "epoch": 2.3303316494618933, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.0582962036132812, + "learning_rate": 1e-06, + "loss": 0.9748, + "mean_token_accuracy": 0.6988161206245422, + "num_tokens": 549100841.0, + "step": 21220 + }, + { + "epoch": 2.330441467164507, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5064048767089844, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7177678942680359, + "num_tokens": 549125312.0, + "step": 21221 + }, + { + "epoch": 2.330551284867121, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.8006768226623535, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7398370504379272, + "num_tokens": 549144977.0, + "step": 21222 + }, + { + "epoch": 2.330661102569734, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.111254930496216, + "learning_rate": 1e-06, + "loss": 0.9902, + "mean_token_accuracy": 0.6973415613174438, + "num_tokens": 549178168.0, + "step": 21223 + }, + { + "epoch": 2.330770920272348, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4114584922790527, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7373713850975037, + "num_tokens": 549203052.0, + "step": 21224 + }, + { + "epoch": 2.3308807379749616, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.454885482788086, + "learning_rate": 1e-06, + "loss": 0.7674, + "mean_token_accuracy": 0.7595045566558838, + "num_tokens": 549226087.0, + "step": 21225 + }, + { + "epoch": 2.3309905556775754, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3771016597747803, + "learning_rate": 1e-06, + "loss": 0.7585, + "mean_token_accuracy": 0.7536518573760986, + "num_tokens": 549251485.0, + "step": 21226 + }, + { + "epoch": 2.3311003733801887, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.240610122680664, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7186058163642883, + "num_tokens": 549280757.0, + "step": 21227 + }, + { + "epoch": 2.3312101910828025, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1029436588287354, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7289524078369141, + "num_tokens": 549311095.0, + "step": 21228 + }, + { + "epoch": 2.3313200087854162, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4092941284179688, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7364211082458496, + "num_tokens": 549336923.0, + "step": 21229 + }, + { + "epoch": 2.33142982648803, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.2343218326568604, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.7163983583450317, + "num_tokens": 549367414.0, + "step": 21230 + }, + { + "epoch": 2.3315396441906433, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5784645080566406, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7413493394851685, + "num_tokens": 549390980.0, + "step": 21231 + }, + { + "epoch": 2.331649461893257, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.435582399368286, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7390018105506897, + "num_tokens": 549415582.0, + "step": 21232 + }, + { + "epoch": 2.331759279595871, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3347885608673096, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7081261873245239, + "num_tokens": 549442023.0, + "step": 21233 + }, + { + "epoch": 2.3318690972984846, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.270871639251709, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7440600991249084, + "num_tokens": 549468923.0, + "step": 21234 + }, + { + "epoch": 2.3319789150010983, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.612180709838867, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7168396711349487, + "num_tokens": 549493366.0, + "step": 21235 + }, + { + "epoch": 2.3320887327037116, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1971633434295654, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7237856984138489, + "num_tokens": 549521336.0, + "step": 21236 + }, + { + "epoch": 2.3321985504063254, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4034512042999268, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7146197557449341, + "num_tokens": 549547172.0, + "step": 21237 + }, + { + "epoch": 2.332308368108939, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.802665948867798, + "learning_rate": 1e-06, + "loss": 0.7646, + "mean_token_accuracy": 0.7544115781784058, + "num_tokens": 549566702.0, + "step": 21238 + }, + { + "epoch": 2.332418185811553, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3774771690368652, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7512809038162231, + "num_tokens": 549591628.0, + "step": 21239 + }, + { + "epoch": 2.3325280035141667, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.268653392791748, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.717725932598114, + "num_tokens": 549618880.0, + "step": 21240 + }, + { + "epoch": 2.33263782121678, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 1.9355688095092773, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7277522087097168, + "num_tokens": 549652810.0, + "step": 21241 + }, + { + "epoch": 2.3327476389193937, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.821226119995117, + "learning_rate": 1e-06, + "loss": 0.7013, + "mean_token_accuracy": 0.7710628509521484, + "num_tokens": 549671444.0, + "step": 21242 + }, + { + "epoch": 2.3328574566220075, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.807112455368042, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.7528551816940308, + "num_tokens": 549691769.0, + "step": 21243 + }, + { + "epoch": 2.3329672743246213, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.54002046585083, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7306501865386963, + "num_tokens": 549715034.0, + "step": 21244 + }, + { + "epoch": 2.333077092027235, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.274209976196289, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7353577613830566, + "num_tokens": 549740360.0, + "step": 21245 + }, + { + "epoch": 2.3331869097298483, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.296980142593384, + "learning_rate": 1e-06, + "loss": 0.7511, + "mean_token_accuracy": 0.7574126124382019, + "num_tokens": 549766285.0, + "step": 21246 + }, + { + "epoch": 2.333296727432462, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1356399059295654, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7099565863609314, + "num_tokens": 549796952.0, + "step": 21247 + }, + { + "epoch": 2.333406545135076, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.7012622356414795, + "learning_rate": 1e-06, + "loss": 0.7687, + "mean_token_accuracy": 0.7516028881072998, + "num_tokens": 549816546.0, + "step": 21248 + }, + { + "epoch": 2.3335163628376896, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.360971212387085, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7360154390335083, + "num_tokens": 549840980.0, + "step": 21249 + }, + { + "epoch": 2.333626180540303, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.537977457046509, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7109343409538269, + "num_tokens": 549865581.0, + "step": 21250 + }, + { + "epoch": 2.3337359982429167, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3849267959594727, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7297344207763672, + "num_tokens": 549893505.0, + "step": 21251 + }, + { + "epoch": 2.3338458159455304, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.7318310737609863, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7376047968864441, + "num_tokens": 549913166.0, + "step": 21252 + }, + { + "epoch": 2.333955633648144, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5022499561309814, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7310041189193726, + "num_tokens": 549936725.0, + "step": 21253 + }, + { + "epoch": 2.334065451350758, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.2285196781158447, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7359135150909424, + "num_tokens": 549965484.0, + "step": 21254 + }, + { + "epoch": 2.3341752690533712, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.282240629196167, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7329083681106567, + "num_tokens": 549990882.0, + "step": 21255 + }, + { + "epoch": 2.334285086755985, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3080062866210938, + "learning_rate": 1e-06, + "loss": 0.9391, + "mean_token_accuracy": 0.7079017162322998, + "num_tokens": 550019734.0, + "step": 21256 + }, + { + "epoch": 2.3343949044585988, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.617194175720215, + "learning_rate": 1e-06, + "loss": 0.7833, + "mean_token_accuracy": 0.7501492500305176, + "num_tokens": 550040980.0, + "step": 21257 + }, + { + "epoch": 2.3345047221612125, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3225386142730713, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7199335694313049, + "num_tokens": 550065783.0, + "step": 21258 + }, + { + "epoch": 2.334614539863826, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1785213947296143, + "learning_rate": 1e-06, + "loss": 0.7683, + "mean_token_accuracy": 0.7498406171798706, + "num_tokens": 550094010.0, + "step": 21259 + }, + { + "epoch": 2.3347243575664396, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.7297208309173584, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7253159284591675, + "num_tokens": 550114344.0, + "step": 21260 + }, + { + "epoch": 2.3348341752690533, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.4220755100250244, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7270615100860596, + "num_tokens": 550142710.0, + "step": 21261 + }, + { + "epoch": 2.334943992971667, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.313513994216919, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7404719591140747, + "num_tokens": 550168823.0, + "step": 21262 + }, + { + "epoch": 2.335053810674281, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.358421802520752, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7340792417526245, + "num_tokens": 550193547.0, + "step": 21263 + }, + { + "epoch": 2.335163628376894, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3238179683685303, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7342910170555115, + "num_tokens": 550219685.0, + "step": 21264 + }, + { + "epoch": 2.335273446079508, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.204641819000244, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7150901556015015, + "num_tokens": 550248353.0, + "step": 21265 + }, + { + "epoch": 2.3353832637821217, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1650989055633545, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7413464784622192, + "num_tokens": 550278798.0, + "step": 21266 + }, + { + "epoch": 2.3354930814847354, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3767917156219482, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.717842698097229, + "num_tokens": 550306111.0, + "step": 21267 + }, + { + "epoch": 2.335602899187349, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.32131290435791, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7424561977386475, + "num_tokens": 550333461.0, + "step": 21268 + }, + { + "epoch": 2.3357127168899625, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5802927017211914, + "learning_rate": 1e-06, + "loss": 0.6659, + "mean_token_accuracy": 0.7785639762878418, + "num_tokens": 550354430.0, + "step": 21269 + }, + { + "epoch": 2.3358225345925763, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5136401653289795, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.728900671005249, + "num_tokens": 550379375.0, + "step": 21270 + }, + { + "epoch": 2.33593235229519, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1900734901428223, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7177730798721313, + "num_tokens": 550408939.0, + "step": 21271 + }, + { + "epoch": 2.336042169997804, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.6355977058410645, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.737092137336731, + "num_tokens": 550430170.0, + "step": 21272 + }, + { + "epoch": 2.3361519877004175, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.865628719329834, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.7561072111129761, + "num_tokens": 550453836.0, + "step": 21273 + }, + { + "epoch": 2.336261805403031, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.0614497661590576, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7212215662002563, + "num_tokens": 550485340.0, + "step": 21274 + }, + { + "epoch": 2.3363716231056446, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.0824134349823, + "learning_rate": 1e-06, + "loss": 0.9676, + "mean_token_accuracy": 0.7045873999595642, + "num_tokens": 550520121.0, + "step": 21275 + }, + { + "epoch": 2.3364814408082584, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5318069458007812, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7384294867515564, + "num_tokens": 550545685.0, + "step": 21276 + }, + { + "epoch": 2.336591258510872, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.196991205215454, + "learning_rate": 1e-06, + "loss": 0.9842, + "mean_token_accuracy": 0.6954087018966675, + "num_tokens": 550576946.0, + "step": 21277 + }, + { + "epoch": 2.3367010762134854, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.5655722618103027, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7501869201660156, + "num_tokens": 550598771.0, + "step": 21278 + }, + { + "epoch": 2.336810893916099, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.3726143836975098, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.722055196762085, + "num_tokens": 550625923.0, + "step": 21279 + }, + { + "epoch": 2.336920711618713, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.2382843494415283, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7230480909347534, + "num_tokens": 550654256.0, + "step": 21280 + }, + { + "epoch": 2.3370305293213267, + "ewc_loss": 1.3649463653564453e-05, + "grad_norm": 2.1516551971435547, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7184411287307739, + "num_tokens": 550684929.0, + "step": 21281 + }, + { + "epoch": 2.33714034702394, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3325910568237305, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7208447456359863, + "num_tokens": 550712460.0, + "step": 21282 + }, + { + "epoch": 2.337250164726554, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.231881618499756, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7328413724899292, + "num_tokens": 550740829.0, + "step": 21283 + }, + { + "epoch": 2.3373599824291675, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.37019681930542, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7306930422782898, + "num_tokens": 550767310.0, + "step": 21284 + }, + { + "epoch": 2.3374698001317813, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2107460498809814, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7435120940208435, + "num_tokens": 550793297.0, + "step": 21285 + }, + { + "epoch": 2.337579617834395, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.511423349380493, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7400160431861877, + "num_tokens": 550814235.0, + "step": 21286 + }, + { + "epoch": 2.3376894355370084, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4130468368530273, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7546073794364929, + "num_tokens": 550838545.0, + "step": 21287 + }, + { + "epoch": 2.337799253239622, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.544337749481201, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7288272380828857, + "num_tokens": 550861019.0, + "step": 21288 + }, + { + "epoch": 2.337909070942236, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4715206623077393, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.7321184873580933, + "num_tokens": 550884329.0, + "step": 21289 + }, + { + "epoch": 2.3380188886448496, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.144587755203247, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.719545304775238, + "num_tokens": 550914916.0, + "step": 21290 + }, + { + "epoch": 2.3381287063474634, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.6381070613861084, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7430378794670105, + "num_tokens": 550937653.0, + "step": 21291 + }, + { + "epoch": 2.3382385240500767, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.320023536682129, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7211787700653076, + "num_tokens": 550964748.0, + "step": 21292 + }, + { + "epoch": 2.3383483417526905, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4797394275665283, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7350020408630371, + "num_tokens": 550988746.0, + "step": 21293 + }, + { + "epoch": 2.338458159455304, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3455069065093994, + "learning_rate": 1e-06, + "loss": 0.7826, + "mean_token_accuracy": 0.7452409267425537, + "num_tokens": 551012690.0, + "step": 21294 + }, + { + "epoch": 2.338567977157918, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.0594871044158936, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7387173175811768, + "num_tokens": 551046199.0, + "step": 21295 + }, + { + "epoch": 2.3386777948605317, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3021597862243652, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7235602140426636, + "num_tokens": 551072392.0, + "step": 21296 + }, + { + "epoch": 2.338787612563145, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4918174743652344, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7193387746810913, + "num_tokens": 551096181.0, + "step": 21297 + }, + { + "epoch": 2.338897430265759, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3759100437164307, + "learning_rate": 1e-06, + "loss": 0.9783, + "mean_token_accuracy": 0.6975703835487366, + "num_tokens": 551126784.0, + "step": 21298 + }, + { + "epoch": 2.3390072479683726, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.1664609909057617, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7175865769386292, + "num_tokens": 551157975.0, + "step": 21299 + }, + { + "epoch": 2.3391170656709863, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.1864185333251953, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7160139679908752, + "num_tokens": 551188296.0, + "step": 21300 + }, + { + "epoch": 2.3392268833735996, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.5358810424804688, + "learning_rate": 1e-06, + "loss": 0.7586, + "mean_token_accuracy": 0.7490361928939819, + "num_tokens": 551209808.0, + "step": 21301 + }, + { + "epoch": 2.3393367010762134, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2089128494262695, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7288368940353394, + "num_tokens": 551237982.0, + "step": 21302 + }, + { + "epoch": 2.339446518778827, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3409759998321533, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7364389896392822, + "num_tokens": 551263064.0, + "step": 21303 + }, + { + "epoch": 2.339556336481441, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.357949733734131, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7409968376159668, + "num_tokens": 551290330.0, + "step": 21304 + }, + { + "epoch": 2.3396661541840547, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2407186031341553, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7379142642021179, + "num_tokens": 551324588.0, + "step": 21305 + }, + { + "epoch": 2.339775971886668, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.5759987831115723, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.728917121887207, + "num_tokens": 551346696.0, + "step": 21306 + }, + { + "epoch": 2.3398857895892817, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.315704107284546, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.708187460899353, + "num_tokens": 551374637.0, + "step": 21307 + }, + { + "epoch": 2.3399956072918955, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2431092262268066, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7267621755599976, + "num_tokens": 551402293.0, + "step": 21308 + }, + { + "epoch": 2.3401054249945092, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3088881969451904, + "learning_rate": 1e-06, + "loss": 0.9012, + "mean_token_accuracy": 0.7188743948936462, + "num_tokens": 551430867.0, + "step": 21309 + }, + { + "epoch": 2.3402152426971226, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.490459680557251, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7126525640487671, + "num_tokens": 551456102.0, + "step": 21310 + }, + { + "epoch": 2.3403250603997363, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.5123438835144043, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7284784317016602, + "num_tokens": 551479576.0, + "step": 21311 + }, + { + "epoch": 2.34043487810235, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.14518404006958, + "learning_rate": 1e-06, + "loss": 0.7464, + "mean_token_accuracy": 0.7646253705024719, + "num_tokens": 551508248.0, + "step": 21312 + }, + { + "epoch": 2.340544695804964, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.6313226222991943, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7286518216133118, + "num_tokens": 551529376.0, + "step": 21313 + }, + { + "epoch": 2.3406545135075776, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.291564702987671, + "learning_rate": 1e-06, + "loss": 0.8187, + "mean_token_accuracy": 0.7402019500732422, + "num_tokens": 551554610.0, + "step": 21314 + }, + { + "epoch": 2.340764331210191, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.633678436279297, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7322033643722534, + "num_tokens": 551579653.0, + "step": 21315 + }, + { + "epoch": 2.3408741489128047, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3930318355560303, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7381042242050171, + "num_tokens": 551603025.0, + "step": 21316 + }, + { + "epoch": 2.3409839666154184, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.329756736755371, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7259650826454163, + "num_tokens": 551631007.0, + "step": 21317 + }, + { + "epoch": 2.341093784318032, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.230396032333374, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6969395875930786, + "num_tokens": 551662782.0, + "step": 21318 + }, + { + "epoch": 2.341203602020646, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.7752413749694824, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7333111763000488, + "num_tokens": 551683837.0, + "step": 21319 + }, + { + "epoch": 2.3413134197232592, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.538059949874878, + "learning_rate": 1e-06, + "loss": 0.7738, + "mean_token_accuracy": 0.7583613991737366, + "num_tokens": 551705496.0, + "step": 21320 + }, + { + "epoch": 2.341423237425873, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.631608247756958, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7265638709068298, + "num_tokens": 551728394.0, + "step": 21321 + }, + { + "epoch": 2.3415330551284868, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.638986587524414, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7442851066589355, + "num_tokens": 551748347.0, + "step": 21322 + }, + { + "epoch": 2.3416428728311005, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4796700477600098, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.741723895072937, + "num_tokens": 551771301.0, + "step": 21323 + }, + { + "epoch": 2.3417526905337143, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.6339635848999023, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7429272532463074, + "num_tokens": 551791860.0, + "step": 21324 + }, + { + "epoch": 2.3418625082363276, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2427358627319336, + "learning_rate": 1e-06, + "loss": 0.8735, + "mean_token_accuracy": 0.7318731546401978, + "num_tokens": 551820556.0, + "step": 21325 + }, + { + "epoch": 2.3419723259389413, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4296326637268066, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.740179181098938, + "num_tokens": 551845083.0, + "step": 21326 + }, + { + "epoch": 2.342082143641555, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.547288179397583, + "learning_rate": 1e-06, + "loss": 0.7452, + "mean_token_accuracy": 0.7569630146026611, + "num_tokens": 551866820.0, + "step": 21327 + }, + { + "epoch": 2.342191961344169, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.416011095046997, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7389084100723267, + "num_tokens": 551890731.0, + "step": 21328 + }, + { + "epoch": 2.342301779046782, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.1732635498046875, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7222053408622742, + "num_tokens": 551919352.0, + "step": 21329 + }, + { + "epoch": 2.342411596749396, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.424499273300171, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.7223348617553711, + "num_tokens": 551946592.0, + "step": 21330 + }, + { + "epoch": 2.3425214144520097, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4174842834472656, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7198114395141602, + "num_tokens": 551972924.0, + "step": 21331 + }, + { + "epoch": 2.3426312321546234, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.5397489070892334, + "learning_rate": 1e-06, + "loss": 0.7645, + "mean_token_accuracy": 0.7557315826416016, + "num_tokens": 551994666.0, + "step": 21332 + }, + { + "epoch": 2.3427410498572367, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4171698093414307, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7176842093467712, + "num_tokens": 552021422.0, + "step": 21333 + }, + { + "epoch": 2.3428508675598505, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2411062717437744, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7265779972076416, + "num_tokens": 552051876.0, + "step": 21334 + }, + { + "epoch": 2.3429606852624643, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.605496406555176, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7433806657791138, + "num_tokens": 552074281.0, + "step": 21335 + }, + { + "epoch": 2.343070502965078, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.312763214111328, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7342733144760132, + "num_tokens": 552101939.0, + "step": 21336 + }, + { + "epoch": 2.3431803206676918, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2011194229125977, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7348911762237549, + "num_tokens": 552133240.0, + "step": 21337 + }, + { + "epoch": 2.343290138370305, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.222942590713501, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7458177804946899, + "num_tokens": 552160652.0, + "step": 21338 + }, + { + "epoch": 2.343399956072919, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3060624599456787, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7232736349105835, + "num_tokens": 552188326.0, + "step": 21339 + }, + { + "epoch": 2.3435097737755326, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4901342391967773, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7505581378936768, + "num_tokens": 552211141.0, + "step": 21340 + }, + { + "epoch": 2.3436195914781464, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.536177635192871, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7315753698348999, + "num_tokens": 552235120.0, + "step": 21341 + }, + { + "epoch": 2.34372940918076, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4899680614471436, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.746789813041687, + "num_tokens": 552257858.0, + "step": 21342 + }, + { + "epoch": 2.3438392268833734, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2638025283813477, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7395247220993042, + "num_tokens": 552286546.0, + "step": 21343 + }, + { + "epoch": 2.343949044585987, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.357874631881714, + "learning_rate": 1e-06, + "loss": 0.763, + "mean_token_accuracy": 0.7541128396987915, + "num_tokens": 552308667.0, + "step": 21344 + }, + { + "epoch": 2.344058862288601, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.1522114276885986, + "learning_rate": 1e-06, + "loss": 0.9848, + "mean_token_accuracy": 0.6921699643135071, + "num_tokens": 552342909.0, + "step": 21345 + }, + { + "epoch": 2.3441686799912147, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2819910049438477, + "learning_rate": 1e-06, + "loss": 0.7674, + "mean_token_accuracy": 0.7520899772644043, + "num_tokens": 552368696.0, + "step": 21346 + }, + { + "epoch": 2.3442784976938285, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.286844253540039, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7069072723388672, + "num_tokens": 552395639.0, + "step": 21347 + }, + { + "epoch": 2.3443883153964418, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.5035674571990967, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7196187376976013, + "num_tokens": 552420544.0, + "step": 21348 + }, + { + "epoch": 2.3444981330990555, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.797513484954834, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7443018555641174, + "num_tokens": 552439038.0, + "step": 21349 + }, + { + "epoch": 2.3446079508016693, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3719348907470703, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7359505891799927, + "num_tokens": 552464579.0, + "step": 21350 + }, + { + "epoch": 2.344717768504283, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.194563865661621, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7233157753944397, + "num_tokens": 552495383.0, + "step": 21351 + }, + { + "epoch": 2.344827586206897, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3062517642974854, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7218934297561646, + "num_tokens": 552520115.0, + "step": 21352 + }, + { + "epoch": 2.34493740390951, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.6328864097595215, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7361950874328613, + "num_tokens": 552541490.0, + "step": 21353 + }, + { + "epoch": 2.345047221612124, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.0934267044067383, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7253544330596924, + "num_tokens": 552571728.0, + "step": 21354 + }, + { + "epoch": 2.3451570393147376, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.2098002433776855, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7333781719207764, + "num_tokens": 552599496.0, + "step": 21355 + }, + { + "epoch": 2.3452668570173514, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.3540995121002197, + "learning_rate": 1e-06, + "loss": 0.8987, + "mean_token_accuracy": 0.7202787399291992, + "num_tokens": 552627444.0, + "step": 21356 + }, + { + "epoch": 2.3453766747199647, + "ewc_loss": 1.3768672943115234e-05, + "grad_norm": 2.4667270183563232, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7333971858024597, + "num_tokens": 552651759.0, + "step": 21357 + }, + { + "epoch": 2.3454864924225785, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.105895757675171, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7348130941390991, + "num_tokens": 552684943.0, + "step": 21358 + }, + { + "epoch": 2.345596310125192, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.487779140472412, + "learning_rate": 1e-06, + "loss": 0.8356, + "mean_token_accuracy": 0.7299374341964722, + "num_tokens": 552708085.0, + "step": 21359 + }, + { + "epoch": 2.345706127827806, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.415286064147949, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7143615484237671, + "num_tokens": 552738251.0, + "step": 21360 + }, + { + "epoch": 2.3458159455304193, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1628665924072266, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7385292053222656, + "num_tokens": 552770263.0, + "step": 21361 + }, + { + "epoch": 2.345925763233033, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1800882816314697, + "learning_rate": 1e-06, + "loss": 0.9526, + "mean_token_accuracy": 0.7148545980453491, + "num_tokens": 552799588.0, + "step": 21362 + }, + { + "epoch": 2.346035580935647, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2813897132873535, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7282733917236328, + "num_tokens": 552825885.0, + "step": 21363 + }, + { + "epoch": 2.3461453986382605, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5984039306640625, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7220761775970459, + "num_tokens": 552849550.0, + "step": 21364 + }, + { + "epoch": 2.3462552163408743, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.751781702041626, + "learning_rate": 1e-06, + "loss": 0.7684, + "mean_token_accuracy": 0.7485529780387878, + "num_tokens": 552870271.0, + "step": 21365 + }, + { + "epoch": 2.3463650340434876, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1466524600982666, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7436888217926025, + "num_tokens": 552897487.0, + "step": 21366 + }, + { + "epoch": 2.3464748517461014, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.185117244720459, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7157186269760132, + "num_tokens": 552931433.0, + "step": 21367 + }, + { + "epoch": 2.346584669448715, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.632467746734619, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7354971170425415, + "num_tokens": 552952389.0, + "step": 21368 + }, + { + "epoch": 2.346694487151329, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2556962966918945, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7300195693969727, + "num_tokens": 552980525.0, + "step": 21369 + }, + { + "epoch": 2.3468043048539426, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.24617600440979, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7285596132278442, + "num_tokens": 553006900.0, + "step": 21370 + }, + { + "epoch": 2.346914122556556, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.31892991065979, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7243771553039551, + "num_tokens": 553032119.0, + "step": 21371 + }, + { + "epoch": 2.3470239402591697, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.7354609966278076, + "learning_rate": 1e-06, + "loss": 0.7743, + "mean_token_accuracy": 0.7547605633735657, + "num_tokens": 553051757.0, + "step": 21372 + }, + { + "epoch": 2.3471337579617835, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 3.0209755897521973, + "learning_rate": 1e-06, + "loss": 0.7497, + "mean_token_accuracy": 0.7548649311065674, + "num_tokens": 553068911.0, + "step": 21373 + }, + { + "epoch": 2.3472435756643972, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3988187313079834, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7425128221511841, + "num_tokens": 553091522.0, + "step": 21374 + }, + { + "epoch": 2.347353393367011, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.6614084243774414, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7264425158500671, + "num_tokens": 553114261.0, + "step": 21375 + }, + { + "epoch": 2.3474632110696243, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3974530696868896, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7225635051727295, + "num_tokens": 553138696.0, + "step": 21376 + }, + { + "epoch": 2.347573028772238, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5362279415130615, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.7112419009208679, + "num_tokens": 553164347.0, + "step": 21377 + }, + { + "epoch": 2.347682846474852, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.073695182800293, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7070780992507935, + "num_tokens": 553196295.0, + "step": 21378 + }, + { + "epoch": 2.3477926641774656, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2312307357788086, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.721251368522644, + "num_tokens": 553224688.0, + "step": 21379 + }, + { + "epoch": 2.347902481880079, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.4146785736083984, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7355002164840698, + "num_tokens": 553248386.0, + "step": 21380 + }, + { + "epoch": 2.3480122995826926, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3529303073883057, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7199143171310425, + "num_tokens": 553275638.0, + "step": 21381 + }, + { + "epoch": 2.3481221172853064, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.7175838947296143, + "learning_rate": 1e-06, + "loss": 0.7135, + "mean_token_accuracy": 0.7604987025260925, + "num_tokens": 553294839.0, + "step": 21382 + }, + { + "epoch": 2.34823193498792, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.602717638015747, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.7536475658416748, + "num_tokens": 553316714.0, + "step": 21383 + }, + { + "epoch": 2.3483417526905335, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1516318321228027, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7213338613510132, + "num_tokens": 553346341.0, + "step": 21384 + }, + { + "epoch": 2.3484515703931472, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5324180126190186, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7250065207481384, + "num_tokens": 553368820.0, + "step": 21385 + }, + { + "epoch": 2.348561388095761, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5660998821258545, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7526751756668091, + "num_tokens": 553390215.0, + "step": 21386 + }, + { + "epoch": 2.3486712057983747, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.39089035987854, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7267393469810486, + "num_tokens": 553416577.0, + "step": 21387 + }, + { + "epoch": 2.3487810235009885, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.255723714828491, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7443239092826843, + "num_tokens": 553443738.0, + "step": 21388 + }, + { + "epoch": 2.348890841203602, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.6069962978363037, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.721705436706543, + "num_tokens": 553465023.0, + "step": 21389 + }, + { + "epoch": 2.3490006589062156, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5018229484558105, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7115827202796936, + "num_tokens": 553489967.0, + "step": 21390 + }, + { + "epoch": 2.3491104766088293, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.296002149581909, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.744843602180481, + "num_tokens": 553516728.0, + "step": 21391 + }, + { + "epoch": 2.349220294311443, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5089471340179443, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7447119951248169, + "num_tokens": 553538784.0, + "step": 21392 + }, + { + "epoch": 2.349330112014057, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2052524089813232, + "learning_rate": 1e-06, + "loss": 0.9678, + "mean_token_accuracy": 0.6942092180252075, + "num_tokens": 553568410.0, + "step": 21393 + }, + { + "epoch": 2.34943992971667, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2300078868865967, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7492862939834595, + "num_tokens": 553597423.0, + "step": 21394 + }, + { + "epoch": 2.349549747419284, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.576547861099243, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.733411431312561, + "num_tokens": 553619223.0, + "step": 21395 + }, + { + "epoch": 2.3496595651218977, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.0972402095794678, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7187081575393677, + "num_tokens": 553648604.0, + "step": 21396 + }, + { + "epoch": 2.3497693828245114, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.7738800048828125, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7419461011886597, + "num_tokens": 553668209.0, + "step": 21397 + }, + { + "epoch": 2.349879200527125, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3549609184265137, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7330160140991211, + "num_tokens": 553693394.0, + "step": 21398 + }, + { + "epoch": 2.3499890182297385, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.6480374336242676, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7347850799560547, + "num_tokens": 553714622.0, + "step": 21399 + }, + { + "epoch": 2.3500988359323522, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3929357528686523, + "learning_rate": 1e-06, + "loss": 0.7506, + "mean_token_accuracy": 0.7529926300048828, + "num_tokens": 553738769.0, + "step": 21400 + }, + { + "epoch": 2.350208653634966, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.539673089981079, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7068685293197632, + "num_tokens": 553762917.0, + "step": 21401 + }, + { + "epoch": 2.3503184713375798, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.008751153945923, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7225409746170044, + "num_tokens": 553798724.0, + "step": 21402 + }, + { + "epoch": 2.3504282890401935, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.7460598945617676, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.736766517162323, + "num_tokens": 553819757.0, + "step": 21403 + }, + { + "epoch": 2.350538106742807, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.33716082572937, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7345110177993774, + "num_tokens": 553846996.0, + "step": 21404 + }, + { + "epoch": 2.3506479244454206, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.4781653881073, + "learning_rate": 1e-06, + "loss": 0.8857, + "mean_token_accuracy": 0.7224242687225342, + "num_tokens": 553871149.0, + "step": 21405 + }, + { + "epoch": 2.3507577421480343, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1329727172851562, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.719241738319397, + "num_tokens": 553904582.0, + "step": 21406 + }, + { + "epoch": 2.350867559850648, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.686215877532959, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7492695450782776, + "num_tokens": 553925932.0, + "step": 21407 + }, + { + "epoch": 2.3509773775532614, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.453221321105957, + "learning_rate": 1e-06, + "loss": 0.7674, + "mean_token_accuracy": 0.7490379810333252, + "num_tokens": 553950760.0, + "step": 21408 + }, + { + "epoch": 2.351087195255875, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2000410556793213, + "learning_rate": 1e-06, + "loss": 0.9166, + "mean_token_accuracy": 0.7127910256385803, + "num_tokens": 553980795.0, + "step": 21409 + }, + { + "epoch": 2.351197012958489, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.269345760345459, + "learning_rate": 1e-06, + "loss": 0.7504, + "mean_token_accuracy": 0.7553788423538208, + "num_tokens": 554007473.0, + "step": 21410 + }, + { + "epoch": 2.3513068306611027, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1689035892486572, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7151532769203186, + "num_tokens": 554036706.0, + "step": 21411 + }, + { + "epoch": 2.351416648363716, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.799130439758301, + "learning_rate": 1e-06, + "loss": 0.7281, + "mean_token_accuracy": 0.7665195465087891, + "num_tokens": 554055256.0, + "step": 21412 + }, + { + "epoch": 2.3515264660663298, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.453702688217163, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7325745820999146, + "num_tokens": 554081341.0, + "step": 21413 + }, + { + "epoch": 2.3516362837689435, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.1207263469696045, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7281039357185364, + "num_tokens": 554110544.0, + "step": 21414 + }, + { + "epoch": 2.3517461014715573, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.292558193206787, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.730238676071167, + "num_tokens": 554136803.0, + "step": 21415 + }, + { + "epoch": 2.351855919174171, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.155996084213257, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7239723205566406, + "num_tokens": 554164323.0, + "step": 21416 + }, + { + "epoch": 2.3519657368767843, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.294415235519409, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7461948990821838, + "num_tokens": 554188916.0, + "step": 21417 + }, + { + "epoch": 2.352075554579398, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5970892906188965, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7376250624656677, + "num_tokens": 554210815.0, + "step": 21418 + }, + { + "epoch": 2.352185372282012, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5863680839538574, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7379370331764221, + "num_tokens": 554233278.0, + "step": 21419 + }, + { + "epoch": 2.3522951899846256, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.522592782974243, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.726309597492218, + "num_tokens": 554256319.0, + "step": 21420 + }, + { + "epoch": 2.3524050076872394, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.7829880714416504, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7263031601905823, + "num_tokens": 554275103.0, + "step": 21421 + }, + { + "epoch": 2.3525148253898527, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.2940762042999268, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7228555083274841, + "num_tokens": 554302972.0, + "step": 21422 + }, + { + "epoch": 2.3526246430924664, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.4264230728149414, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7327991724014282, + "num_tokens": 554326597.0, + "step": 21423 + }, + { + "epoch": 2.35273446079508, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.4281537532806396, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.713529109954834, + "num_tokens": 554350438.0, + "step": 21424 + }, + { + "epoch": 2.352844278497694, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.6698720455169678, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7326918840408325, + "num_tokens": 554372166.0, + "step": 21425 + }, + { + "epoch": 2.3529540962003077, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.544724225997925, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7416468858718872, + "num_tokens": 554394105.0, + "step": 21426 + }, + { + "epoch": 2.353063913902921, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.43276047706604, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7232851982116699, + "num_tokens": 554419285.0, + "step": 21427 + }, + { + "epoch": 2.353173731605535, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.520339012145996, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.740380048751831, + "num_tokens": 554442757.0, + "step": 21428 + }, + { + "epoch": 2.3532835493081485, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.406186103820801, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7140591144561768, + "num_tokens": 554468867.0, + "step": 21429 + }, + { + "epoch": 2.3533933670107623, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.211522340774536, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7266203165054321, + "num_tokens": 554499879.0, + "step": 21430 + }, + { + "epoch": 2.3535031847133756, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.479201555252075, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7357704043388367, + "num_tokens": 554524143.0, + "step": 21431 + }, + { + "epoch": 2.3536130024159894, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.5782926082611084, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7380200624465942, + "num_tokens": 554546452.0, + "step": 21432 + }, + { + "epoch": 2.353722820118603, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.3814337253570557, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7426353693008423, + "num_tokens": 554571337.0, + "step": 21433 + }, + { + "epoch": 2.353832637821217, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.510772705078125, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.733418881893158, + "num_tokens": 554595132.0, + "step": 21434 + }, + { + "epoch": 2.3539424555238306, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.380032777786255, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7255558967590332, + "num_tokens": 554622249.0, + "step": 21435 + }, + { + "epoch": 2.354052273226444, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.058274507522583, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7175731658935547, + "num_tokens": 554654011.0, + "step": 21436 + }, + { + "epoch": 2.3541620909290577, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.318119525909424, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7336794137954712, + "num_tokens": 554681248.0, + "step": 21437 + }, + { + "epoch": 2.3542719086316715, + "ewc_loss": 1.3828277587890625e-05, + "grad_norm": 2.350621223449707, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7451484203338623, + "num_tokens": 554704839.0, + "step": 21438 + }, + { + "epoch": 2.354381726334285, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.0624568462371826, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7266032695770264, + "num_tokens": 554736566.0, + "step": 21439 + }, + { + "epoch": 2.3544915440368985, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.539163589477539, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.733035683631897, + "num_tokens": 554759453.0, + "step": 21440 + }, + { + "epoch": 2.3546013617395123, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.337721824645996, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7299038767814636, + "num_tokens": 554785923.0, + "step": 21441 + }, + { + "epoch": 2.354711179442126, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.6356894969940186, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7250581979751587, + "num_tokens": 554807141.0, + "step": 21442 + }, + { + "epoch": 2.35482099714474, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.381443500518799, + "learning_rate": 1e-06, + "loss": 0.7325, + "mean_token_accuracy": 0.7665042877197266, + "num_tokens": 554830120.0, + "step": 21443 + }, + { + "epoch": 2.3549308148473536, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.350853204727173, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7335191965103149, + "num_tokens": 554854882.0, + "step": 21444 + }, + { + "epoch": 2.355040632549967, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5357110500335693, + "learning_rate": 1e-06, + "loss": 0.9647, + "mean_token_accuracy": 0.70272296667099, + "num_tokens": 554879697.0, + "step": 21445 + }, + { + "epoch": 2.3551504502525806, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4970972537994385, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7551528811454773, + "num_tokens": 554901742.0, + "step": 21446 + }, + { + "epoch": 2.3552602679551944, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3661551475524902, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7280683517456055, + "num_tokens": 554927477.0, + "step": 21447 + }, + { + "epoch": 2.355370085657808, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.8581883907318115, + "learning_rate": 1e-06, + "loss": 0.7389, + "mean_token_accuracy": 0.7673634886741638, + "num_tokens": 554945121.0, + "step": 21448 + }, + { + "epoch": 2.355479903360422, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4896533489227295, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7236548662185669, + "num_tokens": 554971482.0, + "step": 21449 + }, + { + "epoch": 2.355589721063035, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.741272449493408, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7274105548858643, + "num_tokens": 554991702.0, + "step": 21450 + }, + { + "epoch": 2.355699538765649, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5195417404174805, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.7498400807380676, + "num_tokens": 555012775.0, + "step": 21451 + }, + { + "epoch": 2.3558093564682627, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2821731567382812, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7268217206001282, + "num_tokens": 555042151.0, + "step": 21452 + }, + { + "epoch": 2.3559191741708765, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.220365047454834, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7302559614181519, + "num_tokens": 555069862.0, + "step": 21453 + }, + { + "epoch": 2.3560289918734902, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.579328775405884, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7484347224235535, + "num_tokens": 555092922.0, + "step": 21454 + }, + { + "epoch": 2.3561388095761036, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2803187370300293, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7405552864074707, + "num_tokens": 555119169.0, + "step": 21455 + }, + { + "epoch": 2.3562486272787173, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.321620464324951, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7294944524765015, + "num_tokens": 555145956.0, + "step": 21456 + }, + { + "epoch": 2.356358444981331, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2316811084747314, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7190009951591492, + "num_tokens": 555170733.0, + "step": 21457 + }, + { + "epoch": 2.356468262683945, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.620546340942383, + "learning_rate": 1e-06, + "loss": 0.751, + "mean_token_accuracy": 0.757051408290863, + "num_tokens": 555190957.0, + "step": 21458 + }, + { + "epoch": 2.356578080386558, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.775649070739746, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7342251539230347, + "num_tokens": 555211908.0, + "step": 21459 + }, + { + "epoch": 2.356687898089172, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.488046884536743, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7279781699180603, + "num_tokens": 555236327.0, + "step": 21460 + }, + { + "epoch": 2.3567977157917857, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5572054386138916, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7382696866989136, + "num_tokens": 555260806.0, + "step": 21461 + }, + { + "epoch": 2.3569075334943994, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2463204860687256, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7264454364776611, + "num_tokens": 555290160.0, + "step": 21462 + }, + { + "epoch": 2.3570173511970127, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.7147340774536133, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7379544377326965, + "num_tokens": 555310957.0, + "step": 21463 + }, + { + "epoch": 2.3571271688996265, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.484957218170166, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7258991599082947, + "num_tokens": 555335161.0, + "step": 21464 + }, + { + "epoch": 2.3572369866022402, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.274526357650757, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.723538875579834, + "num_tokens": 555362095.0, + "step": 21465 + }, + { + "epoch": 2.357346804304854, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.503296375274658, + "learning_rate": 1e-06, + "loss": 0.7906, + "mean_token_accuracy": 0.7452858686447144, + "num_tokens": 555384986.0, + "step": 21466 + }, + { + "epoch": 2.3574566220074678, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4656550884246826, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7399487495422363, + "num_tokens": 555410007.0, + "step": 21467 + }, + { + "epoch": 2.357566439710081, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.1777641773223877, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7233099937438965, + "num_tokens": 555441338.0, + "step": 21468 + }, + { + "epoch": 2.357676257412695, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.260836601257324, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7153214812278748, + "num_tokens": 555470582.0, + "step": 21469 + }, + { + "epoch": 2.3577860751153086, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.06250262260437, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7318108081817627, + "num_tokens": 555503886.0, + "step": 21470 + }, + { + "epoch": 2.3578958928179223, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.458350419998169, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7350156307220459, + "num_tokens": 555528617.0, + "step": 21471 + }, + { + "epoch": 2.358005710520536, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3249216079711914, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7410618662834167, + "num_tokens": 555555610.0, + "step": 21472 + }, + { + "epoch": 2.3581155282231494, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.344362735748291, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7046697735786438, + "num_tokens": 555581989.0, + "step": 21473 + }, + { + "epoch": 2.358225345925763, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.824125051498413, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7341412901878357, + "num_tokens": 555604084.0, + "step": 21474 + }, + { + "epoch": 2.358335163628377, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2541120052337646, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7446944117546082, + "num_tokens": 555631310.0, + "step": 21475 + }, + { + "epoch": 2.3584449813309907, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2023603916168213, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7199895977973938, + "num_tokens": 555661471.0, + "step": 21476 + }, + { + "epoch": 2.3585547990336044, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.272832155227661, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7522451877593994, + "num_tokens": 555686611.0, + "step": 21477 + }, + { + "epoch": 2.3586646167362177, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.6217844486236572, + "learning_rate": 1e-06, + "loss": 0.784, + "mean_token_accuracy": 0.747286856174469, + "num_tokens": 555706852.0, + "step": 21478 + }, + { + "epoch": 2.3587744344388315, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.328174114227295, + "learning_rate": 1e-06, + "loss": 0.7653, + "mean_token_accuracy": 0.7561143040657043, + "num_tokens": 555731573.0, + "step": 21479 + }, + { + "epoch": 2.3588842521414453, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3992199897766113, + "learning_rate": 1e-06, + "loss": 0.7535, + "mean_token_accuracy": 0.7537358999252319, + "num_tokens": 555754399.0, + "step": 21480 + }, + { + "epoch": 2.358994069844059, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.442986011505127, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7179733514785767, + "num_tokens": 555779355.0, + "step": 21481 + }, + { + "epoch": 2.3591038875466728, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.511030912399292, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7436891794204712, + "num_tokens": 555801736.0, + "step": 21482 + }, + { + "epoch": 2.359213705249286, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4231953620910645, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7264186143875122, + "num_tokens": 555828488.0, + "step": 21483 + }, + { + "epoch": 2.3593235229519, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4375572204589844, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7560096383094788, + "num_tokens": 555853675.0, + "step": 21484 + }, + { + "epoch": 2.3594333406545136, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4621081352233887, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7418076395988464, + "num_tokens": 555877499.0, + "step": 21485 + }, + { + "epoch": 2.3595431583571274, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2826597690582275, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7290014028549194, + "num_tokens": 555904816.0, + "step": 21486 + }, + { + "epoch": 2.3596529760597407, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4107627868652344, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7421222925186157, + "num_tokens": 555930791.0, + "step": 21487 + }, + { + "epoch": 2.3597627937623544, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5165793895721436, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7237581014633179, + "num_tokens": 555954763.0, + "step": 21488 + }, + { + "epoch": 2.359872611464968, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4646599292755127, + "learning_rate": 1e-06, + "loss": 0.7616, + "mean_token_accuracy": 0.7586071491241455, + "num_tokens": 555978907.0, + "step": 21489 + }, + { + "epoch": 2.359982429167582, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.328516721725464, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7283716797828674, + "num_tokens": 556004729.0, + "step": 21490 + }, + { + "epoch": 2.3600922468701953, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.319643974304199, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7242638468742371, + "num_tokens": 556031476.0, + "step": 21491 + }, + { + "epoch": 2.360202064572809, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5073375701904297, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7347950339317322, + "num_tokens": 556054552.0, + "step": 21492 + }, + { + "epoch": 2.3603118822754228, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.603877305984497, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7154169678688049, + "num_tokens": 556079224.0, + "step": 21493 + }, + { + "epoch": 2.3604216999780365, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.710376501083374, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.7451938390731812, + "num_tokens": 556099862.0, + "step": 21494 + }, + { + "epoch": 2.3605315176806503, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.1144933700561523, + "learning_rate": 1e-06, + "loss": 0.9152, + "mean_token_accuracy": 0.7122975587844849, + "num_tokens": 556131139.0, + "step": 21495 + }, + { + "epoch": 2.3606413353832636, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4417712688446045, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7437852025032043, + "num_tokens": 556154231.0, + "step": 21496 + }, + { + "epoch": 2.3607511530858774, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.171696424484253, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7249029278755188, + "num_tokens": 556185699.0, + "step": 21497 + }, + { + "epoch": 2.360860970788491, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.170132875442505, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7412207126617432, + "num_tokens": 556213295.0, + "step": 21498 + }, + { + "epoch": 2.360970788491105, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4707674980163574, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7190067768096924, + "num_tokens": 556236909.0, + "step": 21499 + }, + { + "epoch": 2.3610806061937186, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.640939235687256, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7386704683303833, + "num_tokens": 556260911.0, + "step": 21500 + }, + { + "epoch": 2.361190423896332, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.542121171951294, + "learning_rate": 1e-06, + "loss": 0.7826, + "mean_token_accuracy": 0.7493772506713867, + "num_tokens": 556284155.0, + "step": 21501 + }, + { + "epoch": 2.3613002415989457, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.372462749481201, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7250303626060486, + "num_tokens": 556311007.0, + "step": 21502 + }, + { + "epoch": 2.3614100593015594, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.8959155082702637, + "learning_rate": 1e-06, + "loss": 0.6974, + "mean_token_accuracy": 0.7735658884048462, + "num_tokens": 556327874.0, + "step": 21503 + }, + { + "epoch": 2.361519877004173, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2907369136810303, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7360537052154541, + "num_tokens": 556354663.0, + "step": 21504 + }, + { + "epoch": 2.361629694706787, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.424900770187378, + "learning_rate": 1e-06, + "loss": 0.7553, + "mean_token_accuracy": 0.7616392374038696, + "num_tokens": 556380131.0, + "step": 21505 + }, + { + "epoch": 2.3617395124094003, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.1005165576934814, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7344049215316772, + "num_tokens": 556412535.0, + "step": 21506 + }, + { + "epoch": 2.361849330112014, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2996904850006104, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7267399430274963, + "num_tokens": 556438374.0, + "step": 21507 + }, + { + "epoch": 2.361959147814628, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5408644676208496, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7301913499832153, + "num_tokens": 556461616.0, + "step": 21508 + }, + { + "epoch": 2.3620689655172415, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.467698097229004, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.7414509057998657, + "num_tokens": 556483612.0, + "step": 21509 + }, + { + "epoch": 2.362178783219855, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4189655780792236, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.744886577129364, + "num_tokens": 556508354.0, + "step": 21510 + }, + { + "epoch": 2.3622886009224686, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.422985315322876, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7360718846321106, + "num_tokens": 556533953.0, + "step": 21511 + }, + { + "epoch": 2.3623984186250824, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3690011501312256, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7305364608764648, + "num_tokens": 556559964.0, + "step": 21512 + }, + { + "epoch": 2.362508236327696, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2057769298553467, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.719667911529541, + "num_tokens": 556591345.0, + "step": 21513 + }, + { + "epoch": 2.3626180540303094, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.836350202560425, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7364696264266968, + "num_tokens": 556610964.0, + "step": 21514 + }, + { + "epoch": 2.362727871732923, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3761179447174072, + "learning_rate": 1e-06, + "loss": 0.7792, + "mean_token_accuracy": 0.7516192197799683, + "num_tokens": 556635293.0, + "step": 21515 + }, + { + "epoch": 2.362837689435537, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2433924674987793, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7122591733932495, + "num_tokens": 556663735.0, + "step": 21516 + }, + { + "epoch": 2.3629475071381507, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.686490535736084, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7331104278564453, + "num_tokens": 556686187.0, + "step": 21517 + }, + { + "epoch": 2.3630573248407645, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2970361709594727, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7259725332260132, + "num_tokens": 556714669.0, + "step": 21518 + }, + { + "epoch": 2.363167142543378, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2644574642181396, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7279858589172363, + "num_tokens": 556741696.0, + "step": 21519 + }, + { + "epoch": 2.3632769602459915, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.4798471927642822, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7401305437088013, + "num_tokens": 556763765.0, + "step": 21520 + }, + { + "epoch": 2.3633867779486053, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2849233150482178, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7253721952438354, + "num_tokens": 556790514.0, + "step": 21521 + }, + { + "epoch": 2.363496595651219, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.526852607727051, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7329564690589905, + "num_tokens": 556813619.0, + "step": 21522 + }, + { + "epoch": 2.363606413353833, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.3051140308380127, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7442308664321899, + "num_tokens": 556841250.0, + "step": 21523 + }, + { + "epoch": 2.363716231056446, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5381081104278564, + "learning_rate": 1e-06, + "loss": 0.7937, + "mean_token_accuracy": 0.7507038116455078, + "num_tokens": 556863528.0, + "step": 21524 + }, + { + "epoch": 2.36382604875906, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.203115463256836, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7265756726264954, + "num_tokens": 556892962.0, + "step": 21525 + }, + { + "epoch": 2.3639358664616736, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.2778725624084473, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7241599559783936, + "num_tokens": 556920673.0, + "step": 21526 + }, + { + "epoch": 2.3640456841642874, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5578925609588623, + "learning_rate": 1e-06, + "loss": 0.7408, + "mean_token_accuracy": 0.7582069039344788, + "num_tokens": 556943449.0, + "step": 21527 + }, + { + "epoch": 2.364155501866901, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.190553665161133, + "learning_rate": 1e-06, + "loss": 0.8931, + "mean_token_accuracy": 0.7183394432067871, + "num_tokens": 556972204.0, + "step": 21528 + }, + { + "epoch": 2.3642653195695145, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.261992931365967, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7072598338127136, + "num_tokens": 557002786.0, + "step": 21529 + }, + { + "epoch": 2.3643751372721282, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.561534881591797, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7086194157600403, + "num_tokens": 557028432.0, + "step": 21530 + }, + { + "epoch": 2.364484954974742, + "ewc_loss": 1.3887882232666016e-05, + "grad_norm": 2.5187885761260986, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7120564579963684, + "num_tokens": 557053635.0, + "step": 21531 + }, + { + "epoch": 2.3645947726773557, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3742969036102295, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7364336848258972, + "num_tokens": 557080238.0, + "step": 21532 + }, + { + "epoch": 2.3647045903799695, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4699108600616455, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7412925958633423, + "num_tokens": 557102780.0, + "step": 21533 + }, + { + "epoch": 2.364814408082583, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.288524627685547, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7217260599136353, + "num_tokens": 557131178.0, + "step": 21534 + }, + { + "epoch": 2.3649242257851966, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6153225898742676, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7422837018966675, + "num_tokens": 557153394.0, + "step": 21535 + }, + { + "epoch": 2.3650340434878103, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2813098430633545, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7307652235031128, + "num_tokens": 557180490.0, + "step": 21536 + }, + { + "epoch": 2.365143861190424, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.094265937805176, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7304770946502686, + "num_tokens": 557213714.0, + "step": 21537 + }, + { + "epoch": 2.3652536788930374, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4155478477478027, + "learning_rate": 1e-06, + "loss": 0.8768, + "mean_token_accuracy": 0.7399157285690308, + "num_tokens": 557238581.0, + "step": 21538 + }, + { + "epoch": 2.365363496595651, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3664443492889404, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.7197925448417664, + "num_tokens": 557264846.0, + "step": 21539 + }, + { + "epoch": 2.365473314298265, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3770012855529785, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7254399061203003, + "num_tokens": 557291147.0, + "step": 21540 + }, + { + "epoch": 2.3655831320008787, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5639092922210693, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7335674166679382, + "num_tokens": 557312834.0, + "step": 21541 + }, + { + "epoch": 2.365692949703492, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.418759346008301, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7451894283294678, + "num_tokens": 557339964.0, + "step": 21542 + }, + { + "epoch": 2.3658027674061057, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3539013862609863, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7435941696166992, + "num_tokens": 557366144.0, + "step": 21543 + }, + { + "epoch": 2.3659125851087195, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.351091146469116, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7239567041397095, + "num_tokens": 557391270.0, + "step": 21544 + }, + { + "epoch": 2.3660224028113332, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3106794357299805, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.730035662651062, + "num_tokens": 557417639.0, + "step": 21545 + }, + { + "epoch": 2.366132220513947, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4773287773132324, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7353988885879517, + "num_tokens": 557444990.0, + "step": 21546 + }, + { + "epoch": 2.3662420382165603, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6096036434173584, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7050966024398804, + "num_tokens": 557467022.0, + "step": 21547 + }, + { + "epoch": 2.366351855919174, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2691681385040283, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7232916355133057, + "num_tokens": 557495147.0, + "step": 21548 + }, + { + "epoch": 2.366461673621788, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.502915382385254, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7403701543807983, + "num_tokens": 557519017.0, + "step": 21549 + }, + { + "epoch": 2.3665714913244016, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3931398391723633, + "learning_rate": 1e-06, + "loss": 0.7373, + "mean_token_accuracy": 0.7671911716461182, + "num_tokens": 557543080.0, + "step": 21550 + }, + { + "epoch": 2.3666813090270153, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3511850833892822, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7311516404151917, + "num_tokens": 557569638.0, + "step": 21551 + }, + { + "epoch": 2.3667911267296287, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1987156867980957, + "learning_rate": 1e-06, + "loss": 0.797, + "mean_token_accuracy": 0.7442212104797363, + "num_tokens": 557598536.0, + "step": 21552 + }, + { + "epoch": 2.3669009444322424, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 1.9995708465576172, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.719928503036499, + "num_tokens": 557634436.0, + "step": 21553 + }, + { + "epoch": 2.367010762134856, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.8023128509521484, + "learning_rate": 1e-06, + "loss": 0.7684, + "mean_token_accuracy": 0.7537524700164795, + "num_tokens": 557654703.0, + "step": 21554 + }, + { + "epoch": 2.36712057983747, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.324115514755249, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7256923913955688, + "num_tokens": 557683104.0, + "step": 21555 + }, + { + "epoch": 2.3672303975400837, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2300431728363037, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7152372598648071, + "num_tokens": 557715197.0, + "step": 21556 + }, + { + "epoch": 2.367340215242697, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2831871509552, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7307971715927124, + "num_tokens": 557742148.0, + "step": 21557 + }, + { + "epoch": 2.3674500329453108, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.182245969772339, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7249266505241394, + "num_tokens": 557769927.0, + "step": 21558 + }, + { + "epoch": 2.3675598506479245, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.528851270675659, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7313661575317383, + "num_tokens": 557793071.0, + "step": 21559 + }, + { + "epoch": 2.3676696683505383, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1534574031829834, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7478740811347961, + "num_tokens": 557822869.0, + "step": 21560 + }, + { + "epoch": 2.3677794860531516, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.219902276992798, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7326861619949341, + "num_tokens": 557852209.0, + "step": 21561 + }, + { + "epoch": 2.3678893037557653, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.426809787750244, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.724044680595398, + "num_tokens": 557876654.0, + "step": 21562 + }, + { + "epoch": 2.367999121458379, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2437663078308105, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.739874005317688, + "num_tokens": 557906940.0, + "step": 21563 + }, + { + "epoch": 2.368108939160993, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1578891277313232, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7245127558708191, + "num_tokens": 557938472.0, + "step": 21564 + }, + { + "epoch": 2.368218756863606, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.620748996734619, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7298843860626221, + "num_tokens": 557960421.0, + "step": 21565 + }, + { + "epoch": 2.36832857456622, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.702392101287842, + "learning_rate": 1e-06, + "loss": 0.8222, + "mean_token_accuracy": 0.7433624267578125, + "num_tokens": 557981998.0, + "step": 21566 + }, + { + "epoch": 2.3684383922688337, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.514178514480591, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7395710945129395, + "num_tokens": 558004848.0, + "step": 21567 + }, + { + "epoch": 2.3685482099714474, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5101466178894043, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.742629885673523, + "num_tokens": 558028237.0, + "step": 21568 + }, + { + "epoch": 2.368658027674061, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2121570110321045, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7245050668716431, + "num_tokens": 558058594.0, + "step": 21569 + }, + { + "epoch": 2.3687678453766745, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3289642333984375, + "learning_rate": 1e-06, + "loss": 0.7758, + "mean_token_accuracy": 0.7542527914047241, + "num_tokens": 558084242.0, + "step": 21570 + }, + { + "epoch": 2.3688776630792883, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.101747751235962, + "learning_rate": 1e-06, + "loss": 0.9763, + "mean_token_accuracy": 0.7005659341812134, + "num_tokens": 558115959.0, + "step": 21571 + }, + { + "epoch": 2.368987480781902, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3679962158203125, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7219538688659668, + "num_tokens": 558141050.0, + "step": 21572 + }, + { + "epoch": 2.369097298484516, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.124804735183716, + "learning_rate": 1e-06, + "loss": 0.9883, + "mean_token_accuracy": 0.6930845975875854, + "num_tokens": 558173423.0, + "step": 21573 + }, + { + "epoch": 2.3692071161871295, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3971526622772217, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7364280223846436, + "num_tokens": 558197401.0, + "step": 21574 + }, + { + "epoch": 2.369316933889743, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6374130249023438, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.738494336605072, + "num_tokens": 558220872.0, + "step": 21575 + }, + { + "epoch": 2.3694267515923566, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.452415943145752, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7305558323860168, + "num_tokens": 558245340.0, + "step": 21576 + }, + { + "epoch": 2.3695365692949704, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.358347177505493, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7515459656715393, + "num_tokens": 558269052.0, + "step": 21577 + }, + { + "epoch": 2.369646386997584, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.257575511932373, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7386569380760193, + "num_tokens": 558296135.0, + "step": 21578 + }, + { + "epoch": 2.369756204700198, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.547799825668335, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7245837450027466, + "num_tokens": 558319289.0, + "step": 21579 + }, + { + "epoch": 2.369866022402811, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.54288387298584, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7431133985519409, + "num_tokens": 558341681.0, + "step": 21580 + }, + { + "epoch": 2.369975840105425, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2566845417022705, + "learning_rate": 1e-06, + "loss": 0.944, + "mean_token_accuracy": 0.7086910605430603, + "num_tokens": 558369106.0, + "step": 21581 + }, + { + "epoch": 2.3700856578080387, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.213907480239868, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7340748310089111, + "num_tokens": 558398945.0, + "step": 21582 + }, + { + "epoch": 2.3701954755106525, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.733858346939087, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.7418414354324341, + "num_tokens": 558422580.0, + "step": 21583 + }, + { + "epoch": 2.370305293213266, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2341573238372803, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7381798624992371, + "num_tokens": 558450847.0, + "step": 21584 + }, + { + "epoch": 2.3704151109158795, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.351311445236206, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7335889935493469, + "num_tokens": 558475667.0, + "step": 21585 + }, + { + "epoch": 2.3705249286184933, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2286529541015625, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7426064014434814, + "num_tokens": 558504062.0, + "step": 21586 + }, + { + "epoch": 2.370634746321107, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.538545846939087, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7400466203689575, + "num_tokens": 558527171.0, + "step": 21587 + }, + { + "epoch": 2.370744564023721, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.798755168914795, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.743821918964386, + "num_tokens": 558546295.0, + "step": 21588 + }, + { + "epoch": 2.370854381726334, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.522963762283325, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7225419282913208, + "num_tokens": 558568236.0, + "step": 21589 + }, + { + "epoch": 2.370964199428948, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1513519287109375, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7040328979492188, + "num_tokens": 558598369.0, + "step": 21590 + }, + { + "epoch": 2.3710740171315616, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6255340576171875, + "learning_rate": 1e-06, + "loss": 0.7552, + "mean_token_accuracy": 0.7610608339309692, + "num_tokens": 558620381.0, + "step": 21591 + }, + { + "epoch": 2.3711838348341754, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.619518280029297, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7413673400878906, + "num_tokens": 558642173.0, + "step": 21592 + }, + { + "epoch": 2.3712936525367887, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.464580774307251, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7257530093193054, + "num_tokens": 558669174.0, + "step": 21593 + }, + { + "epoch": 2.3714034702394025, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5542919635772705, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7374545335769653, + "num_tokens": 558691184.0, + "step": 21594 + }, + { + "epoch": 2.371513287942016, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.9237470626831055, + "learning_rate": 1e-06, + "loss": 0.7358, + "mean_token_accuracy": 0.757678747177124, + "num_tokens": 558707598.0, + "step": 21595 + }, + { + "epoch": 2.37162310564463, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0394411087036133, + "learning_rate": 1e-06, + "loss": 0.9926, + "mean_token_accuracy": 0.6934967637062073, + "num_tokens": 558744396.0, + "step": 21596 + }, + { + "epoch": 2.3717329233472437, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.459951877593994, + "learning_rate": 1e-06, + "loss": 0.7922, + "mean_token_accuracy": 0.7442635297775269, + "num_tokens": 558769673.0, + "step": 21597 + }, + { + "epoch": 2.371842741049857, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4405643939971924, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7328375577926636, + "num_tokens": 558795749.0, + "step": 21598 + }, + { + "epoch": 2.371952558752471, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.125091552734375, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.7459535002708435, + "num_tokens": 558826193.0, + "step": 21599 + }, + { + "epoch": 2.3720623764550846, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.476451873779297, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7487140893936157, + "num_tokens": 558850118.0, + "step": 21600 + }, + { + "epoch": 2.3721721941576983, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7012221813201904, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7314306497573853, + "num_tokens": 558870899.0, + "step": 21601 + }, + { + "epoch": 2.372282011860312, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2787342071533203, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7234413623809814, + "num_tokens": 558897874.0, + "step": 21602 + }, + { + "epoch": 2.3723918295629254, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2585504055023193, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7224401235580444, + "num_tokens": 558925023.0, + "step": 21603 + }, + { + "epoch": 2.372501647265539, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4109718799591064, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7368456125259399, + "num_tokens": 558950567.0, + "step": 21604 + }, + { + "epoch": 2.372611464968153, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3764638900756836, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7427924275398254, + "num_tokens": 558974871.0, + "step": 21605 + }, + { + "epoch": 2.3727212826707667, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.186277151107788, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7198607921600342, + "num_tokens": 559005912.0, + "step": 21606 + }, + { + "epoch": 2.3728311003733804, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.329237222671509, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7286968231201172, + "num_tokens": 559032318.0, + "step": 21607 + }, + { + "epoch": 2.3729409180759937, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 3.128140926361084, + "learning_rate": 1e-06, + "loss": 0.7852, + "mean_token_accuracy": 0.7521277666091919, + "num_tokens": 559050923.0, + "step": 21608 + }, + { + "epoch": 2.3730507357786075, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0367114543914795, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7509876489639282, + "num_tokens": 559081779.0, + "step": 21609 + }, + { + "epoch": 2.3731605534812212, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.687405824661255, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7202534675598145, + "num_tokens": 559101853.0, + "step": 21610 + }, + { + "epoch": 2.373270371183835, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6676878929138184, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7148561477661133, + "num_tokens": 559125924.0, + "step": 21611 + }, + { + "epoch": 2.3733801888864483, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1961820125579834, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7432839870452881, + "num_tokens": 559155653.0, + "step": 21612 + }, + { + "epoch": 2.373490006589062, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.689802885055542, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7343317270278931, + "num_tokens": 559179068.0, + "step": 21613 + }, + { + "epoch": 2.373599824291676, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.284177780151367, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7223532199859619, + "num_tokens": 559206875.0, + "step": 21614 + }, + { + "epoch": 2.3737096419942896, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1946332454681396, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.729075014591217, + "num_tokens": 559235853.0, + "step": 21615 + }, + { + "epoch": 2.3738194596969033, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.717209577560425, + "learning_rate": 1e-06, + "loss": 0.7275, + "mean_token_accuracy": 0.7605428099632263, + "num_tokens": 559256191.0, + "step": 21616 + }, + { + "epoch": 2.3739292773995166, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.625380516052246, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7165284156799316, + "num_tokens": 559278183.0, + "step": 21617 + }, + { + "epoch": 2.3740390951021304, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6221585273742676, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7406612634658813, + "num_tokens": 559299964.0, + "step": 21618 + }, + { + "epoch": 2.374148912804744, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.446769952774048, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7206915616989136, + "num_tokens": 559324445.0, + "step": 21619 + }, + { + "epoch": 2.374258730507358, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4013640880584717, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7280986309051514, + "num_tokens": 559349323.0, + "step": 21620 + }, + { + "epoch": 2.3743685482099712, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2331483364105225, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7256594300270081, + "num_tokens": 559378716.0, + "step": 21621 + }, + { + "epoch": 2.374478365912585, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0272719860076904, + "learning_rate": 1e-06, + "loss": 0.9358, + "mean_token_accuracy": 0.708764374256134, + "num_tokens": 559412640.0, + "step": 21622 + }, + { + "epoch": 2.3745881836151987, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.758300542831421, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7470772862434387, + "num_tokens": 559431472.0, + "step": 21623 + }, + { + "epoch": 2.3746980013178125, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7137458324432373, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7454639673233032, + "num_tokens": 559453465.0, + "step": 21624 + }, + { + "epoch": 2.3748078190204263, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4814109802246094, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7171091437339783, + "num_tokens": 559479660.0, + "step": 21625 + }, + { + "epoch": 2.3749176367230396, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1321589946746826, + "learning_rate": 1e-06, + "loss": 0.8019, + "mean_token_accuracy": 0.7442396879196167, + "num_tokens": 559509927.0, + "step": 21626 + }, + { + "epoch": 2.3750274544256533, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.383429765701294, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7273016571998596, + "num_tokens": 559535535.0, + "step": 21627 + }, + { + "epoch": 2.375137272128267, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.274998188018799, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7392414808273315, + "num_tokens": 559563278.0, + "step": 21628 + }, + { + "epoch": 2.375247089830881, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3234453201293945, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7300752997398376, + "num_tokens": 559588563.0, + "step": 21629 + }, + { + "epoch": 2.3753569075334946, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2465860843658447, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7124258279800415, + "num_tokens": 559619051.0, + "step": 21630 + }, + { + "epoch": 2.375466725236108, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.632329225540161, + "learning_rate": 1e-06, + "loss": 0.7424, + "mean_token_accuracy": 0.7567598819732666, + "num_tokens": 559639925.0, + "step": 21631 + }, + { + "epoch": 2.3755765429387217, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.178138017654419, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7528244853019714, + "num_tokens": 559671210.0, + "step": 21632 + }, + { + "epoch": 2.3756863606413354, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.609304904937744, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7314165830612183, + "num_tokens": 559694236.0, + "step": 21633 + }, + { + "epoch": 2.375796178343949, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7800254821777344, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.722938060760498, + "num_tokens": 559714548.0, + "step": 21634 + }, + { + "epoch": 2.375905996046563, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2013611793518066, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7204400300979614, + "num_tokens": 559744762.0, + "step": 21635 + }, + { + "epoch": 2.3760158137491763, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3805694580078125, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7320025563240051, + "num_tokens": 559770635.0, + "step": 21636 + }, + { + "epoch": 2.37612563145179, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7743492126464844, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7332923412322998, + "num_tokens": 559791610.0, + "step": 21637 + }, + { + "epoch": 2.3762354491544038, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5296576023101807, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7312606573104858, + "num_tokens": 559814186.0, + "step": 21638 + }, + { + "epoch": 2.3763452668570175, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.608577013015747, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7228963971138, + "num_tokens": 559835190.0, + "step": 21639 + }, + { + "epoch": 2.376455084559631, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.10249662399292, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7201460599899292, + "num_tokens": 559866682.0, + "step": 21640 + }, + { + "epoch": 2.3765649022622446, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2107932567596436, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7238842844963074, + "num_tokens": 559895962.0, + "step": 21641 + }, + { + "epoch": 2.3766747199648584, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.562424659729004, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7551476955413818, + "num_tokens": 559918730.0, + "step": 21642 + }, + { + "epoch": 2.376784537667472, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4656708240509033, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7284240126609802, + "num_tokens": 559942407.0, + "step": 21643 + }, + { + "epoch": 2.3768943553700854, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4274611473083496, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7128170728683472, + "num_tokens": 559969006.0, + "step": 21644 + }, + { + "epoch": 2.377004173072699, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.506161689758301, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7313532829284668, + "num_tokens": 559992947.0, + "step": 21645 + }, + { + "epoch": 2.377113990775313, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.318000316619873, + "learning_rate": 1e-06, + "loss": 0.9412, + "mean_token_accuracy": 0.7069576978683472, + "num_tokens": 560020395.0, + "step": 21646 + }, + { + "epoch": 2.3772238084779267, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2302255630493164, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7367187738418579, + "num_tokens": 560048467.0, + "step": 21647 + }, + { + "epoch": 2.3773336261805404, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6198384761810303, + "learning_rate": 1e-06, + "loss": 0.8453, + "mean_token_accuracy": 0.7347291707992554, + "num_tokens": 560070449.0, + "step": 21648 + }, + { + "epoch": 2.3774434438831538, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0953328609466553, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.706440269947052, + "num_tokens": 560102206.0, + "step": 21649 + }, + { + "epoch": 2.3775532615857675, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.458879232406616, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7291003465652466, + "num_tokens": 560124865.0, + "step": 21650 + }, + { + "epoch": 2.3776630792883813, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.434823751449585, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.742906928062439, + "num_tokens": 560146976.0, + "step": 21651 + }, + { + "epoch": 2.377772896990995, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6379551887512207, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7463744878768921, + "num_tokens": 560166722.0, + "step": 21652 + }, + { + "epoch": 2.377882714693609, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5611488819122314, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7393932342529297, + "num_tokens": 560189172.0, + "step": 21653 + }, + { + "epoch": 2.377992532396222, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4107298851013184, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7213683724403381, + "num_tokens": 560214066.0, + "step": 21654 + }, + { + "epoch": 2.378102350098836, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0908567905426025, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7372019290924072, + "num_tokens": 560246232.0, + "step": 21655 + }, + { + "epoch": 2.3782121678014496, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.35325026512146, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7343475222587585, + "num_tokens": 560272804.0, + "step": 21656 + }, + { + "epoch": 2.3783219855040634, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4997339248657227, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7353747487068176, + "num_tokens": 560296297.0, + "step": 21657 + }, + { + "epoch": 2.378431803206677, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3477962017059326, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7182304859161377, + "num_tokens": 560319923.0, + "step": 21658 + }, + { + "epoch": 2.3785416209092904, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.442291021347046, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7254025936126709, + "num_tokens": 560345057.0, + "step": 21659 + }, + { + "epoch": 2.378651438611904, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6861519813537598, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7334166765213013, + "num_tokens": 560365551.0, + "step": 21660 + }, + { + "epoch": 2.378761256314518, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3490023612976074, + "learning_rate": 1e-06, + "loss": 0.7631, + "mean_token_accuracy": 0.7561652660369873, + "num_tokens": 560390724.0, + "step": 21661 + }, + { + "epoch": 2.3788710740171317, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.62713360786438, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7212297916412354, + "num_tokens": 560412951.0, + "step": 21662 + }, + { + "epoch": 2.3789808917197455, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4518251419067383, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.729877233505249, + "num_tokens": 560437061.0, + "step": 21663 + }, + { + "epoch": 2.379090709422359, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.454685926437378, + "learning_rate": 1e-06, + "loss": 0.7592, + "mean_token_accuracy": 0.7577450275421143, + "num_tokens": 560459323.0, + "step": 21664 + }, + { + "epoch": 2.3792005271249725, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7181625366210938, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7214652895927429, + "num_tokens": 560480023.0, + "step": 21665 + }, + { + "epoch": 2.3793103448275863, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.485581398010254, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7191922068595886, + "num_tokens": 560507057.0, + "step": 21666 + }, + { + "epoch": 2.3794201625302, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.390111207962036, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7210534811019897, + "num_tokens": 560531806.0, + "step": 21667 + }, + { + "epoch": 2.3795299802328134, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.0103538036346436, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7351254820823669, + "num_tokens": 560562967.0, + "step": 21668 + }, + { + "epoch": 2.379639797935427, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.468817710876465, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7457000017166138, + "num_tokens": 560587734.0, + "step": 21669 + }, + { + "epoch": 2.379749615638041, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.6106040477752686, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7243878245353699, + "num_tokens": 560612771.0, + "step": 21670 + }, + { + "epoch": 2.3798594333406546, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.805830717086792, + "learning_rate": 1e-06, + "loss": 0.681, + "mean_token_accuracy": 0.7861771583557129, + "num_tokens": 560631954.0, + "step": 21671 + }, + { + "epoch": 2.379969251043268, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.065214157104492, + "learning_rate": 1e-06, + "loss": 0.9637, + "mean_token_accuracy": 0.7003782987594604, + "num_tokens": 560664443.0, + "step": 21672 + }, + { + "epoch": 2.3800790687458817, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.430062770843506, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7312895059585571, + "num_tokens": 560689391.0, + "step": 21673 + }, + { + "epoch": 2.3801888864484955, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.210561990737915, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7229489088058472, + "num_tokens": 560717244.0, + "step": 21674 + }, + { + "epoch": 2.3802987041511092, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5009822845458984, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7463527917861938, + "num_tokens": 560738910.0, + "step": 21675 + }, + { + "epoch": 2.380408521853723, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1113009452819824, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7293323278427124, + "num_tokens": 560769723.0, + "step": 21676 + }, + { + "epoch": 2.3805183395563363, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 1.8906632661819458, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7200764417648315, + "num_tokens": 560806610.0, + "step": 21677 + }, + { + "epoch": 2.38062815725895, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5369980335235596, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7277762293815613, + "num_tokens": 560829502.0, + "step": 21678 + }, + { + "epoch": 2.380737974961564, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.281094789505005, + "learning_rate": 1e-06, + "loss": 0.7527, + "mean_token_accuracy": 0.7549078464508057, + "num_tokens": 560855250.0, + "step": 21679 + }, + { + "epoch": 2.3808477926641776, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.655487060546875, + "learning_rate": 1e-06, + "loss": 0.7894, + "mean_token_accuracy": 0.7494295835494995, + "num_tokens": 560876357.0, + "step": 21680 + }, + { + "epoch": 2.3809576103667913, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4527995586395264, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7153629660606384, + "num_tokens": 560902717.0, + "step": 21681 + }, + { + "epoch": 2.3810674280694046, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2132632732391357, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.7405291795730591, + "num_tokens": 560930540.0, + "step": 21682 + }, + { + "epoch": 2.3811772457720184, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.428262948989868, + "learning_rate": 1e-06, + "loss": 0.797, + "mean_token_accuracy": 0.7432156801223755, + "num_tokens": 560954786.0, + "step": 21683 + }, + { + "epoch": 2.381287063474632, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.266683578491211, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7318645119667053, + "num_tokens": 560981734.0, + "step": 21684 + }, + { + "epoch": 2.381396881177246, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4229819774627686, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7396672964096069, + "num_tokens": 561004489.0, + "step": 21685 + }, + { + "epoch": 2.3815066988798597, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.659989595413208, + "learning_rate": 1e-06, + "loss": 0.7898, + "mean_token_accuracy": 0.7483892440795898, + "num_tokens": 561024652.0, + "step": 21686 + }, + { + "epoch": 2.381616516582473, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4161131381988525, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7366436719894409, + "num_tokens": 561049401.0, + "step": 21687 + }, + { + "epoch": 2.3817263342850867, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.626119375228882, + "learning_rate": 1e-06, + "loss": 0.7749, + "mean_token_accuracy": 0.7459508776664734, + "num_tokens": 561071127.0, + "step": 21688 + }, + { + "epoch": 2.3818361519877005, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4002649784088135, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7118082046508789, + "num_tokens": 561095712.0, + "step": 21689 + }, + { + "epoch": 2.3819459696903142, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5216360092163086, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7423352003097534, + "num_tokens": 561118285.0, + "step": 21690 + }, + { + "epoch": 2.3820557873929276, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.201460838317871, + "learning_rate": 1e-06, + "loss": 0.9485, + "mean_token_accuracy": 0.7122342586517334, + "num_tokens": 561147901.0, + "step": 21691 + }, + { + "epoch": 2.3821656050955413, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.189614772796631, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7249688506126404, + "num_tokens": 561176893.0, + "step": 21692 + }, + { + "epoch": 2.382275422798155, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1946358680725098, + "learning_rate": 1e-06, + "loss": 0.9967, + "mean_token_accuracy": 0.6915676593780518, + "num_tokens": 561207332.0, + "step": 21693 + }, + { + "epoch": 2.382385240500769, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.5969841480255127, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7419309020042419, + "num_tokens": 561228511.0, + "step": 21694 + }, + { + "epoch": 2.382495058203382, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1937568187713623, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7246265411376953, + "num_tokens": 561256955.0, + "step": 21695 + }, + { + "epoch": 2.382604875905996, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.512897491455078, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7385620474815369, + "num_tokens": 561279513.0, + "step": 21696 + }, + { + "epoch": 2.3827146936086097, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.254129409790039, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7228678464889526, + "num_tokens": 561307164.0, + "step": 21697 + }, + { + "epoch": 2.3828245113112234, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1443421840667725, + "learning_rate": 1e-06, + "loss": 1.0051, + "mean_token_accuracy": 0.6933943629264832, + "num_tokens": 561337569.0, + "step": 21698 + }, + { + "epoch": 2.382934329013837, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2482194900512695, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7214407920837402, + "num_tokens": 561367097.0, + "step": 21699 + }, + { + "epoch": 2.3830441467164505, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.2882747650146484, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7246460914611816, + "num_tokens": 561394499.0, + "step": 21700 + }, + { + "epoch": 2.3831539644190642, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.273402452468872, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7300790548324585, + "num_tokens": 561422943.0, + "step": 21701 + }, + { + "epoch": 2.383263782121678, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.43179988861084, + "learning_rate": 1e-06, + "loss": 0.8036, + "mean_token_accuracy": 0.742901086807251, + "num_tokens": 561446362.0, + "step": 21702 + }, + { + "epoch": 2.3833735998242918, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.467883825302124, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7189086079597473, + "num_tokens": 561471673.0, + "step": 21703 + }, + { + "epoch": 2.3834834175269055, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.7917301654815674, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7372263669967651, + "num_tokens": 561491792.0, + "step": 21704 + }, + { + "epoch": 2.383593235229519, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1357109546661377, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7090227603912354, + "num_tokens": 561523490.0, + "step": 21705 + }, + { + "epoch": 2.3837030529321326, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.512021541595459, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7318977117538452, + "num_tokens": 561546471.0, + "step": 21706 + }, + { + "epoch": 2.3838128706347463, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.561418056488037, + "learning_rate": 1e-06, + "loss": 0.7625, + "mean_token_accuracy": 0.7511993646621704, + "num_tokens": 561568689.0, + "step": 21707 + }, + { + "epoch": 2.38392268833736, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3444221019744873, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7246583700180054, + "num_tokens": 561595060.0, + "step": 21708 + }, + { + "epoch": 2.384032506039974, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.3764891624450684, + "learning_rate": 1e-06, + "loss": 0.791, + "mean_token_accuracy": 0.7472347617149353, + "num_tokens": 561619795.0, + "step": 21709 + }, + { + "epoch": 2.384142323742587, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 3.249427318572998, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7460549473762512, + "num_tokens": 561636235.0, + "step": 21710 + }, + { + "epoch": 2.384252141445201, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.118943691253662, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.7517168521881104, + "num_tokens": 561665324.0, + "step": 21711 + }, + { + "epoch": 2.3843619591478147, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.525596857070923, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7113251686096191, + "num_tokens": 561686989.0, + "step": 21712 + }, + { + "epoch": 2.3844717768504284, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.385979413986206, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7202153205871582, + "num_tokens": 561713475.0, + "step": 21713 + }, + { + "epoch": 2.384581594553042, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.217609167098999, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7312405109405518, + "num_tokens": 561740807.0, + "step": 21714 + }, + { + "epoch": 2.3846914122556555, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.4851772785186768, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7457197904586792, + "num_tokens": 561764139.0, + "step": 21715 + }, + { + "epoch": 2.3848012299582693, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.1762053966522217, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7480819225311279, + "num_tokens": 561791526.0, + "step": 21716 + }, + { + "epoch": 2.384911047660883, + "ewc_loss": 1.4007091522216797e-05, + "grad_norm": 2.688291311264038, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7337675094604492, + "num_tokens": 561813069.0, + "step": 21717 + }, + { + "epoch": 2.385020865363497, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2807297706604004, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7424290180206299, + "num_tokens": 561838720.0, + "step": 21718 + }, + { + "epoch": 2.38513068306611, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4155006408691406, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7411494255065918, + "num_tokens": 561864126.0, + "step": 21719 + }, + { + "epoch": 2.385240500768724, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.40842604637146, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7341952323913574, + "num_tokens": 561889744.0, + "step": 21720 + }, + { + "epoch": 2.3853503184713376, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3429219722747803, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7313144207000732, + "num_tokens": 561914889.0, + "step": 21721 + }, + { + "epoch": 2.3854601361739514, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.484605312347412, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7308081388473511, + "num_tokens": 561939297.0, + "step": 21722 + }, + { + "epoch": 2.3855699538765647, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3708560466766357, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7320206165313721, + "num_tokens": 561965447.0, + "step": 21723 + }, + { + "epoch": 2.3856797715791784, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2793402671813965, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7008587121963501, + "num_tokens": 561994561.0, + "step": 21724 + }, + { + "epoch": 2.385789589281792, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 1.9644099473953247, + "learning_rate": 1e-06, + "loss": 0.7694, + "mean_token_accuracy": 0.746684730052948, + "num_tokens": 562028541.0, + "step": 21725 + }, + { + "epoch": 2.385899406984406, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.1795215606689453, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7334574460983276, + "num_tokens": 562059217.0, + "step": 21726 + }, + { + "epoch": 2.3860092246870197, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.304227828979492, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7407253980636597, + "num_tokens": 562085241.0, + "step": 21727 + }, + { + "epoch": 2.386119042389633, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4891586303710938, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7476283311843872, + "num_tokens": 562109433.0, + "step": 21728 + }, + { + "epoch": 2.3862288600922468, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.7737555503845215, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7364833354949951, + "num_tokens": 562130464.0, + "step": 21729 + }, + { + "epoch": 2.3863386777948605, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5077288150787354, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7444546222686768, + "num_tokens": 562154334.0, + "step": 21730 + }, + { + "epoch": 2.3864484954974743, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3369548320770264, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7512657642364502, + "num_tokens": 562179353.0, + "step": 21731 + }, + { + "epoch": 2.386558313200088, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2085280418395996, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7331868410110474, + "num_tokens": 562209174.0, + "step": 21732 + }, + { + "epoch": 2.3866681309027014, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2754359245300293, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7447806596755981, + "num_tokens": 562238300.0, + "step": 21733 + }, + { + "epoch": 2.386777948605315, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.248358726501465, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7183592915534973, + "num_tokens": 562267018.0, + "step": 21734 + }, + { + "epoch": 2.386887766307929, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.091024160385132, + "learning_rate": 1e-06, + "loss": 0.7713, + "mean_token_accuracy": 0.7571541666984558, + "num_tokens": 562296651.0, + "step": 21735 + }, + { + "epoch": 2.3869975840105426, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.483459949493408, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7216975688934326, + "num_tokens": 562322156.0, + "step": 21736 + }, + { + "epoch": 2.3871074017131564, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4204163551330566, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7383112907409668, + "num_tokens": 562344786.0, + "step": 21737 + }, + { + "epoch": 2.3872172194157697, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5139522552490234, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7322688102722168, + "num_tokens": 562368680.0, + "step": 21738 + }, + { + "epoch": 2.3873270371183835, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2368810176849365, + "learning_rate": 1e-06, + "loss": 0.946, + "mean_token_accuracy": 0.7109949588775635, + "num_tokens": 562396086.0, + "step": 21739 + }, + { + "epoch": 2.387436854820997, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.289414644241333, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7311019897460938, + "num_tokens": 562424322.0, + "step": 21740 + }, + { + "epoch": 2.387546672523611, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3937575817108154, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7076712846755981, + "num_tokens": 562449818.0, + "step": 21741 + }, + { + "epoch": 2.3876564902262243, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.193493604660034, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7289265394210815, + "num_tokens": 562479687.0, + "step": 21742 + }, + { + "epoch": 2.387766307928838, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.6070237159729004, + "learning_rate": 1e-06, + "loss": 0.7642, + "mean_token_accuracy": 0.7541177868843079, + "num_tokens": 562500708.0, + "step": 21743 + }, + { + "epoch": 2.387876125631452, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5685794353485107, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7183378338813782, + "num_tokens": 562524041.0, + "step": 21744 + }, + { + "epoch": 2.3879859433340656, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4306557178497314, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7269008159637451, + "num_tokens": 562551157.0, + "step": 21745 + }, + { + "epoch": 2.388095761036679, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.058258295059204, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7390009760856628, + "num_tokens": 562583952.0, + "step": 21746 + }, + { + "epoch": 2.3882055787392926, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5950844287872314, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7407585382461548, + "num_tokens": 562606927.0, + "step": 21747 + }, + { + "epoch": 2.3883153964419064, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.733412742614746, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7311924695968628, + "num_tokens": 562627104.0, + "step": 21748 + }, + { + "epoch": 2.38842521414452, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5902624130249023, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7267308235168457, + "num_tokens": 562649867.0, + "step": 21749 + }, + { + "epoch": 2.388535031847134, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.653566837310791, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7167534232139587, + "num_tokens": 562672739.0, + "step": 21750 + }, + { + "epoch": 2.388644849549747, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4094674587249756, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7147020697593689, + "num_tokens": 562699666.0, + "step": 21751 + }, + { + "epoch": 2.388754667252361, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.6530768871307373, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7476407289505005, + "num_tokens": 562719402.0, + "step": 21752 + }, + { + "epoch": 2.3888644849549747, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5849997997283936, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7237253189086914, + "num_tokens": 562743460.0, + "step": 21753 + }, + { + "epoch": 2.3889743026575885, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5675621032714844, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.729756236076355, + "num_tokens": 562766855.0, + "step": 21754 + }, + { + "epoch": 2.3890841203602022, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5595293045043945, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7378733158111572, + "num_tokens": 562789578.0, + "step": 21755 + }, + { + "epoch": 2.3891939380628155, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2582712173461914, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7411903142929077, + "num_tokens": 562816256.0, + "step": 21756 + }, + { + "epoch": 2.3893037557654293, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4375572204589844, + "learning_rate": 1e-06, + "loss": 0.9839, + "mean_token_accuracy": 0.697094738483429, + "num_tokens": 562842188.0, + "step": 21757 + }, + { + "epoch": 2.389413573468043, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2747347354888916, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7361034154891968, + "num_tokens": 562870493.0, + "step": 21758 + }, + { + "epoch": 2.389523391170657, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.341393232345581, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7337520718574524, + "num_tokens": 562896126.0, + "step": 21759 + }, + { + "epoch": 2.3896332088732706, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.695460081100464, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.7466185092926025, + "num_tokens": 562916736.0, + "step": 21760 + }, + { + "epoch": 2.389743026575884, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.0675370693206787, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7197290658950806, + "num_tokens": 562948370.0, + "step": 21761 + }, + { + "epoch": 2.3898528442784976, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2265965938568115, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7458745241165161, + "num_tokens": 562974389.0, + "step": 21762 + }, + { + "epoch": 2.3899626619811114, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2211802005767822, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7324948310852051, + "num_tokens": 563001419.0, + "step": 21763 + }, + { + "epoch": 2.390072479683725, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.4595372676849365, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.712023138999939, + "num_tokens": 563028210.0, + "step": 21764 + }, + { + "epoch": 2.390182297386339, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.334472417831421, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7410761117935181, + "num_tokens": 563052099.0, + "step": 21765 + }, + { + "epoch": 2.3902921150889522, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.666619300842285, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7203453779220581, + "num_tokens": 563074501.0, + "step": 21766 + }, + { + "epoch": 2.390401932791566, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.1157965660095215, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7295191287994385, + "num_tokens": 563105438.0, + "step": 21767 + }, + { + "epoch": 2.3905117504941797, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.5033116340637207, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7337875366210938, + "num_tokens": 563127829.0, + "step": 21768 + }, + { + "epoch": 2.3906215681967935, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.550888776779175, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7371105551719666, + "num_tokens": 563150649.0, + "step": 21769 + }, + { + "epoch": 2.390731385899407, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.406184196472168, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.722381591796875, + "num_tokens": 563174548.0, + "step": 21770 + }, + { + "epoch": 2.3908412036020206, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 1.9664946794509888, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7103461027145386, + "num_tokens": 563211815.0, + "step": 21771 + }, + { + "epoch": 2.3909510213046343, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.1930506229400635, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7161089181900024, + "num_tokens": 563240838.0, + "step": 21772 + }, + { + "epoch": 2.391060839007248, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.376197576522827, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7343709468841553, + "num_tokens": 563266701.0, + "step": 21773 + }, + { + "epoch": 2.3911706567098614, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.430093765258789, + "learning_rate": 1e-06, + "loss": 0.7594, + "mean_token_accuracy": 0.7552855014801025, + "num_tokens": 563290876.0, + "step": 21774 + }, + { + "epoch": 2.391280474412475, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.869597911834717, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.7511408925056458, + "num_tokens": 563310841.0, + "step": 21775 + }, + { + "epoch": 2.391390292115089, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4236831665039062, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7487738132476807, + "num_tokens": 563335458.0, + "step": 21776 + }, + { + "epoch": 2.3915001098177027, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2548017501831055, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7544741630554199, + "num_tokens": 563362466.0, + "step": 21777 + }, + { + "epoch": 2.3916099275203164, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.26928448677063, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7334854602813721, + "num_tokens": 563391871.0, + "step": 21778 + }, + { + "epoch": 2.3917197452229297, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4116029739379883, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7247402667999268, + "num_tokens": 563417280.0, + "step": 21779 + }, + { + "epoch": 2.3918295629255435, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.208263635635376, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7144975066184998, + "num_tokens": 563446860.0, + "step": 21780 + }, + { + "epoch": 2.3919393806281573, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.340663194656372, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7294107675552368, + "num_tokens": 563476694.0, + "step": 21781 + }, + { + "epoch": 2.392049198330771, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4205453395843506, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7470843195915222, + "num_tokens": 563501967.0, + "step": 21782 + }, + { + "epoch": 2.3921590160333848, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.1800670623779297, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7218995690345764, + "num_tokens": 563533006.0, + "step": 21783 + }, + { + "epoch": 2.392268833735998, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4773318767547607, + "learning_rate": 1e-06, + "loss": 0.7617, + "mean_token_accuracy": 0.7534340023994446, + "num_tokens": 563556085.0, + "step": 21784 + }, + { + "epoch": 2.392378651438612, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.283989429473877, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7348586320877075, + "num_tokens": 563583818.0, + "step": 21785 + }, + { + "epoch": 2.3924884691412256, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2513561248779297, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7208707928657532, + "num_tokens": 563611700.0, + "step": 21786 + }, + { + "epoch": 2.3925982868438394, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.8662567138671875, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.728134274482727, + "num_tokens": 563635091.0, + "step": 21787 + }, + { + "epoch": 2.392708104546453, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.223505973815918, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7333428859710693, + "num_tokens": 563664243.0, + "step": 21788 + }, + { + "epoch": 2.3928179222490664, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.620457887649536, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7426413297653198, + "num_tokens": 563687430.0, + "step": 21789 + }, + { + "epoch": 2.39292773995168, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.243346929550171, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7268573045730591, + "num_tokens": 563716757.0, + "step": 21790 + }, + { + "epoch": 2.393037557654294, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4963772296905518, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7469412088394165, + "num_tokens": 563739681.0, + "step": 21791 + }, + { + "epoch": 2.3931473753569077, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.414797067642212, + "learning_rate": 1e-06, + "loss": 0.8088, + "mean_token_accuracy": 0.7421427965164185, + "num_tokens": 563763013.0, + "step": 21792 + }, + { + "epoch": 2.393257193059521, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.379385471343994, + "learning_rate": 1e-06, + "loss": 0.9233, + "mean_token_accuracy": 0.7107787132263184, + "num_tokens": 563789203.0, + "step": 21793 + }, + { + "epoch": 2.3933670107621348, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.4596974849700928, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7365652322769165, + "num_tokens": 563813546.0, + "step": 21794 + }, + { + "epoch": 2.3934768284647485, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.4006423950195312, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7228111028671265, + "num_tokens": 563840301.0, + "step": 21795 + }, + { + "epoch": 2.3935866461673623, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.380601644515991, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7222490310668945, + "num_tokens": 563867125.0, + "step": 21796 + }, + { + "epoch": 2.393696463869976, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.600571870803833, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7343271970748901, + "num_tokens": 563888995.0, + "step": 21797 + }, + { + "epoch": 2.3938062815725893, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.2830312252044678, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7380478978157043, + "num_tokens": 563914370.0, + "step": 21798 + }, + { + "epoch": 2.393916099275203, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3104686737060547, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7415122389793396, + "num_tokens": 563941515.0, + "step": 21799 + }, + { + "epoch": 2.394025916977817, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.2526893615722656, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7510541677474976, + "num_tokens": 563967176.0, + "step": 21800 + }, + { + "epoch": 2.3941357346804306, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.5655109882354736, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7461336255073547, + "num_tokens": 563989351.0, + "step": 21801 + }, + { + "epoch": 2.394245552383044, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.249483346939087, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7383483052253723, + "num_tokens": 564015982.0, + "step": 21802 + }, + { + "epoch": 2.3943553700856577, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.170767307281494, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7410452961921692, + "num_tokens": 564043171.0, + "step": 21803 + }, + { + "epoch": 2.3944651877882714, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.131093740463257, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.719588577747345, + "num_tokens": 564073133.0, + "step": 21804 + }, + { + "epoch": 2.394575005490885, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.30385422706604, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7631640434265137, + "num_tokens": 564099305.0, + "step": 21805 + }, + { + "epoch": 2.394684823193499, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.263366460800171, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7461341619491577, + "num_tokens": 564124842.0, + "step": 21806 + }, + { + "epoch": 2.3947946408961123, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.1598751544952393, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7108913660049438, + "num_tokens": 564158542.0, + "step": 21807 + }, + { + "epoch": 2.394904458598726, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3069357872009277, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.72872394323349, + "num_tokens": 564185746.0, + "step": 21808 + }, + { + "epoch": 2.39501427630134, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4690229892730713, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7389282584190369, + "num_tokens": 564208325.0, + "step": 21809 + }, + { + "epoch": 2.3951240940039535, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.170621633529663, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7293492555618286, + "num_tokens": 564237384.0, + "step": 21810 + }, + { + "epoch": 2.3952339117065673, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.467963695526123, + "learning_rate": 1e-06, + "loss": 0.8843, + "mean_token_accuracy": 0.719819962978363, + "num_tokens": 564263098.0, + "step": 21811 + }, + { + "epoch": 2.3953437294091806, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.4474446773529053, + "learning_rate": 1e-06, + "loss": 0.8997, + "mean_token_accuracy": 0.7145535945892334, + "num_tokens": 564287499.0, + "step": 21812 + }, + { + "epoch": 2.3954535471117944, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.7415502071380615, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7376806735992432, + "num_tokens": 564307598.0, + "step": 21813 + }, + { + "epoch": 2.395563364814408, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.206608295440674, + "learning_rate": 1e-06, + "loss": 0.9449, + "mean_token_accuracy": 0.7118749618530273, + "num_tokens": 564338375.0, + "step": 21814 + }, + { + "epoch": 2.395673182517022, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.439588785171509, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7353354692459106, + "num_tokens": 564361749.0, + "step": 21815 + }, + { + "epoch": 2.3957830002196356, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3408713340759277, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7271546125411987, + "num_tokens": 564387269.0, + "step": 21816 + }, + { + "epoch": 2.395892817922249, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.3864176273345947, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7195462584495544, + "num_tokens": 564411993.0, + "step": 21817 + }, + { + "epoch": 2.3960026356248627, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.2672367095947266, + "learning_rate": 1e-06, + "loss": 0.8274, + "mean_token_accuracy": 0.7407552003860474, + "num_tokens": 564437303.0, + "step": 21818 + }, + { + "epoch": 2.3961124533274765, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.220831871032715, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.734089732170105, + "num_tokens": 564467197.0, + "step": 21819 + }, + { + "epoch": 2.3962222710300902, + "ewc_loss": 1.4066696166992188e-05, + "grad_norm": 2.0579965114593506, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7233662605285645, + "num_tokens": 564501039.0, + "step": 21820 + }, + { + "epoch": 2.3963320887327035, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.187554359436035, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.724097728729248, + "num_tokens": 564529656.0, + "step": 21821 + }, + { + "epoch": 2.3964419064353173, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.0584306716918945, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7346295118331909, + "num_tokens": 564560411.0, + "step": 21822 + }, + { + "epoch": 2.396551724137931, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.1336817741394043, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7347933650016785, + "num_tokens": 564589332.0, + "step": 21823 + }, + { + "epoch": 2.396661541840545, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.1810450553894043, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.742801308631897, + "num_tokens": 564617369.0, + "step": 21824 + }, + { + "epoch": 2.396771359543158, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.581541061401367, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7266121506690979, + "num_tokens": 564640388.0, + "step": 21825 + }, + { + "epoch": 2.396881177245772, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.508992910385132, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.7140049934387207, + "num_tokens": 564664886.0, + "step": 21826 + }, + { + "epoch": 2.3969909949483856, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.2973766326904297, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7199558615684509, + "num_tokens": 564693525.0, + "step": 21827 + }, + { + "epoch": 2.3971008126509994, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3537864685058594, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7239996194839478, + "num_tokens": 564721609.0, + "step": 21828 + }, + { + "epoch": 2.397210630353613, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.379589557647705, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.731706440448761, + "num_tokens": 564747590.0, + "step": 21829 + }, + { + "epoch": 2.3973204480562265, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.0901689529418945, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7030825614929199, + "num_tokens": 564778965.0, + "step": 21830 + }, + { + "epoch": 2.39743026575884, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.5035791397094727, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.741693913936615, + "num_tokens": 564800943.0, + "step": 21831 + }, + { + "epoch": 2.397540083461454, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3864307403564453, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7093158960342407, + "num_tokens": 564826490.0, + "step": 21832 + }, + { + "epoch": 2.3976499011640677, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.412623405456543, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7184332013130188, + "num_tokens": 564852336.0, + "step": 21833 + }, + { + "epoch": 2.3977597188666815, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.410142421722412, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7398545742034912, + "num_tokens": 564876595.0, + "step": 21834 + }, + { + "epoch": 2.397869536569295, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3090617656707764, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7292330265045166, + "num_tokens": 564906207.0, + "step": 21835 + }, + { + "epoch": 2.3979793542719086, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.7634963989257812, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7433013916015625, + "num_tokens": 564925607.0, + "step": 21836 + }, + { + "epoch": 2.3980891719745223, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.470679759979248, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7458182573318481, + "num_tokens": 564948981.0, + "step": 21837 + }, + { + "epoch": 2.398198989677136, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.1038572788238525, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7202442288398743, + "num_tokens": 564980853.0, + "step": 21838 + }, + { + "epoch": 2.39830880737975, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.1978042125701904, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.7119432091712952, + "num_tokens": 565010521.0, + "step": 21839 + }, + { + "epoch": 2.398418625082363, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.7132022380828857, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.755730152130127, + "num_tokens": 565030037.0, + "step": 21840 + }, + { + "epoch": 2.398528442784977, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.4363229274749756, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7245278358459473, + "num_tokens": 565056268.0, + "step": 21841 + }, + { + "epoch": 2.3986382604875907, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3450450897216797, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7170606255531311, + "num_tokens": 565082378.0, + "step": 21842 + }, + { + "epoch": 2.3987480781902044, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.9266934394836426, + "learning_rate": 1e-06, + "loss": 0.7719, + "mean_token_accuracy": 0.7456324696540833, + "num_tokens": 565100237.0, + "step": 21843 + }, + { + "epoch": 2.398857895892818, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3548011779785156, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7185835838317871, + "num_tokens": 565125361.0, + "step": 21844 + }, + { + "epoch": 2.3989677135954315, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.599339723587036, + "learning_rate": 1e-06, + "loss": 0.728, + "mean_token_accuracy": 0.7582417726516724, + "num_tokens": 565146362.0, + "step": 21845 + }, + { + "epoch": 2.3990775312980452, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.529825448989868, + "learning_rate": 1e-06, + "loss": 0.7597, + "mean_token_accuracy": 0.7521512508392334, + "num_tokens": 565168609.0, + "step": 21846 + }, + { + "epoch": 2.399187349000659, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.5178189277648926, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7211308479309082, + "num_tokens": 565192456.0, + "step": 21847 + }, + { + "epoch": 2.3992971667032728, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.277279853820801, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7188898921012878, + "num_tokens": 565221728.0, + "step": 21848 + }, + { + "epoch": 2.399406984405886, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.119159698486328, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7384095191955566, + "num_tokens": 565253334.0, + "step": 21849 + }, + { + "epoch": 2.3995168021085, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.3314898014068604, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.717729926109314, + "num_tokens": 565280534.0, + "step": 21850 + }, + { + "epoch": 2.3996266198111136, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.2951862812042236, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.731569766998291, + "num_tokens": 565307008.0, + "step": 21851 + }, + { + "epoch": 2.3997364375137273, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.5213193893432617, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7400848865509033, + "num_tokens": 565329287.0, + "step": 21852 + }, + { + "epoch": 2.3998462552163407, + "ewc_loss": 1.4185905456542969e-05, + "grad_norm": 2.197422504425049, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7421255707740784, + "num_tokens": 565358958.0, + "step": 21853 + }, + { + "epoch": 2.3999560729189544, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.2600085735321045, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7482441663742065, + "num_tokens": 565388515.0, + "step": 21854 + }, + { + "epoch": 2.400065890621568, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.592092752456665, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7324496507644653, + "num_tokens": 565411671.0, + "step": 21855 + }, + { + "epoch": 2.400175708324182, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.475673198699951, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.744577169418335, + "num_tokens": 565434986.0, + "step": 21856 + }, + { + "epoch": 2.4002855260267957, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.351482629776001, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7281039953231812, + "num_tokens": 565460261.0, + "step": 21857 + }, + { + "epoch": 2.400395343729409, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.2231686115264893, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7188403606414795, + "num_tokens": 565489292.0, + "step": 21858 + }, + { + "epoch": 2.4005051614320227, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.2927608489990234, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7586084604263306, + "num_tokens": 565512923.0, + "step": 21859 + }, + { + "epoch": 2.4006149791346365, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.301805019378662, + "learning_rate": 1e-06, + "loss": 0.9563, + "mean_token_accuracy": 0.6984165906906128, + "num_tokens": 565542048.0, + "step": 21860 + }, + { + "epoch": 2.4007247968372503, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.2274110317230225, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7163326740264893, + "num_tokens": 565571858.0, + "step": 21861 + }, + { + "epoch": 2.400834614539864, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4037179946899414, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.720115065574646, + "num_tokens": 565597292.0, + "step": 21862 + }, + { + "epoch": 2.4009444322424773, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.282886505126953, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.7422553300857544, + "num_tokens": 565622850.0, + "step": 21863 + }, + { + "epoch": 2.401054249945091, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.5498111248016357, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7242313623428345, + "num_tokens": 565645293.0, + "step": 21864 + }, + { + "epoch": 2.401164067647705, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.368654727935791, + "learning_rate": 1e-06, + "loss": 0.7516, + "mean_token_accuracy": 0.7706329822540283, + "num_tokens": 565673669.0, + "step": 21865 + }, + { + "epoch": 2.4012738853503186, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.1603128910064697, + "learning_rate": 1e-06, + "loss": 0.9571, + "mean_token_accuracy": 0.7098844647407532, + "num_tokens": 565703567.0, + "step": 21866 + }, + { + "epoch": 2.4013837030529324, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.3441858291625977, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7183624505996704, + "num_tokens": 565728942.0, + "step": 21867 + }, + { + "epoch": 2.4014935207555457, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.1241207122802734, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7304857969284058, + "num_tokens": 565760420.0, + "step": 21868 + }, + { + "epoch": 2.4016033384581594, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.5265021324157715, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7449280023574829, + "num_tokens": 565782604.0, + "step": 21869 + }, + { + "epoch": 2.401713156160773, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.6665611267089844, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7212347984313965, + "num_tokens": 565803210.0, + "step": 21870 + }, + { + "epoch": 2.401822973863387, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.24303936958313, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7288249731063843, + "num_tokens": 565830695.0, + "step": 21871 + }, + { + "epoch": 2.4019327915660003, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.434797763824463, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.730921745300293, + "num_tokens": 565855945.0, + "step": 21872 + }, + { + "epoch": 2.402042609268614, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.756845474243164, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7436216473579407, + "num_tokens": 565875323.0, + "step": 21873 + }, + { + "epoch": 2.4021524269712278, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.3310978412628174, + "learning_rate": 1e-06, + "loss": 0.9674, + "mean_token_accuracy": 0.7062888741493225, + "num_tokens": 565904005.0, + "step": 21874 + }, + { + "epoch": 2.4022622446738415, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.1634342670440674, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.71992027759552, + "num_tokens": 565933325.0, + "step": 21875 + }, + { + "epoch": 2.402372062376455, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.8057515621185303, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7311276793479919, + "num_tokens": 565952176.0, + "step": 21876 + }, + { + "epoch": 2.4024818800790686, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.484434127807617, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7461534142494202, + "num_tokens": 565975579.0, + "step": 21877 + }, + { + "epoch": 2.4025916977816824, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.1970901489257812, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7172094583511353, + "num_tokens": 566003823.0, + "step": 21878 + }, + { + "epoch": 2.402701515484296, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.3404381275177, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7384049892425537, + "num_tokens": 566031231.0, + "step": 21879 + }, + { + "epoch": 2.40281133318691, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.608790159225464, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7369903922080994, + "num_tokens": 566053359.0, + "step": 21880 + }, + { + "epoch": 2.402921150889523, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.3348324298858643, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7382993102073669, + "num_tokens": 566081184.0, + "step": 21881 + }, + { + "epoch": 2.403030968592137, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4170877933502197, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7220106720924377, + "num_tokens": 566107003.0, + "step": 21882 + }, + { + "epoch": 2.4031407862947507, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4051225185394287, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7207571268081665, + "num_tokens": 566131996.0, + "step": 21883 + }, + { + "epoch": 2.4032506039973645, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.320582628250122, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7421469688415527, + "num_tokens": 566156993.0, + "step": 21884 + }, + { + "epoch": 2.403360421699978, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.3971428871154785, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7167165279388428, + "num_tokens": 566181627.0, + "step": 21885 + }, + { + "epoch": 2.4034702394025915, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.443157911300659, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7288815975189209, + "num_tokens": 566207643.0, + "step": 21886 + }, + { + "epoch": 2.4035800571052053, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4254798889160156, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.727796196937561, + "num_tokens": 566232623.0, + "step": 21887 + }, + { + "epoch": 2.403689874807819, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4020261764526367, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7229253649711609, + "num_tokens": 566259132.0, + "step": 21888 + }, + { + "epoch": 2.403799692510433, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.5114331245422363, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7511669397354126, + "num_tokens": 566283076.0, + "step": 21889 + }, + { + "epoch": 2.4039095102130466, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.4439697265625, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7424408793449402, + "num_tokens": 566307400.0, + "step": 21890 + }, + { + "epoch": 2.40401932791566, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.8592679500579834, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7383748292922974, + "num_tokens": 566326041.0, + "step": 21891 + }, + { + "epoch": 2.4041291456182736, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.2137649059295654, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.732467532157898, + "num_tokens": 566355322.0, + "step": 21892 + }, + { + "epoch": 2.4042389633208874, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.314319610595703, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7323007583618164, + "num_tokens": 566381855.0, + "step": 21893 + }, + { + "epoch": 2.404348781023501, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 3.0904629230499268, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7469087839126587, + "num_tokens": 566405157.0, + "step": 21894 + }, + { + "epoch": 2.404458598726115, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.167112112045288, + "learning_rate": 1e-06, + "loss": 0.7596, + "mean_token_accuracy": 0.7552947402000427, + "num_tokens": 566432532.0, + "step": 21895 + }, + { + "epoch": 2.404568416428728, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.82308292388916, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7448257207870483, + "num_tokens": 566452057.0, + "step": 21896 + }, + { + "epoch": 2.404678234131342, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.7696146965026855, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7442857623100281, + "num_tokens": 566472312.0, + "step": 21897 + }, + { + "epoch": 2.4047880518339557, + "ewc_loss": 1.424551010131836e-05, + "grad_norm": 2.468017816543579, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7398281097412109, + "num_tokens": 566498063.0, + "step": 21898 + }, + { + "epoch": 2.4048978695365695, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5321106910705566, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.735101044178009, + "num_tokens": 566520570.0, + "step": 21899 + }, + { + "epoch": 2.405007687239183, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1587889194488525, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7283665537834167, + "num_tokens": 566547892.0, + "step": 21900 + }, + { + "epoch": 2.4051175049417965, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5059268474578857, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7421501278877258, + "num_tokens": 566569457.0, + "step": 21901 + }, + { + "epoch": 2.4052273226444103, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1109392642974854, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7237330675125122, + "num_tokens": 566601355.0, + "step": 21902 + }, + { + "epoch": 2.405337140347024, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.529906988143921, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.7446416616439819, + "num_tokens": 566622944.0, + "step": 21903 + }, + { + "epoch": 2.4054469580496374, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.346245765686035, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7171652317047119, + "num_tokens": 566649726.0, + "step": 21904 + }, + { + "epoch": 2.405556775752251, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4357521533966064, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7142733335494995, + "num_tokens": 566676527.0, + "step": 21905 + }, + { + "epoch": 2.405666593454865, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.236088991165161, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.735052227973938, + "num_tokens": 566703545.0, + "step": 21906 + }, + { + "epoch": 2.4057764111574786, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2629599571228027, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7498341798782349, + "num_tokens": 566730234.0, + "step": 21907 + }, + { + "epoch": 2.4058862288600924, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.259644031524658, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7316557168960571, + "num_tokens": 566755985.0, + "step": 21908 + }, + { + "epoch": 2.4059960465627057, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.405346155166626, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7280029058456421, + "num_tokens": 566780924.0, + "step": 21909 + }, + { + "epoch": 2.4061058642653195, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3904776573181152, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7444450259208679, + "num_tokens": 566805480.0, + "step": 21910 + }, + { + "epoch": 2.4062156819679332, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.476344108581543, + "learning_rate": 1e-06, + "loss": 0.9718, + "mean_token_accuracy": 0.6961968541145325, + "num_tokens": 566832750.0, + "step": 21911 + }, + { + "epoch": 2.406325499670547, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.485399007797241, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.739569902420044, + "num_tokens": 566855160.0, + "step": 21912 + }, + { + "epoch": 2.4064353173731607, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.592568874359131, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.741754949092865, + "num_tokens": 566876672.0, + "step": 21913 + }, + { + "epoch": 2.406545135075774, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.249131917953491, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7142616510391235, + "num_tokens": 566906330.0, + "step": 21914 + }, + { + "epoch": 2.406654952778388, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4891152381896973, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.71851646900177, + "num_tokens": 566930922.0, + "step": 21915 + }, + { + "epoch": 2.4067647704810016, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.250756025314331, + "learning_rate": 1e-06, + "loss": 0.7844, + "mean_token_accuracy": 0.7469323873519897, + "num_tokens": 566957596.0, + "step": 21916 + }, + { + "epoch": 2.4068745881836153, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3935062885284424, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7295776605606079, + "num_tokens": 566982957.0, + "step": 21917 + }, + { + "epoch": 2.406984405886229, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4340591430664062, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7312405109405518, + "num_tokens": 567006482.0, + "step": 21918 + }, + { + "epoch": 2.4070942235888424, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3385369777679443, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7380095720291138, + "num_tokens": 567032029.0, + "step": 21919 + }, + { + "epoch": 2.407204041291456, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4435157775878906, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7352151870727539, + "num_tokens": 567057629.0, + "step": 21920 + }, + { + "epoch": 2.40731385899407, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.293863296508789, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7243148684501648, + "num_tokens": 567085534.0, + "step": 21921 + }, + { + "epoch": 2.4074236766966837, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1007609367370605, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.718589186668396, + "num_tokens": 567118239.0, + "step": 21922 + }, + { + "epoch": 2.407533494399297, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.598320245742798, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.745763897895813, + "num_tokens": 567141682.0, + "step": 21923 + }, + { + "epoch": 2.4076433121019107, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1054162979125977, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.730736494064331, + "num_tokens": 567174548.0, + "step": 21924 + }, + { + "epoch": 2.4077531298045245, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.484421968460083, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7442930936813354, + "num_tokens": 567197996.0, + "step": 21925 + }, + { + "epoch": 2.4078629475071383, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.054370641708374, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7224386930465698, + "num_tokens": 567231224.0, + "step": 21926 + }, + { + "epoch": 2.4079727652097516, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.383603811264038, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7247574329376221, + "num_tokens": 567257919.0, + "step": 21927 + }, + { + "epoch": 2.4080825829123653, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4991610050201416, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7249662280082703, + "num_tokens": 567283142.0, + "step": 21928 + }, + { + "epoch": 2.408192400614979, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3185741901397705, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7172476053237915, + "num_tokens": 567308445.0, + "step": 21929 + }, + { + "epoch": 2.408302218317593, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1600470542907715, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7282682061195374, + "num_tokens": 567337026.0, + "step": 21930 + }, + { + "epoch": 2.4084120360202066, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.461639642715454, + "learning_rate": 1e-06, + "loss": 0.9435, + "mean_token_accuracy": 0.7102546691894531, + "num_tokens": 567362751.0, + "step": 21931 + }, + { + "epoch": 2.40852185372282, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2549400329589844, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.737170934677124, + "num_tokens": 567389259.0, + "step": 21932 + }, + { + "epoch": 2.4086316714254337, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.402348279953003, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.733771562576294, + "num_tokens": 567413614.0, + "step": 21933 + }, + { + "epoch": 2.4087414891280474, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 1.9518643617630005, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7351609468460083, + "num_tokens": 567449268.0, + "step": 21934 + }, + { + "epoch": 2.408851306830661, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6835010051727295, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.75154048204422, + "num_tokens": 567469816.0, + "step": 21935 + }, + { + "epoch": 2.408961124533275, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.415048599243164, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7256027460098267, + "num_tokens": 567495710.0, + "step": 21936 + }, + { + "epoch": 2.4090709422358882, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.555675745010376, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7440474033355713, + "num_tokens": 567518594.0, + "step": 21937 + }, + { + "epoch": 2.409180759938502, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3295063972473145, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7347397804260254, + "num_tokens": 567545070.0, + "step": 21938 + }, + { + "epoch": 2.4092905776411158, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2848267555236816, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7396940588951111, + "num_tokens": 567572442.0, + "step": 21939 + }, + { + "epoch": 2.4094003953437295, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5431835651397705, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7222774028778076, + "num_tokens": 567597535.0, + "step": 21940 + }, + { + "epoch": 2.4095102130463433, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3713185787200928, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7053425908088684, + "num_tokens": 567625135.0, + "step": 21941 + }, + { + "epoch": 2.4096200307489566, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5374910831451416, + "learning_rate": 1e-06, + "loss": 0.7651, + "mean_token_accuracy": 0.749250054359436, + "num_tokens": 567647853.0, + "step": 21942 + }, + { + "epoch": 2.4097298484515703, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.227843761444092, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7198666930198669, + "num_tokens": 567676905.0, + "step": 21943 + }, + { + "epoch": 2.409839666154184, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5913045406341553, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7537643909454346, + "num_tokens": 567698642.0, + "step": 21944 + }, + { + "epoch": 2.409949483856798, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.15220046043396, + "learning_rate": 1e-06, + "loss": 0.8623, + "mean_token_accuracy": 0.7374300360679626, + "num_tokens": 567729030.0, + "step": 21945 + }, + { + "epoch": 2.4100593015594116, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6813182830810547, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7542151808738708, + "num_tokens": 567747945.0, + "step": 21946 + }, + { + "epoch": 2.410169119262025, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.355891227722168, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7286884188652039, + "num_tokens": 567773337.0, + "step": 21947 + }, + { + "epoch": 2.4102789369646387, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.480795383453369, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7156668901443481, + "num_tokens": 567798204.0, + "step": 21948 + }, + { + "epoch": 2.4103887546672524, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.399007558822632, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7217581868171692, + "num_tokens": 567824427.0, + "step": 21949 + }, + { + "epoch": 2.410498572369866, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.177614688873291, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7509616613388062, + "num_tokens": 567853385.0, + "step": 21950 + }, + { + "epoch": 2.4106083900724795, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7205183506011963, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7383804321289062, + "num_tokens": 567874455.0, + "step": 21951 + }, + { + "epoch": 2.4107182077750933, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1498069763183594, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7302790880203247, + "num_tokens": 567904011.0, + "step": 21952 + }, + { + "epoch": 2.410828025477707, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3767006397247314, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7130045890808105, + "num_tokens": 567929546.0, + "step": 21953 + }, + { + "epoch": 2.410937843180321, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3337960243225098, + "learning_rate": 1e-06, + "loss": 0.7422, + "mean_token_accuracy": 0.765943706035614, + "num_tokens": 567955058.0, + "step": 21954 + }, + { + "epoch": 2.411047660882934, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3382551670074463, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7437134385108948, + "num_tokens": 567982081.0, + "step": 21955 + }, + { + "epoch": 2.411157478585548, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3848390579223633, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7341107130050659, + "num_tokens": 568007640.0, + "step": 21956 + }, + { + "epoch": 2.4112672962881616, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4915223121643066, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7275545001029968, + "num_tokens": 568031630.0, + "step": 21957 + }, + { + "epoch": 2.4113771139907754, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3987622261047363, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7349379062652588, + "num_tokens": 568054579.0, + "step": 21958 + }, + { + "epoch": 2.411486931693389, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.481393337249756, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7494858503341675, + "num_tokens": 568078160.0, + "step": 21959 + }, + { + "epoch": 2.4115967493960024, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3941757678985596, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7255096435546875, + "num_tokens": 568102673.0, + "step": 21960 + }, + { + "epoch": 2.411706567098616, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.254952907562256, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7340809106826782, + "num_tokens": 568129707.0, + "step": 21961 + }, + { + "epoch": 2.41181638480123, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.136810302734375, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7360867261886597, + "num_tokens": 568159400.0, + "step": 21962 + }, + { + "epoch": 2.4119262025038437, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3768932819366455, + "learning_rate": 1e-06, + "loss": 0.905, + "mean_token_accuracy": 0.7243205308914185, + "num_tokens": 568184122.0, + "step": 21963 + }, + { + "epoch": 2.4120360202064575, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.466752529144287, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7202622294425964, + "num_tokens": 568207931.0, + "step": 21964 + }, + { + "epoch": 2.412145837909071, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.076964855194092, + "learning_rate": 1e-06, + "loss": 0.8553, + "mean_token_accuracy": 0.7317084074020386, + "num_tokens": 568239815.0, + "step": 21965 + }, + { + "epoch": 2.4122556556116845, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 1.9278101921081543, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7145676612854004, + "num_tokens": 568277237.0, + "step": 21966 + }, + { + "epoch": 2.4123654733142983, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1124343872070312, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7196961641311646, + "num_tokens": 568311920.0, + "step": 21967 + }, + { + "epoch": 2.412475291016912, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.682074546813965, + "learning_rate": 1e-06, + "loss": 0.671, + "mean_token_accuracy": 0.7732964158058167, + "num_tokens": 568330760.0, + "step": 21968 + }, + { + "epoch": 2.412585108719526, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4880764484405518, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7192685008049011, + "num_tokens": 568355507.0, + "step": 21969 + }, + { + "epoch": 2.412694926422139, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3164637088775635, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7221390604972839, + "num_tokens": 568382024.0, + "step": 21970 + }, + { + "epoch": 2.412804744124753, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5911033153533936, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7293024659156799, + "num_tokens": 568405147.0, + "step": 21971 + }, + { + "epoch": 2.4129145618273666, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.250713586807251, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7244118452072144, + "num_tokens": 568433274.0, + "step": 21972 + }, + { + "epoch": 2.4130243795299804, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4759981632232666, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7333434820175171, + "num_tokens": 568457178.0, + "step": 21973 + }, + { + "epoch": 2.4131341972325937, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0354199409484863, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7474508285522461, + "num_tokens": 568487821.0, + "step": 21974 + }, + { + "epoch": 2.4132440149352075, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5619945526123047, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7259202599525452, + "num_tokens": 568511891.0, + "step": 21975 + }, + { + "epoch": 2.413353832637821, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6253294944763184, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7308593988418579, + "num_tokens": 568533065.0, + "step": 21976 + }, + { + "epoch": 2.413463650340435, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.216921091079712, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.709989607334137, + "num_tokens": 568562977.0, + "step": 21977 + }, + { + "epoch": 2.4135734680430487, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2877681255340576, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7173610329627991, + "num_tokens": 568588649.0, + "step": 21978 + }, + { + "epoch": 2.413683285745662, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7654104232788086, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7323466539382935, + "num_tokens": 568609408.0, + "step": 21979 + }, + { + "epoch": 2.413793103448276, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2688686847686768, + "learning_rate": 1e-06, + "loss": 0.9231, + "mean_token_accuracy": 0.7129247188568115, + "num_tokens": 568635159.0, + "step": 21980 + }, + { + "epoch": 2.4139029211508896, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4820892810821533, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7075161933898926, + "num_tokens": 568662805.0, + "step": 21981 + }, + { + "epoch": 2.4140127388535033, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.472925901412964, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7419449090957642, + "num_tokens": 568686176.0, + "step": 21982 + }, + { + "epoch": 2.4141225565561166, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.760230779647827, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7216939330101013, + "num_tokens": 568706853.0, + "step": 21983 + }, + { + "epoch": 2.4142323742587304, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.182152509689331, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7181130051612854, + "num_tokens": 568736656.0, + "step": 21984 + }, + { + "epoch": 2.414342191961344, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5363097190856934, + "learning_rate": 1e-06, + "loss": 0.7375, + "mean_token_accuracy": 0.7624932527542114, + "num_tokens": 568758166.0, + "step": 21985 + }, + { + "epoch": 2.414452009663958, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.597435712814331, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7425779104232788, + "num_tokens": 568778611.0, + "step": 21986 + }, + { + "epoch": 2.4145618273665717, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7722060680389404, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7390674948692322, + "num_tokens": 568799349.0, + "step": 21987 + }, + { + "epoch": 2.414671645069185, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1268539428710938, + "learning_rate": 1e-06, + "loss": 0.9784, + "mean_token_accuracy": 0.7134028077125549, + "num_tokens": 568830941.0, + "step": 21988 + }, + { + "epoch": 2.4147814627717987, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0649051666259766, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7143450379371643, + "num_tokens": 568861571.0, + "step": 21989 + }, + { + "epoch": 2.4148912804744125, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.239227056503296, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7386754751205444, + "num_tokens": 568889511.0, + "step": 21990 + }, + { + "epoch": 2.4150010981770262, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.244378089904785, + "learning_rate": 1e-06, + "loss": 0.8001, + "mean_token_accuracy": 0.744901180267334, + "num_tokens": 568916393.0, + "step": 21991 + }, + { + "epoch": 2.41511091587964, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2593536376953125, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7227628231048584, + "num_tokens": 568943616.0, + "step": 21992 + }, + { + "epoch": 2.4152207335822533, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.607316732406616, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7438716888427734, + "num_tokens": 568966047.0, + "step": 21993 + }, + { + "epoch": 2.415330551284867, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5285279750823975, + "learning_rate": 1e-06, + "loss": 0.7437, + "mean_token_accuracy": 0.7574365139007568, + "num_tokens": 568987775.0, + "step": 21994 + }, + { + "epoch": 2.415440368987481, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1189863681793213, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7294126749038696, + "num_tokens": 569017181.0, + "step": 21995 + }, + { + "epoch": 2.4155501866900946, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2391364574432373, + "learning_rate": 1e-06, + "loss": 0.7745, + "mean_token_accuracy": 0.747116208076477, + "num_tokens": 569043843.0, + "step": 21996 + }, + { + "epoch": 2.4156600043927083, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4783287048339844, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7250678539276123, + "num_tokens": 569070037.0, + "step": 21997 + }, + { + "epoch": 2.4157698220953217, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7403488159179688, + "learning_rate": 1e-06, + "loss": 0.778, + "mean_token_accuracy": 0.7405840754508972, + "num_tokens": 569090111.0, + "step": 21998 + }, + { + "epoch": 2.4158796397979354, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 3.149975061416626, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7152380347251892, + "num_tokens": 569108485.0, + "step": 21999 + }, + { + "epoch": 2.415989457500549, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.437100410461426, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.713856041431427, + "num_tokens": 569134939.0, + "step": 22000 + }, + { + "epoch": 2.416099275203163, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.283202648162842, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7366593480110168, + "num_tokens": 569163020.0, + "step": 22001 + }, + { + "epoch": 2.4162090929057762, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4509387016296387, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7287982106208801, + "num_tokens": 569188176.0, + "step": 22002 + }, + { + "epoch": 2.41631891060839, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1720733642578125, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7424308657646179, + "num_tokens": 569216999.0, + "step": 22003 + }, + { + "epoch": 2.4164287283110037, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.077791452407837, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7337682843208313, + "num_tokens": 569247007.0, + "step": 22004 + }, + { + "epoch": 2.4165385460136175, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.8435306549072266, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7489509582519531, + "num_tokens": 569264811.0, + "step": 22005 + }, + { + "epoch": 2.416648363716231, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5370547771453857, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7251807451248169, + "num_tokens": 569289177.0, + "step": 22006 + }, + { + "epoch": 2.4167581814188446, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.086688280105591, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7177061438560486, + "num_tokens": 569320467.0, + "step": 22007 + }, + { + "epoch": 2.4168679991214583, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5764753818511963, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.727851152420044, + "num_tokens": 569344687.0, + "step": 22008 + }, + { + "epoch": 2.416977816824072, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0578527450561523, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7256965637207031, + "num_tokens": 569376539.0, + "step": 22009 + }, + { + "epoch": 2.417087634526686, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.423008680343628, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7488707304000854, + "num_tokens": 569399611.0, + "step": 22010 + }, + { + "epoch": 2.417197452229299, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5335850715637207, + "learning_rate": 1e-06, + "loss": 0.9373, + "mean_token_accuracy": 0.7095397710800171, + "num_tokens": 569425410.0, + "step": 22011 + }, + { + "epoch": 2.417307269931913, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5101330280303955, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.738061249256134, + "num_tokens": 569447821.0, + "step": 22012 + }, + { + "epoch": 2.4174170876345267, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.605984687805176, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7237229943275452, + "num_tokens": 569471760.0, + "step": 22013 + }, + { + "epoch": 2.4175269053371404, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.443108320236206, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7344960570335388, + "num_tokens": 569495548.0, + "step": 22014 + }, + { + "epoch": 2.417636723039754, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4627134799957275, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7295845746994019, + "num_tokens": 569518643.0, + "step": 22015 + }, + { + "epoch": 2.4177465407423675, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.775118350982666, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7376638650894165, + "num_tokens": 569538571.0, + "step": 22016 + }, + { + "epoch": 2.4178563584449813, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1068012714385986, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7212260365486145, + "num_tokens": 569569174.0, + "step": 22017 + }, + { + "epoch": 2.417966176147595, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.325942039489746, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7154543399810791, + "num_tokens": 569595496.0, + "step": 22018 + }, + { + "epoch": 2.4180759938502088, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.048121213912964, + "learning_rate": 1e-06, + "loss": 0.9124, + "mean_token_accuracy": 0.7138242721557617, + "num_tokens": 569629327.0, + "step": 22019 + }, + { + "epoch": 2.4181858115528225, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3270630836486816, + "learning_rate": 1e-06, + "loss": 0.7533, + "mean_token_accuracy": 0.7557640075683594, + "num_tokens": 569654413.0, + "step": 22020 + }, + { + "epoch": 2.418295629255436, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.572359561920166, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.7543359398841858, + "num_tokens": 569676924.0, + "step": 22021 + }, + { + "epoch": 2.4184054469580496, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.261457920074463, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7313632965087891, + "num_tokens": 569706225.0, + "step": 22022 + }, + { + "epoch": 2.4185152646606634, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.408080577850342, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7364070415496826, + "num_tokens": 569729987.0, + "step": 22023 + }, + { + "epoch": 2.418625082363277, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.616533041000366, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7178007364273071, + "num_tokens": 569751708.0, + "step": 22024 + }, + { + "epoch": 2.418734900065891, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.57877254486084, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7323347330093384, + "num_tokens": 569775611.0, + "step": 22025 + }, + { + "epoch": 2.418844717768504, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0403215885162354, + "learning_rate": 1e-06, + "loss": 1.0451, + "mean_token_accuracy": 0.6948897838592529, + "num_tokens": 569811089.0, + "step": 22026 + }, + { + "epoch": 2.418954535471118, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.332507610321045, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7269219160079956, + "num_tokens": 569836572.0, + "step": 22027 + }, + { + "epoch": 2.4190643531737317, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.801724910736084, + "learning_rate": 1e-06, + "loss": 0.8012, + "mean_token_accuracy": 0.7474334239959717, + "num_tokens": 569856953.0, + "step": 22028 + }, + { + "epoch": 2.4191741708763455, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.264981269836426, + "learning_rate": 1e-06, + "loss": 0.7746, + "mean_token_accuracy": 0.7493281364440918, + "num_tokens": 569883376.0, + "step": 22029 + }, + { + "epoch": 2.4192839885789588, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.489482879638672, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7278766632080078, + "num_tokens": 569907999.0, + "step": 22030 + }, + { + "epoch": 2.4193938062815725, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2007369995117188, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7272121906280518, + "num_tokens": 569938111.0, + "step": 22031 + }, + { + "epoch": 2.4195036239841863, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.249485492706299, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7163475155830383, + "num_tokens": 569966184.0, + "step": 22032 + }, + { + "epoch": 2.4196134416868, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.824352502822876, + "learning_rate": 1e-06, + "loss": 0.7184, + "mean_token_accuracy": 0.7668591141700745, + "num_tokens": 569985005.0, + "step": 22033 + }, + { + "epoch": 2.4197232593894134, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4210691452026367, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7241853475570679, + "num_tokens": 570010300.0, + "step": 22034 + }, + { + "epoch": 2.419833077092027, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4930334091186523, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.7406284213066101, + "num_tokens": 570032860.0, + "step": 22035 + }, + { + "epoch": 2.419942894794641, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.288299083709717, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.7484203577041626, + "num_tokens": 570061059.0, + "step": 22036 + }, + { + "epoch": 2.4200527124972546, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2996530532836914, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7387902736663818, + "num_tokens": 570087969.0, + "step": 22037 + }, + { + "epoch": 2.4201625301998684, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.294442892074585, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7087117433547974, + "num_tokens": 570115084.0, + "step": 22038 + }, + { + "epoch": 2.4202723479024817, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0825252532958984, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7292747497558594, + "num_tokens": 570143869.0, + "step": 22039 + }, + { + "epoch": 2.4203821656050954, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.129711866378784, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7265689969062805, + "num_tokens": 570173988.0, + "step": 22040 + }, + { + "epoch": 2.420491983307709, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2595834732055664, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.715272068977356, + "num_tokens": 570202500.0, + "step": 22041 + }, + { + "epoch": 2.420601801010323, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.616909980773926, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7307824492454529, + "num_tokens": 570226143.0, + "step": 22042 + }, + { + "epoch": 2.4207116187129367, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5885848999023438, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7438952922821045, + "num_tokens": 570246966.0, + "step": 22043 + }, + { + "epoch": 2.42082143641555, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.141465663909912, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.749457597732544, + "num_tokens": 570276122.0, + "step": 22044 + }, + { + "epoch": 2.420931254118164, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.489576578140259, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.725756049156189, + "num_tokens": 570299470.0, + "step": 22045 + }, + { + "epoch": 2.4210410718207775, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3288040161132812, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7334764003753662, + "num_tokens": 570326235.0, + "step": 22046 + }, + { + "epoch": 2.4211508895233913, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.63865327835083, + "learning_rate": 1e-06, + "loss": 0.7632, + "mean_token_accuracy": 0.7464518547058105, + "num_tokens": 570346722.0, + "step": 22047 + }, + { + "epoch": 2.421260707226005, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7448089122772217, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7169975638389587, + "num_tokens": 570368090.0, + "step": 22048 + }, + { + "epoch": 2.4213705249286184, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.626434564590454, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7375484108924866, + "num_tokens": 570391613.0, + "step": 22049 + }, + { + "epoch": 2.421480342631232, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6236307621002197, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7498499751091003, + "num_tokens": 570411032.0, + "step": 22050 + }, + { + "epoch": 2.421590160333846, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4273037910461426, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7239959239959717, + "num_tokens": 570435697.0, + "step": 22051 + }, + { + "epoch": 2.4216999780364596, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.645030975341797, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7245610952377319, + "num_tokens": 570457632.0, + "step": 22052 + }, + { + "epoch": 2.421809795739073, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.351679563522339, + "learning_rate": 1e-06, + "loss": 0.7792, + "mean_token_accuracy": 0.7510534524917603, + "num_tokens": 570482577.0, + "step": 22053 + }, + { + "epoch": 2.4219196134416867, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.154158115386963, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7202777862548828, + "num_tokens": 570515113.0, + "step": 22054 + }, + { + "epoch": 2.4220294311443005, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1643314361572266, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7211997509002686, + "num_tokens": 570546724.0, + "step": 22055 + }, + { + "epoch": 2.4221392488469142, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3291642665863037, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.736037015914917, + "num_tokens": 570572622.0, + "step": 22056 + }, + { + "epoch": 2.4222490665495275, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7252042293548584, + "learning_rate": 1e-06, + "loss": 0.7566, + "mean_token_accuracy": 0.753349781036377, + "num_tokens": 570592893.0, + "step": 22057 + }, + { + "epoch": 2.4223588842521413, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4297678470611572, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.72569340467453, + "num_tokens": 570617906.0, + "step": 22058 + }, + { + "epoch": 2.422468701954755, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6292996406555176, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.731414258480072, + "num_tokens": 570640269.0, + "step": 22059 + }, + { + "epoch": 2.422578519657369, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.7870686054229736, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.74654620885849, + "num_tokens": 570661526.0, + "step": 22060 + }, + { + "epoch": 2.4226883373599826, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.134371519088745, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7178167104721069, + "num_tokens": 570692755.0, + "step": 22061 + }, + { + "epoch": 2.422798155062596, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.592153310775757, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7362260818481445, + "num_tokens": 570715163.0, + "step": 22062 + }, + { + "epoch": 2.4229079727652096, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.099350690841675, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7122012376785278, + "num_tokens": 570746436.0, + "step": 22063 + }, + { + "epoch": 2.4230177904678234, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6722748279571533, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7206470966339111, + "num_tokens": 570768185.0, + "step": 22064 + }, + { + "epoch": 2.423127608170437, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.456976890563965, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7335445880889893, + "num_tokens": 570793782.0, + "step": 22065 + }, + { + "epoch": 2.423237425873051, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.324908494949341, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7491943836212158, + "num_tokens": 570817523.0, + "step": 22066 + }, + { + "epoch": 2.4233472435756642, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2252726554870605, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7254390716552734, + "num_tokens": 570847349.0, + "step": 22067 + }, + { + "epoch": 2.423457061278278, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2044951915740967, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7449826002120972, + "num_tokens": 570874899.0, + "step": 22068 + }, + { + "epoch": 2.4235668789808917, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.353748321533203, + "learning_rate": 1e-06, + "loss": 0.6697, + "mean_token_accuracy": 0.775567889213562, + "num_tokens": 570898312.0, + "step": 22069 + }, + { + "epoch": 2.4236766966835055, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4324746131896973, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7298851609230042, + "num_tokens": 570923526.0, + "step": 22070 + }, + { + "epoch": 2.4237865143861193, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1058287620544434, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7347239255905151, + "num_tokens": 570955857.0, + "step": 22071 + }, + { + "epoch": 2.4238963320887326, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3141791820526123, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7470278739929199, + "num_tokens": 570982408.0, + "step": 22072 + }, + { + "epoch": 2.4240061497913463, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2981982231140137, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7275865077972412, + "num_tokens": 571012542.0, + "step": 22073 + }, + { + "epoch": 2.42411596749396, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3614039421081543, + "learning_rate": 1e-06, + "loss": 0.9085, + "mean_token_accuracy": 0.7156593799591064, + "num_tokens": 571038782.0, + "step": 22074 + }, + { + "epoch": 2.424225785196574, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 1.995513916015625, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7491610050201416, + "num_tokens": 571071042.0, + "step": 22075 + }, + { + "epoch": 2.4243356028991876, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.338059663772583, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7289371490478516, + "num_tokens": 571097481.0, + "step": 22076 + }, + { + "epoch": 2.424445420601801, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3797450065612793, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7335947155952454, + "num_tokens": 571123289.0, + "step": 22077 + }, + { + "epoch": 2.4245552383044147, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.306018829345703, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7360067367553711, + "num_tokens": 571150458.0, + "step": 22078 + }, + { + "epoch": 2.4246650560070284, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4583349227905273, + "learning_rate": 1e-06, + "loss": 0.7727, + "mean_token_accuracy": 0.7560136318206787, + "num_tokens": 571174379.0, + "step": 22079 + }, + { + "epoch": 2.424774873709642, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.22859525680542, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7350717782974243, + "num_tokens": 571203623.0, + "step": 22080 + }, + { + "epoch": 2.4248846914122555, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0996220111846924, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.7088955640792847, + "num_tokens": 571236688.0, + "step": 22081 + }, + { + "epoch": 2.4249945091148692, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1033382415771484, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7215431332588196, + "num_tokens": 571266124.0, + "step": 22082 + }, + { + "epoch": 2.425104326817483, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.490318775177002, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7193952798843384, + "num_tokens": 571289727.0, + "step": 22083 + }, + { + "epoch": 2.4252141445200968, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.233858823776245, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7333051562309265, + "num_tokens": 571317444.0, + "step": 22084 + }, + { + "epoch": 2.42532396222271, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.312432050704956, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7216478586196899, + "num_tokens": 571346632.0, + "step": 22085 + }, + { + "epoch": 2.425433779925324, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3416836261749268, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7244468927383423, + "num_tokens": 571374381.0, + "step": 22086 + }, + { + "epoch": 2.4255435976279376, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2654876708984375, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.722809910774231, + "num_tokens": 571400742.0, + "step": 22087 + }, + { + "epoch": 2.4256534153305513, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.332641839981079, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7506293058395386, + "num_tokens": 571427455.0, + "step": 22088 + }, + { + "epoch": 2.425763233033165, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3408615589141846, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7190430164337158, + "num_tokens": 571452419.0, + "step": 22089 + }, + { + "epoch": 2.4258730507357784, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3184192180633545, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7305010557174683, + "num_tokens": 571479041.0, + "step": 22090 + }, + { + "epoch": 2.425982868438392, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.483978509902954, + "learning_rate": 1e-06, + "loss": 0.7962, + "mean_token_accuracy": 0.7456147074699402, + "num_tokens": 571501627.0, + "step": 22091 + }, + { + "epoch": 2.426092686141006, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.34101939201355, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7277731895446777, + "num_tokens": 571527922.0, + "step": 22092 + }, + { + "epoch": 2.4262025038436197, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.470944404602051, + "learning_rate": 1e-06, + "loss": 0.7888, + "mean_token_accuracy": 0.754697322845459, + "num_tokens": 571552399.0, + "step": 22093 + }, + { + "epoch": 2.4263123215462334, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4006388187408447, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7351201772689819, + "num_tokens": 571577848.0, + "step": 22094 + }, + { + "epoch": 2.4264221392488468, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1879305839538574, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7202835083007812, + "num_tokens": 571607660.0, + "step": 22095 + }, + { + "epoch": 2.4265319569514605, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3381056785583496, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7385618686676025, + "num_tokens": 571633207.0, + "step": 22096 + }, + { + "epoch": 2.4266417746540743, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2407310009002686, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7199657559394836, + "num_tokens": 571660053.0, + "step": 22097 + }, + { + "epoch": 2.426751592356688, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.188584804534912, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7330657243728638, + "num_tokens": 571689648.0, + "step": 22098 + }, + { + "epoch": 2.426861410059302, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3072452545166016, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7090707421302795, + "num_tokens": 571718070.0, + "step": 22099 + }, + { + "epoch": 2.426971227761915, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.0157535076141357, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7289581298828125, + "num_tokens": 571755031.0, + "step": 22100 + }, + { + "epoch": 2.427081045464529, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.3882527351379395, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7334141135215759, + "num_tokens": 571779782.0, + "step": 22101 + }, + { + "epoch": 2.4271908631671426, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.6699702739715576, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.7503477334976196, + "num_tokens": 571800827.0, + "step": 22102 + }, + { + "epoch": 2.4273006808697564, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.4354732036590576, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7251995801925659, + "num_tokens": 571823948.0, + "step": 22103 + }, + { + "epoch": 2.4274104985723697, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.813236713409424, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7478289604187012, + "num_tokens": 571844500.0, + "step": 22104 + }, + { + "epoch": 2.4275203162749834, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.338981866836548, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.725287675857544, + "num_tokens": 571871878.0, + "step": 22105 + }, + { + "epoch": 2.427630133977597, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2731223106384277, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7182329893112183, + "num_tokens": 571899440.0, + "step": 22106 + }, + { + "epoch": 2.427739951680211, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.252939224243164, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7334691286087036, + "num_tokens": 571930948.0, + "step": 22107 + }, + { + "epoch": 2.4278497693828243, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.1231048107147217, + "learning_rate": 1e-06, + "loss": 0.9194, + "mean_token_accuracy": 0.7225738763809204, + "num_tokens": 571960027.0, + "step": 22108 + }, + { + "epoch": 2.427959587085438, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.2998056411743164, + "learning_rate": 1e-06, + "loss": 0.9557, + "mean_token_accuracy": 0.699165940284729, + "num_tokens": 571989873.0, + "step": 22109 + }, + { + "epoch": 2.4280694047880518, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.325310468673706, + "learning_rate": 1e-06, + "loss": 0.856, + "mean_token_accuracy": 0.7266196012496948, + "num_tokens": 572017896.0, + "step": 22110 + }, + { + "epoch": 2.4281792224906655, + "ewc_loss": 1.430511474609375e-05, + "grad_norm": 2.5455360412597656, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.732412576675415, + "num_tokens": 572042412.0, + "step": 22111 + }, + { + "epoch": 2.4282890401932793, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.296078681945801, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7448577880859375, + "num_tokens": 572068675.0, + "step": 22112 + }, + { + "epoch": 2.4283988578958926, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.4813671112060547, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7322568297386169, + "num_tokens": 572092858.0, + "step": 22113 + }, + { + "epoch": 2.4285086755985064, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.2891249656677246, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7370595932006836, + "num_tokens": 572122554.0, + "step": 22114 + }, + { + "epoch": 2.42861849330112, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.549424886703491, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7227636575698853, + "num_tokens": 572145919.0, + "step": 22115 + }, + { + "epoch": 2.428728311003734, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.1989593505859375, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7227593064308167, + "num_tokens": 572174593.0, + "step": 22116 + }, + { + "epoch": 2.4288381287063476, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.7471046447753906, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.739660382270813, + "num_tokens": 572194188.0, + "step": 22117 + }, + { + "epoch": 2.428947946408961, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.753225326538086, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7378333806991577, + "num_tokens": 572213760.0, + "step": 22118 + }, + { + "epoch": 2.4290577641115747, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.3132991790771484, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.722176730632782, + "num_tokens": 572240846.0, + "step": 22119 + }, + { + "epoch": 2.4291675818141885, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.798969030380249, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7372720837593079, + "num_tokens": 572260410.0, + "step": 22120 + }, + { + "epoch": 2.429277399516802, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.2746851444244385, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7144261002540588, + "num_tokens": 572288503.0, + "step": 22121 + }, + { + "epoch": 2.429387217219416, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.3000707626342773, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7248501777648926, + "num_tokens": 572314762.0, + "step": 22122 + }, + { + "epoch": 2.4294970349220293, + "ewc_loss": 1.4424324035644531e-05, + "grad_norm": 2.4578254222869873, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7291626930236816, + "num_tokens": 572338449.0, + "step": 22123 + }, + { + "epoch": 2.429606852624643, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1738078594207764, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7274834513664246, + "num_tokens": 572371814.0, + "step": 22124 + }, + { + "epoch": 2.429716670327257, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3279545307159424, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7178853750228882, + "num_tokens": 572401330.0, + "step": 22125 + }, + { + "epoch": 2.4298264880298706, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1053202152252197, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7428852319717407, + "num_tokens": 572431993.0, + "step": 22126 + }, + { + "epoch": 2.4299363057324843, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.154060125350952, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7498867511749268, + "num_tokens": 572461968.0, + "step": 22127 + }, + { + "epoch": 2.4300461234350976, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.416241407394409, + "learning_rate": 1e-06, + "loss": 0.9422, + "mean_token_accuracy": 0.7134438753128052, + "num_tokens": 572487597.0, + "step": 22128 + }, + { + "epoch": 2.4301559411377114, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2679412364959717, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7396926879882812, + "num_tokens": 572513688.0, + "step": 22129 + }, + { + "epoch": 2.430265758840325, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4308061599731445, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.726629376411438, + "num_tokens": 572537892.0, + "step": 22130 + }, + { + "epoch": 2.430375576542939, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5155696868896484, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.74139404296875, + "num_tokens": 572562504.0, + "step": 22131 + }, + { + "epoch": 2.430485394245552, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5128560066223145, + "learning_rate": 1e-06, + "loss": 0.7062, + "mean_token_accuracy": 0.7669447064399719, + "num_tokens": 572585240.0, + "step": 22132 + }, + { + "epoch": 2.430595211948166, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.594491720199585, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7284595966339111, + "num_tokens": 572608644.0, + "step": 22133 + }, + { + "epoch": 2.4307050296507797, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.402845859527588, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7470067739486694, + "num_tokens": 572634180.0, + "step": 22134 + }, + { + "epoch": 2.4308148473533935, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.701028823852539, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7345404028892517, + "num_tokens": 572655918.0, + "step": 22135 + }, + { + "epoch": 2.430924665056007, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.221024990081787, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7412028312683105, + "num_tokens": 572683866.0, + "step": 22136 + }, + { + "epoch": 2.4310344827586206, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5795674324035645, + "learning_rate": 1e-06, + "loss": 0.7671, + "mean_token_accuracy": 0.7495664358139038, + "num_tokens": 572706081.0, + "step": 22137 + }, + { + "epoch": 2.4311443004612343, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.566556930541992, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7475082874298096, + "num_tokens": 572729135.0, + "step": 22138 + }, + { + "epoch": 2.431254118163848, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.542853593826294, + "learning_rate": 1e-06, + "loss": 0.7526, + "mean_token_accuracy": 0.7628320455551147, + "num_tokens": 572752702.0, + "step": 22139 + }, + { + "epoch": 2.431363935866462, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3255088329315186, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7419533133506775, + "num_tokens": 572778951.0, + "step": 22140 + }, + { + "epoch": 2.431473753569075, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.8070034980773926, + "learning_rate": 1e-06, + "loss": 0.7102, + "mean_token_accuracy": 0.7721620798110962, + "num_tokens": 572797376.0, + "step": 22141 + }, + { + "epoch": 2.431583571271689, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4141056537628174, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7271969318389893, + "num_tokens": 572821225.0, + "step": 22142 + }, + { + "epoch": 2.4316933889743026, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1972594261169434, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7455384731292725, + "num_tokens": 572851640.0, + "step": 22143 + }, + { + "epoch": 2.4318032066769164, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.413320779800415, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.722441554069519, + "num_tokens": 572876784.0, + "step": 22144 + }, + { + "epoch": 2.43191302437953, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.63077449798584, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7372597455978394, + "num_tokens": 572901633.0, + "step": 22145 + }, + { + "epoch": 2.4320228420821435, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3946900367736816, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7244766354560852, + "num_tokens": 572925554.0, + "step": 22146 + }, + { + "epoch": 2.4321326597847572, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.754572868347168, + "learning_rate": 1e-06, + "loss": 0.763, + "mean_token_accuracy": 0.7516993284225464, + "num_tokens": 572945689.0, + "step": 22147 + }, + { + "epoch": 2.432242477487371, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4272165298461914, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7334915399551392, + "num_tokens": 572969929.0, + "step": 22148 + }, + { + "epoch": 2.4323522951899847, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3926048278808594, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7480501532554626, + "num_tokens": 572996094.0, + "step": 22149 + }, + { + "epoch": 2.4324621128925985, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 3.0019192695617676, + "learning_rate": 1e-06, + "loss": 0.7779, + "mean_token_accuracy": 0.7457242608070374, + "num_tokens": 573013678.0, + "step": 22150 + }, + { + "epoch": 2.432571930595212, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.279696464538574, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7222942113876343, + "num_tokens": 573044076.0, + "step": 22151 + }, + { + "epoch": 2.4326817482978256, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.268660545349121, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7184195518493652, + "num_tokens": 573075911.0, + "step": 22152 + }, + { + "epoch": 2.4327915660004393, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.278000593185425, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7212165594100952, + "num_tokens": 573104044.0, + "step": 22153 + }, + { + "epoch": 2.432901383703053, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5020015239715576, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.727296769618988, + "num_tokens": 573128751.0, + "step": 22154 + }, + { + "epoch": 2.4330112014056664, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0444836616516113, + "learning_rate": 1e-06, + "loss": 0.7602, + "mean_token_accuracy": 0.75341796875, + "num_tokens": 573161004.0, + "step": 22155 + }, + { + "epoch": 2.43312101910828, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.513500213623047, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7462868094444275, + "num_tokens": 573183890.0, + "step": 22156 + }, + { + "epoch": 2.433230836810894, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.13489031791687, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7095077037811279, + "num_tokens": 573214899.0, + "step": 22157 + }, + { + "epoch": 2.4333406545135077, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2108099460601807, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7264058589935303, + "num_tokens": 573244349.0, + "step": 22158 + }, + { + "epoch": 2.4334504722161214, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.541398048400879, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7348170280456543, + "num_tokens": 573266804.0, + "step": 22159 + }, + { + "epoch": 2.4335602899187347, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.800788164138794, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7372948527336121, + "num_tokens": 573285840.0, + "step": 22160 + }, + { + "epoch": 2.4336701076213485, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.187896490097046, + "learning_rate": 1e-06, + "loss": 0.9053, + "mean_token_accuracy": 0.724958062171936, + "num_tokens": 573313675.0, + "step": 22161 + }, + { + "epoch": 2.4337799253239623, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2593512535095215, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7313134074211121, + "num_tokens": 573341634.0, + "step": 22162 + }, + { + "epoch": 2.433889743026576, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.142383098602295, + "learning_rate": 1e-06, + "loss": 0.9137, + "mean_token_accuracy": 0.7139146327972412, + "num_tokens": 573372386.0, + "step": 22163 + }, + { + "epoch": 2.4339995607291893, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5735890865325928, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7478896379470825, + "num_tokens": 573393531.0, + "step": 22164 + }, + { + "epoch": 2.434109378431803, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3794987201690674, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7392682433128357, + "num_tokens": 573421354.0, + "step": 22165 + }, + { + "epoch": 2.434219196134417, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.161017894744873, + "learning_rate": 1e-06, + "loss": 0.9338, + "mean_token_accuracy": 0.7112470269203186, + "num_tokens": 573450837.0, + "step": 22166 + }, + { + "epoch": 2.4343290138370306, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.647461414337158, + "learning_rate": 1e-06, + "loss": 0.8097, + "mean_token_accuracy": 0.7421955466270447, + "num_tokens": 573472620.0, + "step": 22167 + }, + { + "epoch": 2.4344388315396444, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.328888416290283, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7150413393974304, + "num_tokens": 573499618.0, + "step": 22168 + }, + { + "epoch": 2.4345486492422577, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4284963607788086, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7402629256248474, + "num_tokens": 573525270.0, + "step": 22169 + }, + { + "epoch": 2.4346584669448714, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1963412761688232, + "learning_rate": 1e-06, + "loss": 0.9716, + "mean_token_accuracy": 0.6950024962425232, + "num_tokens": 573558375.0, + "step": 22170 + }, + { + "epoch": 2.434768284647485, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.553013801574707, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7306821942329407, + "num_tokens": 573580819.0, + "step": 22171 + }, + { + "epoch": 2.434878102350099, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5930144786834717, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7207481861114502, + "num_tokens": 573603032.0, + "step": 22172 + }, + { + "epoch": 2.4349879200527127, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.550992727279663, + "learning_rate": 1e-06, + "loss": 0.7783, + "mean_token_accuracy": 0.7531672716140747, + "num_tokens": 573624413.0, + "step": 22173 + }, + { + "epoch": 2.435097737755326, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.488701581954956, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7375873327255249, + "num_tokens": 573647899.0, + "step": 22174 + }, + { + "epoch": 2.4352075554579398, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6322224140167236, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.7085008025169373, + "num_tokens": 573672654.0, + "step": 22175 + }, + { + "epoch": 2.4353173731605535, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.698880672454834, + "learning_rate": 1e-06, + "loss": 0.7385, + "mean_token_accuracy": 0.7650266885757446, + "num_tokens": 573692131.0, + "step": 22176 + }, + { + "epoch": 2.4354271908631673, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.9408624172210693, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7290250062942505, + "num_tokens": 573712666.0, + "step": 22177 + }, + { + "epoch": 2.435537008565781, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.7035300731658936, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7545697689056396, + "num_tokens": 573732616.0, + "step": 22178 + }, + { + "epoch": 2.4356468262683943, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.436319589614868, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7322677969932556, + "num_tokens": 573755985.0, + "step": 22179 + }, + { + "epoch": 2.435756643971008, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3746132850646973, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7141500115394592, + "num_tokens": 573781635.0, + "step": 22180 + }, + { + "epoch": 2.435866461673622, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5676803588867188, + "learning_rate": 1e-06, + "loss": 0.7676, + "mean_token_accuracy": 0.7563906908035278, + "num_tokens": 573803827.0, + "step": 22181 + }, + { + "epoch": 2.4359762793762356, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.464398145675659, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7478938102722168, + "num_tokens": 573828707.0, + "step": 22182 + }, + { + "epoch": 2.436086097078849, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0582194328308105, + "learning_rate": 1e-06, + "loss": 0.776, + "mean_token_accuracy": 0.7510372400283813, + "num_tokens": 573858685.0, + "step": 22183 + }, + { + "epoch": 2.4361959147814627, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.323988914489746, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7306873202323914, + "num_tokens": 573885476.0, + "step": 22184 + }, + { + "epoch": 2.4363057324840764, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.54887056350708, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7298026084899902, + "num_tokens": 573910411.0, + "step": 22185 + }, + { + "epoch": 2.43641555018669, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.179225444793701, + "learning_rate": 1e-06, + "loss": 0.8315, + "mean_token_accuracy": 0.7397466897964478, + "num_tokens": 573940680.0, + "step": 22186 + }, + { + "epoch": 2.4365253678893035, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 1.983120083808899, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7445228099822998, + "num_tokens": 573974782.0, + "step": 22187 + }, + { + "epoch": 2.4366351855919173, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3178341388702393, + "learning_rate": 1e-06, + "loss": 0.9709, + "mean_token_accuracy": 0.7014206647872925, + "num_tokens": 574001647.0, + "step": 22188 + }, + { + "epoch": 2.436745003294531, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3792996406555176, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7249056696891785, + "num_tokens": 574027374.0, + "step": 22189 + }, + { + "epoch": 2.436854820997145, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.7893757820129395, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7389310598373413, + "num_tokens": 574047848.0, + "step": 22190 + }, + { + "epoch": 2.4369646386997585, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4422554969787598, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7442045211791992, + "num_tokens": 574071840.0, + "step": 22191 + }, + { + "epoch": 2.437074456402372, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.474677085876465, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7225127220153809, + "num_tokens": 574096632.0, + "step": 22192 + }, + { + "epoch": 2.4371842741049856, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.461587905883789, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.726130485534668, + "num_tokens": 574121632.0, + "step": 22193 + }, + { + "epoch": 2.4372940918075994, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5774989128112793, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7413231134414673, + "num_tokens": 574143569.0, + "step": 22194 + }, + { + "epoch": 2.437403909510213, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6657063961029053, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7459058165550232, + "num_tokens": 574165403.0, + "step": 22195 + }, + { + "epoch": 2.437513727212827, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.282329559326172, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7218524217605591, + "num_tokens": 574192232.0, + "step": 22196 + }, + { + "epoch": 2.43762354491544, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2911365032196045, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7391249537467957, + "num_tokens": 574218583.0, + "step": 22197 + }, + { + "epoch": 2.437733362618054, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.204096794128418, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7342336177825928, + "num_tokens": 574247284.0, + "step": 22198 + }, + { + "epoch": 2.4378431803206677, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3494386672973633, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7317680716514587, + "num_tokens": 574272439.0, + "step": 22199 + }, + { + "epoch": 2.4379529980232815, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.311655044555664, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7391696572303772, + "num_tokens": 574300708.0, + "step": 22200 + }, + { + "epoch": 2.4380628157258952, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1830663681030273, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7276663780212402, + "num_tokens": 574329878.0, + "step": 22201 + }, + { + "epoch": 2.4381726334285085, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0860564708709717, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7282241582870483, + "num_tokens": 574362905.0, + "step": 22202 + }, + { + "epoch": 2.4382824511311223, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3176345825195312, + "learning_rate": 1e-06, + "loss": 0.9793, + "mean_token_accuracy": 0.6991285085678101, + "num_tokens": 574389659.0, + "step": 22203 + }, + { + "epoch": 2.438392268833736, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.509507179260254, + "learning_rate": 1e-06, + "loss": 0.87, + "mean_token_accuracy": 0.7238136529922485, + "num_tokens": 574415611.0, + "step": 22204 + }, + { + "epoch": 2.43850208653635, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2038192749023438, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7163088321685791, + "num_tokens": 574444974.0, + "step": 22205 + }, + { + "epoch": 2.4386119042389636, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.497150421142578, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7472327947616577, + "num_tokens": 574467654.0, + "step": 22206 + }, + { + "epoch": 2.438721721941577, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.8715150356292725, + "learning_rate": 1e-06, + "loss": 0.7717, + "mean_token_accuracy": 0.7515478134155273, + "num_tokens": 574487929.0, + "step": 22207 + }, + { + "epoch": 2.4388315396441906, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6561379432678223, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7348741292953491, + "num_tokens": 574510732.0, + "step": 22208 + }, + { + "epoch": 2.4389413573468044, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.399852752685547, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7271997928619385, + "num_tokens": 574536138.0, + "step": 22209 + }, + { + "epoch": 2.439051175049418, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.225954532623291, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7423694133758545, + "num_tokens": 574563499.0, + "step": 22210 + }, + { + "epoch": 2.4391609927520315, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.335803985595703, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7185885906219482, + "num_tokens": 574589829.0, + "step": 22211 + }, + { + "epoch": 2.439270810454645, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6543397903442383, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.738349199295044, + "num_tokens": 574611065.0, + "step": 22212 + }, + { + "epoch": 2.439380628157259, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.308992862701416, + "learning_rate": 1e-06, + "loss": 0.9481, + "mean_token_accuracy": 0.7130606174468994, + "num_tokens": 574637669.0, + "step": 22213 + }, + { + "epoch": 2.4394904458598727, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.8783979415893555, + "learning_rate": 1e-06, + "loss": 0.7665, + "mean_token_accuracy": 0.7564869523048401, + "num_tokens": 574656869.0, + "step": 22214 + }, + { + "epoch": 2.439600263562486, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3122994899749756, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.731390655040741, + "num_tokens": 574684604.0, + "step": 22215 + }, + { + "epoch": 2.4397100812651, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5178534984588623, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7157516479492188, + "num_tokens": 574708139.0, + "step": 22216 + }, + { + "epoch": 2.4398198989677136, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5038299560546875, + "learning_rate": 1e-06, + "loss": 0.7235, + "mean_token_accuracy": 0.7649279832839966, + "num_tokens": 574732374.0, + "step": 22217 + }, + { + "epoch": 2.4399297166703273, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2737348079681396, + "learning_rate": 1e-06, + "loss": 0.9727, + "mean_token_accuracy": 0.7066400647163391, + "num_tokens": 574760582.0, + "step": 22218 + }, + { + "epoch": 2.440039534372941, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3911874294281006, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7353600263595581, + "num_tokens": 574786011.0, + "step": 22219 + }, + { + "epoch": 2.4401493520755544, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4002630710601807, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7281438112258911, + "num_tokens": 574812478.0, + "step": 22220 + }, + { + "epoch": 2.440259169778168, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.2531564235687256, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7388193011283875, + "num_tokens": 574838868.0, + "step": 22221 + }, + { + "epoch": 2.440368987480782, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.213714838027954, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.713064432144165, + "num_tokens": 574867871.0, + "step": 22222 + }, + { + "epoch": 2.4404788051833957, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4283032417297363, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7313928604125977, + "num_tokens": 574890716.0, + "step": 22223 + }, + { + "epoch": 2.4405886228860094, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3380801677703857, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7317965030670166, + "num_tokens": 574916637.0, + "step": 22224 + }, + { + "epoch": 2.4406984405886227, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.395421028137207, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7354920506477356, + "num_tokens": 574943381.0, + "step": 22225 + }, + { + "epoch": 2.4408082582912365, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.196743965148926, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7261984348297119, + "num_tokens": 574972351.0, + "step": 22226 + }, + { + "epoch": 2.4409180759938502, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.188410520553589, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7407324314117432, + "num_tokens": 575000134.0, + "step": 22227 + }, + { + "epoch": 2.441027893696464, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.011864423751831, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7316350340843201, + "num_tokens": 575035356.0, + "step": 22228 + }, + { + "epoch": 2.4411377113990778, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6864190101623535, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7351812124252319, + "num_tokens": 575057505.0, + "step": 22229 + }, + { + "epoch": 2.441247529101691, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.401090621948242, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.7443919777870178, + "num_tokens": 575081014.0, + "step": 22230 + }, + { + "epoch": 2.441357346804305, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0283052921295166, + "learning_rate": 1e-06, + "loss": 0.9747, + "mean_token_accuracy": 0.7022464275360107, + "num_tokens": 575115581.0, + "step": 22231 + }, + { + "epoch": 2.4414671645069186, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4832913875579834, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7235790491104126, + "num_tokens": 575139436.0, + "step": 22232 + }, + { + "epoch": 2.4415769822095323, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.137096881866455, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7105304002761841, + "num_tokens": 575171593.0, + "step": 22233 + }, + { + "epoch": 2.4416867999121457, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3733348846435547, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7446199655532837, + "num_tokens": 575197302.0, + "step": 22234 + }, + { + "epoch": 2.4417966176147594, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.270613431930542, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7260894775390625, + "num_tokens": 575225667.0, + "step": 22235 + }, + { + "epoch": 2.441906435317373, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3978683948516846, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7333790063858032, + "num_tokens": 575250392.0, + "step": 22236 + }, + { + "epoch": 2.442016253019987, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.6968400478363037, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7315568923950195, + "num_tokens": 575270867.0, + "step": 22237 + }, + { + "epoch": 2.4421260707226002, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.0367836952209473, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7139032483100891, + "num_tokens": 575303943.0, + "step": 22238 + }, + { + "epoch": 2.442235888425214, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.633430004119873, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7335343360900879, + "num_tokens": 575325586.0, + "step": 22239 + }, + { + "epoch": 2.4423457061278278, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.769061803817749, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.739192008972168, + "num_tokens": 575345010.0, + "step": 22240 + }, + { + "epoch": 2.4424555238304415, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.319263458251953, + "learning_rate": 1e-06, + "loss": 0.9372, + "mean_token_accuracy": 0.7065044641494751, + "num_tokens": 575371827.0, + "step": 22241 + }, + { + "epoch": 2.4425653415330553, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.396820068359375, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7450193166732788, + "num_tokens": 575397851.0, + "step": 22242 + }, + { + "epoch": 2.4426751592356686, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.530225992202759, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7276537418365479, + "num_tokens": 575420803.0, + "step": 22243 + }, + { + "epoch": 2.4427849769382823, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.645059823989868, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.747129499912262, + "num_tokens": 575441226.0, + "step": 22244 + }, + { + "epoch": 2.442894794640896, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.334972858428955, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7315396666526794, + "num_tokens": 575467323.0, + "step": 22245 + }, + { + "epoch": 2.44300461234351, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3837227821350098, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7311358451843262, + "num_tokens": 575492186.0, + "step": 22246 + }, + { + "epoch": 2.4431144300461236, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.7131919860839844, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7391551733016968, + "num_tokens": 575512859.0, + "step": 22247 + }, + { + "epoch": 2.443224247748737, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.510777711868286, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.744342029094696, + "num_tokens": 575534950.0, + "step": 22248 + }, + { + "epoch": 2.4433340654513507, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3934566974639893, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7464137077331543, + "num_tokens": 575560010.0, + "step": 22249 + }, + { + "epoch": 2.4434438831539644, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.361828327178955, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7136410474777222, + "num_tokens": 575587780.0, + "step": 22250 + }, + { + "epoch": 2.443553700856578, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4136173725128174, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7221874594688416, + "num_tokens": 575613729.0, + "step": 22251 + }, + { + "epoch": 2.443663518559192, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.374656915664673, + "learning_rate": 1e-06, + "loss": 0.9751, + "mean_token_accuracy": 0.6960771679878235, + "num_tokens": 575641897.0, + "step": 22252 + }, + { + "epoch": 2.4437733362618053, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.525825262069702, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7383817434310913, + "num_tokens": 575663939.0, + "step": 22253 + }, + { + "epoch": 2.443883153964419, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.331779718399048, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7234766483306885, + "num_tokens": 575693398.0, + "step": 22254 + }, + { + "epoch": 2.4439929716670328, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.488151788711548, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7347004413604736, + "num_tokens": 575716003.0, + "step": 22255 + }, + { + "epoch": 2.4441027893696465, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.5782980918884277, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7338299751281738, + "num_tokens": 575737645.0, + "step": 22256 + }, + { + "epoch": 2.4442126070722603, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3396706581115723, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7364470958709717, + "num_tokens": 575764482.0, + "step": 22257 + }, + { + "epoch": 2.4443224247748736, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.650740623474121, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7225643396377563, + "num_tokens": 575785339.0, + "step": 22258 + }, + { + "epoch": 2.4444322424774874, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.3737735748291016, + "learning_rate": 1e-06, + "loss": 0.766, + "mean_token_accuracy": 0.7612987756729126, + "num_tokens": 575810353.0, + "step": 22259 + }, + { + "epoch": 2.444542060180101, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.4328877925872803, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7257546186447144, + "num_tokens": 575837686.0, + "step": 22260 + }, + { + "epoch": 2.444651877882715, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.063840866088867, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7068524360656738, + "num_tokens": 575873131.0, + "step": 22261 + }, + { + "epoch": 2.444761695585328, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.189668655395508, + "learning_rate": 1e-06, + "loss": 0.9243, + "mean_token_accuracy": 0.7109498977661133, + "num_tokens": 575901282.0, + "step": 22262 + }, + { + "epoch": 2.444871513287942, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.604222297668457, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7355076670646667, + "num_tokens": 575923327.0, + "step": 22263 + }, + { + "epoch": 2.4449813309905557, + "ewc_loss": 1.4543533325195312e-05, + "grad_norm": 2.1345415115356445, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7183459401130676, + "num_tokens": 575955089.0, + "step": 22264 + }, + { + "epoch": 2.4450911486931695, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.354886770248413, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7104736566543579, + "num_tokens": 575982716.0, + "step": 22265 + }, + { + "epoch": 2.4452009663957828, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2352941036224365, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.7059540748596191, + "num_tokens": 576012552.0, + "step": 22266 + }, + { + "epoch": 2.4453107840983965, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6641385555267334, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7418807744979858, + "num_tokens": 576033064.0, + "step": 22267 + }, + { + "epoch": 2.4454206018010103, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1429779529571533, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7142271995544434, + "num_tokens": 576063360.0, + "step": 22268 + }, + { + "epoch": 2.445530419503624, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.280282735824585, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7481927275657654, + "num_tokens": 576088138.0, + "step": 22269 + }, + { + "epoch": 2.445640237206238, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.0345394611358643, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.7082761526107788, + "num_tokens": 576120804.0, + "step": 22270 + }, + { + "epoch": 2.445750054908851, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2818174362182617, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7320181131362915, + "num_tokens": 576146306.0, + "step": 22271 + }, + { + "epoch": 2.445859872611465, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.488126516342163, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7247425317764282, + "num_tokens": 576171014.0, + "step": 22272 + }, + { + "epoch": 2.4459696903140786, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.355776786804199, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7213423252105713, + "num_tokens": 576197866.0, + "step": 22273 + }, + { + "epoch": 2.4460795080166924, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.385995864868164, + "learning_rate": 1e-06, + "loss": 0.9619, + "mean_token_accuracy": 0.6998420357704163, + "num_tokens": 576225642.0, + "step": 22274 + }, + { + "epoch": 2.446189325719306, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2404370307922363, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7338122129440308, + "num_tokens": 576253010.0, + "step": 22275 + }, + { + "epoch": 2.4462991434219195, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3192026615142822, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7238158583641052, + "num_tokens": 576279767.0, + "step": 22276 + }, + { + "epoch": 2.446408961124533, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.181100606918335, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7119449377059937, + "num_tokens": 576310196.0, + "step": 22277 + }, + { + "epoch": 2.446518778827147, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.387352228164673, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7061260938644409, + "num_tokens": 576336773.0, + "step": 22278 + }, + { + "epoch": 2.4466285965297607, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6332478523254395, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7346314191818237, + "num_tokens": 576358569.0, + "step": 22279 + }, + { + "epoch": 2.4467384142323745, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2874107360839844, + "learning_rate": 1e-06, + "loss": 0.8354, + "mean_token_accuracy": 0.7351740002632141, + "num_tokens": 576386669.0, + "step": 22280 + }, + { + "epoch": 2.446848231934988, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5002222061157227, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.740993320941925, + "num_tokens": 576409761.0, + "step": 22281 + }, + { + "epoch": 2.4469580496376016, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6392428874969482, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7399152517318726, + "num_tokens": 576432019.0, + "step": 22282 + }, + { + "epoch": 2.4470678673402153, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.381690502166748, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7317452430725098, + "num_tokens": 576459040.0, + "step": 22283 + }, + { + "epoch": 2.447177685042829, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3438568115234375, + "learning_rate": 1e-06, + "loss": 0.7812, + "mean_token_accuracy": 0.7503620386123657, + "num_tokens": 576485630.0, + "step": 22284 + }, + { + "epoch": 2.4472875027454424, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4106390476226807, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7453470230102539, + "num_tokens": 576510185.0, + "step": 22285 + }, + { + "epoch": 2.447397320448056, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6006569862365723, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7295920848846436, + "num_tokens": 576533097.0, + "step": 22286 + }, + { + "epoch": 2.44750713815067, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.896341562271118, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.7359172105789185, + "num_tokens": 576552424.0, + "step": 22287 + }, + { + "epoch": 2.4476169558532836, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6177070140838623, + "learning_rate": 1e-06, + "loss": 0.7793, + "mean_token_accuracy": 0.7508257031440735, + "num_tokens": 576572851.0, + "step": 22288 + }, + { + "epoch": 2.447726773555897, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2898716926574707, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7385222911834717, + "num_tokens": 576599655.0, + "step": 22289 + }, + { + "epoch": 2.4478365912585107, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1226956844329834, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7306129932403564, + "num_tokens": 576628544.0, + "step": 22290 + }, + { + "epoch": 2.4479464089611245, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.169750928878784, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7435052394866943, + "num_tokens": 576657072.0, + "step": 22291 + }, + { + "epoch": 2.4480562266637382, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3609676361083984, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7285909652709961, + "num_tokens": 576683784.0, + "step": 22292 + }, + { + "epoch": 2.448166044366352, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5059404373168945, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7439256310462952, + "num_tokens": 576706149.0, + "step": 22293 + }, + { + "epoch": 2.4482758620689653, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5493552684783936, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7296517491340637, + "num_tokens": 576727558.0, + "step": 22294 + }, + { + "epoch": 2.448385679771579, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.232306718826294, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7329256534576416, + "num_tokens": 576754030.0, + "step": 22295 + }, + { + "epoch": 2.448495497474193, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.288611888885498, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7318693399429321, + "num_tokens": 576782912.0, + "step": 22296 + }, + { + "epoch": 2.4486053151768066, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.355713367462158, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7326956987380981, + "num_tokens": 576807119.0, + "step": 22297 + }, + { + "epoch": 2.4487151328794203, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5071299076080322, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7283686399459839, + "num_tokens": 576830037.0, + "step": 22298 + }, + { + "epoch": 2.4488249505820336, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5449979305267334, + "learning_rate": 1e-06, + "loss": 0.7214, + "mean_token_accuracy": 0.768271803855896, + "num_tokens": 576851476.0, + "step": 22299 + }, + { + "epoch": 2.4489347682846474, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5622899532318115, + "learning_rate": 1e-06, + "loss": 0.9331, + "mean_token_accuracy": 0.7189767956733704, + "num_tokens": 576875256.0, + "step": 22300 + }, + { + "epoch": 2.449044585987261, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2206268310546875, + "learning_rate": 1e-06, + "loss": 0.8806, + "mean_token_accuracy": 0.7279945611953735, + "num_tokens": 576905309.0, + "step": 22301 + }, + { + "epoch": 2.449154403689875, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4192049503326416, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7260589003562927, + "num_tokens": 576930590.0, + "step": 22302 + }, + { + "epoch": 2.4492642213924887, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6563894748687744, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7358559370040894, + "num_tokens": 576952543.0, + "step": 22303 + }, + { + "epoch": 2.449374039095102, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4050850868225098, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7222467064857483, + "num_tokens": 576978843.0, + "step": 22304 + }, + { + "epoch": 2.4494838567977157, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5136046409606934, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7247867584228516, + "num_tokens": 577004525.0, + "step": 22305 + }, + { + "epoch": 2.4495936745003295, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.407559394836426, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7339751720428467, + "num_tokens": 577028989.0, + "step": 22306 + }, + { + "epoch": 2.4497034922029433, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.618253707885742, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7514464259147644, + "num_tokens": 577049797.0, + "step": 22307 + }, + { + "epoch": 2.449813309905557, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.564969301223755, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.735653281211853, + "num_tokens": 577072989.0, + "step": 22308 + }, + { + "epoch": 2.4499231276081703, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.356936454772949, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7298166751861572, + "num_tokens": 577097721.0, + "step": 22309 + }, + { + "epoch": 2.450032945310784, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2133002281188965, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7093640565872192, + "num_tokens": 577125235.0, + "step": 22310 + }, + { + "epoch": 2.450142763013398, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.571413516998291, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7503998279571533, + "num_tokens": 577146621.0, + "step": 22311 + }, + { + "epoch": 2.4502525807160116, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.831022262573242, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.752190113067627, + "num_tokens": 577165943.0, + "step": 22312 + }, + { + "epoch": 2.450362398418625, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.617075204849243, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7227322459220886, + "num_tokens": 577187725.0, + "step": 22313 + }, + { + "epoch": 2.4504722161212387, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3277361392974854, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7407646179199219, + "num_tokens": 577214425.0, + "step": 22314 + }, + { + "epoch": 2.4505820338238524, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3055849075317383, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7365668416023254, + "num_tokens": 577241342.0, + "step": 22315 + }, + { + "epoch": 2.450691851526466, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4352591037750244, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7451385259628296, + "num_tokens": 577264693.0, + "step": 22316 + }, + { + "epoch": 2.4508016692290795, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.224644184112549, + "learning_rate": 1e-06, + "loss": 0.7976, + "mean_token_accuracy": 0.7381811141967773, + "num_tokens": 577293109.0, + "step": 22317 + }, + { + "epoch": 2.4509114869316933, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.252307415008545, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7339829802513123, + "num_tokens": 577322748.0, + "step": 22318 + }, + { + "epoch": 2.451021304634307, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.441880464553833, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7309308052062988, + "num_tokens": 577346093.0, + "step": 22319 + }, + { + "epoch": 2.4511311223369208, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1661734580993652, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.7336732745170593, + "num_tokens": 577376360.0, + "step": 22320 + }, + { + "epoch": 2.4512409400395345, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.223966121673584, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7237961292266846, + "num_tokens": 577406450.0, + "step": 22321 + }, + { + "epoch": 2.451350757742148, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.657275438308716, + "learning_rate": 1e-06, + "loss": 0.893, + "mean_token_accuracy": 0.7185570597648621, + "num_tokens": 577428068.0, + "step": 22322 + }, + { + "epoch": 2.4514605754447616, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.47090482711792, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7401634454727173, + "num_tokens": 577451745.0, + "step": 22323 + }, + { + "epoch": 2.4515703931473753, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6480600833892822, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7132828831672668, + "num_tokens": 577473353.0, + "step": 22324 + }, + { + "epoch": 2.451680210849989, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 1.9326460361480713, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7252427935600281, + "num_tokens": 577511202.0, + "step": 22325 + }, + { + "epoch": 2.451790028552603, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2226603031158447, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7439440488815308, + "num_tokens": 577540374.0, + "step": 22326 + }, + { + "epoch": 2.451899846255216, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1613709926605225, + "learning_rate": 1e-06, + "loss": 0.9381, + "mean_token_accuracy": 0.7075732946395874, + "num_tokens": 577571983.0, + "step": 22327 + }, + { + "epoch": 2.45200966395783, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5383756160736084, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.724166750907898, + "num_tokens": 577596579.0, + "step": 22328 + }, + { + "epoch": 2.4521194816604437, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.173473596572876, + "learning_rate": 1e-06, + "loss": 0.7423, + "mean_token_accuracy": 0.7513614892959595, + "num_tokens": 577626159.0, + "step": 22329 + }, + { + "epoch": 2.4522292993630574, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2759156227111816, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7208225727081299, + "num_tokens": 577653838.0, + "step": 22330 + }, + { + "epoch": 2.452339117065671, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2579023838043213, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7254263758659363, + "num_tokens": 577681468.0, + "step": 22331 + }, + { + "epoch": 2.4524489347682845, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.220120668411255, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7270428538322449, + "num_tokens": 577711263.0, + "step": 22332 + }, + { + "epoch": 2.4525587524708983, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.462132692337036, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7313960790634155, + "num_tokens": 577734729.0, + "step": 22333 + }, + { + "epoch": 2.452668570173512, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.680654525756836, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7415882349014282, + "num_tokens": 577756749.0, + "step": 22334 + }, + { + "epoch": 2.452778387876126, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2038917541503906, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7275262475013733, + "num_tokens": 577785506.0, + "step": 22335 + }, + { + "epoch": 2.452888205578739, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4548721313476562, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7330494523048401, + "num_tokens": 577808889.0, + "step": 22336 + }, + { + "epoch": 2.452998023281353, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.310757637023926, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7174531817436218, + "num_tokens": 577837719.0, + "step": 22337 + }, + { + "epoch": 2.4531078409839666, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.36625075340271, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7361178398132324, + "num_tokens": 577863447.0, + "step": 22338 + }, + { + "epoch": 2.4532176586865804, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6749918460845947, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.728769838809967, + "num_tokens": 577883916.0, + "step": 22339 + }, + { + "epoch": 2.453327476389194, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.37374210357666, + "learning_rate": 1e-06, + "loss": 0.9204, + "mean_token_accuracy": 0.7128830552101135, + "num_tokens": 577910642.0, + "step": 22340 + }, + { + "epoch": 2.4534372940918074, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.487837314605713, + "learning_rate": 1e-06, + "loss": 0.7401, + "mean_token_accuracy": 0.7580268383026123, + "num_tokens": 577934895.0, + "step": 22341 + }, + { + "epoch": 2.453547111794421, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.409548282623291, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7276689410209656, + "num_tokens": 577960405.0, + "step": 22342 + }, + { + "epoch": 2.453656929497035, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1139559745788574, + "learning_rate": 1e-06, + "loss": 0.9813, + "mean_token_accuracy": 0.6945849061012268, + "num_tokens": 577991669.0, + "step": 22343 + }, + { + "epoch": 2.4537667471996487, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5065817832946777, + "learning_rate": 1e-06, + "loss": 0.7562, + "mean_token_accuracy": 0.7547155022621155, + "num_tokens": 578012901.0, + "step": 22344 + }, + { + "epoch": 2.453876564902262, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.6512680053710938, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.751417875289917, + "num_tokens": 578033330.0, + "step": 22345 + }, + { + "epoch": 2.453986382604876, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.803516149520874, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7488296031951904, + "num_tokens": 578053690.0, + "step": 22346 + }, + { + "epoch": 2.4540962003074895, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.215625762939453, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7295048832893372, + "num_tokens": 578084004.0, + "step": 22347 + }, + { + "epoch": 2.4542060180101033, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3149492740631104, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.740125298500061, + "num_tokens": 578109929.0, + "step": 22348 + }, + { + "epoch": 2.454315835712717, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1233110427856445, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7231526374816895, + "num_tokens": 578143144.0, + "step": 22349 + }, + { + "epoch": 2.4544256534153304, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.11021089553833, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.719896137714386, + "num_tokens": 578172629.0, + "step": 22350 + }, + { + "epoch": 2.454535471117944, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.398871898651123, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7350254058837891, + "num_tokens": 578196976.0, + "step": 22351 + }, + { + "epoch": 2.454645288820558, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.296323537826538, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7161248922348022, + "num_tokens": 578226395.0, + "step": 22352 + }, + { + "epoch": 2.4547551065231716, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2951974868774414, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7245189547538757, + "num_tokens": 578254367.0, + "step": 22353 + }, + { + "epoch": 2.4548649242257854, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.352489709854126, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.733055591583252, + "num_tokens": 578279572.0, + "step": 22354 + }, + { + "epoch": 2.4549747419283987, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4695301055908203, + "learning_rate": 1e-06, + "loss": 0.6827, + "mean_token_accuracy": 0.775890588760376, + "num_tokens": 578302458.0, + "step": 22355 + }, + { + "epoch": 2.4550845596310125, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2824854850769043, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7142608165740967, + "num_tokens": 578333129.0, + "step": 22356 + }, + { + "epoch": 2.455194377333626, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4291739463806152, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7455617189407349, + "num_tokens": 578357560.0, + "step": 22357 + }, + { + "epoch": 2.45530419503624, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2236616611480713, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7436829805374146, + "num_tokens": 578384080.0, + "step": 22358 + }, + { + "epoch": 2.4554140127388537, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.236743450164795, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7173370122909546, + "num_tokens": 578413855.0, + "step": 22359 + }, + { + "epoch": 2.455523830441467, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.409864902496338, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7291083335876465, + "num_tokens": 578441063.0, + "step": 22360 + }, + { + "epoch": 2.455633648144081, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2939319610595703, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7160648703575134, + "num_tokens": 578469577.0, + "step": 22361 + }, + { + "epoch": 2.4557434658466946, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4133567810058594, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7482143044471741, + "num_tokens": 578492684.0, + "step": 22362 + }, + { + "epoch": 2.4558532835493083, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.452152967453003, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7286921143531799, + "num_tokens": 578516977.0, + "step": 22363 + }, + { + "epoch": 2.4559631012519216, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5093727111816406, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7157701253890991, + "num_tokens": 578542266.0, + "step": 22364 + }, + { + "epoch": 2.4560729189545354, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5396101474761963, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7357125878334045, + "num_tokens": 578564821.0, + "step": 22365 + }, + { + "epoch": 2.456182736657149, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.7291769981384277, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7371435165405273, + "num_tokens": 578586215.0, + "step": 22366 + }, + { + "epoch": 2.456292554359763, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.402498245239258, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7244212627410889, + "num_tokens": 578612609.0, + "step": 22367 + }, + { + "epoch": 2.456402372062376, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4679388999938965, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7090644240379333, + "num_tokens": 578637678.0, + "step": 22368 + }, + { + "epoch": 2.45651218976499, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.074059009552002, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7153775691986084, + "num_tokens": 578669710.0, + "step": 22369 + }, + { + "epoch": 2.4566220074676037, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4543867111206055, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.7483770847320557, + "num_tokens": 578693789.0, + "step": 22370 + }, + { + "epoch": 2.4567318251702175, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1778512001037598, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7286524772644043, + "num_tokens": 578724447.0, + "step": 22371 + }, + { + "epoch": 2.4568416428728312, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3210232257843018, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7239742875099182, + "num_tokens": 578750989.0, + "step": 22372 + }, + { + "epoch": 2.4569514605754446, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5575673580169678, + "learning_rate": 1e-06, + "loss": 0.8017, + "mean_token_accuracy": 0.7455164194107056, + "num_tokens": 578772778.0, + "step": 22373 + }, + { + "epoch": 2.4570612782780583, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2720940113067627, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7224963307380676, + "num_tokens": 578800895.0, + "step": 22374 + }, + { + "epoch": 2.457171095980672, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.183342933654785, + "learning_rate": 1e-06, + "loss": 0.7452, + "mean_token_accuracy": 0.755273699760437, + "num_tokens": 578829230.0, + "step": 22375 + }, + { + "epoch": 2.457280913683286, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2030539512634277, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.716507613658905, + "num_tokens": 578860437.0, + "step": 22376 + }, + { + "epoch": 2.4573907313858996, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3201613426208496, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7087955474853516, + "num_tokens": 578887887.0, + "step": 22377 + }, + { + "epoch": 2.457500549088513, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.248298406600952, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7196179628372192, + "num_tokens": 578916142.0, + "step": 22378 + }, + { + "epoch": 2.4576103667911267, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.365663528442383, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7321301698684692, + "num_tokens": 578943239.0, + "step": 22379 + }, + { + "epoch": 2.4577201844937404, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2260494232177734, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7430846095085144, + "num_tokens": 578970595.0, + "step": 22380 + }, + { + "epoch": 2.457830002196354, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2472612857818604, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7277518510818481, + "num_tokens": 578999841.0, + "step": 22381 + }, + { + "epoch": 2.457939819898968, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2374725341796875, + "learning_rate": 1e-06, + "loss": 0.9555, + "mean_token_accuracy": 0.6957664489746094, + "num_tokens": 579028000.0, + "step": 22382 + }, + { + "epoch": 2.4580496376015812, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.507678270339966, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7272130250930786, + "num_tokens": 579051268.0, + "step": 22383 + }, + { + "epoch": 2.458159455304195, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.375742197036743, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.7094506025314331, + "num_tokens": 579079968.0, + "step": 22384 + }, + { + "epoch": 2.4582692730068088, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.7234861850738525, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7307493686676025, + "num_tokens": 579100722.0, + "step": 22385 + }, + { + "epoch": 2.4583790907094225, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.239621877670288, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7310408353805542, + "num_tokens": 579129843.0, + "step": 22386 + }, + { + "epoch": 2.4584889084120363, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.0887374877929688, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.721679151058197, + "num_tokens": 579164875.0, + "step": 22387 + }, + { + "epoch": 2.4585987261146496, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.436014175415039, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7155314683914185, + "num_tokens": 579190306.0, + "step": 22388 + }, + { + "epoch": 2.4587085438172633, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.7506461143493652, + "learning_rate": 1e-06, + "loss": 0.7297, + "mean_token_accuracy": 0.7689921855926514, + "num_tokens": 579209165.0, + "step": 22389 + }, + { + "epoch": 2.458818361519877, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.389878988265991, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7105151414871216, + "num_tokens": 579236012.0, + "step": 22390 + }, + { + "epoch": 2.458928179222491, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1962687969207764, + "learning_rate": 1e-06, + "loss": 0.9218, + "mean_token_accuracy": 0.7129316329956055, + "num_tokens": 579266172.0, + "step": 22391 + }, + { + "epoch": 2.459037996925104, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2918381690979004, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.732377290725708, + "num_tokens": 579293066.0, + "step": 22392 + }, + { + "epoch": 2.459147814627718, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4209556579589844, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7224187850952148, + "num_tokens": 579318838.0, + "step": 22393 + }, + { + "epoch": 2.4592576323303317, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.0821475982666016, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7161950469017029, + "num_tokens": 579352649.0, + "step": 22394 + }, + { + "epoch": 2.4593674500329454, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3482017517089844, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7401700019836426, + "num_tokens": 579376862.0, + "step": 22395 + }, + { + "epoch": 2.4594772677355587, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2301106452941895, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7462832927703857, + "num_tokens": 579404419.0, + "step": 22396 + }, + { + "epoch": 2.4595870854381725, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3954453468322754, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7296578288078308, + "num_tokens": 579431207.0, + "step": 22397 + }, + { + "epoch": 2.4596969031407863, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 1.9741995334625244, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.731960654258728, + "num_tokens": 579463351.0, + "step": 22398 + }, + { + "epoch": 2.4598067208434, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.33956241607666, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7443698644638062, + "num_tokens": 579490210.0, + "step": 22399 + }, + { + "epoch": 2.4599165385460138, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4198644161224365, + "learning_rate": 1e-06, + "loss": 0.7627, + "mean_token_accuracy": 0.7621619701385498, + "num_tokens": 579514549.0, + "step": 22400 + }, + { + "epoch": 2.460026356248627, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.505899429321289, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7360793948173523, + "num_tokens": 579538738.0, + "step": 22401 + }, + { + "epoch": 2.460136173951241, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.1859729290008545, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7200374603271484, + "num_tokens": 579568753.0, + "step": 22402 + }, + { + "epoch": 2.4602459916538546, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.411228895187378, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.7464858293533325, + "num_tokens": 579593085.0, + "step": 22403 + }, + { + "epoch": 2.4603558093564684, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5799143314361572, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7163146734237671, + "num_tokens": 579615851.0, + "step": 22404 + }, + { + "epoch": 2.460465627059082, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2804017066955566, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7509303092956543, + "num_tokens": 579641659.0, + "step": 22405 + }, + { + "epoch": 2.4605754447616954, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5325164794921875, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7402446269989014, + "num_tokens": 579665632.0, + "step": 22406 + }, + { + "epoch": 2.460685262464309, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.38346266746521, + "learning_rate": 1e-06, + "loss": 0.7969, + "mean_token_accuracy": 0.7559302449226379, + "num_tokens": 579689755.0, + "step": 22407 + }, + { + "epoch": 2.460795080166923, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4334938526153564, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7373976111412048, + "num_tokens": 579713267.0, + "step": 22408 + }, + { + "epoch": 2.4609048978695367, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.3737058639526367, + "learning_rate": 1e-06, + "loss": 0.916, + "mean_token_accuracy": 0.7101892232894897, + "num_tokens": 579739361.0, + "step": 22409 + }, + { + "epoch": 2.4610147155721505, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.5582282543182373, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7139173150062561, + "num_tokens": 579761589.0, + "step": 22410 + }, + { + "epoch": 2.4611245332747638, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.472337245941162, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7181838154792786, + "num_tokens": 579788253.0, + "step": 22411 + }, + { + "epoch": 2.4612343509773775, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.215735912322998, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7051939368247986, + "num_tokens": 579817323.0, + "step": 22412 + }, + { + "epoch": 2.4613441686799913, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.4757988452911377, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7319375276565552, + "num_tokens": 579840291.0, + "step": 22413 + }, + { + "epoch": 2.461453986382605, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.218869686126709, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7033380270004272, + "num_tokens": 579872806.0, + "step": 22414 + }, + { + "epoch": 2.4615638040852184, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.0631000995635986, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7233496308326721, + "num_tokens": 579905609.0, + "step": 22415 + }, + { + "epoch": 2.461673621787832, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.2367300987243652, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7269890904426575, + "num_tokens": 579934455.0, + "step": 22416 + }, + { + "epoch": 2.461783439490446, + "ewc_loss": 1.4603137969970703e-05, + "grad_norm": 2.631436586380005, + "learning_rate": 1e-06, + "loss": 0.7593, + "mean_token_accuracy": 0.7546988725662231, + "num_tokens": 579956473.0, + "step": 22417 + }, + { + "epoch": 2.4618932571930596, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3700337409973145, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7141245603561401, + "num_tokens": 579983569.0, + "step": 22418 + }, + { + "epoch": 2.462003074895673, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.245178461074829, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7300695180892944, + "num_tokens": 580010210.0, + "step": 22419 + }, + { + "epoch": 2.4621128925982867, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4683070182800293, + "learning_rate": 1e-06, + "loss": 0.9683, + "mean_token_accuracy": 0.7005558013916016, + "num_tokens": 580035267.0, + "step": 22420 + }, + { + "epoch": 2.4622227103009005, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4268929958343506, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7389644384384155, + "num_tokens": 580059049.0, + "step": 22421 + }, + { + "epoch": 2.462332528003514, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3517587184906006, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7475681304931641, + "num_tokens": 580083726.0, + "step": 22422 + }, + { + "epoch": 2.462442345706128, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4853341579437256, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7189230918884277, + "num_tokens": 580109057.0, + "step": 22423 + }, + { + "epoch": 2.4625521634087413, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.522899866104126, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7314242720603943, + "num_tokens": 580131527.0, + "step": 22424 + }, + { + "epoch": 2.462661981111355, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4085187911987305, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7507579326629639, + "num_tokens": 580158296.0, + "step": 22425 + }, + { + "epoch": 2.462771798813969, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3050761222839355, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.721000075340271, + "num_tokens": 580185587.0, + "step": 22426 + }, + { + "epoch": 2.4628816165165826, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.304445743560791, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7319570779800415, + "num_tokens": 580211264.0, + "step": 22427 + }, + { + "epoch": 2.4629914342191963, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2488560676574707, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7445840835571289, + "num_tokens": 580238128.0, + "step": 22428 + }, + { + "epoch": 2.4631012519218096, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3342020511627197, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7316428422927856, + "num_tokens": 580263380.0, + "step": 22429 + }, + { + "epoch": 2.4632110696244234, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3017656803131104, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.6978532671928406, + "num_tokens": 580291588.0, + "step": 22430 + }, + { + "epoch": 2.463320887327037, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3171396255493164, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.7200517654418945, + "num_tokens": 580321155.0, + "step": 22431 + }, + { + "epoch": 2.463430705029651, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3780739307403564, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7310000658035278, + "num_tokens": 580346786.0, + "step": 22432 + }, + { + "epoch": 2.4635405227322646, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5167503356933594, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7423491477966309, + "num_tokens": 580368009.0, + "step": 22433 + }, + { + "epoch": 2.463650340434878, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.484130620956421, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7456440329551697, + "num_tokens": 580391179.0, + "step": 22434 + }, + { + "epoch": 2.4637601581374917, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.248591184616089, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7263265252113342, + "num_tokens": 580417400.0, + "step": 22435 + }, + { + "epoch": 2.4638699758401055, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.543952465057373, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7354512214660645, + "num_tokens": 580440859.0, + "step": 22436 + }, + { + "epoch": 2.4639797935427192, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3666129112243652, + "learning_rate": 1e-06, + "loss": 0.8568, + "mean_token_accuracy": 0.7263069748878479, + "num_tokens": 580465329.0, + "step": 22437 + }, + { + "epoch": 2.464089611245333, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.169116735458374, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7490412592887878, + "num_tokens": 580494802.0, + "step": 22438 + }, + { + "epoch": 2.4641994289479463, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.332357168197632, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7431102991104126, + "num_tokens": 580520357.0, + "step": 22439 + }, + { + "epoch": 2.46430924665056, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.337071418762207, + "learning_rate": 1e-06, + "loss": 0.7004, + "mean_token_accuracy": 0.7707861065864563, + "num_tokens": 580544172.0, + "step": 22440 + }, + { + "epoch": 2.464419064353174, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5842182636260986, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.7478101253509521, + "num_tokens": 580565362.0, + "step": 22441 + }, + { + "epoch": 2.4645288820557876, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.257622241973877, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7338910102844238, + "num_tokens": 580592144.0, + "step": 22442 + }, + { + "epoch": 2.464638699758401, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.316087245941162, + "learning_rate": 1e-06, + "loss": 0.8941, + "mean_token_accuracy": 0.7245993614196777, + "num_tokens": 580619409.0, + "step": 22443 + }, + { + "epoch": 2.4647485174610146, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2119226455688477, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7304983139038086, + "num_tokens": 580647254.0, + "step": 22444 + }, + { + "epoch": 2.4648583351636284, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5470168590545654, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7377852201461792, + "num_tokens": 580669445.0, + "step": 22445 + }, + { + "epoch": 2.464968152866242, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.240072727203369, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.742722749710083, + "num_tokens": 580698773.0, + "step": 22446 + }, + { + "epoch": 2.4650779705688555, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.150648355484009, + "learning_rate": 1e-06, + "loss": 0.9297, + "mean_token_accuracy": 0.7089633941650391, + "num_tokens": 580728787.0, + "step": 22447 + }, + { + "epoch": 2.4651877882714692, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1581778526306152, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7287153005599976, + "num_tokens": 580758320.0, + "step": 22448 + }, + { + "epoch": 2.465297605974083, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1762709617614746, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7290361523628235, + "num_tokens": 580787451.0, + "step": 22449 + }, + { + "epoch": 2.4654074236766967, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2239694595336914, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7359427213668823, + "num_tokens": 580813681.0, + "step": 22450 + }, + { + "epoch": 2.4655172413793105, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.333571434020996, + "learning_rate": 1e-06, + "loss": 0.9188, + "mean_token_accuracy": 0.7213246822357178, + "num_tokens": 580838606.0, + "step": 22451 + }, + { + "epoch": 2.465627059081924, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2773876190185547, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7359440326690674, + "num_tokens": 580866549.0, + "step": 22452 + }, + { + "epoch": 2.4657368767845376, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.293144941329956, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7168914675712585, + "num_tokens": 580893306.0, + "step": 22453 + }, + { + "epoch": 2.4658466944871513, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5451107025146484, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7342405915260315, + "num_tokens": 580916349.0, + "step": 22454 + }, + { + "epoch": 2.465956512189765, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6620352268218994, + "learning_rate": 1e-06, + "loss": 0.77, + "mean_token_accuracy": 0.7541782855987549, + "num_tokens": 580936238.0, + "step": 22455 + }, + { + "epoch": 2.466066329892379, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0466716289520264, + "learning_rate": 1e-06, + "loss": 0.935, + "mean_token_accuracy": 0.706072986125946, + "num_tokens": 580970218.0, + "step": 22456 + }, + { + "epoch": 2.466176147594992, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5918080806732178, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7449113130569458, + "num_tokens": 580993275.0, + "step": 22457 + }, + { + "epoch": 2.466285965297606, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 1.9885250329971313, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7451303601264954, + "num_tokens": 581023208.0, + "step": 22458 + }, + { + "epoch": 2.4663957830002197, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.36745285987854, + "learning_rate": 1e-06, + "loss": 0.7704, + "mean_token_accuracy": 0.7597173452377319, + "num_tokens": 581047242.0, + "step": 22459 + }, + { + "epoch": 2.4665056007028334, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.597917318344116, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7503374814987183, + "num_tokens": 581069604.0, + "step": 22460 + }, + { + "epoch": 2.466615418405447, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.512446403503418, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7413811683654785, + "num_tokens": 581093665.0, + "step": 22461 + }, + { + "epoch": 2.4667252361080605, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5024406909942627, + "learning_rate": 1e-06, + "loss": 0.731, + "mean_token_accuracy": 0.7607568502426147, + "num_tokens": 581117503.0, + "step": 22462 + }, + { + "epoch": 2.4668350538106742, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 3.3131465911865234, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7428832054138184, + "num_tokens": 581141075.0, + "step": 22463 + }, + { + "epoch": 2.466944871513288, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2786107063293457, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7185791730880737, + "num_tokens": 581168850.0, + "step": 22464 + }, + { + "epoch": 2.4670546892159018, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3854973316192627, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7336381673812866, + "num_tokens": 581194901.0, + "step": 22465 + }, + { + "epoch": 2.467164506918515, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.242385149002075, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7249348759651184, + "num_tokens": 581223253.0, + "step": 22466 + }, + { + "epoch": 2.467274324621129, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0722625255584717, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.709469199180603, + "num_tokens": 581257657.0, + "step": 22467 + }, + { + "epoch": 2.4673841423237426, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6221728324890137, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.7330917119979858, + "num_tokens": 581280242.0, + "step": 22468 + }, + { + "epoch": 2.4674939600263563, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2743284702301025, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7268744707107544, + "num_tokens": 581309213.0, + "step": 22469 + }, + { + "epoch": 2.4676037777289697, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.7940688133239746, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7345041036605835, + "num_tokens": 581328185.0, + "step": 22470 + }, + { + "epoch": 2.4677135954315834, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6543562412261963, + "learning_rate": 1e-06, + "loss": 0.7374, + "mean_token_accuracy": 0.7650340795516968, + "num_tokens": 581348910.0, + "step": 22471 + }, + { + "epoch": 2.467823413134197, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3537144660949707, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7344251871109009, + "num_tokens": 581375189.0, + "step": 22472 + }, + { + "epoch": 2.467933230836811, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5394437313079834, + "learning_rate": 1e-06, + "loss": 0.7623, + "mean_token_accuracy": 0.7509033679962158, + "num_tokens": 581396012.0, + "step": 22473 + }, + { + "epoch": 2.4680430485394247, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2684991359710693, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7350484728813171, + "num_tokens": 581423460.0, + "step": 22474 + }, + { + "epoch": 2.468152866242038, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4787495136260986, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7467278838157654, + "num_tokens": 581449750.0, + "step": 22475 + }, + { + "epoch": 2.4682626839446518, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.554514169692993, + "learning_rate": 1e-06, + "loss": 0.789, + "mean_token_accuracy": 0.742801308631897, + "num_tokens": 581470226.0, + "step": 22476 + }, + { + "epoch": 2.4683725016472655, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.349872350692749, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.727435827255249, + "num_tokens": 581496730.0, + "step": 22477 + }, + { + "epoch": 2.4684823193498793, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3285140991210938, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7411110401153564, + "num_tokens": 581520234.0, + "step": 22478 + }, + { + "epoch": 2.468592137052493, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.389031410217285, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7367463707923889, + "num_tokens": 581548145.0, + "step": 22479 + }, + { + "epoch": 2.4687019547551063, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6702537536621094, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.7397909760475159, + "num_tokens": 581570008.0, + "step": 22480 + }, + { + "epoch": 2.46881177245772, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3255066871643066, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7445698976516724, + "num_tokens": 581594675.0, + "step": 22481 + }, + { + "epoch": 2.468921590160334, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.247032880783081, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7312542200088501, + "num_tokens": 581621903.0, + "step": 22482 + }, + { + "epoch": 2.4690314078629476, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.176607370376587, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7262375354766846, + "num_tokens": 581653424.0, + "step": 22483 + }, + { + "epoch": 2.4691412255655614, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.398144245147705, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7309982776641846, + "num_tokens": 581680659.0, + "step": 22484 + }, + { + "epoch": 2.4692510432681747, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.327537775039673, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7274385094642639, + "num_tokens": 581706034.0, + "step": 22485 + }, + { + "epoch": 2.4693608609707884, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.091315984725952, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7218809127807617, + "num_tokens": 581736537.0, + "step": 22486 + }, + { + "epoch": 2.469470678673402, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6499712467193604, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7532055377960205, + "num_tokens": 581758212.0, + "step": 22487 + }, + { + "epoch": 2.469580496376016, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.782604455947876, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7403872013092041, + "num_tokens": 581777728.0, + "step": 22488 + }, + { + "epoch": 2.4696903140786297, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3905270099639893, + "learning_rate": 1e-06, + "loss": 0.7814, + "mean_token_accuracy": 0.7605222463607788, + "num_tokens": 581802384.0, + "step": 22489 + }, + { + "epoch": 2.469800131781243, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2426843643188477, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7193195819854736, + "num_tokens": 581831897.0, + "step": 22490 + }, + { + "epoch": 2.469909949483857, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.7687110900878906, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7499139308929443, + "num_tokens": 581852383.0, + "step": 22491 + }, + { + "epoch": 2.4700197671864705, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.430424690246582, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7311866879463196, + "num_tokens": 581876337.0, + "step": 22492 + }, + { + "epoch": 2.4701295848890843, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.218183994293213, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.713779866695404, + "num_tokens": 581904782.0, + "step": 22493 + }, + { + "epoch": 2.4702394025916976, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0494463443756104, + "learning_rate": 1e-06, + "loss": 0.9855, + "mean_token_accuracy": 0.6929948329925537, + "num_tokens": 581937627.0, + "step": 22494 + }, + { + "epoch": 2.4703492202943114, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4396979808807373, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7301486730575562, + "num_tokens": 581961556.0, + "step": 22495 + }, + { + "epoch": 2.470459037996925, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.54701566696167, + "learning_rate": 1e-06, + "loss": 0.7512, + "mean_token_accuracy": 0.7563026547431946, + "num_tokens": 581982982.0, + "step": 22496 + }, + { + "epoch": 2.470568855699539, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4878718852996826, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7467926740646362, + "num_tokens": 582007050.0, + "step": 22497 + }, + { + "epoch": 2.470678673402152, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.160951852798462, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7375048398971558, + "num_tokens": 582037396.0, + "step": 22498 + }, + { + "epoch": 2.470788491104766, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.536003589630127, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7108554840087891, + "num_tokens": 582060656.0, + "step": 22499 + }, + { + "epoch": 2.4708983088073797, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0052082538604736, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7004245519638062, + "num_tokens": 582098147.0, + "step": 22500 + }, + { + "epoch": 2.4710081265099935, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5565319061279297, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7138248085975647, + "num_tokens": 582120838.0, + "step": 22501 + }, + { + "epoch": 2.471117944212607, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5769567489624023, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7334283590316772, + "num_tokens": 582142587.0, + "step": 22502 + }, + { + "epoch": 2.4712277619152205, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6955201625823975, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7245433330535889, + "num_tokens": 582164094.0, + "step": 22503 + }, + { + "epoch": 2.4713375796178343, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.161939859390259, + "learning_rate": 1e-06, + "loss": 0.8862, + "mean_token_accuracy": 0.7222463488578796, + "num_tokens": 582191897.0, + "step": 22504 + }, + { + "epoch": 2.471447397320448, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.466576099395752, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7328406572341919, + "num_tokens": 582215152.0, + "step": 22505 + }, + { + "epoch": 2.471557215023062, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.750115394592285, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7370925545692444, + "num_tokens": 582235363.0, + "step": 22506 + }, + { + "epoch": 2.4716670327256756, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0458409786224365, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7280125021934509, + "num_tokens": 582268099.0, + "step": 22507 + }, + { + "epoch": 2.471776850428289, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.677680015563965, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7349227070808411, + "num_tokens": 582287577.0, + "step": 22508 + }, + { + "epoch": 2.4718866681309026, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.254591941833496, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7305049896240234, + "num_tokens": 582316791.0, + "step": 22509 + }, + { + "epoch": 2.4719964858335164, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.36969256401062, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7385104894638062, + "num_tokens": 582343571.0, + "step": 22510 + }, + { + "epoch": 2.47210630353613, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3991787433624268, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7286178469657898, + "num_tokens": 582370286.0, + "step": 22511 + }, + { + "epoch": 2.472216121238744, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.143786907196045, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7504339218139648, + "num_tokens": 582396706.0, + "step": 22512 + }, + { + "epoch": 2.472325938941357, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.652810573577881, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7240861058235168, + "num_tokens": 582419388.0, + "step": 22513 + }, + { + "epoch": 2.472435756643971, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.165980815887451, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.716657280921936, + "num_tokens": 582449128.0, + "step": 22514 + }, + { + "epoch": 2.4725455743465847, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1189980506896973, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.730964183807373, + "num_tokens": 582481366.0, + "step": 22515 + }, + { + "epoch": 2.4726553920491985, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2608680725097656, + "learning_rate": 1e-06, + "loss": 0.9062, + "mean_token_accuracy": 0.7164696455001831, + "num_tokens": 582509599.0, + "step": 22516 + }, + { + "epoch": 2.472765209751812, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3644979000091553, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7255420088768005, + "num_tokens": 582537510.0, + "step": 22517 + }, + { + "epoch": 2.4728750274544256, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1676974296569824, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7301430702209473, + "num_tokens": 582565833.0, + "step": 22518 + }, + { + "epoch": 2.4729848451570393, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1401333808898926, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7300999760627747, + "num_tokens": 582595069.0, + "step": 22519 + }, + { + "epoch": 2.473094662859653, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.1824164390563965, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7240194082260132, + "num_tokens": 582626140.0, + "step": 22520 + }, + { + "epoch": 2.473204480562267, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0931732654571533, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7378022074699402, + "num_tokens": 582656988.0, + "step": 22521 + }, + { + "epoch": 2.47331429826488, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.100808620452881, + "learning_rate": 1e-06, + "loss": 0.984, + "mean_token_accuracy": 0.7018737196922302, + "num_tokens": 582690171.0, + "step": 22522 + }, + { + "epoch": 2.473424115967494, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3324573040008545, + "learning_rate": 1e-06, + "loss": 0.7746, + "mean_token_accuracy": 0.7494173645973206, + "num_tokens": 582714233.0, + "step": 22523 + }, + { + "epoch": 2.4735339336701077, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3895087242126465, + "learning_rate": 1e-06, + "loss": 0.7729, + "mean_token_accuracy": 0.7515217661857605, + "num_tokens": 582738901.0, + "step": 22524 + }, + { + "epoch": 2.4736437513727214, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4253346920013428, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7519997358322144, + "num_tokens": 582761168.0, + "step": 22525 + }, + { + "epoch": 2.4737535690753347, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.334378957748413, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7294728755950928, + "num_tokens": 582786137.0, + "step": 22526 + }, + { + "epoch": 2.4738633867779485, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.295168161392212, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.718856692314148, + "num_tokens": 582812937.0, + "step": 22527 + }, + { + "epoch": 2.4739732044805622, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3745806217193604, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7219198346138, + "num_tokens": 582840990.0, + "step": 22528 + }, + { + "epoch": 2.474083022183176, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.6828389167785645, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7440969347953796, + "num_tokens": 582862034.0, + "step": 22529 + }, + { + "epoch": 2.4741928398857898, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4085233211517334, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7282402515411377, + "num_tokens": 582887359.0, + "step": 22530 + }, + { + "epoch": 2.474302657588403, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3224806785583496, + "learning_rate": 1e-06, + "loss": 0.7697, + "mean_token_accuracy": 0.7528072595596313, + "num_tokens": 582911724.0, + "step": 22531 + }, + { + "epoch": 2.474412475291017, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.212247371673584, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7335267066955566, + "num_tokens": 582940079.0, + "step": 22532 + }, + { + "epoch": 2.4745222929936306, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.389686346054077, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7290533781051636, + "num_tokens": 582966645.0, + "step": 22533 + }, + { + "epoch": 2.4746321106962443, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.4712467193603516, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7309211492538452, + "num_tokens": 582990337.0, + "step": 22534 + }, + { + "epoch": 2.474741928398858, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.270453929901123, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7290441989898682, + "num_tokens": 583019007.0, + "step": 22535 + }, + { + "epoch": 2.4748517461014714, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.3091087341308594, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7356759309768677, + "num_tokens": 583046525.0, + "step": 22536 + }, + { + "epoch": 2.474961563804085, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.299309015274048, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7124308347702026, + "num_tokens": 583075457.0, + "step": 22537 + }, + { + "epoch": 2.475071381506699, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.5906269550323486, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7331427931785583, + "num_tokens": 583097935.0, + "step": 22538 + }, + { + "epoch": 2.4751811992093127, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.8034987449645996, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7468263506889343, + "num_tokens": 583116913.0, + "step": 22539 + }, + { + "epoch": 2.4752910169119264, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 1.9822466373443604, + "learning_rate": 1e-06, + "loss": 0.8145, + "mean_token_accuracy": 0.744549572467804, + "num_tokens": 583152270.0, + "step": 22540 + }, + { + "epoch": 2.4754008346145397, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.2522168159484863, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7214696407318115, + "num_tokens": 583181986.0, + "step": 22541 + }, + { + "epoch": 2.4755106523171535, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.0701024532318115, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.722940981388092, + "num_tokens": 583212767.0, + "step": 22542 + }, + { + "epoch": 2.4756204700197673, + "ewc_loss": 1.4662742614746094e-05, + "grad_norm": 2.836139678955078, + "learning_rate": 1e-06, + "loss": 0.7587, + "mean_token_accuracy": 0.7560068964958191, + "num_tokens": 583230748.0, + "step": 22543 + }, + { + "epoch": 2.475730287722381, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6860129833221436, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7576274871826172, + "num_tokens": 583250752.0, + "step": 22544 + }, + { + "epoch": 2.4758401054249943, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.361912965774536, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7080131769180298, + "num_tokens": 583278437.0, + "step": 22545 + }, + { + "epoch": 2.475949923127608, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.250903844833374, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7114263772964478, + "num_tokens": 583305175.0, + "step": 22546 + }, + { + "epoch": 2.476059740830222, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.568671703338623, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.72907555103302, + "num_tokens": 583327026.0, + "step": 22547 + }, + { + "epoch": 2.4761695585328356, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1949174404144287, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7226149439811707, + "num_tokens": 583356237.0, + "step": 22548 + }, + { + "epoch": 2.476279376235449, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1530582904815674, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7291237115859985, + "num_tokens": 583384805.0, + "step": 22549 + }, + { + "epoch": 2.4763891939380627, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.380183696746826, + "learning_rate": 1e-06, + "loss": 0.7995, + "mean_token_accuracy": 0.7402395009994507, + "num_tokens": 583409634.0, + "step": 22550 + }, + { + "epoch": 2.4764990116406764, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.138019323348999, + "learning_rate": 1e-06, + "loss": 0.9371, + "mean_token_accuracy": 0.710801362991333, + "num_tokens": 583439142.0, + "step": 22551 + }, + { + "epoch": 2.47660882934329, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.521289110183716, + "learning_rate": 1e-06, + "loss": 0.9002, + "mean_token_accuracy": 0.7193378806114197, + "num_tokens": 583462208.0, + "step": 22552 + }, + { + "epoch": 2.476718647045904, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4203474521636963, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7391244173049927, + "num_tokens": 583485821.0, + "step": 22553 + }, + { + "epoch": 2.4768284647485173, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2278687953948975, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7093432545661926, + "num_tokens": 583516572.0, + "step": 22554 + }, + { + "epoch": 2.476938282451131, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.30442214012146, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7273738384246826, + "num_tokens": 583542215.0, + "step": 22555 + }, + { + "epoch": 2.4770481001537448, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.452232599258423, + "learning_rate": 1e-06, + "loss": 0.9225, + "mean_token_accuracy": 0.7144591808319092, + "num_tokens": 583568246.0, + "step": 22556 + }, + { + "epoch": 2.4771579178563585, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4754326343536377, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7418969869613647, + "num_tokens": 583593046.0, + "step": 22557 + }, + { + "epoch": 2.4772677355589723, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.464578866958618, + "learning_rate": 1e-06, + "loss": 0.7956, + "mean_token_accuracy": 0.7486317753791809, + "num_tokens": 583617182.0, + "step": 22558 + }, + { + "epoch": 2.4773775532615856, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.543755054473877, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7278515100479126, + "num_tokens": 583640409.0, + "step": 22559 + }, + { + "epoch": 2.4774873709641994, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.320094347000122, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7312721610069275, + "num_tokens": 583667501.0, + "step": 22560 + }, + { + "epoch": 2.477597188666813, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.465956211090088, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7339093685150146, + "num_tokens": 583691981.0, + "step": 22561 + }, + { + "epoch": 2.477707006369427, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.088916540145874, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7268208861351013, + "num_tokens": 583723544.0, + "step": 22562 + }, + { + "epoch": 2.4778168240720406, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6492042541503906, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7370919585227966, + "num_tokens": 583745249.0, + "step": 22563 + }, + { + "epoch": 2.477926641774654, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4996235370635986, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7462785243988037, + "num_tokens": 583767713.0, + "step": 22564 + }, + { + "epoch": 2.4780364594772677, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2463479042053223, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7345530986785889, + "num_tokens": 583796351.0, + "step": 22565 + }, + { + "epoch": 2.4781462771798815, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.653148889541626, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7452881932258606, + "num_tokens": 583818522.0, + "step": 22566 + }, + { + "epoch": 2.478256094882495, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.736689805984497, + "learning_rate": 1e-06, + "loss": 0.7836, + "mean_token_accuracy": 0.7493597269058228, + "num_tokens": 583841306.0, + "step": 22567 + }, + { + "epoch": 2.478365912585109, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.599010467529297, + "learning_rate": 1e-06, + "loss": 0.752, + "mean_token_accuracy": 0.7591731548309326, + "num_tokens": 583862720.0, + "step": 22568 + }, + { + "epoch": 2.4784757302877223, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2738399505615234, + "learning_rate": 1e-06, + "loss": 0.9584, + "mean_token_accuracy": 0.7012661099433899, + "num_tokens": 583891557.0, + "step": 22569 + }, + { + "epoch": 2.478585547990336, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.419808864593506, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7313042879104614, + "num_tokens": 583915524.0, + "step": 22570 + }, + { + "epoch": 2.47869536569295, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.0672879219055176, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7392226457595825, + "num_tokens": 583947311.0, + "step": 22571 + }, + { + "epoch": 2.4788051833955635, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3819169998168945, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7251490354537964, + "num_tokens": 583971538.0, + "step": 22572 + }, + { + "epoch": 2.478915001098177, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.801300048828125, + "learning_rate": 1e-06, + "loss": 0.794, + "mean_token_accuracy": 0.7488821744918823, + "num_tokens": 583990662.0, + "step": 22573 + }, + { + "epoch": 2.4790248188007906, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.230048418045044, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7259725332260132, + "num_tokens": 584020069.0, + "step": 22574 + }, + { + "epoch": 2.4791346365034044, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.525791645050049, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.750413179397583, + "num_tokens": 584043942.0, + "step": 22575 + }, + { + "epoch": 2.479244454206018, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6606032848358154, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7384850978851318, + "num_tokens": 584064359.0, + "step": 22576 + }, + { + "epoch": 2.4793542719086314, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3811821937561035, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7163687944412231, + "num_tokens": 584089446.0, + "step": 22577 + }, + { + "epoch": 2.479464089611245, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6048717498779297, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7572950124740601, + "num_tokens": 584112276.0, + "step": 22578 + }, + { + "epoch": 2.479573907313859, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.900278329849243, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7200238704681396, + "num_tokens": 584131282.0, + "step": 22579 + }, + { + "epoch": 2.4796837250164727, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.451758861541748, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7253753542900085, + "num_tokens": 584157027.0, + "step": 22580 + }, + { + "epoch": 2.4797935427190865, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2183122634887695, + "learning_rate": 1e-06, + "loss": 0.9679, + "mean_token_accuracy": 0.7015907764434814, + "num_tokens": 584187271.0, + "step": 22581 + }, + { + "epoch": 2.4799033604217, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.344484806060791, + "learning_rate": 1e-06, + "loss": 0.8583, + "mean_token_accuracy": 0.7299817204475403, + "num_tokens": 584214126.0, + "step": 22582 + }, + { + "epoch": 2.4800131781243135, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.429382801055908, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7456459999084473, + "num_tokens": 584238308.0, + "step": 22583 + }, + { + "epoch": 2.4801229958269273, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5973470211029053, + "learning_rate": 1e-06, + "loss": 0.9368, + "mean_token_accuracy": 0.7085223197937012, + "num_tokens": 584260562.0, + "step": 22584 + }, + { + "epoch": 2.480232813529541, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.529927968978882, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7250778675079346, + "num_tokens": 584285600.0, + "step": 22585 + }, + { + "epoch": 2.480342631232155, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4261293411254883, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.736616849899292, + "num_tokens": 584309999.0, + "step": 22586 + }, + { + "epoch": 2.480452448934768, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.7503740787506104, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7431885004043579, + "num_tokens": 584331339.0, + "step": 22587 + }, + { + "epoch": 2.480562266637382, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6531524658203125, + "learning_rate": 1e-06, + "loss": 0.7087, + "mean_token_accuracy": 0.7699971199035645, + "num_tokens": 584352018.0, + "step": 22588 + }, + { + "epoch": 2.4806720843399956, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.520115852355957, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7677863836288452, + "num_tokens": 584374314.0, + "step": 22589 + }, + { + "epoch": 2.4807819020426094, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5460426807403564, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7060933709144592, + "num_tokens": 584396902.0, + "step": 22590 + }, + { + "epoch": 2.480891719745223, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5177571773529053, + "learning_rate": 1e-06, + "loss": 0.7852, + "mean_token_accuracy": 0.745984673500061, + "num_tokens": 584419211.0, + "step": 22591 + }, + { + "epoch": 2.4810015374478365, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.617682456970215, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7321396470069885, + "num_tokens": 584442109.0, + "step": 22592 + }, + { + "epoch": 2.4811113551504502, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.541125774383545, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7386802434921265, + "num_tokens": 584466087.0, + "step": 22593 + }, + { + "epoch": 2.481221172853064, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4376869201660156, + "learning_rate": 1e-06, + "loss": 0.8812, + "mean_token_accuracy": 0.7223938703536987, + "num_tokens": 584490736.0, + "step": 22594 + }, + { + "epoch": 2.4813309905556777, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.0060536861419678, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7254949808120728, + "num_tokens": 584525952.0, + "step": 22595 + }, + { + "epoch": 2.481440808258291, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2036287784576416, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7278052568435669, + "num_tokens": 584554175.0, + "step": 22596 + }, + { + "epoch": 2.481550625960905, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3205130100250244, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7361229062080383, + "num_tokens": 584581818.0, + "step": 22597 + }, + { + "epoch": 2.4816604436635186, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2784664630889893, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7322390675544739, + "num_tokens": 584608837.0, + "step": 22598 + }, + { + "epoch": 2.4817702613661323, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2742421627044678, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.718006432056427, + "num_tokens": 584637585.0, + "step": 22599 + }, + { + "epoch": 2.4818800790687456, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4294495582580566, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7337722778320312, + "num_tokens": 584661031.0, + "step": 22600 + }, + { + "epoch": 2.4819898967713594, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3889145851135254, + "learning_rate": 1e-06, + "loss": 0.799, + "mean_token_accuracy": 0.7437019348144531, + "num_tokens": 584684536.0, + "step": 22601 + }, + { + "epoch": 2.482099714473973, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.449589490890503, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.7472638487815857, + "num_tokens": 584708199.0, + "step": 22602 + }, + { + "epoch": 2.482209532176587, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3383824825286865, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7116620540618896, + "num_tokens": 584733161.0, + "step": 22603 + }, + { + "epoch": 2.4823193498792007, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.467761754989624, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7417474985122681, + "num_tokens": 584757816.0, + "step": 22604 + }, + { + "epoch": 2.482429167581814, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1131269931793213, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7495957612991333, + "num_tokens": 584787031.0, + "step": 22605 + }, + { + "epoch": 2.4825389852844277, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5674808025360107, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7472381591796875, + "num_tokens": 584808485.0, + "step": 22606 + }, + { + "epoch": 2.4826488029870415, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.731337070465088, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7347584962844849, + "num_tokens": 584830143.0, + "step": 22607 + }, + { + "epoch": 2.4827586206896552, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4166829586029053, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7318145036697388, + "num_tokens": 584854495.0, + "step": 22608 + }, + { + "epoch": 2.482868438392269, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.536987781524658, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7303928136825562, + "num_tokens": 584877030.0, + "step": 22609 + }, + { + "epoch": 2.4829782560948823, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.583566665649414, + "learning_rate": 1e-06, + "loss": 0.7436, + "mean_token_accuracy": 0.7587778568267822, + "num_tokens": 584899034.0, + "step": 22610 + }, + { + "epoch": 2.483088073797496, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5120887756347656, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7268888354301453, + "num_tokens": 584924205.0, + "step": 22611 + }, + { + "epoch": 2.48319789150011, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4024055004119873, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7115834951400757, + "num_tokens": 584949529.0, + "step": 22612 + }, + { + "epoch": 2.4833077092027236, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.666600227355957, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7374193668365479, + "num_tokens": 584972765.0, + "step": 22613 + }, + { + "epoch": 2.4834175269053373, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.0876030921936035, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7118443250656128, + "num_tokens": 585004051.0, + "step": 22614 + }, + { + "epoch": 2.4835273446079507, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.08150053024292, + "learning_rate": 1e-06, + "loss": 0.9617, + "mean_token_accuracy": 0.6989189982414246, + "num_tokens": 585035043.0, + "step": 22615 + }, + { + "epoch": 2.4836371623105644, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3021156787872314, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7351590394973755, + "num_tokens": 585063833.0, + "step": 22616 + }, + { + "epoch": 2.483746980013178, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4405250549316406, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.719243049621582, + "num_tokens": 585089772.0, + "step": 22617 + }, + { + "epoch": 2.483856797715792, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.382244348526001, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7182185053825378, + "num_tokens": 585113862.0, + "step": 22618 + }, + { + "epoch": 2.4839666154184057, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2422873973846436, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7256618738174438, + "num_tokens": 585141848.0, + "step": 22619 + }, + { + "epoch": 2.484076433121019, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.506331205368042, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7275481224060059, + "num_tokens": 585165372.0, + "step": 22620 + }, + { + "epoch": 2.4841862508236328, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4180350303649902, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7257223725318909, + "num_tokens": 585190186.0, + "step": 22621 + }, + { + "epoch": 2.4842960685262465, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4592807292938232, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7425980567932129, + "num_tokens": 585215818.0, + "step": 22622 + }, + { + "epoch": 2.4844058862288603, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.427401065826416, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7520887851715088, + "num_tokens": 585238265.0, + "step": 22623 + }, + { + "epoch": 2.4845157039314736, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1687121391296387, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7440227270126343, + "num_tokens": 585266467.0, + "step": 22624 + }, + { + "epoch": 2.4846255216340873, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2722527980804443, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.7442097067832947, + "num_tokens": 585295652.0, + "step": 22625 + }, + { + "epoch": 2.484735339336701, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2336225509643555, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7429240942001343, + "num_tokens": 585323200.0, + "step": 22626 + }, + { + "epoch": 2.484845157039315, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.415428638458252, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7360387444496155, + "num_tokens": 585347754.0, + "step": 22627 + }, + { + "epoch": 2.484954974741928, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.347003698348999, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7336661219596863, + "num_tokens": 585373003.0, + "step": 22628 + }, + { + "epoch": 2.485064792444542, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.284203290939331, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7239430546760559, + "num_tokens": 585399580.0, + "step": 22629 + }, + { + "epoch": 2.4851746101471557, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5342328548431396, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7417452335357666, + "num_tokens": 585423217.0, + "step": 22630 + }, + { + "epoch": 2.4852844278497694, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.242889165878296, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7276915907859802, + "num_tokens": 585451117.0, + "step": 22631 + }, + { + "epoch": 2.485394245552383, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.150364398956299, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7289813160896301, + "num_tokens": 585480424.0, + "step": 22632 + }, + { + "epoch": 2.4855040632549965, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2274935245513916, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7451032400131226, + "num_tokens": 585508401.0, + "step": 22633 + }, + { + "epoch": 2.4856138809576103, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2274932861328125, + "learning_rate": 1e-06, + "loss": 0.9211, + "mean_token_accuracy": 0.7170060873031616, + "num_tokens": 585537066.0, + "step": 22634 + }, + { + "epoch": 2.485723698660224, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3498857021331787, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7233926057815552, + "num_tokens": 585564235.0, + "step": 22635 + }, + { + "epoch": 2.485833516362838, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.6738321781158447, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7399968504905701, + "num_tokens": 585585897.0, + "step": 22636 + }, + { + "epoch": 2.4859433340654515, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1644234657287598, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7274264693260193, + "num_tokens": 585613902.0, + "step": 22637 + }, + { + "epoch": 2.486053151768065, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3448007106781006, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7312303781509399, + "num_tokens": 585640191.0, + "step": 22638 + }, + { + "epoch": 2.4861629694706786, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.41450572013855, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7303725481033325, + "num_tokens": 585668203.0, + "step": 22639 + }, + { + "epoch": 2.4862727871732924, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5467374324798584, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7371161580085754, + "num_tokens": 585692082.0, + "step": 22640 + }, + { + "epoch": 2.486382604875906, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4968581199645996, + "learning_rate": 1e-06, + "loss": 0.7366, + "mean_token_accuracy": 0.7578771710395813, + "num_tokens": 585713766.0, + "step": 22641 + }, + { + "epoch": 2.48649242257852, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3355519771575928, + "learning_rate": 1e-06, + "loss": 0.7718, + "mean_token_accuracy": 0.748242974281311, + "num_tokens": 585739024.0, + "step": 22642 + }, + { + "epoch": 2.486602240281133, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1874897480010986, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.723625659942627, + "num_tokens": 585768397.0, + "step": 22643 + }, + { + "epoch": 2.486712057983747, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.553659439086914, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7310640215873718, + "num_tokens": 585790832.0, + "step": 22644 + }, + { + "epoch": 2.4868218756863607, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.382506847381592, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7204645872116089, + "num_tokens": 585816244.0, + "step": 22645 + }, + { + "epoch": 2.4869316933889745, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.342481851577759, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7442992925643921, + "num_tokens": 585841796.0, + "step": 22646 + }, + { + "epoch": 2.4870415110915878, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3265087604522705, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7064450979232788, + "num_tokens": 585868691.0, + "step": 22647 + }, + { + "epoch": 2.4871513287942015, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.249457836151123, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7256028652191162, + "num_tokens": 585896525.0, + "step": 22648 + }, + { + "epoch": 2.4872611464968153, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4318981170654297, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7308983206748962, + "num_tokens": 585922581.0, + "step": 22649 + }, + { + "epoch": 2.487370964199429, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2387869358062744, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7416428327560425, + "num_tokens": 585952214.0, + "step": 22650 + }, + { + "epoch": 2.487480781902043, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.158714771270752, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7333125472068787, + "num_tokens": 585981483.0, + "step": 22651 + }, + { + "epoch": 2.487590599604656, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.324214458465576, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7370102405548096, + "num_tokens": 586006401.0, + "step": 22652 + }, + { + "epoch": 2.48770041730727, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1427252292633057, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7184668779373169, + "num_tokens": 586035868.0, + "step": 22653 + }, + { + "epoch": 2.4878102350098836, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3482439517974854, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.721854567527771, + "num_tokens": 586063210.0, + "step": 22654 + }, + { + "epoch": 2.4879200527124974, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.428924798965454, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7468864321708679, + "num_tokens": 586088749.0, + "step": 22655 + }, + { + "epoch": 2.4880298704151107, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2746808528900146, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7375590801239014, + "num_tokens": 586114721.0, + "step": 22656 + }, + { + "epoch": 2.4881396881177245, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3862361907958984, + "learning_rate": 1e-06, + "loss": 0.7884, + "mean_token_accuracy": 0.7428447604179382, + "num_tokens": 586138740.0, + "step": 22657 + }, + { + "epoch": 2.488249505820338, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.423717498779297, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7231063842773438, + "num_tokens": 586163806.0, + "step": 22658 + }, + { + "epoch": 2.488359323522952, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3852529525756836, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7150638103485107, + "num_tokens": 586189894.0, + "step": 22659 + }, + { + "epoch": 2.4884691412255657, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.045990228652954, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7241005301475525, + "num_tokens": 586218294.0, + "step": 22660 + }, + { + "epoch": 2.488578958928179, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5082592964172363, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7372660636901855, + "num_tokens": 586241428.0, + "step": 22661 + }, + { + "epoch": 2.488688776630793, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.687840461730957, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7485024929046631, + "num_tokens": 586259610.0, + "step": 22662 + }, + { + "epoch": 2.4887985943334066, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.363377571105957, + "learning_rate": 1e-06, + "loss": 0.8992, + "mean_token_accuracy": 0.7165684103965759, + "num_tokens": 586287264.0, + "step": 22663 + }, + { + "epoch": 2.4889084120360203, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.174367666244507, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7439662218093872, + "num_tokens": 586315879.0, + "step": 22664 + }, + { + "epoch": 2.489018229738634, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.7091424465179443, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7162768840789795, + "num_tokens": 586336976.0, + "step": 22665 + }, + { + "epoch": 2.4891280474412474, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.098067045211792, + "learning_rate": 1e-06, + "loss": 0.94, + "mean_token_accuracy": 0.7149204015731812, + "num_tokens": 586369199.0, + "step": 22666 + }, + { + "epoch": 2.489237865143861, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.95559024810791, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7388651371002197, + "num_tokens": 586387568.0, + "step": 22667 + }, + { + "epoch": 2.489347682846475, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2651236057281494, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7227914333343506, + "num_tokens": 586415242.0, + "step": 22668 + }, + { + "epoch": 2.4894575005490887, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3338472843170166, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7331967949867249, + "num_tokens": 586441756.0, + "step": 22669 + }, + { + "epoch": 2.4895673182517024, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2150540351867676, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7396708726882935, + "num_tokens": 586468933.0, + "step": 22670 + }, + { + "epoch": 2.4896771359543157, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.443143606185913, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7198441624641418, + "num_tokens": 586495660.0, + "step": 22671 + }, + { + "epoch": 2.4897869536569295, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3170199394226074, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7215057015419006, + "num_tokens": 586522449.0, + "step": 22672 + }, + { + "epoch": 2.4898967713595432, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2094595432281494, + "learning_rate": 1e-06, + "loss": 0.7685, + "mean_token_accuracy": 0.7633084058761597, + "num_tokens": 586549179.0, + "step": 22673 + }, + { + "epoch": 2.490006589062157, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4535183906555176, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7337050437927246, + "num_tokens": 586574623.0, + "step": 22674 + }, + { + "epoch": 2.4901164067647703, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4843270778656006, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7198992967605591, + "num_tokens": 586600188.0, + "step": 22675 + }, + { + "epoch": 2.490226224467384, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 3.116807222366333, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7310160398483276, + "num_tokens": 586623978.0, + "step": 22676 + }, + { + "epoch": 2.490336042169998, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5549755096435547, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7201846837997437, + "num_tokens": 586647401.0, + "step": 22677 + }, + { + "epoch": 2.4904458598726116, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.7364561557769775, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7289055585861206, + "num_tokens": 586668775.0, + "step": 22678 + }, + { + "epoch": 2.490555677575225, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.381484270095825, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7367786765098572, + "num_tokens": 586694421.0, + "step": 22679 + }, + { + "epoch": 2.4906654952778386, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2701454162597656, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7245194911956787, + "num_tokens": 586721979.0, + "step": 22680 + }, + { + "epoch": 2.4907753129804524, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.339810848236084, + "learning_rate": 1e-06, + "loss": 0.9997, + "mean_token_accuracy": 0.6910573840141296, + "num_tokens": 586752442.0, + "step": 22681 + }, + { + "epoch": 2.490885130683066, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.320673704147339, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.7087231874465942, + "num_tokens": 586780628.0, + "step": 22682 + }, + { + "epoch": 2.49099494838568, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.161048412322998, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7476521730422974, + "num_tokens": 586809136.0, + "step": 22683 + }, + { + "epoch": 2.4911047660882932, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.146498918533325, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7285850048065186, + "num_tokens": 586838550.0, + "step": 22684 + }, + { + "epoch": 2.491214583790907, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.0682756900787354, + "learning_rate": 1e-06, + "loss": 0.9843, + "mean_token_accuracy": 0.6957367658615112, + "num_tokens": 586872032.0, + "step": 22685 + }, + { + "epoch": 2.4913244014935207, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.108205795288086, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.720239520072937, + "num_tokens": 586906818.0, + "step": 22686 + }, + { + "epoch": 2.4914342191961345, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.735957622528076, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7412083148956299, + "num_tokens": 586928153.0, + "step": 22687 + }, + { + "epoch": 2.4915440368987483, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.290347099304199, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7221907377243042, + "num_tokens": 586954338.0, + "step": 22688 + }, + { + "epoch": 2.4916538546013616, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.303956985473633, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7258371114730835, + "num_tokens": 586981709.0, + "step": 22689 + }, + { + "epoch": 2.4917636723039753, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.295222043991089, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7470241785049438, + "num_tokens": 587009605.0, + "step": 22690 + }, + { + "epoch": 2.491873490006589, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4739623069763184, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7461545467376709, + "num_tokens": 587032841.0, + "step": 22691 + }, + { + "epoch": 2.491983307709203, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.583378314971924, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7314162850379944, + "num_tokens": 587056958.0, + "step": 22692 + }, + { + "epoch": 2.4920931254118166, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3038480281829834, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7347425818443298, + "num_tokens": 587083782.0, + "step": 22693 + }, + { + "epoch": 2.49220294311443, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3877997398376465, + "learning_rate": 1e-06, + "loss": 0.9352, + "mean_token_accuracy": 0.7115119695663452, + "num_tokens": 587110823.0, + "step": 22694 + }, + { + "epoch": 2.4923127608170437, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.148625612258911, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7144638299942017, + "num_tokens": 587140248.0, + "step": 22695 + }, + { + "epoch": 2.4924225785196574, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.297657012939453, + "learning_rate": 1e-06, + "loss": 0.9455, + "mean_token_accuracy": 0.7068482637405396, + "num_tokens": 587169688.0, + "step": 22696 + }, + { + "epoch": 2.492532396222271, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.572740316390991, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7421725988388062, + "num_tokens": 587192285.0, + "step": 22697 + }, + { + "epoch": 2.4926422139248845, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5210349559783936, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.74465012550354, + "num_tokens": 587213982.0, + "step": 22698 + }, + { + "epoch": 2.4927520316274983, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2358102798461914, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7327209711074829, + "num_tokens": 587241666.0, + "step": 22699 + }, + { + "epoch": 2.492861849330112, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.409914970397949, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7404861450195312, + "num_tokens": 587265561.0, + "step": 22700 + }, + { + "epoch": 2.4929716670327258, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2607173919677734, + "learning_rate": 1e-06, + "loss": 0.9298, + "mean_token_accuracy": 0.7116143107414246, + "num_tokens": 587295396.0, + "step": 22701 + }, + { + "epoch": 2.4930814847353395, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3329849243164062, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7535656690597534, + "num_tokens": 587321031.0, + "step": 22702 + }, + { + "epoch": 2.493191302437953, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.874256134033203, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.743722677230835, + "num_tokens": 587339850.0, + "step": 22703 + }, + { + "epoch": 2.4933011201405666, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1040873527526855, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.7038781642913818, + "num_tokens": 587369105.0, + "step": 22704 + }, + { + "epoch": 2.4934109378431804, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3662943840026855, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.713654637336731, + "num_tokens": 587397579.0, + "step": 22705 + }, + { + "epoch": 2.493520755545794, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.380992889404297, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7343124747276306, + "num_tokens": 587423378.0, + "step": 22706 + }, + { + "epoch": 2.4936305732484074, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.541091203689575, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7343350648880005, + "num_tokens": 587446342.0, + "step": 22707 + }, + { + "epoch": 2.493740390951021, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.08895206451416, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7233628034591675, + "num_tokens": 587476983.0, + "step": 22708 + }, + { + "epoch": 2.493850208653635, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.8129491806030273, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7337087988853455, + "num_tokens": 587501236.0, + "step": 22709 + }, + { + "epoch": 2.4939600263562487, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.7639405727386475, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.734831690788269, + "num_tokens": 587521020.0, + "step": 22710 + }, + { + "epoch": 2.4940698440588625, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.272671937942505, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7311596870422363, + "num_tokens": 587547275.0, + "step": 22711 + }, + { + "epoch": 2.4941796617614758, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.431701183319092, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7407688498497009, + "num_tokens": 587573084.0, + "step": 22712 + }, + { + "epoch": 2.4942894794640895, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.234029769897461, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7221788167953491, + "num_tokens": 587603257.0, + "step": 22713 + }, + { + "epoch": 2.4943992971667033, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.302863359451294, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7239642143249512, + "num_tokens": 587630187.0, + "step": 22714 + }, + { + "epoch": 2.494509114869317, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.726907730102539, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7366739511489868, + "num_tokens": 587649842.0, + "step": 22715 + }, + { + "epoch": 2.494618932571931, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3540008068084717, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7258846759796143, + "num_tokens": 587675550.0, + "step": 22716 + }, + { + "epoch": 2.494728750274544, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.477966785430908, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7363470196723938, + "num_tokens": 587698137.0, + "step": 22717 + }, + { + "epoch": 2.494838567977158, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2961018085479736, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7353737354278564, + "num_tokens": 587726673.0, + "step": 22718 + }, + { + "epoch": 2.4949483856797716, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.1220316886901855, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7328987121582031, + "num_tokens": 587756002.0, + "step": 22719 + }, + { + "epoch": 2.4950582033823854, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.317120313644409, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7177718877792358, + "num_tokens": 587784302.0, + "step": 22720 + }, + { + "epoch": 2.495168021084999, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2452712059020996, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.728508472442627, + "num_tokens": 587813162.0, + "step": 22721 + }, + { + "epoch": 2.4952778387876124, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.100186347961426, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7434810996055603, + "num_tokens": 587844242.0, + "step": 22722 + }, + { + "epoch": 2.495387656490226, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.2296714782714844, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7172424793243408, + "num_tokens": 587872383.0, + "step": 22723 + }, + { + "epoch": 2.49549747419284, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.4649405479431152, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7296032309532166, + "num_tokens": 587895424.0, + "step": 22724 + }, + { + "epoch": 2.4956072918954537, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.035311698913574, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.728615403175354, + "num_tokens": 587929377.0, + "step": 22725 + }, + { + "epoch": 2.495717109598067, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.484562397003174, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7393261790275574, + "num_tokens": 587952057.0, + "step": 22726 + }, + { + "epoch": 2.495826927300681, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.5337295532226562, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.7561602592468262, + "num_tokens": 587973851.0, + "step": 22727 + }, + { + "epoch": 2.4959367450032945, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.3981199264526367, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7343082427978516, + "num_tokens": 588000156.0, + "step": 22728 + }, + { + "epoch": 2.4960465627059083, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.216452121734619, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7317134141921997, + "num_tokens": 588027864.0, + "step": 22729 + }, + { + "epoch": 2.4961563804085216, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.494460344314575, + "learning_rate": 1e-06, + "loss": 0.9102, + "mean_token_accuracy": 0.7060722708702087, + "num_tokens": 588053574.0, + "step": 22730 + }, + { + "epoch": 2.4962661981111354, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.074134349822998, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7431941032409668, + "num_tokens": 588083717.0, + "step": 22731 + }, + { + "epoch": 2.496376015813749, + "ewc_loss": 1.4722347259521484e-05, + "grad_norm": 2.687865972518921, + "learning_rate": 1e-06, + "loss": 0.8206, + "mean_token_accuracy": 0.7355247735977173, + "num_tokens": 588106376.0, + "step": 22732 + }, + { + "epoch": 2.496485833516363, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3078951835632324, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7337529063224792, + "num_tokens": 588134677.0, + "step": 22733 + }, + { + "epoch": 2.4965956512189766, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.140974521636963, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7214410305023193, + "num_tokens": 588163054.0, + "step": 22734 + }, + { + "epoch": 2.49670546892159, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.192167282104492, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7255291938781738, + "num_tokens": 588192624.0, + "step": 22735 + }, + { + "epoch": 2.4968152866242037, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3870246410369873, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.728489875793457, + "num_tokens": 588217520.0, + "step": 22736 + }, + { + "epoch": 2.4969251043268175, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.258033037185669, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7181441783905029, + "num_tokens": 588246847.0, + "step": 22737 + }, + { + "epoch": 2.4970349220294312, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.540134906768799, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7275948524475098, + "num_tokens": 588271089.0, + "step": 22738 + }, + { + "epoch": 2.497144739732045, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3543052673339844, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7303462624549866, + "num_tokens": 588296509.0, + "step": 22739 + }, + { + "epoch": 2.4972545574346583, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.574465036392212, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7289654016494751, + "num_tokens": 588320644.0, + "step": 22740 + }, + { + "epoch": 2.497364375137272, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.120368242263794, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7377495169639587, + "num_tokens": 588351365.0, + "step": 22741 + }, + { + "epoch": 2.497474192839886, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.4447312355041504, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7385141849517822, + "num_tokens": 588374301.0, + "step": 22742 + }, + { + "epoch": 2.4975840105424996, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.612626314163208, + "learning_rate": 1e-06, + "loss": 0.8258, + "mean_token_accuracy": 0.7378634214401245, + "num_tokens": 588394979.0, + "step": 22743 + }, + { + "epoch": 2.4976938282451133, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.5500426292419434, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7281463146209717, + "num_tokens": 588418678.0, + "step": 22744 + }, + { + "epoch": 2.4978036459477266, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.4221737384796143, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7252996563911438, + "num_tokens": 588442045.0, + "step": 22745 + }, + { + "epoch": 2.4979134636503404, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.6969993114471436, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7377625703811646, + "num_tokens": 588464154.0, + "step": 22746 + }, + { + "epoch": 2.498023281352954, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3295140266418457, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7451343536376953, + "num_tokens": 588490177.0, + "step": 22747 + }, + { + "epoch": 2.498133099055568, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.4543063640594482, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.721819281578064, + "num_tokens": 588514830.0, + "step": 22748 + }, + { + "epoch": 2.4982429167581817, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.8411636352539062, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7459564208984375, + "num_tokens": 588533952.0, + "step": 22749 + }, + { + "epoch": 2.498352734460795, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.810542583465576, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7319557666778564, + "num_tokens": 588553512.0, + "step": 22750 + }, + { + "epoch": 2.4984625521634087, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.871725559234619, + "learning_rate": 1e-06, + "loss": 0.7288, + "mean_token_accuracy": 0.7639291286468506, + "num_tokens": 588570616.0, + "step": 22751 + }, + { + "epoch": 2.4985723698660225, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.441896915435791, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7218784689903259, + "num_tokens": 588595651.0, + "step": 22752 + }, + { + "epoch": 2.4986821875686362, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.654526710510254, + "learning_rate": 1e-06, + "loss": 0.7735, + "mean_token_accuracy": 0.7487500905990601, + "num_tokens": 588617392.0, + "step": 22753 + }, + { + "epoch": 2.4987920052712496, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.103572368621826, + "learning_rate": 1e-06, + "loss": 0.9671, + "mean_token_accuracy": 0.6996214389801025, + "num_tokens": 588649690.0, + "step": 22754 + }, + { + "epoch": 2.4989018229738633, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.246607780456543, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7207487225532532, + "num_tokens": 588680395.0, + "step": 22755 + }, + { + "epoch": 2.499011640676477, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 1.9849004745483398, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7152324318885803, + "num_tokens": 588715486.0, + "step": 22756 + }, + { + "epoch": 2.499121458379091, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4537973403930664, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.722685694694519, + "num_tokens": 588741555.0, + "step": 22757 + }, + { + "epoch": 2.499231276081704, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.1682322025299072, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7322777509689331, + "num_tokens": 588771413.0, + "step": 22758 + }, + { + "epoch": 2.499341093784318, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.6065025329589844, + "learning_rate": 1e-06, + "loss": 0.7602, + "mean_token_accuracy": 0.7554688453674316, + "num_tokens": 588791492.0, + "step": 22759 + }, + { + "epoch": 2.4994509114869317, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.9684693813323975, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7399917840957642, + "num_tokens": 588811461.0, + "step": 22760 + }, + { + "epoch": 2.4995607291895454, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.4076545238494873, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7200942039489746, + "num_tokens": 588840279.0, + "step": 22761 + }, + { + "epoch": 2.499670546892159, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.320035457611084, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7338172197341919, + "num_tokens": 588866667.0, + "step": 22762 + }, + { + "epoch": 2.4997803645947725, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.251603126525879, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7231043577194214, + "num_tokens": 588894255.0, + "step": 22763 + }, + { + "epoch": 2.4998901822973862, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3813743591308594, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.73017418384552, + "num_tokens": 588921334.0, + "step": 22764 + }, + { + "epoch": 2.5, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.6535544395446777, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7261004447937012, + "num_tokens": 588942704.0, + "step": 22765 + }, + { + "epoch": 2.5001098177026138, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.1901378631591797, + "learning_rate": 1e-06, + "loss": 0.8754, + "mean_token_accuracy": 0.7200711965560913, + "num_tokens": 588971823.0, + "step": 22766 + }, + { + "epoch": 2.5002196354052275, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.6766717433929443, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7467969655990601, + "num_tokens": 588993280.0, + "step": 22767 + }, + { + "epoch": 2.500329453107841, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.5973434448242188, + "learning_rate": 1e-06, + "loss": 0.9299, + "mean_token_accuracy": 0.7174288034439087, + "num_tokens": 589016288.0, + "step": 22768 + }, + { + "epoch": 2.5004392708104546, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3104474544525146, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7348867058753967, + "num_tokens": 589042750.0, + "step": 22769 + }, + { + "epoch": 2.5005490885130683, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.421476125717163, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7166882753372192, + "num_tokens": 589068047.0, + "step": 22770 + }, + { + "epoch": 2.500658906215682, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.49554705619812, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.741142749786377, + "num_tokens": 589090217.0, + "step": 22771 + }, + { + "epoch": 2.500768723918296, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.2680323123931885, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7292598485946655, + "num_tokens": 589120073.0, + "step": 22772 + }, + { + "epoch": 2.500878541620909, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.38787841796875, + "learning_rate": 1e-06, + "loss": 0.7541, + "mean_token_accuracy": 0.7551996111869812, + "num_tokens": 589144251.0, + "step": 22773 + }, + { + "epoch": 2.500988359323523, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.41520357131958, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.731019139289856, + "num_tokens": 589168790.0, + "step": 22774 + }, + { + "epoch": 2.5010981770261367, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.393425226211548, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7413045167922974, + "num_tokens": 589193331.0, + "step": 22775 + }, + { + "epoch": 2.5012079947287504, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.2672340869903564, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7286131381988525, + "num_tokens": 589221114.0, + "step": 22776 + }, + { + "epoch": 2.501317812431364, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3207521438598633, + "learning_rate": 1e-06, + "loss": 0.7298, + "mean_token_accuracy": 0.7622484564781189, + "num_tokens": 589246004.0, + "step": 22777 + }, + { + "epoch": 2.5014276301339775, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.8924288749694824, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7350954413414001, + "num_tokens": 589265631.0, + "step": 22778 + }, + { + "epoch": 2.5015374478365913, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.521552562713623, + "learning_rate": 1e-06, + "loss": 0.792, + "mean_token_accuracy": 0.7552067637443542, + "num_tokens": 589287728.0, + "step": 22779 + }, + { + "epoch": 2.501647265539205, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.299975633621216, + "learning_rate": 1e-06, + "loss": 0.8371, + "mean_token_accuracy": 0.7311027646064758, + "num_tokens": 589313655.0, + "step": 22780 + }, + { + "epoch": 2.5017570832418183, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.46899151802063, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7328750491142273, + "num_tokens": 589337141.0, + "step": 22781 + }, + { + "epoch": 2.501866900944432, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.4844257831573486, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7325296998023987, + "num_tokens": 589359967.0, + "step": 22782 + }, + { + "epoch": 2.501976718647046, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.195331573486328, + "learning_rate": 1e-06, + "loss": 0.7559, + "mean_token_accuracy": 0.7575801014900208, + "num_tokens": 589387337.0, + "step": 22783 + }, + { + "epoch": 2.5020865363496596, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.3038082122802734, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7282886505126953, + "num_tokens": 589414927.0, + "step": 22784 + }, + { + "epoch": 2.5021963540522734, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.7225327491760254, + "learning_rate": 1e-06, + "loss": 0.849, + "mean_token_accuracy": 0.7425708770751953, + "num_tokens": 589436169.0, + "step": 22785 + }, + { + "epoch": 2.5023061717548867, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.188802480697632, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7263834476470947, + "num_tokens": 589464166.0, + "step": 22786 + }, + { + "epoch": 2.5024159894575004, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.167207717895508, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7134624719619751, + "num_tokens": 589495499.0, + "step": 22787 + }, + { + "epoch": 2.502525807160114, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.274430513381958, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7252432107925415, + "num_tokens": 589522298.0, + "step": 22788 + }, + { + "epoch": 2.502635624862728, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.7760653495788574, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7415522336959839, + "num_tokens": 589542211.0, + "step": 22789 + }, + { + "epoch": 2.5027454425653417, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.829951524734497, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7130315899848938, + "num_tokens": 589563880.0, + "step": 22790 + }, + { + "epoch": 2.502855260267955, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.240306854248047, + "learning_rate": 1e-06, + "loss": 0.6943, + "mean_token_accuracy": 0.7728196382522583, + "num_tokens": 589590574.0, + "step": 22791 + }, + { + "epoch": 2.5029650779705688, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.267854928970337, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7382350564002991, + "num_tokens": 589619617.0, + "step": 22792 + }, + { + "epoch": 2.5030748956731825, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.7887682914733887, + "learning_rate": 1e-06, + "loss": 0.8756, + "mean_token_accuracy": 0.7217416167259216, + "num_tokens": 589640033.0, + "step": 22793 + }, + { + "epoch": 2.5031847133757963, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.2752466201782227, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7451049089431763, + "num_tokens": 589665609.0, + "step": 22794 + }, + { + "epoch": 2.50329453107841, + "ewc_loss": 1.4781951904296875e-05, + "grad_norm": 2.5665581226348877, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7400643825531006, + "num_tokens": 589687734.0, + "step": 22795 + }, + { + "epoch": 2.5034043487810234, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.621922731399536, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7602675557136536, + "num_tokens": 589707580.0, + "step": 22796 + }, + { + "epoch": 2.503514166483637, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.281062602996826, + "learning_rate": 1e-06, + "loss": 0.7415, + "mean_token_accuracy": 0.7636972665786743, + "num_tokens": 589734307.0, + "step": 22797 + }, + { + "epoch": 2.503623984186251, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.606674909591675, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.738570511341095, + "num_tokens": 589757196.0, + "step": 22798 + }, + { + "epoch": 2.5037338018888646, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.372835636138916, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7430920004844666, + "num_tokens": 589780928.0, + "step": 22799 + }, + { + "epoch": 2.5038436195914784, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2447702884674072, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7387316226959229, + "num_tokens": 589810612.0, + "step": 22800 + }, + { + "epoch": 2.5039534372940917, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.7022223472595215, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7292654514312744, + "num_tokens": 589831931.0, + "step": 22801 + }, + { + "epoch": 2.5040632549967055, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2058961391448975, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7239115834236145, + "num_tokens": 589861413.0, + "step": 22802 + }, + { + "epoch": 2.504173072699319, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.1848573684692383, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7262459397315979, + "num_tokens": 589891007.0, + "step": 22803 + }, + { + "epoch": 2.5042828904019325, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2496397495269775, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7250295877456665, + "num_tokens": 589918620.0, + "step": 22804 + }, + { + "epoch": 2.5043927081045467, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.317662000656128, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.731614351272583, + "num_tokens": 589946586.0, + "step": 22805 + }, + { + "epoch": 2.50450252580716, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.244229555130005, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.739054799079895, + "num_tokens": 589974246.0, + "step": 22806 + }, + { + "epoch": 2.504612343509774, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.302309989929199, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7257169485092163, + "num_tokens": 590002770.0, + "step": 22807 + }, + { + "epoch": 2.5047221612123876, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4767587184906006, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7287527322769165, + "num_tokens": 590027131.0, + "step": 22808 + }, + { + "epoch": 2.504831978915001, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.8400375843048096, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7354066371917725, + "num_tokens": 590046283.0, + "step": 22809 + }, + { + "epoch": 2.5049417966176146, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.045340061187744, + "learning_rate": 1e-06, + "loss": 0.9325, + "mean_token_accuracy": 0.7067877650260925, + "num_tokens": 590076302.0, + "step": 22810 + }, + { + "epoch": 2.5050516143202284, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2965211868286133, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7167112827301025, + "num_tokens": 590102808.0, + "step": 22811 + }, + { + "epoch": 2.505161432022842, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.3477022647857666, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7272046804428101, + "num_tokens": 590128825.0, + "step": 22812 + }, + { + "epoch": 2.505271249725456, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.3107409477233887, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7457323670387268, + "num_tokens": 590153176.0, + "step": 22813 + }, + { + "epoch": 2.505381067428069, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.343275785446167, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7317889928817749, + "num_tokens": 590179251.0, + "step": 22814 + }, + { + "epoch": 2.505490885130683, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.563758134841919, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7394076585769653, + "num_tokens": 590199334.0, + "step": 22815 + }, + { + "epoch": 2.5056007028332967, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.095057964324951, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.7307885885238647, + "num_tokens": 590230521.0, + "step": 22816 + }, + { + "epoch": 2.5057105205359105, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4158830642700195, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7183066010475159, + "num_tokens": 590256396.0, + "step": 22817 + }, + { + "epoch": 2.5058203382385242, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4598240852355957, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7321841716766357, + "num_tokens": 590280641.0, + "step": 22818 + }, + { + "epoch": 2.5059301559411375, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.3404014110565186, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7181421518325806, + "num_tokens": 590307933.0, + "step": 22819 + }, + { + "epoch": 2.5060399736437513, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5756146907806396, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.721049427986145, + "num_tokens": 590328960.0, + "step": 22820 + }, + { + "epoch": 2.506149791346365, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4085140228271484, + "learning_rate": 1e-06, + "loss": 0.7704, + "mean_token_accuracy": 0.7492676377296448, + "num_tokens": 590350891.0, + "step": 22821 + }, + { + "epoch": 2.506259609048979, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.6763088703155518, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7463662624359131, + "num_tokens": 590371449.0, + "step": 22822 + }, + { + "epoch": 2.5063694267515926, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5729494094848633, + "learning_rate": 1e-06, + "loss": 0.7199, + "mean_token_accuracy": 0.7639001607894897, + "num_tokens": 590393247.0, + "step": 22823 + }, + { + "epoch": 2.506479244454206, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.533818006515503, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7325179576873779, + "num_tokens": 590416572.0, + "step": 22824 + }, + { + "epoch": 2.5065890621568196, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5448861122131348, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7435342073440552, + "num_tokens": 590442330.0, + "step": 22825 + }, + { + "epoch": 2.5066988798594334, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2468528747558594, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7101573944091797, + "num_tokens": 590474719.0, + "step": 22826 + }, + { + "epoch": 2.506808697562047, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.0945425033569336, + "learning_rate": 1e-06, + "loss": 0.7199, + "mean_token_accuracy": 0.7681995630264282, + "num_tokens": 590502613.0, + "step": 22827 + }, + { + "epoch": 2.506918515264661, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.311150312423706, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7357711791992188, + "num_tokens": 590528490.0, + "step": 22828 + }, + { + "epoch": 2.5070283329672742, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.314056634902954, + "learning_rate": 1e-06, + "loss": 0.9677, + "mean_token_accuracy": 0.7090900540351868, + "num_tokens": 590556084.0, + "step": 22829 + }, + { + "epoch": 2.507138150669888, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.600919723510742, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7294962406158447, + "num_tokens": 590579798.0, + "step": 22830 + }, + { + "epoch": 2.5072479683725017, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.596862316131592, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7371399402618408, + "num_tokens": 590601789.0, + "step": 22831 + }, + { + "epoch": 2.507357786075115, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.6004629135131836, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.731930136680603, + "num_tokens": 590626173.0, + "step": 22832 + }, + { + "epoch": 2.507467603777729, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.3847005367279053, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7244278192520142, + "num_tokens": 590653382.0, + "step": 22833 + }, + { + "epoch": 2.5075774214803426, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.262941837310791, + "learning_rate": 1e-06, + "loss": 0.7672, + "mean_token_accuracy": 0.7525149583816528, + "num_tokens": 590681835.0, + "step": 22834 + }, + { + "epoch": 2.5076872391829563, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5802364349365234, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.741233229637146, + "num_tokens": 590703115.0, + "step": 22835 + }, + { + "epoch": 2.50779705688557, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.285860300064087, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7433655858039856, + "num_tokens": 590729143.0, + "step": 22836 + }, + { + "epoch": 2.5079068745881834, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.277766466140747, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7159384489059448, + "num_tokens": 590757784.0, + "step": 22837 + }, + { + "epoch": 2.508016692290797, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.9261298179626465, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7462894916534424, + "num_tokens": 590777166.0, + "step": 22838 + }, + { + "epoch": 2.508126509993411, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.4659459590911865, + "learning_rate": 1e-06, + "loss": 0.7489, + "mean_token_accuracy": 0.7607550621032715, + "num_tokens": 590801084.0, + "step": 22839 + }, + { + "epoch": 2.5082363276960247, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.6067299842834473, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7377833127975464, + "num_tokens": 590823667.0, + "step": 22840 + }, + { + "epoch": 2.5083461453986384, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2269184589385986, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7075402140617371, + "num_tokens": 590853423.0, + "step": 22841 + }, + { + "epoch": 2.5084559631012517, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.6190810203552246, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7359697818756104, + "num_tokens": 590876373.0, + "step": 22842 + }, + { + "epoch": 2.5085657808038655, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.426374912261963, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7308272123336792, + "num_tokens": 590900751.0, + "step": 22843 + }, + { + "epoch": 2.5086755985064793, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2219245433807373, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7231281399726868, + "num_tokens": 590929624.0, + "step": 22844 + }, + { + "epoch": 2.508785416209093, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.266660213470459, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7280231714248657, + "num_tokens": 590957559.0, + "step": 22845 + }, + { + "epoch": 2.5088952339117068, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.7713546752929688, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7379042506217957, + "num_tokens": 590977189.0, + "step": 22846 + }, + { + "epoch": 2.50900505161432, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.1942954063415527, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7373164296150208, + "num_tokens": 591006782.0, + "step": 22847 + }, + { + "epoch": 2.509114869316934, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.444746971130371, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.7425391674041748, + "num_tokens": 591032821.0, + "step": 22848 + }, + { + "epoch": 2.5092246870195476, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.3414177894592285, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7406121492385864, + "num_tokens": 591063078.0, + "step": 22849 + }, + { + "epoch": 2.5093345047221614, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.2321829795837402, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7262305617332458, + "num_tokens": 591093057.0, + "step": 22850 + }, + { + "epoch": 2.509444322424775, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.369483709335327, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.7500882148742676, + "num_tokens": 591117707.0, + "step": 22851 + }, + { + "epoch": 2.5095541401273884, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.7241814136505127, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7385338544845581, + "num_tokens": 591139261.0, + "step": 22852 + }, + { + "epoch": 2.509663957830002, + "ewc_loss": 1.4901161193847656e-05, + "grad_norm": 2.5155446529388428, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.7355297207832336, + "num_tokens": 591163001.0, + "step": 22853 + }, + { + "epoch": 2.509773775532616, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2231338024139404, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7377895712852478, + "num_tokens": 591191088.0, + "step": 22854 + }, + { + "epoch": 2.5098835932352292, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3624401092529297, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7070251107215881, + "num_tokens": 591217394.0, + "step": 22855 + }, + { + "epoch": 2.5099934109378434, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.6447336673736572, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7448652982711792, + "num_tokens": 591238614.0, + "step": 22856 + }, + { + "epoch": 2.5101032286404568, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4917473793029785, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7272679805755615, + "num_tokens": 591264604.0, + "step": 22857 + }, + { + "epoch": 2.5102130463430705, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.5626728534698486, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7530184984207153, + "num_tokens": 591287228.0, + "step": 22858 + }, + { + "epoch": 2.5103228640456843, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3504464626312256, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.722520112991333, + "num_tokens": 591312753.0, + "step": 22859 + }, + { + "epoch": 2.5104326817482976, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.411147356033325, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7293003797531128, + "num_tokens": 591338562.0, + "step": 22860 + }, + { + "epoch": 2.5105424994509113, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2375357151031494, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7050818204879761, + "num_tokens": 591366615.0, + "step": 22861 + }, + { + "epoch": 2.510652317153525, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4831275939941406, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7248170375823975, + "num_tokens": 591390433.0, + "step": 22862 + }, + { + "epoch": 2.510762134856139, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.467284917831421, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7209031581878662, + "num_tokens": 591412696.0, + "step": 22863 + }, + { + "epoch": 2.5108719525587526, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2764041423797607, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7360376119613647, + "num_tokens": 591439559.0, + "step": 22864 + }, + { + "epoch": 2.510981770261366, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.6001224517822266, + "learning_rate": 1e-06, + "loss": 0.7192, + "mean_token_accuracy": 0.7646785974502563, + "num_tokens": 591459784.0, + "step": 22865 + }, + { + "epoch": 2.5110915879639797, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.451707601547241, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.7355719208717346, + "num_tokens": 591483730.0, + "step": 22866 + }, + { + "epoch": 2.5112014056665934, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.453439950942993, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.735518217086792, + "num_tokens": 591509318.0, + "step": 22867 + }, + { + "epoch": 2.511311223369207, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.5706348419189453, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7353678941726685, + "num_tokens": 591532774.0, + "step": 22868 + }, + { + "epoch": 2.511421041071821, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.647547721862793, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7247322797775269, + "num_tokens": 591553811.0, + "step": 22869 + }, + { + "epoch": 2.5115308587744343, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.0228402614593506, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7150587439537048, + "num_tokens": 591587246.0, + "step": 22870 + }, + { + "epoch": 2.511640676477048, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3176538944244385, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7067270874977112, + "num_tokens": 591616083.0, + "step": 22871 + }, + { + "epoch": 2.511750494179662, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.24180269241333, + "learning_rate": 1e-06, + "loss": 0.7614, + "mean_token_accuracy": 0.7558280825614929, + "num_tokens": 591642943.0, + "step": 22872 + }, + { + "epoch": 2.5118603118822755, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.5492920875549316, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7517034411430359, + "num_tokens": 591668235.0, + "step": 22873 + }, + { + "epoch": 2.5119701295848893, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.207858085632324, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7264307141304016, + "num_tokens": 591696854.0, + "step": 22874 + }, + { + "epoch": 2.5120799472875026, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.391268730163574, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7232028245925903, + "num_tokens": 591722220.0, + "step": 22875 + }, + { + "epoch": 2.5121897649901164, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3066728115081787, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7175167798995972, + "num_tokens": 591749674.0, + "step": 22876 + }, + { + "epoch": 2.51229958269273, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2970969676971436, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.7562546133995056, + "num_tokens": 591777655.0, + "step": 22877 + }, + { + "epoch": 2.512409400395344, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.116825819015503, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7280524969100952, + "num_tokens": 591810654.0, + "step": 22878 + }, + { + "epoch": 2.5125192180979576, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.260732650756836, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7321364879608154, + "num_tokens": 591838576.0, + "step": 22879 + }, + { + "epoch": 2.512629035800571, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.423445701599121, + "learning_rate": 1e-06, + "loss": 0.7432, + "mean_token_accuracy": 0.7590762376785278, + "num_tokens": 591860586.0, + "step": 22880 + }, + { + "epoch": 2.5127388535031847, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4787838459014893, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7394634485244751, + "num_tokens": 591885342.0, + "step": 22881 + }, + { + "epoch": 2.5128486712057985, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.610243320465088, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7438139319419861, + "num_tokens": 591907658.0, + "step": 22882 + }, + { + "epoch": 2.512958488908412, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.371565580368042, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7308608293533325, + "num_tokens": 591933944.0, + "step": 22883 + }, + { + "epoch": 2.5130683066110255, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.1157925128936768, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7238317131996155, + "num_tokens": 591965979.0, + "step": 22884 + }, + { + "epoch": 2.5131781243136393, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.7815845012664795, + "learning_rate": 1e-06, + "loss": 0.7581, + "mean_token_accuracy": 0.7519150972366333, + "num_tokens": 591985122.0, + "step": 22885 + }, + { + "epoch": 2.513287942016253, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.5121958255767822, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7310467958450317, + "num_tokens": 592009719.0, + "step": 22886 + }, + { + "epoch": 2.513397759718867, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3269643783569336, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7327600121498108, + "num_tokens": 592036278.0, + "step": 22887 + }, + { + "epoch": 2.51350757742148, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.25653338432312, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7229881882667542, + "num_tokens": 592062284.0, + "step": 22888 + }, + { + "epoch": 2.513617395124094, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 1.8912367820739746, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7184740304946899, + "num_tokens": 592098411.0, + "step": 22889 + }, + { + "epoch": 2.5137272128267076, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.437718629837036, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7408989071846008, + "num_tokens": 592124820.0, + "step": 22890 + }, + { + "epoch": 2.5138370305293214, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.850553035736084, + "learning_rate": 1e-06, + "loss": 0.7722, + "mean_token_accuracy": 0.7440860867500305, + "num_tokens": 592143372.0, + "step": 22891 + }, + { + "epoch": 2.513946848231935, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2360470294952393, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7274124622344971, + "num_tokens": 592171269.0, + "step": 22892 + }, + { + "epoch": 2.5140566659345485, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.441835641860962, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7264463305473328, + "num_tokens": 592196793.0, + "step": 22893 + }, + { + "epoch": 2.514166483637162, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3308112621307373, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7393569946289062, + "num_tokens": 592221733.0, + "step": 22894 + }, + { + "epoch": 2.514276301339776, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4080922603607178, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7202355861663818, + "num_tokens": 592247370.0, + "step": 22895 + }, + { + "epoch": 2.5143861190423897, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.1581850051879883, + "learning_rate": 1e-06, + "loss": 0.941, + "mean_token_accuracy": 0.705226480960846, + "num_tokens": 592276498.0, + "step": 22896 + }, + { + "epoch": 2.5144959367450035, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.364220380783081, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7334703803062439, + "num_tokens": 592300544.0, + "step": 22897 + }, + { + "epoch": 2.514605754447617, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.200289726257324, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.712585985660553, + "num_tokens": 592330526.0, + "step": 22898 + }, + { + "epoch": 2.5147155721502306, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2563459873199463, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7395752668380737, + "num_tokens": 592358300.0, + "step": 22899 + }, + { + "epoch": 2.5148253898528443, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.212383270263672, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7358270287513733, + "num_tokens": 592385102.0, + "step": 22900 + }, + { + "epoch": 2.514935207555458, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.1951379776000977, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7391247749328613, + "num_tokens": 592415240.0, + "step": 22901 + }, + { + "epoch": 2.515045025258072, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.887256383895874, + "learning_rate": 1e-06, + "loss": 0.7495, + "mean_token_accuracy": 0.7506698369979858, + "num_tokens": 592432834.0, + "step": 22902 + }, + { + "epoch": 2.515154842960685, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4045114517211914, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7487067580223083, + "num_tokens": 592457682.0, + "step": 22903 + }, + { + "epoch": 2.515264660663299, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3814697265625, + "learning_rate": 1e-06, + "loss": 0.7298, + "mean_token_accuracy": 0.7596268653869629, + "num_tokens": 592481379.0, + "step": 22904 + }, + { + "epoch": 2.5153744783659127, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.094092607498169, + "learning_rate": 1e-06, + "loss": 0.7677, + "mean_token_accuracy": 0.7529588937759399, + "num_tokens": 592511498.0, + "step": 22905 + }, + { + "epoch": 2.515484296068526, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.5394983291625977, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7290530204772949, + "num_tokens": 592535201.0, + "step": 22906 + }, + { + "epoch": 2.51559411377114, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 1.9515209197998047, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.726677417755127, + "num_tokens": 592569210.0, + "step": 22907 + }, + { + "epoch": 2.5157039314737535, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 1.8954263925552368, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7141901850700378, + "num_tokens": 592606097.0, + "step": 22908 + }, + { + "epoch": 2.5158137491763672, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.87705397605896, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7335830330848694, + "num_tokens": 592625761.0, + "step": 22909 + }, + { + "epoch": 2.515923566878981, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.411414623260498, + "learning_rate": 1e-06, + "loss": 0.7758, + "mean_token_accuracy": 0.7528847455978394, + "num_tokens": 592649864.0, + "step": 22910 + }, + { + "epoch": 2.5160333845815943, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4187421798706055, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.727499783039093, + "num_tokens": 592674280.0, + "step": 22911 + }, + { + "epoch": 2.516143202284208, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.595048427581787, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7163240909576416, + "num_tokens": 592696943.0, + "step": 22912 + }, + { + "epoch": 2.516253019986822, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.386368989944458, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7356293201446533, + "num_tokens": 592722978.0, + "step": 22913 + }, + { + "epoch": 2.5163628376894356, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.485762119293213, + "learning_rate": 1e-06, + "loss": 0.7832, + "mean_token_accuracy": 0.7504086494445801, + "num_tokens": 592746688.0, + "step": 22914 + }, + { + "epoch": 2.5164726553920493, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2152695655822754, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7314620018005371, + "num_tokens": 592776387.0, + "step": 22915 + }, + { + "epoch": 2.5165824730946627, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4305286407470703, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7224760055541992, + "num_tokens": 592800075.0, + "step": 22916 + }, + { + "epoch": 2.5166922907972764, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.254772663116455, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7510505318641663, + "num_tokens": 592826613.0, + "step": 22917 + }, + { + "epoch": 2.51680210849989, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.333136796951294, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7384566068649292, + "num_tokens": 592851609.0, + "step": 22918 + }, + { + "epoch": 2.516911926202504, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3144311904907227, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7274700403213501, + "num_tokens": 592879143.0, + "step": 22919 + }, + { + "epoch": 2.5170217439051177, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.319946527481079, + "learning_rate": 1e-06, + "loss": 0.8402, + "mean_token_accuracy": 0.7354410886764526, + "num_tokens": 592905359.0, + "step": 22920 + }, + { + "epoch": 2.517131561607731, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.1903786659240723, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7193236351013184, + "num_tokens": 592934634.0, + "step": 22921 + }, + { + "epoch": 2.5172413793103448, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.368023157119751, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7434650659561157, + "num_tokens": 592961039.0, + "step": 22922 + }, + { + "epoch": 2.5173511970129585, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3513271808624268, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7212311029434204, + "num_tokens": 592988378.0, + "step": 22923 + }, + { + "epoch": 2.5174610147155723, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2475686073303223, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7210109233856201, + "num_tokens": 593017157.0, + "step": 22924 + }, + { + "epoch": 2.517570832418186, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2005579471588135, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7048367857933044, + "num_tokens": 593052808.0, + "step": 22925 + }, + { + "epoch": 2.5176806501207993, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4564690589904785, + "learning_rate": 1e-06, + "loss": 0.7827, + "mean_token_accuracy": 0.7535773515701294, + "num_tokens": 593077480.0, + "step": 22926 + }, + { + "epoch": 2.517790467823413, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.027501344680786, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7359107732772827, + "num_tokens": 593111996.0, + "step": 22927 + }, + { + "epoch": 2.517900285526027, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.261664390563965, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7087008953094482, + "num_tokens": 593139114.0, + "step": 22928 + }, + { + "epoch": 2.5180101032286406, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.2457737922668457, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7296313047409058, + "num_tokens": 593167042.0, + "step": 22929 + }, + { + "epoch": 2.5181199209312544, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.3817522525787354, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7246804237365723, + "num_tokens": 593191849.0, + "step": 22930 + }, + { + "epoch": 2.5182297386338677, + "ewc_loss": 1.4960765838623047e-05, + "grad_norm": 2.4193739891052246, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7350829243659973, + "num_tokens": 593218714.0, + "step": 22931 + }, + { + "epoch": 2.5183395563364814, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2394628524780273, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7271681427955627, + "num_tokens": 593247208.0, + "step": 22932 + }, + { + "epoch": 2.518449374039095, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4301769733428955, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7178787589073181, + "num_tokens": 593272635.0, + "step": 22933 + }, + { + "epoch": 2.5185591917417085, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4772069454193115, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7275744676589966, + "num_tokens": 593297594.0, + "step": 22934 + }, + { + "epoch": 2.5186690094443223, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2295982837677, + "learning_rate": 1e-06, + "loss": 0.9049, + "mean_token_accuracy": 0.7240539789199829, + "num_tokens": 593329123.0, + "step": 22935 + }, + { + "epoch": 2.518778827146936, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2513301372528076, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7296668291091919, + "num_tokens": 593358256.0, + "step": 22936 + }, + { + "epoch": 2.5188886448495498, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.673253297805786, + "learning_rate": 1e-06, + "loss": 0.7446, + "mean_token_accuracy": 0.7581285238265991, + "num_tokens": 593379323.0, + "step": 22937 + }, + { + "epoch": 2.5189984625521635, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.0954110622406006, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7411912679672241, + "num_tokens": 593412384.0, + "step": 22938 + }, + { + "epoch": 2.519108280254777, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3868072032928467, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7248988151550293, + "num_tokens": 593439200.0, + "step": 22939 + }, + { + "epoch": 2.5192180979573906, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.245574474334717, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7125720977783203, + "num_tokens": 593467494.0, + "step": 22940 + }, + { + "epoch": 2.5193279156600044, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.14803147315979, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7322057485580444, + "num_tokens": 593497676.0, + "step": 22941 + }, + { + "epoch": 2.519437733362618, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.6281213760375977, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.7533183097839355, + "num_tokens": 593519999.0, + "step": 22942 + }, + { + "epoch": 2.519547551065232, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.342418670654297, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7117494940757751, + "num_tokens": 593546497.0, + "step": 22943 + }, + { + "epoch": 2.519657368767845, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.269767999649048, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7330628633499146, + "num_tokens": 593571869.0, + "step": 22944 + }, + { + "epoch": 2.519767186470459, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4659390449523926, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7356836795806885, + "num_tokens": 593595902.0, + "step": 22945 + }, + { + "epoch": 2.5198770041730727, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.840960741043091, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7537028789520264, + "num_tokens": 593616021.0, + "step": 22946 + }, + { + "epoch": 2.5199868218756865, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3747777938842773, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.718147873878479, + "num_tokens": 593640677.0, + "step": 22947 + }, + { + "epoch": 2.5200966395783, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 1.9953734874725342, + "learning_rate": 1e-06, + "loss": 0.9383, + "mean_token_accuracy": 0.7117314338684082, + "num_tokens": 593675273.0, + "step": 22948 + }, + { + "epoch": 2.5202064572809135, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.614318370819092, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.724345326423645, + "num_tokens": 593698371.0, + "step": 22949 + }, + { + "epoch": 2.5203162749835273, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.6246402263641357, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7112936973571777, + "num_tokens": 593719487.0, + "step": 22950 + }, + { + "epoch": 2.520426092686141, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.221038341522217, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7323892116546631, + "num_tokens": 593748209.0, + "step": 22951 + }, + { + "epoch": 2.520535910388755, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.075681447982788, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7214797139167786, + "num_tokens": 593779992.0, + "step": 22952 + }, + { + "epoch": 2.5206457280913686, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3227956295013428, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7379651665687561, + "num_tokens": 593805162.0, + "step": 22953 + }, + { + "epoch": 2.520755545793982, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.5118720531463623, + "learning_rate": 1e-06, + "loss": 0.7723, + "mean_token_accuracy": 0.7525993585586548, + "num_tokens": 593826809.0, + "step": 22954 + }, + { + "epoch": 2.5208653634965956, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3387017250061035, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7322226762771606, + "num_tokens": 593854502.0, + "step": 22955 + }, + { + "epoch": 2.5209751811992094, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3299460411071777, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7103539705276489, + "num_tokens": 593882364.0, + "step": 22956 + }, + { + "epoch": 2.521084998901823, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.0832622051239014, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7226157188415527, + "num_tokens": 593914002.0, + "step": 22957 + }, + { + "epoch": 2.521194816604437, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2494747638702393, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7494915723800659, + "num_tokens": 593942545.0, + "step": 22958 + }, + { + "epoch": 2.52130463430705, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2730395793914795, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.72887122631073, + "num_tokens": 593969663.0, + "step": 22959 + }, + { + "epoch": 2.521414452009664, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4726290702819824, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7201392650604248, + "num_tokens": 593993735.0, + "step": 22960 + }, + { + "epoch": 2.5215242697122777, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.210740089416504, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.7324821949005127, + "num_tokens": 594022149.0, + "step": 22961 + }, + { + "epoch": 2.521634087414891, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.359253406524658, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7381446361541748, + "num_tokens": 594048254.0, + "step": 22962 + }, + { + "epoch": 2.521743905117505, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.6870570182800293, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7449662685394287, + "num_tokens": 594069964.0, + "step": 22963 + }, + { + "epoch": 2.5218537228201185, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3454244136810303, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7453650236129761, + "num_tokens": 594094639.0, + "step": 22964 + }, + { + "epoch": 2.5219635405227323, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.7722389698028564, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.7515250444412231, + "num_tokens": 594115174.0, + "step": 22965 + }, + { + "epoch": 2.522073358225346, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3957533836364746, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.735413670539856, + "num_tokens": 594140095.0, + "step": 22966 + }, + { + "epoch": 2.5221831759279594, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6028990745544434, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7265476584434509, + "num_tokens": 594162787.0, + "step": 22967 + }, + { + "epoch": 2.522292993630573, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.475038766860962, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7328557372093201, + "num_tokens": 594188094.0, + "step": 22968 + }, + { + "epoch": 2.522402811333187, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3664915561676025, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7310482859611511, + "num_tokens": 594213783.0, + "step": 22969 + }, + { + "epoch": 2.5225126290358006, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.24082088470459, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7220025062561035, + "num_tokens": 594240306.0, + "step": 22970 + }, + { + "epoch": 2.5226224467384144, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1308107376098633, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7304480671882629, + "num_tokens": 594269709.0, + "step": 22971 + }, + { + "epoch": 2.5227322644410277, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6777212619781494, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7524647116661072, + "num_tokens": 594289564.0, + "step": 22972 + }, + { + "epoch": 2.5228420821436415, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2591140270233154, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7292450666427612, + "num_tokens": 594317962.0, + "step": 22973 + }, + { + "epoch": 2.5229518998462552, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.5401411056518555, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7466671466827393, + "num_tokens": 594340334.0, + "step": 22974 + }, + { + "epoch": 2.523061717548869, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.505688428878784, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7185484170913696, + "num_tokens": 594364305.0, + "step": 22975 + }, + { + "epoch": 2.5231715352514827, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.6631033420562744, + "learning_rate": 1e-06, + "loss": 0.7469, + "mean_token_accuracy": 0.757530152797699, + "num_tokens": 594383972.0, + "step": 22976 + }, + { + "epoch": 2.523281352954096, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.107861042022705, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7372509837150574, + "num_tokens": 594413752.0, + "step": 22977 + }, + { + "epoch": 2.52339117065671, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2939672470092773, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.715990424156189, + "num_tokens": 594442079.0, + "step": 22978 + }, + { + "epoch": 2.5235009883593236, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.5780961513519287, + "learning_rate": 1e-06, + "loss": 0.9207, + "mean_token_accuracy": 0.7094727754592896, + "num_tokens": 594465901.0, + "step": 22979 + }, + { + "epoch": 2.5236108060619373, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2345802783966064, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7196486592292786, + "num_tokens": 594496278.0, + "step": 22980 + }, + { + "epoch": 2.523720623764551, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.3785240650177, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7385948300361633, + "num_tokens": 594520948.0, + "step": 22981 + }, + { + "epoch": 2.5238304414671644, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2201504707336426, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7412909865379333, + "num_tokens": 594548810.0, + "step": 22982 + }, + { + "epoch": 2.523940259169778, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4540281295776367, + "learning_rate": 1e-06, + "loss": 0.9491, + "mean_token_accuracy": 0.7055892944335938, + "num_tokens": 594574861.0, + "step": 22983 + }, + { + "epoch": 2.524050076872392, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.251422643661499, + "learning_rate": 1e-06, + "loss": 0.9224, + "mean_token_accuracy": 0.711808979511261, + "num_tokens": 594604307.0, + "step": 22984 + }, + { + "epoch": 2.5241598945750052, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.406261920928955, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7164479494094849, + "num_tokens": 594629850.0, + "step": 22985 + }, + { + "epoch": 2.5242697122776194, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.9444518089294434, + "learning_rate": 1e-06, + "loss": 0.8211, + "mean_token_accuracy": 0.7405993342399597, + "num_tokens": 594649519.0, + "step": 22986 + }, + { + "epoch": 2.5243795299802327, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.766458511352539, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7350361347198486, + "num_tokens": 594669664.0, + "step": 22987 + }, + { + "epoch": 2.5244893476828465, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.2394566535949707, + "learning_rate": 1e-06, + "loss": 0.9268, + "mean_token_accuracy": 0.7213153839111328, + "num_tokens": 594698318.0, + "step": 22988 + }, + { + "epoch": 2.5245991653854603, + "ewc_loss": 1.5020370483398438e-05, + "grad_norm": 2.4447078704833984, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7237846255302429, + "num_tokens": 594724229.0, + "step": 22989 + }, + { + "epoch": 2.5247089830880736, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5231268405914307, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.726588249206543, + "num_tokens": 594750082.0, + "step": 22990 + }, + { + "epoch": 2.5248188007906873, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.844924211502075, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.745409369468689, + "num_tokens": 594769454.0, + "step": 22991 + }, + { + "epoch": 2.524928618493301, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.650989294052124, + "learning_rate": 1e-06, + "loss": 0.7936, + "mean_token_accuracy": 0.7468016147613525, + "num_tokens": 594790967.0, + "step": 22992 + }, + { + "epoch": 2.525038436195915, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6756691932678223, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7183797955513, + "num_tokens": 594813218.0, + "step": 22993 + }, + { + "epoch": 2.5251482538985286, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.0496044158935547, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7197693586349487, + "num_tokens": 594844603.0, + "step": 22994 + }, + { + "epoch": 2.525258071601142, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.322279691696167, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7284885048866272, + "num_tokens": 594871719.0, + "step": 22995 + }, + { + "epoch": 2.5253678893037557, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4712555408477783, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7256450057029724, + "num_tokens": 594895488.0, + "step": 22996 + }, + { + "epoch": 2.5254777070063694, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.534231424331665, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7378932237625122, + "num_tokens": 594919364.0, + "step": 22997 + }, + { + "epoch": 2.525587524708983, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.450305461883545, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7305138111114502, + "num_tokens": 594946118.0, + "step": 22998 + }, + { + "epoch": 2.525697342411597, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.465864896774292, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7195194363594055, + "num_tokens": 594969450.0, + "step": 22999 + }, + { + "epoch": 2.5258071601142102, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.7729296684265137, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7395883798599243, + "num_tokens": 594988795.0, + "step": 23000 + }, + { + "epoch": 2.525916977816824, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3139734268188477, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7462738752365112, + "num_tokens": 595016239.0, + "step": 23001 + }, + { + "epoch": 2.5260267955194378, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3988568782806396, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7464550733566284, + "num_tokens": 595041564.0, + "step": 23002 + }, + { + "epoch": 2.5261366132220515, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.468108654022217, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7418191432952881, + "num_tokens": 595064933.0, + "step": 23003 + }, + { + "epoch": 2.5262464309246653, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2137558460235596, + "learning_rate": 1e-06, + "loss": 0.8885, + "mean_token_accuracy": 0.7200433015823364, + "num_tokens": 595094820.0, + "step": 23004 + }, + { + "epoch": 2.5263562486272786, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.437016248703003, + "learning_rate": 1e-06, + "loss": 0.7703, + "mean_token_accuracy": 0.7558026909828186, + "num_tokens": 595120478.0, + "step": 23005 + }, + { + "epoch": 2.5264660663298923, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3639743328094482, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7442106008529663, + "num_tokens": 595144689.0, + "step": 23006 + }, + { + "epoch": 2.526575884032506, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2475810050964355, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.7314110994338989, + "num_tokens": 595171859.0, + "step": 23007 + }, + { + "epoch": 2.52668570173512, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4184529781341553, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7383114099502563, + "num_tokens": 595196206.0, + "step": 23008 + }, + { + "epoch": 2.5267955194377336, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5200319290161133, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7147388458251953, + "num_tokens": 595219237.0, + "step": 23009 + }, + { + "epoch": 2.526905337140347, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2318766117095947, + "learning_rate": 1e-06, + "loss": 0.9905, + "mean_token_accuracy": 0.6973377466201782, + "num_tokens": 595249726.0, + "step": 23010 + }, + { + "epoch": 2.5270151548429607, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.391113042831421, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7311928272247314, + "num_tokens": 595274027.0, + "step": 23011 + }, + { + "epoch": 2.5271249725455744, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.558579206466675, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7193941473960876, + "num_tokens": 595296661.0, + "step": 23012 + }, + { + "epoch": 2.5272347902481878, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5255541801452637, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7414377927780151, + "num_tokens": 595319437.0, + "step": 23013 + }, + { + "epoch": 2.5273446079508015, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.0862364768981934, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7110979557037354, + "num_tokens": 595351096.0, + "step": 23014 + }, + { + "epoch": 2.5274544256534153, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4586360454559326, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7283939719200134, + "num_tokens": 595376735.0, + "step": 23015 + }, + { + "epoch": 2.527564243356029, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3365139961242676, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.725634753704071, + "num_tokens": 595404045.0, + "step": 23016 + }, + { + "epoch": 2.527674061058643, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.422898292541504, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7259936332702637, + "num_tokens": 595429491.0, + "step": 23017 + }, + { + "epoch": 2.527783878761256, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.316676616668701, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.711950421333313, + "num_tokens": 595457524.0, + "step": 23018 + }, + { + "epoch": 2.52789369646387, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.0299625396728516, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.734539806842804, + "num_tokens": 595490527.0, + "step": 23019 + }, + { + "epoch": 2.5280035141664836, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.474297523498535, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7270970344543457, + "num_tokens": 595515326.0, + "step": 23020 + }, + { + "epoch": 2.5281133318690974, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.04433274269104, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7199088335037231, + "num_tokens": 595547889.0, + "step": 23021 + }, + { + "epoch": 2.528223149571711, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2337937355041504, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.719758152961731, + "num_tokens": 595576016.0, + "step": 23022 + }, + { + "epoch": 2.5283329672743244, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5879769325256348, + "learning_rate": 1e-06, + "loss": 0.9072, + "mean_token_accuracy": 0.7200506925582886, + "num_tokens": 595599972.0, + "step": 23023 + }, + { + "epoch": 2.528442784976938, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1915955543518066, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7383982539176941, + "num_tokens": 595629268.0, + "step": 23024 + }, + { + "epoch": 2.528552602679552, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.466881513595581, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7339915633201599, + "num_tokens": 595653835.0, + "step": 23025 + }, + { + "epoch": 2.5286624203821657, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4262969493865967, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7260832190513611, + "num_tokens": 595679525.0, + "step": 23026 + }, + { + "epoch": 2.5287722380847795, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.543870687484741, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7176996469497681, + "num_tokens": 595704461.0, + "step": 23027 + }, + { + "epoch": 2.528882055787393, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5908241271972656, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7355571985244751, + "num_tokens": 595726467.0, + "step": 23028 + }, + { + "epoch": 2.5289918734900065, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5216434001922607, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7402610778808594, + "num_tokens": 595749917.0, + "step": 23029 + }, + { + "epoch": 2.5291016911926203, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1587300300598145, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7155828475952148, + "num_tokens": 595778799.0, + "step": 23030 + }, + { + "epoch": 2.529211508895234, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.625495433807373, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7325654029846191, + "num_tokens": 595798731.0, + "step": 23031 + }, + { + "epoch": 2.529321326597848, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.497877359390259, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7212874293327332, + "num_tokens": 595824073.0, + "step": 23032 + }, + { + "epoch": 2.529431144300461, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.288774013519287, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7244483232498169, + "num_tokens": 595849564.0, + "step": 23033 + }, + { + "epoch": 2.529540962003075, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4430882930755615, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7291461229324341, + "num_tokens": 595878108.0, + "step": 23034 + }, + { + "epoch": 2.5296507797056886, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.181506633758545, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7177161574363708, + "num_tokens": 595906890.0, + "step": 23035 + }, + { + "epoch": 2.529760597408302, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.800739288330078, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.736472487449646, + "num_tokens": 595926522.0, + "step": 23036 + }, + { + "epoch": 2.529870415110916, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.394991397857666, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7250486612319946, + "num_tokens": 595949645.0, + "step": 23037 + }, + { + "epoch": 2.5299802328135295, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2369771003723145, + "learning_rate": 1e-06, + "loss": 0.8884, + "mean_token_accuracy": 0.723864734172821, + "num_tokens": 595977614.0, + "step": 23038 + }, + { + "epoch": 2.530090050516143, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.134718656539917, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7192649245262146, + "num_tokens": 596007884.0, + "step": 23039 + }, + { + "epoch": 2.530199868218757, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4157800674438477, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7516674995422363, + "num_tokens": 596031739.0, + "step": 23040 + }, + { + "epoch": 2.5303096859213703, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.834897756576538, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7397521734237671, + "num_tokens": 596050106.0, + "step": 23041 + }, + { + "epoch": 2.530419503623984, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.69461989402771, + "learning_rate": 1e-06, + "loss": 0.6918, + "mean_token_accuracy": 0.7742304801940918, + "num_tokens": 596069132.0, + "step": 23042 + }, + { + "epoch": 2.530529321326598, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1972556114196777, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7368841171264648, + "num_tokens": 596097303.0, + "step": 23043 + }, + { + "epoch": 2.5306391390292116, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.439946174621582, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7473673820495605, + "num_tokens": 596121083.0, + "step": 23044 + }, + { + "epoch": 2.5307489567318253, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.319572687149048, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7227878570556641, + "num_tokens": 596148564.0, + "step": 23045 + }, + { + "epoch": 2.5308587744344386, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.298475980758667, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7445716857910156, + "num_tokens": 596174298.0, + "step": 23046 + }, + { + "epoch": 2.5309685921370524, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3861289024353027, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7086976766586304, + "num_tokens": 596198003.0, + "step": 23047 + }, + { + "epoch": 2.531078409839666, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6227025985717773, + "learning_rate": 1e-06, + "loss": 0.7467, + "mean_token_accuracy": 0.7563376426696777, + "num_tokens": 596220404.0, + "step": 23048 + }, + { + "epoch": 2.53118822754228, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1584551334381104, + "learning_rate": 1e-06, + "loss": 0.9073, + "mean_token_accuracy": 0.7150411605834961, + "num_tokens": 596253105.0, + "step": 23049 + }, + { + "epoch": 2.5312980452448937, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4085488319396973, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7305910587310791, + "num_tokens": 596277084.0, + "step": 23050 + }, + { + "epoch": 2.531407862947507, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.593883752822876, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7144265174865723, + "num_tokens": 596300341.0, + "step": 23051 + }, + { + "epoch": 2.5315176806501207, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5991392135620117, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7485396862030029, + "num_tokens": 596322781.0, + "step": 23052 + }, + { + "epoch": 2.5316274983527345, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5072720050811768, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7326910495758057, + "num_tokens": 596345388.0, + "step": 23053 + }, + { + "epoch": 2.5317373160553482, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.205223321914673, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.725548505783081, + "num_tokens": 596373827.0, + "step": 23054 + }, + { + "epoch": 2.531847133757962, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.449765205383301, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.748632550239563, + "num_tokens": 596396903.0, + "step": 23055 + }, + { + "epoch": 2.5319569514605753, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.289947748184204, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7159896492958069, + "num_tokens": 596424384.0, + "step": 23056 + }, + { + "epoch": 2.532066769163189, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4331696033477783, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7307579517364502, + "num_tokens": 596448517.0, + "step": 23057 + }, + { + "epoch": 2.532176586865803, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2500977516174316, + "learning_rate": 1e-06, + "loss": 0.9077, + "mean_token_accuracy": 0.7091375589370728, + "num_tokens": 596476470.0, + "step": 23058 + }, + { + "epoch": 2.5322864045684166, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.283695936203003, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7358995676040649, + "num_tokens": 596504634.0, + "step": 23059 + }, + { + "epoch": 2.5323962222710303, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.149801015853882, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7159522771835327, + "num_tokens": 596533937.0, + "step": 23060 + }, + { + "epoch": 2.5325060399736437, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1343283653259277, + "learning_rate": 1e-06, + "loss": 0.9456, + "mean_token_accuracy": 0.7019537091255188, + "num_tokens": 596566767.0, + "step": 23061 + }, + { + "epoch": 2.5326158576762574, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4133496284484863, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7206922769546509, + "num_tokens": 596592004.0, + "step": 23062 + }, + { + "epoch": 2.532725675378871, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3687314987182617, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7365775108337402, + "num_tokens": 596617623.0, + "step": 23063 + }, + { + "epoch": 2.5328354930814845, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1435327529907227, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.733185887336731, + "num_tokens": 596647337.0, + "step": 23064 + }, + { + "epoch": 2.5329453107840982, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6132614612579346, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7343960404396057, + "num_tokens": 596668598.0, + "step": 23065 + }, + { + "epoch": 2.533055128486712, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1480743885040283, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.7272915840148926, + "num_tokens": 596699776.0, + "step": 23066 + }, + { + "epoch": 2.5331649461893258, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4346070289611816, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7171105146408081, + "num_tokens": 596725281.0, + "step": 23067 + }, + { + "epoch": 2.5332747638919395, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.7153546810150146, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7489082217216492, + "num_tokens": 596747043.0, + "step": 23068 + }, + { + "epoch": 2.533384581594553, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.632451057434082, + "learning_rate": 1e-06, + "loss": 0.9302, + "mean_token_accuracy": 0.709208607673645, + "num_tokens": 596771535.0, + "step": 23069 + }, + { + "epoch": 2.5334943992971666, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.316648006439209, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7216596007347107, + "num_tokens": 596798443.0, + "step": 23070 + }, + { + "epoch": 2.5336042169997803, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.711514711380005, + "learning_rate": 1e-06, + "loss": 0.7784, + "mean_token_accuracy": 0.7475204467773438, + "num_tokens": 596819292.0, + "step": 23071 + }, + { + "epoch": 2.533714034702394, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2759451866149902, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7411277890205383, + "num_tokens": 596848840.0, + "step": 23072 + }, + { + "epoch": 2.533823852405008, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2687268257141113, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7340807318687439, + "num_tokens": 596876828.0, + "step": 23073 + }, + { + "epoch": 2.533933670107621, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3104023933410645, + "learning_rate": 1e-06, + "loss": 0.9025, + "mean_token_accuracy": 0.7196599841117859, + "num_tokens": 596903668.0, + "step": 23074 + }, + { + "epoch": 2.534043487810235, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.015218496322632, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7162696719169617, + "num_tokens": 596939752.0, + "step": 23075 + }, + { + "epoch": 2.5341533055128487, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3780648708343506, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7214363813400269, + "num_tokens": 596965149.0, + "step": 23076 + }, + { + "epoch": 2.5342631232154624, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4009506702423096, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.736548900604248, + "num_tokens": 596988327.0, + "step": 23077 + }, + { + "epoch": 2.534372940918076, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.577659845352173, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7295104265213013, + "num_tokens": 597013645.0, + "step": 23078 + }, + { + "epoch": 2.5344827586206895, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.466696262359619, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7367025017738342, + "num_tokens": 597036964.0, + "step": 23079 + }, + { + "epoch": 2.5345925763233033, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.689393997192383, + "learning_rate": 1e-06, + "loss": 0.7625, + "mean_token_accuracy": 0.7512431144714355, + "num_tokens": 597055558.0, + "step": 23080 + }, + { + "epoch": 2.534702394025917, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3141868114471436, + "learning_rate": 1e-06, + "loss": 0.9003, + "mean_token_accuracy": 0.718584418296814, + "num_tokens": 597081249.0, + "step": 23081 + }, + { + "epoch": 2.5348122117285308, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.397212505340576, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7431182861328125, + "num_tokens": 597104520.0, + "step": 23082 + }, + { + "epoch": 2.5349220294311445, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.023513078689575, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7330961227416992, + "num_tokens": 597136510.0, + "step": 23083 + }, + { + "epoch": 2.535031847133758, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.695114850997925, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7247767448425293, + "num_tokens": 597157658.0, + "step": 23084 + }, + { + "epoch": 2.5351416648363716, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.0774474143981934, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.729187548160553, + "num_tokens": 597189398.0, + "step": 23085 + }, + { + "epoch": 2.5352514825389854, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6151068210601807, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7502291202545166, + "num_tokens": 597210921.0, + "step": 23086 + }, + { + "epoch": 2.5353613002415987, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2787907123565674, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.744560718536377, + "num_tokens": 597237123.0, + "step": 23087 + }, + { + "epoch": 2.535471117944213, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.474020481109619, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7406608462333679, + "num_tokens": 597260578.0, + "step": 23088 + }, + { + "epoch": 2.535580935646826, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.250598430633545, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7413886785507202, + "num_tokens": 597289215.0, + "step": 23089 + }, + { + "epoch": 2.53569075334944, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.304291248321533, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7295079231262207, + "num_tokens": 597316786.0, + "step": 23090 + }, + { + "epoch": 2.5358005710520537, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3897011280059814, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7218537926673889, + "num_tokens": 597340930.0, + "step": 23091 + }, + { + "epoch": 2.535910388754667, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.538421154022217, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7272720336914062, + "num_tokens": 597364706.0, + "step": 23092 + }, + { + "epoch": 2.5360202064572808, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.490584373474121, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.725614607334137, + "num_tokens": 597386630.0, + "step": 23093 + }, + { + "epoch": 2.5361300241598945, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4442102909088135, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.735031247138977, + "num_tokens": 597410659.0, + "step": 23094 + }, + { + "epoch": 2.5362398418625083, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.3855667114257812, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7300009727478027, + "num_tokens": 597437572.0, + "step": 23095 + }, + { + "epoch": 2.536349659565122, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.40407133102417, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7327539324760437, + "num_tokens": 597460736.0, + "step": 23096 + }, + { + "epoch": 2.5364594772677354, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4474213123321533, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7310043573379517, + "num_tokens": 597485885.0, + "step": 23097 + }, + { + "epoch": 2.536569294970349, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.357699155807495, + "learning_rate": 1e-06, + "loss": 0.8324, + "mean_token_accuracy": 0.7339226603507996, + "num_tokens": 597511911.0, + "step": 23098 + }, + { + "epoch": 2.536679112672963, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2181789875030518, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7356845140457153, + "num_tokens": 597540549.0, + "step": 23099 + }, + { + "epoch": 2.5367889303755766, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.1934866905212402, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7283477783203125, + "num_tokens": 597568988.0, + "step": 23100 + }, + { + "epoch": 2.5368987480781904, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.104017496109009, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7245512008666992, + "num_tokens": 597599355.0, + "step": 23101 + }, + { + "epoch": 2.5370085657808037, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.2140095233917236, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7330160140991211, + "num_tokens": 597626892.0, + "step": 23102 + }, + { + "epoch": 2.5371183834834174, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.0715043544769287, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7207920551300049, + "num_tokens": 597656696.0, + "step": 23103 + }, + { + "epoch": 2.537228201186031, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.136831760406494, + "learning_rate": 1e-06, + "loss": 1.0104, + "mean_token_accuracy": 0.6944336891174316, + "num_tokens": 597686492.0, + "step": 23104 + }, + { + "epoch": 2.537338018888645, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.061020851135254, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7076116800308228, + "num_tokens": 597720272.0, + "step": 23105 + }, + { + "epoch": 2.5374478365912587, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5197863578796387, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7305810451507568, + "num_tokens": 597745245.0, + "step": 23106 + }, + { + "epoch": 2.537557654293872, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5008544921875, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7328797578811646, + "num_tokens": 597769483.0, + "step": 23107 + }, + { + "epoch": 2.537667471996486, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2915525436401367, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7391108870506287, + "num_tokens": 597794789.0, + "step": 23108 + }, + { + "epoch": 2.5377772896990995, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.226285457611084, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7308417558670044, + "num_tokens": 597823774.0, + "step": 23109 + }, + { + "epoch": 2.5378871074017133, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.241560220718384, + "learning_rate": 1e-06, + "loss": 0.8966, + "mean_token_accuracy": 0.7238418459892273, + "num_tokens": 597854226.0, + "step": 23110 + }, + { + "epoch": 2.537996925104327, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4243836402893066, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7448214888572693, + "num_tokens": 597879666.0, + "step": 23111 + }, + { + "epoch": 2.5381067428069404, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5103697776794434, + "learning_rate": 1e-06, + "loss": 0.8095, + "mean_token_accuracy": 0.7408014535903931, + "num_tokens": 597903409.0, + "step": 23112 + }, + { + "epoch": 2.538216560509554, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2881438732147217, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7407610416412354, + "num_tokens": 597932638.0, + "step": 23113 + }, + { + "epoch": 2.538326378212168, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.5676512718200684, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7457576990127563, + "num_tokens": 597954380.0, + "step": 23114 + }, + { + "epoch": 2.538436195914781, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4619176387786865, + "learning_rate": 1e-06, + "loss": 0.9142, + "mean_token_accuracy": 0.725987434387207, + "num_tokens": 597980225.0, + "step": 23115 + }, + { + "epoch": 2.538546013617395, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.601905345916748, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.737559974193573, + "num_tokens": 598001818.0, + "step": 23116 + }, + { + "epoch": 2.5386558313200087, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.185131311416626, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.73997962474823, + "num_tokens": 598031330.0, + "step": 23117 + }, + { + "epoch": 2.5387656490226225, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.593123435974121, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7414078712463379, + "num_tokens": 598052679.0, + "step": 23118 + }, + { + "epoch": 2.5388754667252362, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.609224796295166, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.7408530712127686, + "num_tokens": 598073395.0, + "step": 23119 + }, + { + "epoch": 2.5389852844278495, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.4030039310455322, + "learning_rate": 1e-06, + "loss": 0.8433, + "mean_token_accuracy": 0.7318360805511475, + "num_tokens": 598099590.0, + "step": 23120 + }, + { + "epoch": 2.5390951021304633, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.9870340824127197, + "learning_rate": 1e-06, + "loss": 0.7374, + "mean_token_accuracy": 0.758770763874054, + "num_tokens": 598117357.0, + "step": 23121 + }, + { + "epoch": 2.539204919833077, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.109722852706909, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7472606897354126, + "num_tokens": 598146407.0, + "step": 23122 + }, + { + "epoch": 2.539314737535691, + "ewc_loss": 1.5079975128173828e-05, + "grad_norm": 2.6099767684936523, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7307663559913635, + "num_tokens": 598168807.0, + "step": 23123 + }, + { + "epoch": 2.5394245552383046, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4734878540039062, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7333618402481079, + "num_tokens": 598193948.0, + "step": 23124 + }, + { + "epoch": 2.539534372940918, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.323970079421997, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7357819676399231, + "num_tokens": 598220707.0, + "step": 23125 + }, + { + "epoch": 2.5396441906435316, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3300654888153076, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7440544962882996, + "num_tokens": 598246641.0, + "step": 23126 + }, + { + "epoch": 2.5397540083461454, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3215954303741455, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.734981894493103, + "num_tokens": 598275816.0, + "step": 23127 + }, + { + "epoch": 2.539863826048759, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2836759090423584, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7296609878540039, + "num_tokens": 598304609.0, + "step": 23128 + }, + { + "epoch": 2.539973643751373, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1494481563568115, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7163122296333313, + "num_tokens": 598332952.0, + "step": 23129 + }, + { + "epoch": 2.5400834614539862, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1808066368103027, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.713866114616394, + "num_tokens": 598361065.0, + "step": 23130 + }, + { + "epoch": 2.5401932791566, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1623735427856445, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7073488235473633, + "num_tokens": 598394400.0, + "step": 23131 + }, + { + "epoch": 2.5403030968592137, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.40330171585083, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7368749380111694, + "num_tokens": 598421433.0, + "step": 23132 + }, + { + "epoch": 2.5404129145618275, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.686676263809204, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7304961681365967, + "num_tokens": 598441169.0, + "step": 23133 + }, + { + "epoch": 2.5405227322644413, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.542274236679077, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.7357922792434692, + "num_tokens": 598463886.0, + "step": 23134 + }, + { + "epoch": 2.5406325499670546, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.7293331623077393, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.725673258304596, + "num_tokens": 598485379.0, + "step": 23135 + }, + { + "epoch": 2.5407423676696683, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4973528385162354, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7339303493499756, + "num_tokens": 598511676.0, + "step": 23136 + }, + { + "epoch": 2.540852185372282, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.630194664001465, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7185522317886353, + "num_tokens": 598533470.0, + "step": 23137 + }, + { + "epoch": 2.540962003074896, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2994649410247803, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7211310863494873, + "num_tokens": 598560612.0, + "step": 23138 + }, + { + "epoch": 2.5410718207775096, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.560920476913452, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7456570863723755, + "num_tokens": 598583473.0, + "step": 23139 + }, + { + "epoch": 2.541181638480123, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.224682331085205, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7297336459159851, + "num_tokens": 598611489.0, + "step": 23140 + }, + { + "epoch": 2.5412914561827367, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4287843704223633, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.732252299785614, + "num_tokens": 598633506.0, + "step": 23141 + }, + { + "epoch": 2.5414012738853504, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5577595233917236, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7455400228500366, + "num_tokens": 598655695.0, + "step": 23142 + }, + { + "epoch": 2.5415110915879637, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.439253330230713, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7487821578979492, + "num_tokens": 598678323.0, + "step": 23143 + }, + { + "epoch": 2.5416209092905775, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.323273181915283, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7397733330726624, + "num_tokens": 598702544.0, + "step": 23144 + }, + { + "epoch": 2.5417307269931912, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3961870670318604, + "learning_rate": 1e-06, + "loss": 0.9107, + "mean_token_accuracy": 0.7128472328186035, + "num_tokens": 598728721.0, + "step": 23145 + }, + { + "epoch": 2.541840544695805, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.748322010040283, + "learning_rate": 1e-06, + "loss": 0.7569, + "mean_token_accuracy": 0.7535157203674316, + "num_tokens": 598748373.0, + "step": 23146 + }, + { + "epoch": 2.5419503623984188, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4346907138824463, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7309954166412354, + "num_tokens": 598772238.0, + "step": 23147 + }, + { + "epoch": 2.542060180101032, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3729283809661865, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7339240312576294, + "num_tokens": 598798213.0, + "step": 23148 + }, + { + "epoch": 2.542169997803646, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3111793994903564, + "learning_rate": 1e-06, + "loss": 0.9308, + "mean_token_accuracy": 0.7110363245010376, + "num_tokens": 598825979.0, + "step": 23149 + }, + { + "epoch": 2.5422798155062596, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6035878658294678, + "learning_rate": 1e-06, + "loss": 0.8109, + "mean_token_accuracy": 0.7437078356742859, + "num_tokens": 598848689.0, + "step": 23150 + }, + { + "epoch": 2.5423896332088733, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1370582580566406, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7305194735527039, + "num_tokens": 598879821.0, + "step": 23151 + }, + { + "epoch": 2.542499450911487, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.0718464851379395, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7131936550140381, + "num_tokens": 598912920.0, + "step": 23152 + }, + { + "epoch": 2.5426092686141004, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1769144535064697, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7330310940742493, + "num_tokens": 598941236.0, + "step": 23153 + }, + { + "epoch": 2.542719086316714, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.0550389289855957, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7368162870407104, + "num_tokens": 598974363.0, + "step": 23154 + }, + { + "epoch": 2.542828904019328, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4536383152008057, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7233482599258423, + "num_tokens": 598998809.0, + "step": 23155 + }, + { + "epoch": 2.5429387217219417, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.0884950160980225, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7264693975448608, + "num_tokens": 599030383.0, + "step": 23156 + }, + { + "epoch": 2.5430485394245554, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.165262222290039, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7265152335166931, + "num_tokens": 599061889.0, + "step": 23157 + }, + { + "epoch": 2.5431583571271688, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.375074863433838, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7168264985084534, + "num_tokens": 599087343.0, + "step": 23158 + }, + { + "epoch": 2.5432681748297825, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.239687204360962, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7198743224143982, + "num_tokens": 599116440.0, + "step": 23159 + }, + { + "epoch": 2.5433779925323963, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.7493860721588135, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7508450746536255, + "num_tokens": 599137004.0, + "step": 23160 + }, + { + "epoch": 2.54348781023501, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.350146532058716, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.723121166229248, + "num_tokens": 599164010.0, + "step": 23161 + }, + { + "epoch": 2.543597627937624, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.483859062194824, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.749360203742981, + "num_tokens": 599186751.0, + "step": 23162 + }, + { + "epoch": 2.543707445640237, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3221967220306396, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.714533805847168, + "num_tokens": 599216747.0, + "step": 23163 + }, + { + "epoch": 2.543817263342851, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.379051685333252, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7325865030288696, + "num_tokens": 599242956.0, + "step": 23164 + }, + { + "epoch": 2.5439270810454646, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4243574142456055, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7324932813644409, + "num_tokens": 599268162.0, + "step": 23165 + }, + { + "epoch": 2.544036898748078, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5032269954681396, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7327519655227661, + "num_tokens": 599292277.0, + "step": 23166 + }, + { + "epoch": 2.544146716450692, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.7202160358428955, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.74454665184021, + "num_tokens": 599314496.0, + "step": 23167 + }, + { + "epoch": 2.5442565341533054, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3725738525390625, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7106762528419495, + "num_tokens": 599341428.0, + "step": 23168 + }, + { + "epoch": 2.544366351855919, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.9458727836608887, + "learning_rate": 1e-06, + "loss": 0.7659, + "mean_token_accuracy": 0.7611302137374878, + "num_tokens": 599362376.0, + "step": 23169 + }, + { + "epoch": 2.544476169558533, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6188929080963135, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.7535697817802429, + "num_tokens": 599382189.0, + "step": 23170 + }, + { + "epoch": 2.5445859872611463, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6435813903808594, + "learning_rate": 1e-06, + "loss": 0.8114, + "mean_token_accuracy": 0.7429580688476562, + "num_tokens": 599404613.0, + "step": 23171 + }, + { + "epoch": 2.54469580496376, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.915074348449707, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7338423728942871, + "num_tokens": 599422510.0, + "step": 23172 + }, + { + "epoch": 2.544805622666374, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.461228132247925, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7250864505767822, + "num_tokens": 599448293.0, + "step": 23173 + }, + { + "epoch": 2.5449154403689875, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2570858001708984, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7132431268692017, + "num_tokens": 599478848.0, + "step": 23174 + }, + { + "epoch": 2.5450252580716013, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.434011220932007, + "learning_rate": 1e-06, + "loss": 0.7835, + "mean_token_accuracy": 0.751531720161438, + "num_tokens": 599502001.0, + "step": 23175 + }, + { + "epoch": 2.5451350757742146, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2764194011688232, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7298054695129395, + "num_tokens": 599528639.0, + "step": 23176 + }, + { + "epoch": 2.5452448934768284, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3654823303222656, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7013553380966187, + "num_tokens": 599556815.0, + "step": 23177 + }, + { + "epoch": 2.545354711179442, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.420942783355713, + "learning_rate": 1e-06, + "loss": 0.8026, + "mean_token_accuracy": 0.7500847578048706, + "num_tokens": 599580546.0, + "step": 23178 + }, + { + "epoch": 2.545464528882056, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.156099319458008, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7251932621002197, + "num_tokens": 599610573.0, + "step": 23179 + }, + { + "epoch": 2.5455743465846696, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6099436283111572, + "learning_rate": 1e-06, + "loss": 0.9066, + "mean_token_accuracy": 0.7192352414131165, + "num_tokens": 599632809.0, + "step": 23180 + }, + { + "epoch": 2.545684164287283, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.388002634048462, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7141815423965454, + "num_tokens": 599660754.0, + "step": 23181 + }, + { + "epoch": 2.5457939819898967, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.114048957824707, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7232100367546082, + "num_tokens": 599692272.0, + "step": 23182 + }, + { + "epoch": 2.5459037996925105, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.265298843383789, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7174137830734253, + "num_tokens": 599722418.0, + "step": 23183 + }, + { + "epoch": 2.546013617395124, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.236891508102417, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7353909611701965, + "num_tokens": 599750615.0, + "step": 23184 + }, + { + "epoch": 2.546123435097738, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.39189076423645, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7424440979957581, + "num_tokens": 599779424.0, + "step": 23185 + }, + { + "epoch": 2.5462332528003513, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.376342296600342, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7383670210838318, + "num_tokens": 599804042.0, + "step": 23186 + }, + { + "epoch": 2.546343070502965, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.384197473526001, + "learning_rate": 1e-06, + "loss": 0.9945, + "mean_token_accuracy": 0.694286048412323, + "num_tokens": 599831383.0, + "step": 23187 + }, + { + "epoch": 2.546452888205579, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4706883430480957, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7438102960586548, + "num_tokens": 599855286.0, + "step": 23188 + }, + { + "epoch": 2.5465627059081926, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.567173480987549, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7418583631515503, + "num_tokens": 599878271.0, + "step": 23189 + }, + { + "epoch": 2.5466725236108063, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3069350719451904, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.732627809047699, + "num_tokens": 599905729.0, + "step": 23190 + }, + { + "epoch": 2.5467823413134196, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5994057655334473, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7602500915527344, + "num_tokens": 599928077.0, + "step": 23191 + }, + { + "epoch": 2.5468921590160334, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.523876905441284, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.7458351850509644, + "num_tokens": 599951375.0, + "step": 23192 + }, + { + "epoch": 2.547001976718647, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2800936698913574, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7278957962989807, + "num_tokens": 599976493.0, + "step": 23193 + }, + { + "epoch": 2.5471117944212605, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5273690223693848, + "learning_rate": 1e-06, + "loss": 0.8142, + "mean_token_accuracy": 0.7406139969825745, + "num_tokens": 600000689.0, + "step": 23194 + }, + { + "epoch": 2.547221612123874, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.169928550720215, + "learning_rate": 1e-06, + "loss": 0.9922, + "mean_token_accuracy": 0.6968188285827637, + "num_tokens": 600031697.0, + "step": 23195 + }, + { + "epoch": 2.547331429826488, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2671477794647217, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.745550274848938, + "num_tokens": 600056997.0, + "step": 23196 + }, + { + "epoch": 2.5474412475291017, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4488439559936523, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7447044849395752, + "num_tokens": 600081186.0, + "step": 23197 + }, + { + "epoch": 2.5475510652317155, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3777918815612793, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7193566560745239, + "num_tokens": 600106506.0, + "step": 23198 + }, + { + "epoch": 2.547660882934329, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1826770305633545, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.714748740196228, + "num_tokens": 600137370.0, + "step": 23199 + }, + { + "epoch": 2.5477707006369426, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2972755432128906, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7369416952133179, + "num_tokens": 600164594.0, + "step": 23200 + }, + { + "epoch": 2.5478805183395563, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3364126682281494, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7283568382263184, + "num_tokens": 600189231.0, + "step": 23201 + }, + { + "epoch": 2.54799033604217, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.376376152038574, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7320557832717896, + "num_tokens": 600214379.0, + "step": 23202 + }, + { + "epoch": 2.548100153744784, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.215817451477051, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.7477283477783203, + "num_tokens": 600241873.0, + "step": 23203 + }, + { + "epoch": 2.548209971447397, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.390873908996582, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7419312000274658, + "num_tokens": 600265780.0, + "step": 23204 + }, + { + "epoch": 2.548319789150011, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2737176418304443, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7371804714202881, + "num_tokens": 600294184.0, + "step": 23205 + }, + { + "epoch": 2.5484296068526247, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2590510845184326, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7439166903495789, + "num_tokens": 600322499.0, + "step": 23206 + }, + { + "epoch": 2.5485394245552384, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.276798963546753, + "learning_rate": 1e-06, + "loss": 0.887, + "mean_token_accuracy": 0.7246546745300293, + "num_tokens": 600351212.0, + "step": 23207 + }, + { + "epoch": 2.548649242257852, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3557651042938232, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7205238342285156, + "num_tokens": 600376114.0, + "step": 23208 + }, + { + "epoch": 2.5487590599604655, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6743223667144775, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7303436994552612, + "num_tokens": 600397781.0, + "step": 23209 + }, + { + "epoch": 2.5488688776630792, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.230262517929077, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.748130738735199, + "num_tokens": 600425370.0, + "step": 23210 + }, + { + "epoch": 2.548978695365693, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.379586935043335, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7144936919212341, + "num_tokens": 600451785.0, + "step": 23211 + }, + { + "epoch": 2.5490885130683067, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.0664539337158203, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7345043420791626, + "num_tokens": 600485089.0, + "step": 23212 + }, + { + "epoch": 2.5491983307709205, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.754318952560425, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7462332248687744, + "num_tokens": 600505764.0, + "step": 23213 + }, + { + "epoch": 2.549308148473534, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5563442707061768, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7397819757461548, + "num_tokens": 600529607.0, + "step": 23214 + }, + { + "epoch": 2.5494179661761476, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.8799684047698975, + "learning_rate": 1e-06, + "loss": 0.7478, + "mean_token_accuracy": 0.7630620002746582, + "num_tokens": 600549122.0, + "step": 23215 + }, + { + "epoch": 2.5495277838787613, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.142305374145508, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7337793111801147, + "num_tokens": 600578980.0, + "step": 23216 + }, + { + "epoch": 2.5496376015813746, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4544804096221924, + "learning_rate": 1e-06, + "loss": 0.7414, + "mean_token_accuracy": 0.7635633945465088, + "num_tokens": 600602436.0, + "step": 23217 + }, + { + "epoch": 2.549747419283989, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1063687801361084, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7450742721557617, + "num_tokens": 600631297.0, + "step": 23218 + }, + { + "epoch": 2.549857236986602, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.357919216156006, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7290592193603516, + "num_tokens": 600656923.0, + "step": 23219 + }, + { + "epoch": 2.549967054689216, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2421910762786865, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7293127775192261, + "num_tokens": 600683939.0, + "step": 23220 + }, + { + "epoch": 2.5500768723918297, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2758500576019287, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7247653007507324, + "num_tokens": 600712750.0, + "step": 23221 + }, + { + "epoch": 2.550186690094443, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.444086790084839, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7230220437049866, + "num_tokens": 600736385.0, + "step": 23222 + }, + { + "epoch": 2.5502965077970567, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.6902852058410645, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7472071647644043, + "num_tokens": 600756943.0, + "step": 23223 + }, + { + "epoch": 2.5504063254996705, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.286015272140503, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7265444993972778, + "num_tokens": 600784891.0, + "step": 23224 + }, + { + "epoch": 2.5505161432022843, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.1159651279449463, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7236926555633545, + "num_tokens": 600815435.0, + "step": 23225 + }, + { + "epoch": 2.550625960904898, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3703489303588867, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7106775641441345, + "num_tokens": 600842585.0, + "step": 23226 + }, + { + "epoch": 2.5507357786075113, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2154393196105957, + "learning_rate": 1e-06, + "loss": 0.7712, + "mean_token_accuracy": 0.7536191940307617, + "num_tokens": 600870637.0, + "step": 23227 + }, + { + "epoch": 2.550845596310125, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.503765344619751, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7357218861579895, + "num_tokens": 600894183.0, + "step": 23228 + }, + { + "epoch": 2.550955414012739, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.4828941822052, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7343051433563232, + "num_tokens": 600917597.0, + "step": 23229 + }, + { + "epoch": 2.5510652317153526, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.0336859226226807, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7249812483787537, + "num_tokens": 600950127.0, + "step": 23230 + }, + { + "epoch": 2.5511750494179664, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3086445331573486, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7308334708213806, + "num_tokens": 600976942.0, + "step": 23231 + }, + { + "epoch": 2.5512848671205797, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5089995861053467, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7349357008934021, + "num_tokens": 601000981.0, + "step": 23232 + }, + { + "epoch": 2.5513946848231934, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.387824535369873, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7488179802894592, + "num_tokens": 601027115.0, + "step": 23233 + }, + { + "epoch": 2.551504502525807, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3600876331329346, + "learning_rate": 1e-06, + "loss": 0.7357, + "mean_token_accuracy": 0.7613552808761597, + "num_tokens": 601050975.0, + "step": 23234 + }, + { + "epoch": 2.551614320228421, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.239718437194824, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7419233918190002, + "num_tokens": 601079541.0, + "step": 23235 + }, + { + "epoch": 2.5517241379310347, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.282642126083374, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7211414575576782, + "num_tokens": 601109464.0, + "step": 23236 + }, + { + "epoch": 2.551833955633648, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.211193799972534, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7462737560272217, + "num_tokens": 601137656.0, + "step": 23237 + }, + { + "epoch": 2.5519437733362618, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.503390312194824, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.738335371017456, + "num_tokens": 601160857.0, + "step": 23238 + }, + { + "epoch": 2.5520535910388755, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.279684066772461, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.746338963508606, + "num_tokens": 601187777.0, + "step": 23239 + }, + { + "epoch": 2.5521634087414893, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.3869118690490723, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.7485507726669312, + "num_tokens": 601211846.0, + "step": 23240 + }, + { + "epoch": 2.552273226444103, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.450230598449707, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7264084219932556, + "num_tokens": 601236545.0, + "step": 23241 + }, + { + "epoch": 2.5523830441467164, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2081565856933594, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7312043905258179, + "num_tokens": 601264423.0, + "step": 23242 + }, + { + "epoch": 2.55249286184933, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.2870259284973145, + "learning_rate": 1e-06, + "loss": 0.8543, + "mean_token_accuracy": 0.7219774723052979, + "num_tokens": 601290790.0, + "step": 23243 + }, + { + "epoch": 2.552602679551944, + "ewc_loss": 1.5139579772949219e-05, + "grad_norm": 2.5903167724609375, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7400608062744141, + "num_tokens": 601311762.0, + "step": 23244 + }, + { + "epoch": 2.552712497254557, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.27414870262146, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7357912063598633, + "num_tokens": 601338352.0, + "step": 23245 + }, + { + "epoch": 2.552822314957171, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.589667558670044, + "learning_rate": 1e-06, + "loss": 0.7802, + "mean_token_accuracy": 0.7541016340255737, + "num_tokens": 601360971.0, + "step": 23246 + }, + { + "epoch": 2.5529321326597847, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.5644311904907227, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7423814535140991, + "num_tokens": 601382171.0, + "step": 23247 + }, + { + "epoch": 2.5530419503623984, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.46268892288208, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7384086847305298, + "num_tokens": 601405621.0, + "step": 23248 + }, + { + "epoch": 2.553151768065012, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.283991813659668, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7266393899917603, + "num_tokens": 601433299.0, + "step": 23249 + }, + { + "epoch": 2.5532615857676255, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.162034749984741, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7364085912704468, + "num_tokens": 601461285.0, + "step": 23250 + }, + { + "epoch": 2.5533714034702393, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.2188916206359863, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7113425731658936, + "num_tokens": 601491882.0, + "step": 23251 + }, + { + "epoch": 2.553481221172853, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.7261111736297607, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7325695753097534, + "num_tokens": 601513290.0, + "step": 23252 + }, + { + "epoch": 2.553591038875467, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.401477098464966, + "learning_rate": 1e-06, + "loss": 0.9255, + "mean_token_accuracy": 0.7074936032295227, + "num_tokens": 601538138.0, + "step": 23253 + }, + { + "epoch": 2.5537008565780805, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.243870735168457, + "learning_rate": 1e-06, + "loss": 0.9281, + "mean_token_accuracy": 0.709425687789917, + "num_tokens": 601568430.0, + "step": 23254 + }, + { + "epoch": 2.553810674280694, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.382459878921509, + "learning_rate": 1e-06, + "loss": 0.8859, + "mean_token_accuracy": 0.7339330911636353, + "num_tokens": 601594573.0, + "step": 23255 + }, + { + "epoch": 2.5539204919833076, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.1987202167510986, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7432883977890015, + "num_tokens": 601622798.0, + "step": 23256 + }, + { + "epoch": 2.5540303096859214, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.552767515182495, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.727882981300354, + "num_tokens": 601645628.0, + "step": 23257 + }, + { + "epoch": 2.554140127388535, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.782008647918701, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7236435413360596, + "num_tokens": 601666359.0, + "step": 23258 + }, + { + "epoch": 2.554249945091149, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.556135654449463, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.727306604385376, + "num_tokens": 601687415.0, + "step": 23259 + }, + { + "epoch": 2.554359762793762, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.5278940200805664, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7081908583641052, + "num_tokens": 601712944.0, + "step": 23260 + }, + { + "epoch": 2.554469580496376, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.312312126159668, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7290213704109192, + "num_tokens": 601740554.0, + "step": 23261 + }, + { + "epoch": 2.5545793981989897, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.72654390335083, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.729354977607727, + "num_tokens": 601762863.0, + "step": 23262 + }, + { + "epoch": 2.5546892159016035, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.3597471714019775, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7256046533584595, + "num_tokens": 601790171.0, + "step": 23263 + }, + { + "epoch": 2.5547990336042172, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.221961259841919, + "learning_rate": 1e-06, + "loss": 0.7847, + "mean_token_accuracy": 0.7525316476821899, + "num_tokens": 601817728.0, + "step": 23264 + }, + { + "epoch": 2.5549088513068305, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.550780773162842, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.734717071056366, + "num_tokens": 601840399.0, + "step": 23265 + }, + { + "epoch": 2.5550186690094443, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.433201551437378, + "learning_rate": 1e-06, + "loss": 0.8026, + "mean_token_accuracy": 0.7486649751663208, + "num_tokens": 601864120.0, + "step": 23266 + }, + { + "epoch": 2.555128486712058, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.497096538543701, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.728918194770813, + "num_tokens": 601887840.0, + "step": 23267 + }, + { + "epoch": 2.5552383044146714, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.5752227306365967, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7195374965667725, + "num_tokens": 601911940.0, + "step": 23268 + }, + { + "epoch": 2.5553481221172856, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.7415707111358643, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7293477058410645, + "num_tokens": 601932390.0, + "step": 23269 + }, + { + "epoch": 2.555457939819899, + "ewc_loss": 1.519918441772461e-05, + "grad_norm": 2.231734275817871, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7325640916824341, + "num_tokens": 601961872.0, + "step": 23270 + }, + { + "epoch": 2.5555677575225126, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2585556507110596, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.722375750541687, + "num_tokens": 601990854.0, + "step": 23271 + }, + { + "epoch": 2.5556775752251264, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.69893741607666, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7395331859588623, + "num_tokens": 602010988.0, + "step": 23272 + }, + { + "epoch": 2.5557873929277397, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.402815818786621, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.742101788520813, + "num_tokens": 602035175.0, + "step": 23273 + }, + { + "epoch": 2.5558972106303535, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4206550121307373, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7220491170883179, + "num_tokens": 602057633.0, + "step": 23274 + }, + { + "epoch": 2.5560070283329672, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4025495052337646, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7472591400146484, + "num_tokens": 602082040.0, + "step": 23275 + }, + { + "epoch": 2.556116846035581, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.562962532043457, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7314964532852173, + "num_tokens": 602104266.0, + "step": 23276 + }, + { + "epoch": 2.5562266637381947, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.235775947570801, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7141814231872559, + "num_tokens": 602131165.0, + "step": 23277 + }, + { + "epoch": 2.556336481440808, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.355936050415039, + "learning_rate": 1e-06, + "loss": 0.7286, + "mean_token_accuracy": 0.760580837726593, + "num_tokens": 602155657.0, + "step": 23278 + }, + { + "epoch": 2.556446299143422, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.35388445854187, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7331768274307251, + "num_tokens": 602181965.0, + "step": 23279 + }, + { + "epoch": 2.5565561168460356, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.208434581756592, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7212170362472534, + "num_tokens": 602211976.0, + "step": 23280 + }, + { + "epoch": 2.5566659345486493, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.148512125015259, + "learning_rate": 1e-06, + "loss": 0.7848, + "mean_token_accuracy": 0.7533073425292969, + "num_tokens": 602239196.0, + "step": 23281 + }, + { + "epoch": 2.556775752251263, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2252254486083984, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7344647645950317, + "num_tokens": 602267444.0, + "step": 23282 + }, + { + "epoch": 2.5568855699538764, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.073183298110962, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.7292150855064392, + "num_tokens": 602298803.0, + "step": 23283 + }, + { + "epoch": 2.55699538765649, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.203521966934204, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.734378457069397, + "num_tokens": 602328423.0, + "step": 23284 + }, + { + "epoch": 2.557105205359104, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.590070962905884, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7231023907661438, + "num_tokens": 602350943.0, + "step": 23285 + }, + { + "epoch": 2.5572150230617177, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.245728015899658, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.736662745475769, + "num_tokens": 602378886.0, + "step": 23286 + }, + { + "epoch": 2.5573248407643314, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.51865553855896, + "learning_rate": 1e-06, + "loss": 0.754, + "mean_token_accuracy": 0.7569489479064941, + "num_tokens": 602400804.0, + "step": 23287 + }, + { + "epoch": 2.5574346584669447, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.40372633934021, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.7224617600440979, + "num_tokens": 602426240.0, + "step": 23288 + }, + { + "epoch": 2.5575444761695585, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.677255868911743, + "learning_rate": 1e-06, + "loss": 0.8759, + "mean_token_accuracy": 0.7328788042068481, + "num_tokens": 602447651.0, + "step": 23289 + }, + { + "epoch": 2.5576542938721722, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.8127801418304443, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7284641265869141, + "num_tokens": 602469324.0, + "step": 23290 + }, + { + "epoch": 2.557764111574786, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.361628532409668, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7557252645492554, + "num_tokens": 602493436.0, + "step": 23291 + }, + { + "epoch": 2.5578739292773998, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.220426082611084, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7208359241485596, + "num_tokens": 602524451.0, + "step": 23292 + }, + { + "epoch": 2.557983746980013, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2641639709472656, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7367851138114929, + "num_tokens": 602552139.0, + "step": 23293 + }, + { + "epoch": 2.558093564682627, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.1282317638397217, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7341060042381287, + "num_tokens": 602584609.0, + "step": 23294 + }, + { + "epoch": 2.5582033823852406, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2424862384796143, + "learning_rate": 1e-06, + "loss": 0.9487, + "mean_token_accuracy": 0.7088674306869507, + "num_tokens": 602613567.0, + "step": 23295 + }, + { + "epoch": 2.558313200087854, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 3.0311439037323, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7494222521781921, + "num_tokens": 602630565.0, + "step": 23296 + }, + { + "epoch": 2.5584230177904677, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4446492195129395, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7192585468292236, + "num_tokens": 602655516.0, + "step": 23297 + }, + { + "epoch": 2.5585328354930814, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2767333984375, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7366674542427063, + "num_tokens": 602681620.0, + "step": 23298 + }, + { + "epoch": 2.558642653195695, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.246981143951416, + "learning_rate": 1e-06, + "loss": 0.8039, + "mean_token_accuracy": 0.7473389506340027, + "num_tokens": 602707308.0, + "step": 23299 + }, + { + "epoch": 2.558752470898309, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.316528797149658, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7347939014434814, + "num_tokens": 602731336.0, + "step": 23300 + }, + { + "epoch": 2.5588622886009222, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.789165735244751, + "learning_rate": 1e-06, + "loss": 0.7645, + "mean_token_accuracy": 0.7529084086418152, + "num_tokens": 602751478.0, + "step": 23301 + }, + { + "epoch": 2.558972106303536, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.431079626083374, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7131897807121277, + "num_tokens": 602776797.0, + "step": 23302 + }, + { + "epoch": 2.5590819240061498, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4052908420562744, + "learning_rate": 1e-06, + "loss": 0.796, + "mean_token_accuracy": 0.7467387318611145, + "num_tokens": 602799985.0, + "step": 23303 + }, + { + "epoch": 2.5591917417087635, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4421660900115967, + "learning_rate": 1e-06, + "loss": 0.8435, + "mean_token_accuracy": 0.7301815748214722, + "num_tokens": 602825496.0, + "step": 23304 + }, + { + "epoch": 2.5593015594113773, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4681456089019775, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7214164733886719, + "num_tokens": 602849205.0, + "step": 23305 + }, + { + "epoch": 2.5594113771139906, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.1771223545074463, + "learning_rate": 1e-06, + "loss": 0.9179, + "mean_token_accuracy": 0.714109480381012, + "num_tokens": 602879621.0, + "step": 23306 + }, + { + "epoch": 2.5595211948166043, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.1853103637695312, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7405433654785156, + "num_tokens": 602908426.0, + "step": 23307 + }, + { + "epoch": 2.559631012519218, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3714776039123535, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7409179210662842, + "num_tokens": 602934702.0, + "step": 23308 + }, + { + "epoch": 2.559740830221832, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.6041855812072754, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7352446913719177, + "num_tokens": 602956640.0, + "step": 23309 + }, + { + "epoch": 2.5598506479244456, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3631961345672607, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.747505784034729, + "num_tokens": 602980767.0, + "step": 23310 + }, + { + "epoch": 2.559960465627059, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4141294956207275, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7379966974258423, + "num_tokens": 603004907.0, + "step": 23311 + }, + { + "epoch": 2.5600702833296727, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2745447158813477, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7392221689224243, + "num_tokens": 603031709.0, + "step": 23312 + }, + { + "epoch": 2.5601801010322864, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.232055425643921, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7172552347183228, + "num_tokens": 603061858.0, + "step": 23313 + }, + { + "epoch": 2.5602899187349, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.458632707595825, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7395076751708984, + "num_tokens": 603085881.0, + "step": 23314 + }, + { + "epoch": 2.560399736437514, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.9900224208831787, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7408707141876221, + "num_tokens": 603103768.0, + "step": 23315 + }, + { + "epoch": 2.5605095541401273, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.199434518814087, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7334873676300049, + "num_tokens": 603132650.0, + "step": 23316 + }, + { + "epoch": 2.560619371842741, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4724173545837402, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7390636205673218, + "num_tokens": 603157782.0, + "step": 23317 + }, + { + "epoch": 2.560729189545355, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.835580825805664, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7285877466201782, + "num_tokens": 603177854.0, + "step": 23318 + }, + { + "epoch": 2.5608390072479685, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.7181942462921143, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7491570115089417, + "num_tokens": 603199065.0, + "step": 23319 + }, + { + "epoch": 2.5609488249505823, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3434369564056396, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7339017391204834, + "num_tokens": 603225984.0, + "step": 23320 + }, + { + "epoch": 2.5610586426531956, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3117735385894775, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7341117262840271, + "num_tokens": 603253866.0, + "step": 23321 + }, + { + "epoch": 2.5611684603558094, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.5161304473876953, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.745532751083374, + "num_tokens": 603276823.0, + "step": 23322 + }, + { + "epoch": 2.561278278058423, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.498652935028076, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7549676895141602, + "num_tokens": 603298508.0, + "step": 23323 + }, + { + "epoch": 2.5613880957610364, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2345662117004395, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7205054759979248, + "num_tokens": 603329742.0, + "step": 23324 + }, + { + "epoch": 2.56149791346365, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.360995054244995, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7089735269546509, + "num_tokens": 603358198.0, + "step": 23325 + }, + { + "epoch": 2.561607731166264, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.452214002609253, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7407692670822144, + "num_tokens": 603383642.0, + "step": 23326 + }, + { + "epoch": 2.5617175488688777, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4107308387756348, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7467011213302612, + "num_tokens": 603409275.0, + "step": 23327 + }, + { + "epoch": 2.5618273665714915, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.0930275917053223, + "learning_rate": 1e-06, + "loss": 0.9329, + "mean_token_accuracy": 0.71529221534729, + "num_tokens": 603440126.0, + "step": 23328 + }, + { + "epoch": 2.5619371842741048, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.7688333988189697, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7411760687828064, + "num_tokens": 603459754.0, + "step": 23329 + }, + { + "epoch": 2.5620470019767185, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.031369924545288, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7136584520339966, + "num_tokens": 603491305.0, + "step": 23330 + }, + { + "epoch": 2.5621568196793323, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3265607357025146, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7408133745193481, + "num_tokens": 603516967.0, + "step": 23331 + }, + { + "epoch": 2.562266637381946, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2563376426696777, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7379711866378784, + "num_tokens": 603543598.0, + "step": 23332 + }, + { + "epoch": 2.56237645508456, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.1578993797302246, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.7298292517662048, + "num_tokens": 603572869.0, + "step": 23333 + }, + { + "epoch": 2.562486272787173, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.601637601852417, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.721082866191864, + "num_tokens": 603595639.0, + "step": 23334 + }, + { + "epoch": 2.562596090489787, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.629364490509033, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.715120792388916, + "num_tokens": 603619042.0, + "step": 23335 + }, + { + "epoch": 2.5627059081924006, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.0229079723358154, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7126142978668213, + "num_tokens": 603654904.0, + "step": 23336 + }, + { + "epoch": 2.5628157258950144, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4070937633514404, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.729088544845581, + "num_tokens": 603679865.0, + "step": 23337 + }, + { + "epoch": 2.562925543597628, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.590524196624756, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7341486215591431, + "num_tokens": 603705678.0, + "step": 23338 + }, + { + "epoch": 2.5630353613002415, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.549156904220581, + "learning_rate": 1e-06, + "loss": 0.9528, + "mean_token_accuracy": 0.710528552532196, + "num_tokens": 603729771.0, + "step": 23339 + }, + { + "epoch": 2.563145179002855, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.336850643157959, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7242874503135681, + "num_tokens": 603755805.0, + "step": 23340 + }, + { + "epoch": 2.563254996705469, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3574938774108887, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7188646197319031, + "num_tokens": 603784069.0, + "step": 23341 + }, + { + "epoch": 2.5633648144080827, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3706424236297607, + "learning_rate": 1e-06, + "loss": 0.7933, + "mean_token_accuracy": 0.7417064905166626, + "num_tokens": 603809966.0, + "step": 23342 + }, + { + "epoch": 2.5634746321106965, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.5424818992614746, + "learning_rate": 1e-06, + "loss": 0.8918, + "mean_token_accuracy": 0.7160419225692749, + "num_tokens": 603833997.0, + "step": 23343 + }, + { + "epoch": 2.56358444981331, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2459404468536377, + "learning_rate": 1e-06, + "loss": 0.9141, + "mean_token_accuracy": 0.7141749262809753, + "num_tokens": 603862599.0, + "step": 23344 + }, + { + "epoch": 2.5636942675159236, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.383801221847534, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.7123000025749207, + "num_tokens": 603889002.0, + "step": 23345 + }, + { + "epoch": 2.5638040852185373, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.3415818214416504, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7198107242584229, + "num_tokens": 603915766.0, + "step": 23346 + }, + { + "epoch": 2.5639139029211506, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.099691390991211, + "learning_rate": 1e-06, + "loss": 0.9334, + "mean_token_accuracy": 0.7076479196548462, + "num_tokens": 603949935.0, + "step": 23347 + }, + { + "epoch": 2.564023720623765, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.386836528778076, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.750041127204895, + "num_tokens": 603976908.0, + "step": 23348 + }, + { + "epoch": 2.564133538326378, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.4890480041503906, + "learning_rate": 1e-06, + "loss": 0.9008, + "mean_token_accuracy": 0.7254699468612671, + "num_tokens": 604002638.0, + "step": 23349 + }, + { + "epoch": 2.564243356028992, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.223224639892578, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7165600657463074, + "num_tokens": 604032425.0, + "step": 23350 + }, + { + "epoch": 2.5643531737316057, + "ewc_loss": 1.52587890625e-05, + "grad_norm": 2.2173805236816406, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.729114294052124, + "num_tokens": 604062291.0, + "step": 23351 + }, + { + "epoch": 2.564462991434219, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4844133853912354, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7306830883026123, + "num_tokens": 604085436.0, + "step": 23352 + }, + { + "epoch": 2.5645728091368327, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5658135414123535, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7201715111732483, + "num_tokens": 604108637.0, + "step": 23353 + }, + { + "epoch": 2.5646826268394465, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.441343069076538, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7226988077163696, + "num_tokens": 604133226.0, + "step": 23354 + }, + { + "epoch": 2.5647924445420602, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.405158519744873, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7289499044418335, + "num_tokens": 604159178.0, + "step": 23355 + }, + { + "epoch": 2.564902262244674, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6387016773223877, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7411065697669983, + "num_tokens": 604181713.0, + "step": 23356 + }, + { + "epoch": 2.5650120799472873, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.1294875144958496, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.7231081128120422, + "num_tokens": 604211696.0, + "step": 23357 + }, + { + "epoch": 2.565121897649901, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.280144691467285, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7391910552978516, + "num_tokens": 604238132.0, + "step": 23358 + }, + { + "epoch": 2.565231715352515, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2202048301696777, + "learning_rate": 1e-06, + "loss": 0.8488, + "mean_token_accuracy": 0.7315946221351624, + "num_tokens": 604267759.0, + "step": 23359 + }, + { + "epoch": 2.5653415330551286, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.365558624267578, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7286611795425415, + "num_tokens": 604295286.0, + "step": 23360 + }, + { + "epoch": 2.5654513507577423, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5260589122772217, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7386116981506348, + "num_tokens": 604318751.0, + "step": 23361 + }, + { + "epoch": 2.5655611684603556, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2808876037597656, + "learning_rate": 1e-06, + "loss": 0.7511, + "mean_token_accuracy": 0.7562578916549683, + "num_tokens": 604343713.0, + "step": 23362 + }, + { + "epoch": 2.5656709861629694, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.07187819480896, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7162174582481384, + "num_tokens": 604376454.0, + "step": 23363 + }, + { + "epoch": 2.565780803865583, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6402876377105713, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.744662880897522, + "num_tokens": 604397121.0, + "step": 23364 + }, + { + "epoch": 2.565890621568197, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.581284999847412, + "learning_rate": 1e-06, + "loss": 0.7341, + "mean_token_accuracy": 0.7648720145225525, + "num_tokens": 604417671.0, + "step": 23365 + }, + { + "epoch": 2.5660004392708107, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5083229541778564, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7266477346420288, + "num_tokens": 604441561.0, + "step": 23366 + }, + { + "epoch": 2.566110256973424, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4318575859069824, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7246301174163818, + "num_tokens": 604467657.0, + "step": 23367 + }, + { + "epoch": 2.5662200746760377, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3859410285949707, + "learning_rate": 1e-06, + "loss": 0.9118, + "mean_token_accuracy": 0.7163850665092468, + "num_tokens": 604497964.0, + "step": 23368 + }, + { + "epoch": 2.5663298923786515, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.401780843734741, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7305549383163452, + "num_tokens": 604523342.0, + "step": 23369 + }, + { + "epoch": 2.5664397100812653, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.416363477706909, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7227065563201904, + "num_tokens": 604548761.0, + "step": 23370 + }, + { + "epoch": 2.566549527783879, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.475167751312256, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7164605855941772, + "num_tokens": 604574531.0, + "step": 23371 + }, + { + "epoch": 2.5666593454864923, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.059539794921875, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7122496366500854, + "num_tokens": 604605288.0, + "step": 23372 + }, + { + "epoch": 2.566769163189106, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.8771731853485107, + "learning_rate": 1e-06, + "loss": 0.7691, + "mean_token_accuracy": 0.7578172087669373, + "num_tokens": 604624112.0, + "step": 23373 + }, + { + "epoch": 2.56687898089172, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.488154649734497, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7442497611045837, + "num_tokens": 604647904.0, + "step": 23374 + }, + { + "epoch": 2.566988798594333, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.51921010017395, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7297235727310181, + "num_tokens": 604674123.0, + "step": 23375 + }, + { + "epoch": 2.567098616296947, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5300965309143066, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7449746131896973, + "num_tokens": 604698545.0, + "step": 23376 + }, + { + "epoch": 2.5672084339995607, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2349562644958496, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.726474404335022, + "num_tokens": 604728830.0, + "step": 23377 + }, + { + "epoch": 2.5673182517021744, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.428306818008423, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7526201605796814, + "num_tokens": 604753387.0, + "step": 23378 + }, + { + "epoch": 2.567428069404788, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3537166118621826, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7486761212348938, + "num_tokens": 604778137.0, + "step": 23379 + }, + { + "epoch": 2.5675378871074015, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 1.849918007850647, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7337599992752075, + "num_tokens": 604814087.0, + "step": 23380 + }, + { + "epoch": 2.5676477048100153, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4914886951446533, + "learning_rate": 1e-06, + "loss": 0.7459, + "mean_token_accuracy": 0.7599307894706726, + "num_tokens": 604836600.0, + "step": 23381 + }, + { + "epoch": 2.567757522512629, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6944167613983154, + "learning_rate": 1e-06, + "loss": 0.7627, + "mean_token_accuracy": 0.7553716897964478, + "num_tokens": 604855732.0, + "step": 23382 + }, + { + "epoch": 2.5678673402152428, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.342341899871826, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7510861158370972, + "num_tokens": 604883780.0, + "step": 23383 + }, + { + "epoch": 2.5679771579178565, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4939897060394287, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7512009143829346, + "num_tokens": 604906772.0, + "step": 23384 + }, + { + "epoch": 2.56808697562047, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.186668872833252, + "learning_rate": 1e-06, + "loss": 0.9458, + "mean_token_accuracy": 0.7017829418182373, + "num_tokens": 604937195.0, + "step": 23385 + }, + { + "epoch": 2.5681967933230836, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.351215124130249, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7275346517562866, + "num_tokens": 604964163.0, + "step": 23386 + }, + { + "epoch": 2.5683066110256974, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.21505069732666, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7189251184463501, + "num_tokens": 604992272.0, + "step": 23387 + }, + { + "epoch": 2.568416428728311, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.393299102783203, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7306210994720459, + "num_tokens": 605016750.0, + "step": 23388 + }, + { + "epoch": 2.568526246430925, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.291395664215088, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7539018392562866, + "num_tokens": 605041357.0, + "step": 23389 + }, + { + "epoch": 2.568636064133538, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2637646198272705, + "learning_rate": 1e-06, + "loss": 0.7887, + "mean_token_accuracy": 0.7444306015968323, + "num_tokens": 605067758.0, + "step": 23390 + }, + { + "epoch": 2.568745881836152, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3863282203674316, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7391996383666992, + "num_tokens": 605092976.0, + "step": 23391 + }, + { + "epoch": 2.5688556995387657, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.1908252239227295, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.726455807685852, + "num_tokens": 605121236.0, + "step": 23392 + }, + { + "epoch": 2.5689655172413794, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.105722665786743, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7271169424057007, + "num_tokens": 605152048.0, + "step": 23393 + }, + { + "epoch": 2.569075334943993, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6654632091522217, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7467424869537354, + "num_tokens": 605172717.0, + "step": 23394 + }, + { + "epoch": 2.5691851526466065, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.484308958053589, + "learning_rate": 1e-06, + "loss": 0.7618, + "mean_token_accuracy": 0.7534297704696655, + "num_tokens": 605193782.0, + "step": 23395 + }, + { + "epoch": 2.5692949703492203, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.414407730102539, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.729497492313385, + "num_tokens": 605218931.0, + "step": 23396 + }, + { + "epoch": 2.569404788051834, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3486602306365967, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.7250794172286987, + "num_tokens": 605246891.0, + "step": 23397 + }, + { + "epoch": 2.5695146057544473, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3116962909698486, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.739824116230011, + "num_tokens": 605272368.0, + "step": 23398 + }, + { + "epoch": 2.5696244234570615, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.6885087490081787, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7484447956085205, + "num_tokens": 605292360.0, + "step": 23399 + }, + { + "epoch": 2.569734241159675, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.306002616882324, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7154221534729004, + "num_tokens": 605322031.0, + "step": 23400 + }, + { + "epoch": 2.5698440588622886, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3405282497406006, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7387697100639343, + "num_tokens": 605347994.0, + "step": 23401 + }, + { + "epoch": 2.5699538765649024, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5306179523468018, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7266771197319031, + "num_tokens": 605372584.0, + "step": 23402 + }, + { + "epoch": 2.5700636942675157, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3723456859588623, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7362040281295776, + "num_tokens": 605398031.0, + "step": 23403 + }, + { + "epoch": 2.5701735119701294, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.7110841274261475, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7290439605712891, + "num_tokens": 605420220.0, + "step": 23404 + }, + { + "epoch": 2.570283329672743, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.301628589630127, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7141382694244385, + "num_tokens": 605447632.0, + "step": 23405 + }, + { + "epoch": 2.570393147375357, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.5470194816589355, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7369502186775208, + "num_tokens": 605469430.0, + "step": 23406 + }, + { + "epoch": 2.5705029650779707, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.177872896194458, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7077313661575317, + "num_tokens": 605499226.0, + "step": 23407 + }, + { + "epoch": 2.570612782780584, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2296628952026367, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7205605506896973, + "num_tokens": 605527397.0, + "step": 23408 + }, + { + "epoch": 2.570722600483198, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2064449787139893, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7195428609848022, + "num_tokens": 605553009.0, + "step": 23409 + }, + { + "epoch": 2.5708324181858115, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.798917055130005, + "learning_rate": 1e-06, + "loss": 0.7564, + "mean_token_accuracy": 0.7532768249511719, + "num_tokens": 605572554.0, + "step": 23410 + }, + { + "epoch": 2.5709422358884253, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3081603050231934, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7258983850479126, + "num_tokens": 605600291.0, + "step": 23411 + }, + { + "epoch": 2.571052053591039, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.453448534011841, + "learning_rate": 1e-06, + "loss": 0.7039, + "mean_token_accuracy": 0.7721227407455444, + "num_tokens": 605624462.0, + "step": 23412 + }, + { + "epoch": 2.5711618712936524, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.156083583831787, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7378613948822021, + "num_tokens": 605654859.0, + "step": 23413 + }, + { + "epoch": 2.571271688996266, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.550962209701538, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7516618371009827, + "num_tokens": 605677494.0, + "step": 23414 + }, + { + "epoch": 2.57138150669888, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2598562240600586, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7250348329544067, + "num_tokens": 605703942.0, + "step": 23415 + }, + { + "epoch": 2.5714913244014936, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.493903160095215, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7084600925445557, + "num_tokens": 605728956.0, + "step": 23416 + }, + { + "epoch": 2.5716011421041074, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4550957679748535, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.7511885762214661, + "num_tokens": 605752376.0, + "step": 23417 + }, + { + "epoch": 2.5717109598067207, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4680166244506836, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.72609543800354, + "num_tokens": 605776295.0, + "step": 23418 + }, + { + "epoch": 2.5718207775093345, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.503591299057007, + "learning_rate": 1e-06, + "loss": 0.7768, + "mean_token_accuracy": 0.7549657821655273, + "num_tokens": 605798727.0, + "step": 23419 + }, + { + "epoch": 2.5719305952119482, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.197047472000122, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7057441473007202, + "num_tokens": 605830369.0, + "step": 23420 + }, + { + "epoch": 2.572040412914562, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4842450618743896, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.731806755065918, + "num_tokens": 605853381.0, + "step": 23421 + }, + { + "epoch": 2.5721502306171757, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2693803310394287, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7237356305122375, + "num_tokens": 605881397.0, + "step": 23422 + }, + { + "epoch": 2.572260048319789, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4525341987609863, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7299454212188721, + "num_tokens": 605905674.0, + "step": 23423 + }, + { + "epoch": 2.572369866022403, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3934431076049805, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7296328544616699, + "num_tokens": 605931293.0, + "step": 23424 + }, + { + "epoch": 2.5724796837250166, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.8664066791534424, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7319605350494385, + "num_tokens": 605949840.0, + "step": 23425 + }, + { + "epoch": 2.57258950142763, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.7281267642974854, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7536265850067139, + "num_tokens": 605969918.0, + "step": 23426 + }, + { + "epoch": 2.5726993191302436, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.396718740463257, + "learning_rate": 1e-06, + "loss": 0.9672, + "mean_token_accuracy": 0.7040225267410278, + "num_tokens": 605997400.0, + "step": 23427 + }, + { + "epoch": 2.5728091368328574, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.4107167720794678, + "learning_rate": 1e-06, + "loss": 0.9437, + "mean_token_accuracy": 0.7134655714035034, + "num_tokens": 606022978.0, + "step": 23428 + }, + { + "epoch": 2.572918954535471, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.664703845977783, + "learning_rate": 1e-06, + "loss": 0.7342, + "mean_token_accuracy": 0.7659602165222168, + "num_tokens": 606043132.0, + "step": 23429 + }, + { + "epoch": 2.573028772238085, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.463616132736206, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7128812074661255, + "num_tokens": 606070287.0, + "step": 23430 + }, + { + "epoch": 2.573138589940698, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.1685030460357666, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7222108840942383, + "num_tokens": 606100701.0, + "step": 23431 + }, + { + "epoch": 2.573248407643312, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.2740001678466797, + "learning_rate": 1e-06, + "loss": 0.8314, + "mean_token_accuracy": 0.7342987060546875, + "num_tokens": 606128149.0, + "step": 23432 + }, + { + "epoch": 2.5733582253459257, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.644277334213257, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7227587699890137, + "num_tokens": 606150533.0, + "step": 23433 + }, + { + "epoch": 2.5734680430485395, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.7257606983184814, + "learning_rate": 1e-06, + "loss": 0.7178, + "mean_token_accuracy": 0.7688592672348022, + "num_tokens": 606169482.0, + "step": 23434 + }, + { + "epoch": 2.5735778607511532, + "ewc_loss": 1.537799835205078e-05, + "grad_norm": 2.3356456756591797, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7385221719741821, + "num_tokens": 606195832.0, + "step": 23435 + }, + { + "epoch": 2.5736876784537666, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.172563076019287, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7315049767494202, + "num_tokens": 606225471.0, + "step": 23436 + }, + { + "epoch": 2.5737974961563803, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3277833461761475, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7148393988609314, + "num_tokens": 606251218.0, + "step": 23437 + }, + { + "epoch": 2.573907313858994, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.0128562450408936, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7107388973236084, + "num_tokens": 606287133.0, + "step": 23438 + }, + { + "epoch": 2.574017131561608, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.063209056854248, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7160599231719971, + "num_tokens": 606319512.0, + "step": 23439 + }, + { + "epoch": 2.5741269492642216, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.0633349418640137, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7314020395278931, + "num_tokens": 606351132.0, + "step": 23440 + }, + { + "epoch": 2.574236766966835, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2649292945861816, + "learning_rate": 1e-06, + "loss": 0.7455, + "mean_token_accuracy": 0.766872227191925, + "num_tokens": 606377231.0, + "step": 23441 + }, + { + "epoch": 2.5743465846694487, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4423041343688965, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7243324518203735, + "num_tokens": 606401545.0, + "step": 23442 + }, + { + "epoch": 2.5744564023720624, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3690361976623535, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7430902719497681, + "num_tokens": 606426396.0, + "step": 23443 + }, + { + "epoch": 2.574566220074676, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2796707153320312, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7381911277770996, + "num_tokens": 606455352.0, + "step": 23444 + }, + { + "epoch": 2.57467603777729, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.532484292984009, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7390908598899841, + "num_tokens": 606478255.0, + "step": 23445 + }, + { + "epoch": 2.5747858554799032, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.585939884185791, + "learning_rate": 1e-06, + "loss": 0.9159, + "mean_token_accuracy": 0.7186180949211121, + "num_tokens": 606502966.0, + "step": 23446 + }, + { + "epoch": 2.574895673182517, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.850148916244507, + "learning_rate": 1e-06, + "loss": 0.7897, + "mean_token_accuracy": 0.7434440851211548, + "num_tokens": 606520752.0, + "step": 23447 + }, + { + "epoch": 2.5750054908851308, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1689391136169434, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7404524087905884, + "num_tokens": 606549079.0, + "step": 23448 + }, + { + "epoch": 2.575115308587744, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5188302993774414, + "learning_rate": 1e-06, + "loss": 0.8273, + "mean_token_accuracy": 0.7340220212936401, + "num_tokens": 606571157.0, + "step": 23449 + }, + { + "epoch": 2.5752251262903583, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3533682823181152, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7227286100387573, + "num_tokens": 606596777.0, + "step": 23450 + }, + { + "epoch": 2.5753349439929716, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5130345821380615, + "learning_rate": 1e-06, + "loss": 0.836, + "mean_token_accuracy": 0.7341868281364441, + "num_tokens": 606619862.0, + "step": 23451 + }, + { + "epoch": 2.5754447616955853, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.561518669128418, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7350590229034424, + "num_tokens": 606643526.0, + "step": 23452 + }, + { + "epoch": 2.575554579398199, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.650129556655884, + "learning_rate": 1e-06, + "loss": 0.7865, + "mean_token_accuracy": 0.7437252998352051, + "num_tokens": 606665192.0, + "step": 23453 + }, + { + "epoch": 2.5756643971008124, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2769954204559326, + "learning_rate": 1e-06, + "loss": 0.9645, + "mean_token_accuracy": 0.709415078163147, + "num_tokens": 606694666.0, + "step": 23454 + }, + { + "epoch": 2.575774214803426, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2921481132507324, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7165120840072632, + "num_tokens": 606721178.0, + "step": 23455 + }, + { + "epoch": 2.57588403250604, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.207517385482788, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7330014705657959, + "num_tokens": 606751290.0, + "step": 23456 + }, + { + "epoch": 2.5759938502086537, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.366987466812134, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.721940279006958, + "num_tokens": 606777418.0, + "step": 23457 + }, + { + "epoch": 2.5761036679112674, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.413646697998047, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7362461090087891, + "num_tokens": 606802705.0, + "step": 23458 + }, + { + "epoch": 2.5762134856138807, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.395447015762329, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7411154508590698, + "num_tokens": 606826176.0, + "step": 23459 + }, + { + "epoch": 2.5763233033164945, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4348604679107666, + "learning_rate": 1e-06, + "loss": 0.7541, + "mean_token_accuracy": 0.7577311992645264, + "num_tokens": 606851494.0, + "step": 23460 + }, + { + "epoch": 2.5764331210191083, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.200099229812622, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7202363610267639, + "num_tokens": 606880711.0, + "step": 23461 + }, + { + "epoch": 2.576542938721722, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2494044303894043, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7133715748786926, + "num_tokens": 606909157.0, + "step": 23462 + }, + { + "epoch": 2.5766527564243358, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.539781332015991, + "learning_rate": 1e-06, + "loss": 0.8945, + "mean_token_accuracy": 0.7239272594451904, + "num_tokens": 606934181.0, + "step": 23463 + }, + { + "epoch": 2.576762574126949, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2042934894561768, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7011256217956543, + "num_tokens": 606967102.0, + "step": 23464 + }, + { + "epoch": 2.576872391829563, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1420676708221436, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7436215877532959, + "num_tokens": 606994772.0, + "step": 23465 + }, + { + "epoch": 2.5769822095321766, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2300949096679688, + "learning_rate": 1e-06, + "loss": 0.9503, + "mean_token_accuracy": 0.7078372240066528, + "num_tokens": 607024344.0, + "step": 23466 + }, + { + "epoch": 2.5770920272347904, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5983026027679443, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7333165407180786, + "num_tokens": 607045444.0, + "step": 23467 + }, + { + "epoch": 2.577201844937404, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.274489164352417, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7461917996406555, + "num_tokens": 607072488.0, + "step": 23468 + }, + { + "epoch": 2.5773116626400174, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5392508506774902, + "learning_rate": 1e-06, + "loss": 0.9346, + "mean_token_accuracy": 0.7190130949020386, + "num_tokens": 607095180.0, + "step": 23469 + }, + { + "epoch": 2.577421480342631, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4807989597320557, + "learning_rate": 1e-06, + "loss": 0.8068, + "mean_token_accuracy": 0.7479681968688965, + "num_tokens": 607118068.0, + "step": 23470 + }, + { + "epoch": 2.577531298045245, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1988654136657715, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7269909381866455, + "num_tokens": 607151100.0, + "step": 23471 + }, + { + "epoch": 2.5776411157478587, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4487061500549316, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7256243824958801, + "num_tokens": 607176109.0, + "step": 23472 + }, + { + "epoch": 2.5777509334504725, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5072131156921387, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.745071530342102, + "num_tokens": 607200271.0, + "step": 23473 + }, + { + "epoch": 2.5778607511530858, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1351394653320312, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.712989330291748, + "num_tokens": 607231627.0, + "step": 23474 + }, + { + "epoch": 2.5779705688556995, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2823898792266846, + "learning_rate": 1e-06, + "loss": 0.8244, + "mean_token_accuracy": 0.7332895994186401, + "num_tokens": 607261567.0, + "step": 23475 + }, + { + "epoch": 2.5780803865583133, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.539661169052124, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.740551233291626, + "num_tokens": 607283953.0, + "step": 23476 + }, + { + "epoch": 2.5781902042609266, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1831228733062744, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.700772225856781, + "num_tokens": 607314055.0, + "step": 23477 + }, + { + "epoch": 2.578300021963541, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2698402404785156, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7384768724441528, + "num_tokens": 607341826.0, + "step": 23478 + }, + { + "epoch": 2.578409839666154, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4935240745544434, + "learning_rate": 1e-06, + "loss": 0.7848, + "mean_token_accuracy": 0.7472892999649048, + "num_tokens": 607362328.0, + "step": 23479 + }, + { + "epoch": 2.578519657368768, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.6360647678375244, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7339085936546326, + "num_tokens": 607385401.0, + "step": 23480 + }, + { + "epoch": 2.5786294750713816, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2706458568573, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7333104610443115, + "num_tokens": 607411585.0, + "step": 23481 + }, + { + "epoch": 2.578739292773995, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4541451930999756, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7215713262557983, + "num_tokens": 607438405.0, + "step": 23482 + }, + { + "epoch": 2.5788491104766087, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.319772720336914, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7418452501296997, + "num_tokens": 607464513.0, + "step": 23483 + }, + { + "epoch": 2.5789589281792225, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.420548915863037, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7418150305747986, + "num_tokens": 607489036.0, + "step": 23484 + }, + { + "epoch": 2.579068745881836, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.138418436050415, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7234543561935425, + "num_tokens": 607518788.0, + "step": 23485 + }, + { + "epoch": 2.57917856358445, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2327721118927, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7416123151779175, + "num_tokens": 607545199.0, + "step": 23486 + }, + { + "epoch": 2.5792883812870633, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.314307689666748, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7150125503540039, + "num_tokens": 607571705.0, + "step": 23487 + }, + { + "epoch": 2.579398198989677, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.203380823135376, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7445796728134155, + "num_tokens": 607600658.0, + "step": 23488 + }, + { + "epoch": 2.579508016692291, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.566075325012207, + "learning_rate": 1e-06, + "loss": 0.7851, + "mean_token_accuracy": 0.7474068999290466, + "num_tokens": 607622523.0, + "step": 23489 + }, + { + "epoch": 2.5796178343949046, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.237515687942505, + "learning_rate": 1e-06, + "loss": 0.7798, + "mean_token_accuracy": 0.7508503198623657, + "num_tokens": 607649235.0, + "step": 23490 + }, + { + "epoch": 2.5797276520975183, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.378615379333496, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.738061249256134, + "num_tokens": 607674183.0, + "step": 23491 + }, + { + "epoch": 2.5798374698001316, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.556549072265625, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7205491662025452, + "num_tokens": 607697653.0, + "step": 23492 + }, + { + "epoch": 2.5799472875027454, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3897793292999268, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7344907522201538, + "num_tokens": 607720883.0, + "step": 23493 + }, + { + "epoch": 2.580057105205359, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.6117074489593506, + "learning_rate": 1e-06, + "loss": 0.7851, + "mean_token_accuracy": 0.7437622547149658, + "num_tokens": 607742152.0, + "step": 23494 + }, + { + "epoch": 2.580166922907973, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.315777063369751, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7327597737312317, + "num_tokens": 607768335.0, + "step": 23495 + }, + { + "epoch": 2.5802767406105866, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2835488319396973, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.724355936050415, + "num_tokens": 607795566.0, + "step": 23496 + }, + { + "epoch": 2.5803865583132, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2495598793029785, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7146862149238586, + "num_tokens": 607825518.0, + "step": 23497 + }, + { + "epoch": 2.5804963760158137, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3344438076019287, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7404848337173462, + "num_tokens": 607850631.0, + "step": 23498 + }, + { + "epoch": 2.5806061937184275, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.1749463081359863, + "learning_rate": 1e-06, + "loss": 0.9522, + "mean_token_accuracy": 0.7019232511520386, + "num_tokens": 607880161.0, + "step": 23499 + }, + { + "epoch": 2.5807160114210412, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 1.9821971654891968, + "learning_rate": 1e-06, + "loss": 0.9457, + "mean_token_accuracy": 0.7095935344696045, + "num_tokens": 607912535.0, + "step": 23500 + }, + { + "epoch": 2.580825829123655, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 1.9621710777282715, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7289963364601135, + "num_tokens": 607949154.0, + "step": 23501 + }, + { + "epoch": 2.5809356468262683, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2857825756073, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.715267539024353, + "num_tokens": 607978259.0, + "step": 23502 + }, + { + "epoch": 2.581045464528882, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5787363052368164, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7290029525756836, + "num_tokens": 607999828.0, + "step": 23503 + }, + { + "epoch": 2.581155282231496, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.443824529647827, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7244969606399536, + "num_tokens": 608026370.0, + "step": 23504 + }, + { + "epoch": 2.581265099934109, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3301453590393066, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7233937978744507, + "num_tokens": 608052186.0, + "step": 23505 + }, + { + "epoch": 2.581374917636723, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.6945369243621826, + "learning_rate": 1e-06, + "loss": 0.7696, + "mean_token_accuracy": 0.7526475191116333, + "num_tokens": 608071921.0, + "step": 23506 + }, + { + "epoch": 2.5814847353393366, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.6439037322998047, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.736548662185669, + "num_tokens": 608094605.0, + "step": 23507 + }, + { + "epoch": 2.5815945530419504, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2944600582122803, + "learning_rate": 1e-06, + "loss": 0.9587, + "mean_token_accuracy": 0.7045198678970337, + "num_tokens": 608121527.0, + "step": 23508 + }, + { + "epoch": 2.581704370744564, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.313359498977661, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7243520021438599, + "num_tokens": 608148738.0, + "step": 23509 + }, + { + "epoch": 2.5818141884471775, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.2233426570892334, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7179527282714844, + "num_tokens": 608178076.0, + "step": 23510 + }, + { + "epoch": 2.5819240061497912, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.5650510787963867, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7517770528793335, + "num_tokens": 608200723.0, + "step": 23511 + }, + { + "epoch": 2.582033823852405, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.565265655517578, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.7374223470687866, + "num_tokens": 608223552.0, + "step": 23512 + }, + { + "epoch": 2.5821436415550187, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.3366198539733887, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7201042771339417, + "num_tokens": 608248956.0, + "step": 23513 + }, + { + "epoch": 2.5822534592576325, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.10272216796875, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7296286225318909, + "num_tokens": 608281144.0, + "step": 23514 + }, + { + "epoch": 2.582363276960246, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.9256820678710938, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7426672577857971, + "num_tokens": 608298919.0, + "step": 23515 + }, + { + "epoch": 2.5824730946628596, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.7792720794677734, + "learning_rate": 1e-06, + "loss": 0.7433, + "mean_token_accuracy": 0.7628166675567627, + "num_tokens": 608317932.0, + "step": 23516 + }, + { + "epoch": 2.5825829123654733, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.4410486221313477, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7334709763526917, + "num_tokens": 608343042.0, + "step": 23517 + }, + { + "epoch": 2.582692730068087, + "ewc_loss": 1.5497207641601562e-05, + "grad_norm": 2.368638277053833, + "learning_rate": 1e-06, + "loss": 0.7304, + "mean_token_accuracy": 0.7591428160667419, + "num_tokens": 608366394.0, + "step": 23518 + }, + { + "epoch": 2.582802547770701, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.554520845413208, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7204290628433228, + "num_tokens": 608391877.0, + "step": 23519 + }, + { + "epoch": 2.582912365473314, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1956446170806885, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7135095596313477, + "num_tokens": 608422930.0, + "step": 23520 + }, + { + "epoch": 2.583022183175928, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.223918914794922, + "learning_rate": 1e-06, + "loss": 0.9527, + "mean_token_accuracy": 0.7008468508720398, + "num_tokens": 608452170.0, + "step": 23521 + }, + { + "epoch": 2.5831320008785417, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5577690601348877, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7470835447311401, + "num_tokens": 608474138.0, + "step": 23522 + }, + { + "epoch": 2.5832418185811554, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.285249948501587, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7469276189804077, + "num_tokens": 608500710.0, + "step": 23523 + }, + { + "epoch": 2.583351636283769, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3280105590820312, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7462785840034485, + "num_tokens": 608526318.0, + "step": 23524 + }, + { + "epoch": 2.5834614539863825, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.515134334564209, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.7370182871818542, + "num_tokens": 608549646.0, + "step": 23525 + }, + { + "epoch": 2.5835712716889963, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3802380561828613, + "learning_rate": 1e-06, + "loss": 0.9205, + "mean_token_accuracy": 0.7143635749816895, + "num_tokens": 608576690.0, + "step": 23526 + }, + { + "epoch": 2.58368108939161, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2332687377929688, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.721350908279419, + "num_tokens": 608605824.0, + "step": 23527 + }, + { + "epoch": 2.5837909070942233, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.398465156555176, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7310090065002441, + "num_tokens": 608631410.0, + "step": 23528 + }, + { + "epoch": 2.5839007247968375, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3295321464538574, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7281457185745239, + "num_tokens": 608659695.0, + "step": 23529 + }, + { + "epoch": 2.584010542499451, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.385512113571167, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7221982479095459, + "num_tokens": 608684659.0, + "step": 23530 + }, + { + "epoch": 2.5841203602020646, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3369741439819336, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7320283651351929, + "num_tokens": 608712998.0, + "step": 23531 + }, + { + "epoch": 2.5842301779046783, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1061060428619385, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7219381928443909, + "num_tokens": 608742565.0, + "step": 23532 + }, + { + "epoch": 2.5843399956072917, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.161731481552124, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7351407408714294, + "num_tokens": 608772595.0, + "step": 23533 + }, + { + "epoch": 2.5844498133099054, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.296478748321533, + "learning_rate": 1e-06, + "loss": 0.9106, + "mean_token_accuracy": 0.7163295745849609, + "num_tokens": 608800380.0, + "step": 23534 + }, + { + "epoch": 2.584559631012519, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.334231376647949, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7095537185668945, + "num_tokens": 608826483.0, + "step": 23535 + }, + { + "epoch": 2.584669448715133, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4276318550109863, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7231796979904175, + "num_tokens": 608851677.0, + "step": 23536 + }, + { + "epoch": 2.5847792664177467, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3453104496002197, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7256482839584351, + "num_tokens": 608877458.0, + "step": 23537 + }, + { + "epoch": 2.58488908412036, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.421632766723633, + "learning_rate": 1e-06, + "loss": 0.7802, + "mean_token_accuracy": 0.7477821111679077, + "num_tokens": 608903584.0, + "step": 23538 + }, + { + "epoch": 2.5849989018229738, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.259672164916992, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7282887697219849, + "num_tokens": 608931746.0, + "step": 23539 + }, + { + "epoch": 2.5851087195255875, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2167696952819824, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7018587589263916, + "num_tokens": 608960365.0, + "step": 23540 + }, + { + "epoch": 2.5852185372282013, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1824936866760254, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7254804968833923, + "num_tokens": 608990885.0, + "step": 23541 + }, + { + "epoch": 2.585328354930815, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.16969633102417, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.727873682975769, + "num_tokens": 609018899.0, + "step": 23542 + }, + { + "epoch": 2.5854381726334283, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.449289083480835, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7230573892593384, + "num_tokens": 609042820.0, + "step": 23543 + }, + { + "epoch": 2.585547990336042, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1795456409454346, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7305712699890137, + "num_tokens": 609071626.0, + "step": 23544 + }, + { + "epoch": 2.585657808038656, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5867202281951904, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.725375771522522, + "num_tokens": 609094227.0, + "step": 23545 + }, + { + "epoch": 2.5857676257412696, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1804912090301514, + "learning_rate": 1e-06, + "loss": 0.8785, + "mean_token_accuracy": 0.7269166111946106, + "num_tokens": 609125138.0, + "step": 23546 + }, + { + "epoch": 2.5858774434438834, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2766005992889404, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7513948082923889, + "num_tokens": 609153763.0, + "step": 23547 + }, + { + "epoch": 2.5859872611464967, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3236896991729736, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7211124897003174, + "num_tokens": 609181468.0, + "step": 23548 + }, + { + "epoch": 2.5860970788491104, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4951562881469727, + "learning_rate": 1e-06, + "loss": 0.7533, + "mean_token_accuracy": 0.7514191269874573, + "num_tokens": 609203204.0, + "step": 23549 + }, + { + "epoch": 2.586206896551724, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4954564571380615, + "learning_rate": 1e-06, + "loss": 0.7858, + "mean_token_accuracy": 0.7438510656356812, + "num_tokens": 609226339.0, + "step": 23550 + }, + { + "epoch": 2.586316714254338, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.443385124206543, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7252879738807678, + "num_tokens": 609251273.0, + "step": 23551 + }, + { + "epoch": 2.5864265319569517, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.537541151046753, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7268664836883545, + "num_tokens": 609275377.0, + "step": 23552 + }, + { + "epoch": 2.586536349659565, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.426578998565674, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7281036376953125, + "num_tokens": 609302709.0, + "step": 23553 + }, + { + "epoch": 2.586646167362179, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4916281700134277, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7146109938621521, + "num_tokens": 609328684.0, + "step": 23554 + }, + { + "epoch": 2.5867559850647925, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2226548194885254, + "learning_rate": 1e-06, + "loss": 0.98, + "mean_token_accuracy": 0.6973692178726196, + "num_tokens": 609358995.0, + "step": 23555 + }, + { + "epoch": 2.586865802767406, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.419987440109253, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7186765670776367, + "num_tokens": 609383638.0, + "step": 23556 + }, + { + "epoch": 2.5869756204700196, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.511842727661133, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7193809747695923, + "num_tokens": 609406471.0, + "step": 23557 + }, + { + "epoch": 2.5870854381726334, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3531181812286377, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.715139627456665, + "num_tokens": 609432105.0, + "step": 23558 + }, + { + "epoch": 2.587195255875247, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3095359802246094, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7143406271934509, + "num_tokens": 609460303.0, + "step": 23559 + }, + { + "epoch": 2.587305073577861, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5216050148010254, + "learning_rate": 1e-06, + "loss": 0.9318, + "mean_token_accuracy": 0.7111128568649292, + "num_tokens": 609482805.0, + "step": 23560 + }, + { + "epoch": 2.587414891280474, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.501293420791626, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7394872903823853, + "num_tokens": 609508378.0, + "step": 23561 + }, + { + "epoch": 2.587524708983088, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.452284574508667, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7304385900497437, + "num_tokens": 609532991.0, + "step": 23562 + }, + { + "epoch": 2.5876345266857017, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1509103775024414, + "learning_rate": 1e-06, + "loss": 0.943, + "mean_token_accuracy": 0.7084135413169861, + "num_tokens": 609562786.0, + "step": 23563 + }, + { + "epoch": 2.5877443443883155, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5015437602996826, + "learning_rate": 1e-06, + "loss": 0.7257, + "mean_token_accuracy": 0.7587169408798218, + "num_tokens": 609584798.0, + "step": 23564 + }, + { + "epoch": 2.587854162090929, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1732089519500732, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7324559688568115, + "num_tokens": 609613866.0, + "step": 23565 + }, + { + "epoch": 2.5879639797935425, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3487141132354736, + "learning_rate": 1e-06, + "loss": 0.7673, + "mean_token_accuracy": 0.7654260396957397, + "num_tokens": 609637725.0, + "step": 23566 + }, + { + "epoch": 2.5880737974961563, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1997532844543457, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.73353111743927, + "num_tokens": 609666176.0, + "step": 23567 + }, + { + "epoch": 2.58818361519877, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5496838092803955, + "learning_rate": 1e-06, + "loss": 0.7935, + "mean_token_accuracy": 0.7435729503631592, + "num_tokens": 609688773.0, + "step": 23568 + }, + { + "epoch": 2.588293432901384, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.5275633335113525, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7362402081489563, + "num_tokens": 609713472.0, + "step": 23569 + }, + { + "epoch": 2.5884032506039976, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3621082305908203, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7474607229232788, + "num_tokens": 609740360.0, + "step": 23570 + }, + { + "epoch": 2.588513068306611, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4343085289001465, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7438729405403137, + "num_tokens": 609767501.0, + "step": 23571 + }, + { + "epoch": 2.5886228860092246, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4937994480133057, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7379748821258545, + "num_tokens": 609791624.0, + "step": 23572 + }, + { + "epoch": 2.5887327037118384, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.7258238792419434, + "learning_rate": 1e-06, + "loss": 0.7755, + "mean_token_accuracy": 0.7557474374771118, + "num_tokens": 609811024.0, + "step": 23573 + }, + { + "epoch": 2.588842521414452, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.3223514556884766, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7406613826751709, + "num_tokens": 609835938.0, + "step": 23574 + }, + { + "epoch": 2.588952339117066, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1615560054779053, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7314386367797852, + "num_tokens": 609865852.0, + "step": 23575 + }, + { + "epoch": 2.589062156819679, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2525789737701416, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7265855073928833, + "num_tokens": 609892294.0, + "step": 23576 + }, + { + "epoch": 2.589171974522293, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.2427825927734375, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7311135530471802, + "num_tokens": 609918345.0, + "step": 23577 + }, + { + "epoch": 2.5892817922249067, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.505183458328247, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7265486717224121, + "num_tokens": 609943155.0, + "step": 23578 + }, + { + "epoch": 2.58939160992752, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.347550630569458, + "learning_rate": 1e-06, + "loss": 0.7373, + "mean_token_accuracy": 0.7585750222206116, + "num_tokens": 609967669.0, + "step": 23579 + }, + { + "epoch": 2.5895014276301342, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.0643510818481445, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7233183979988098, + "num_tokens": 609999713.0, + "step": 23580 + }, + { + "epoch": 2.5896112453327476, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1370482444763184, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7148144841194153, + "num_tokens": 610029806.0, + "step": 23581 + }, + { + "epoch": 2.5897210630353613, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.12353253364563, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7079233527183533, + "num_tokens": 610060118.0, + "step": 23582 + }, + { + "epoch": 2.589830880737975, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.390869617462158, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7212715148925781, + "num_tokens": 610086735.0, + "step": 23583 + }, + { + "epoch": 2.5899406984405884, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.1898717880249023, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7101539373397827, + "num_tokens": 610115401.0, + "step": 23584 + }, + { + "epoch": 2.590050516143202, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.493062973022461, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7417176961898804, + "num_tokens": 610138102.0, + "step": 23585 + }, + { + "epoch": 2.590160333845816, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.156374454498291, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7210062146186829, + "num_tokens": 610169734.0, + "step": 23586 + }, + { + "epoch": 2.5902701515484297, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.4090116024017334, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.726640522480011, + "num_tokens": 610194219.0, + "step": 23587 + }, + { + "epoch": 2.5903799692510434, + "ewc_loss": 1.5616416931152344e-05, + "grad_norm": 2.433950424194336, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7077332735061646, + "num_tokens": 610218464.0, + "step": 23588 + }, + { + "epoch": 2.5904897869536567, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.42671275138855, + "learning_rate": 1e-06, + "loss": 0.8421, + "mean_token_accuracy": 0.7359607219696045, + "num_tokens": 610244318.0, + "step": 23589 + }, + { + "epoch": 2.5905996046562705, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6832730770111084, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7299219369888306, + "num_tokens": 610265950.0, + "step": 23590 + }, + { + "epoch": 2.5907094223588842, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.558201551437378, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7300551533699036, + "num_tokens": 610289450.0, + "step": 23591 + }, + { + "epoch": 2.590819240061498, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.207174777984619, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7286185026168823, + "num_tokens": 610318546.0, + "step": 23592 + }, + { + "epoch": 2.5909290577641118, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2585701942443848, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7303993701934814, + "num_tokens": 610344578.0, + "step": 23593 + }, + { + "epoch": 2.591038875466725, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5484261512756348, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7465524673461914, + "num_tokens": 610368562.0, + "step": 23594 + }, + { + "epoch": 2.591148693169339, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.437676429748535, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7304055690765381, + "num_tokens": 610393880.0, + "step": 23595 + }, + { + "epoch": 2.5912585108719526, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.096353769302368, + "learning_rate": 1e-06, + "loss": 0.9715, + "mean_token_accuracy": 0.6971749663352966, + "num_tokens": 610427393.0, + "step": 23596 + }, + { + "epoch": 2.5913683285745663, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6349048614501953, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7467899322509766, + "num_tokens": 610449326.0, + "step": 23597 + }, + { + "epoch": 2.59147814627718, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3750081062316895, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7177677154541016, + "num_tokens": 610477099.0, + "step": 23598 + }, + { + "epoch": 2.5915879639797934, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5338134765625, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7181169986724854, + "num_tokens": 610502104.0, + "step": 23599 + }, + { + "epoch": 2.591697781682407, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.466308832168579, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7257031202316284, + "num_tokens": 610526704.0, + "step": 23600 + }, + { + "epoch": 2.591807599385021, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.368224620819092, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7352495193481445, + "num_tokens": 610552280.0, + "step": 23601 + }, + { + "epoch": 2.5919174170876347, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8918306827545166, + "learning_rate": 1e-06, + "loss": 0.7822, + "mean_token_accuracy": 0.7509538531303406, + "num_tokens": 610570278.0, + "step": 23602 + }, + { + "epoch": 2.5920272347902484, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3700640201568604, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7212440967559814, + "num_tokens": 610594914.0, + "step": 23603 + }, + { + "epoch": 2.5921370524928617, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4725258350372314, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7469486594200134, + "num_tokens": 610618667.0, + "step": 23604 + }, + { + "epoch": 2.5922468701954755, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4444072246551514, + "learning_rate": 1e-06, + "loss": 0.8372, + "mean_token_accuracy": 0.7476890683174133, + "num_tokens": 610641748.0, + "step": 23605 + }, + { + "epoch": 2.5923566878980893, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9373624324798584, + "learning_rate": 1e-06, + "loss": 1.0056, + "mean_token_accuracy": 0.6925991177558899, + "num_tokens": 610678830.0, + "step": 23606 + }, + { + "epoch": 2.5924665056007026, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.167525291442871, + "learning_rate": 1e-06, + "loss": 0.7567, + "mean_token_accuracy": 0.7534669637680054, + "num_tokens": 610706433.0, + "step": 23607 + }, + { + "epoch": 2.5925763233033163, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3649415969848633, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.730559229850769, + "num_tokens": 610732306.0, + "step": 23608 + }, + { + "epoch": 2.59268614100593, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5095746517181396, + "learning_rate": 1e-06, + "loss": 0.8067, + "mean_token_accuracy": 0.7405948638916016, + "num_tokens": 610755381.0, + "step": 23609 + }, + { + "epoch": 2.592795958708544, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5883800983428955, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7117735147476196, + "num_tokens": 610779631.0, + "step": 23610 + }, + { + "epoch": 2.5929057764111576, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.481511354446411, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.702041506767273, + "num_tokens": 610807265.0, + "step": 23611 + }, + { + "epoch": 2.593015594113771, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.086141347885132, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7342454791069031, + "num_tokens": 610837654.0, + "step": 23612 + }, + { + "epoch": 2.5931254118163847, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2363827228546143, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7441322207450867, + "num_tokens": 610863735.0, + "step": 23613 + }, + { + "epoch": 2.5932352295189984, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1277573108673096, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.731273889541626, + "num_tokens": 610894086.0, + "step": 23614 + }, + { + "epoch": 2.593345047221612, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.583258628845215, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7358413934707642, + "num_tokens": 610916455.0, + "step": 23615 + }, + { + "epoch": 2.593454864924226, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5967257022857666, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.725338339805603, + "num_tokens": 610939189.0, + "step": 23616 + }, + { + "epoch": 2.5935646826268393, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4768261909484863, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7448143362998962, + "num_tokens": 610962545.0, + "step": 23617 + }, + { + "epoch": 2.593674500329453, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9496562480926514, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7389818429946899, + "num_tokens": 610979802.0, + "step": 23618 + }, + { + "epoch": 2.5937843180320668, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2558858394622803, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7293518781661987, + "num_tokens": 611007790.0, + "step": 23619 + }, + { + "epoch": 2.5938941357346805, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.683509349822998, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7247790098190308, + "num_tokens": 611028766.0, + "step": 23620 + }, + { + "epoch": 2.5940039534372943, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2354204654693604, + "learning_rate": 1e-06, + "loss": 0.9187, + "mean_token_accuracy": 0.715303361415863, + "num_tokens": 611060140.0, + "step": 23621 + }, + { + "epoch": 2.5941137711399076, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4122395515441895, + "learning_rate": 1e-06, + "loss": 0.8311, + "mean_token_accuracy": 0.7370686531066895, + "num_tokens": 611084267.0, + "step": 23622 + }, + { + "epoch": 2.5942235888425214, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4149417877197266, + "learning_rate": 1e-06, + "loss": 0.9138, + "mean_token_accuracy": 0.7168633937835693, + "num_tokens": 611110761.0, + "step": 23623 + }, + { + "epoch": 2.594333406545135, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3911092281341553, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.703582763671875, + "num_tokens": 611136743.0, + "step": 23624 + }, + { + "epoch": 2.594443224247749, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.40096378326416, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7316780090332031, + "num_tokens": 611162697.0, + "step": 23625 + }, + { + "epoch": 2.5945530419503626, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.272712469100952, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7226494550704956, + "num_tokens": 611190809.0, + "step": 23626 + }, + { + "epoch": 2.594662859652976, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3180243968963623, + "learning_rate": 1e-06, + "loss": 0.9369, + "mean_token_accuracy": 0.7005393505096436, + "num_tokens": 611219321.0, + "step": 23627 + }, + { + "epoch": 2.5947726773555897, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7054409980773926, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.737869918346405, + "num_tokens": 611240407.0, + "step": 23628 + }, + { + "epoch": 2.5948824950582035, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.249199151992798, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7245701551437378, + "num_tokens": 611269693.0, + "step": 23629 + }, + { + "epoch": 2.594992312760817, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.397634744644165, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7078946828842163, + "num_tokens": 611294073.0, + "step": 23630 + }, + { + "epoch": 2.595102130463431, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2313759326934814, + "learning_rate": 1e-06, + "loss": 0.8688, + "mean_token_accuracy": 0.7194185853004456, + "num_tokens": 611322370.0, + "step": 23631 + }, + { + "epoch": 2.5952119481660443, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4407646656036377, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7217406034469604, + "num_tokens": 611346265.0, + "step": 23632 + }, + { + "epoch": 2.595321765868658, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4194045066833496, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7345108985900879, + "num_tokens": 611371607.0, + "step": 23633 + }, + { + "epoch": 2.595431583571272, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1907362937927246, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7477449178695679, + "num_tokens": 611399678.0, + "step": 23634 + }, + { + "epoch": 2.595541401273885, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4147019386291504, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.726406991481781, + "num_tokens": 611428886.0, + "step": 23635 + }, + { + "epoch": 2.595651218976499, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6499135494232178, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.726636528968811, + "num_tokens": 611451301.0, + "step": 23636 + }, + { + "epoch": 2.5957610366791126, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2138195037841797, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7294010519981384, + "num_tokens": 611478880.0, + "step": 23637 + }, + { + "epoch": 2.5958708543817264, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2432196140289307, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.727091372013092, + "num_tokens": 611506458.0, + "step": 23638 + }, + { + "epoch": 2.59598067208434, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.291490077972412, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7462490200996399, + "num_tokens": 611532584.0, + "step": 23639 + }, + { + "epoch": 2.5960904897869534, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9264914989471436, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.73403000831604, + "num_tokens": 611568513.0, + "step": 23640 + }, + { + "epoch": 2.596200307489567, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.558793306350708, + "learning_rate": 1e-06, + "loss": 0.924, + "mean_token_accuracy": 0.7120797634124756, + "num_tokens": 611595791.0, + "step": 23641 + }, + { + "epoch": 2.596310125192181, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.32307767868042, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.7163936495780945, + "num_tokens": 611623092.0, + "step": 23642 + }, + { + "epoch": 2.5964199428947947, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4750564098358154, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.737307071685791, + "num_tokens": 611646677.0, + "step": 23643 + }, + { + "epoch": 2.5965297605974085, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3243398666381836, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.7383254766464233, + "num_tokens": 611672403.0, + "step": 23644 + }, + { + "epoch": 2.596639578300022, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2369296550750732, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7333715558052063, + "num_tokens": 611702719.0, + "step": 23645 + }, + { + "epoch": 2.5967493960026355, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9339287281036377, + "learning_rate": 1e-06, + "loss": 0.9389, + "mean_token_accuracy": 0.7075954675674438, + "num_tokens": 611740960.0, + "step": 23646 + }, + { + "epoch": 2.5968592137052493, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.22654128074646, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7313283681869507, + "num_tokens": 611770401.0, + "step": 23647 + }, + { + "epoch": 2.596969031407863, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.533860445022583, + "learning_rate": 1e-06, + "loss": 0.8003, + "mean_token_accuracy": 0.7498766183853149, + "num_tokens": 611791834.0, + "step": 23648 + }, + { + "epoch": 2.597078849110477, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4791994094848633, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7284415364265442, + "num_tokens": 611818006.0, + "step": 23649 + }, + { + "epoch": 2.59718866681309, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.804253101348877, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7361713647842407, + "num_tokens": 611836610.0, + "step": 23650 + }, + { + "epoch": 2.597298484515704, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1457462310791016, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7163503170013428, + "num_tokens": 611868349.0, + "step": 23651 + }, + { + "epoch": 2.5974083022183176, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6583492755889893, + "learning_rate": 1e-06, + "loss": 0.7834, + "mean_token_accuracy": 0.7545228004455566, + "num_tokens": 611888622.0, + "step": 23652 + }, + { + "epoch": 2.5975181199209314, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0438637733459473, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7461870312690735, + "num_tokens": 611919472.0, + "step": 23653 + }, + { + "epoch": 2.597627937623545, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1064372062683105, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7283450365066528, + "num_tokens": 611949302.0, + "step": 23654 + }, + { + "epoch": 2.5977377553261585, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2477853298187256, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7348349094390869, + "num_tokens": 611977106.0, + "step": 23655 + }, + { + "epoch": 2.5978475730287722, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.405229091644287, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7312692403793335, + "num_tokens": 612002776.0, + "step": 23656 + }, + { + "epoch": 2.597957390731386, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3657584190368652, + "learning_rate": 1e-06, + "loss": 0.8928, + "mean_token_accuracy": 0.7171726822853088, + "num_tokens": 612028747.0, + "step": 23657 + }, + { + "epoch": 2.5980672084339993, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6792826652526855, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.745696485042572, + "num_tokens": 612051398.0, + "step": 23658 + }, + { + "epoch": 2.5981770261366135, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5028672218322754, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7384494543075562, + "num_tokens": 612076770.0, + "step": 23659 + }, + { + "epoch": 2.598286843839227, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.390066623687744, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7413524389266968, + "num_tokens": 612100720.0, + "step": 23660 + }, + { + "epoch": 2.5983966615418406, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4883053302764893, + "learning_rate": 1e-06, + "loss": 0.76, + "mean_token_accuracy": 0.7528499960899353, + "num_tokens": 612124654.0, + "step": 23661 + }, + { + "epoch": 2.5985064792444543, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.600973606109619, + "learning_rate": 1e-06, + "loss": 0.7619, + "mean_token_accuracy": 0.7604577541351318, + "num_tokens": 612144913.0, + "step": 23662 + }, + { + "epoch": 2.5986162969470676, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2919085025787354, + "learning_rate": 1e-06, + "loss": 0.7657, + "mean_token_accuracy": 0.7546921372413635, + "num_tokens": 612171849.0, + "step": 23663 + }, + { + "epoch": 2.5987261146496814, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1371614933013916, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.710212767124176, + "num_tokens": 612203713.0, + "step": 23664 + }, + { + "epoch": 2.598835932352295, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1921463012695312, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7281427383422852, + "num_tokens": 612232019.0, + "step": 23665 + }, + { + "epoch": 2.598945750054909, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7019715309143066, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7292183637619019, + "num_tokens": 612254444.0, + "step": 23666 + }, + { + "epoch": 2.5990555677575227, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3301987648010254, + "learning_rate": 1e-06, + "loss": 0.7696, + "mean_token_accuracy": 0.7513694763183594, + "num_tokens": 612280856.0, + "step": 23667 + }, + { + "epoch": 2.599165385460136, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4840810298919678, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7259180545806885, + "num_tokens": 612304793.0, + "step": 23668 + }, + { + "epoch": 2.5992752031627497, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3804702758789062, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.7141796946525574, + "num_tokens": 612330615.0, + "step": 23669 + }, + { + "epoch": 2.5993850208653635, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3687987327575684, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7431950569152832, + "num_tokens": 612356540.0, + "step": 23670 + }, + { + "epoch": 2.5994948385679773, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.63572096824646, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7398667931556702, + "num_tokens": 612378341.0, + "step": 23671 + }, + { + "epoch": 2.599604656270591, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3810110092163086, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7222499251365662, + "num_tokens": 612409321.0, + "step": 23672 + }, + { + "epoch": 2.5997144739732043, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.336888074874878, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7086535692214966, + "num_tokens": 612436828.0, + "step": 23673 + }, + { + "epoch": 2.599824291675818, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4362423419952393, + "learning_rate": 1e-06, + "loss": 0.9065, + "mean_token_accuracy": 0.7265005707740784, + "num_tokens": 612461194.0, + "step": 23674 + }, + { + "epoch": 2.599934109378432, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5646252632141113, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7300788164138794, + "num_tokens": 612484439.0, + "step": 23675 + }, + { + "epoch": 2.6000439270810456, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3138954639434814, + "learning_rate": 1e-06, + "loss": 0.833, + "mean_token_accuracy": 0.7357676029205322, + "num_tokens": 612510719.0, + "step": 23676 + }, + { + "epoch": 2.6001537447836593, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.543635845184326, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.7418138980865479, + "num_tokens": 612531870.0, + "step": 23677 + }, + { + "epoch": 2.6002635624862727, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.378777265548706, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7437006235122681, + "num_tokens": 612559693.0, + "step": 23678 + }, + { + "epoch": 2.6003733801888864, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5255348682403564, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.735539972782135, + "num_tokens": 612582916.0, + "step": 23679 + }, + { + "epoch": 2.6004831978915, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5466766357421875, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7336862087249756, + "num_tokens": 612607430.0, + "step": 23680 + }, + { + "epoch": 2.600593015594114, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.423633337020874, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7321548461914062, + "num_tokens": 612632203.0, + "step": 23681 + }, + { + "epoch": 2.6007028332967277, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.240117073059082, + "learning_rate": 1e-06, + "loss": 0.9602, + "mean_token_accuracy": 0.7038908004760742, + "num_tokens": 612663563.0, + "step": 23682 + }, + { + "epoch": 2.600812650999341, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.559291362762451, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7367619276046753, + "num_tokens": 612687162.0, + "step": 23683 + }, + { + "epoch": 2.6009224687019548, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.726489305496216, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7286062836647034, + "num_tokens": 612708150.0, + "step": 23684 + }, + { + "epoch": 2.6010322864045685, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5090291500091553, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7151528596878052, + "num_tokens": 612732273.0, + "step": 23685 + }, + { + "epoch": 2.601142104107182, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1242482662200928, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.721861720085144, + "num_tokens": 612762367.0, + "step": 23686 + }, + { + "epoch": 2.6012519218097956, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.721555233001709, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7418656945228577, + "num_tokens": 612784479.0, + "step": 23687 + }, + { + "epoch": 2.6013617395124093, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4809303283691406, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.7558051943778992, + "num_tokens": 612807784.0, + "step": 23688 + }, + { + "epoch": 2.601471557215023, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6254239082336426, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7378839254379272, + "num_tokens": 612830441.0, + "step": 23689 + }, + { + "epoch": 2.601581374917637, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.066425323486328, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7530799508094788, + "num_tokens": 612860287.0, + "step": 23690 + }, + { + "epoch": 2.60169119262025, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.51589035987854, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7231467366218567, + "num_tokens": 612885853.0, + "step": 23691 + }, + { + "epoch": 2.601801010322864, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3320095539093018, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.724068284034729, + "num_tokens": 612913970.0, + "step": 23692 + }, + { + "epoch": 2.6019108280254777, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5822606086730957, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7448469996452332, + "num_tokens": 612936848.0, + "step": 23693 + }, + { + "epoch": 2.6020206457280914, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.261005401611328, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7278009653091431, + "num_tokens": 612962683.0, + "step": 23694 + }, + { + "epoch": 2.602130463430705, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2315547466278076, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7556372880935669, + "num_tokens": 612990744.0, + "step": 23695 + }, + { + "epoch": 2.6022402811333185, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2862470149993896, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7415280342102051, + "num_tokens": 613017058.0, + "step": 23696 + }, + { + "epoch": 2.6023500988359323, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6120071411132812, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7503774762153625, + "num_tokens": 613038577.0, + "step": 23697 + }, + { + "epoch": 2.602459916538546, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.271510601043701, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7183690071105957, + "num_tokens": 613067846.0, + "step": 23698 + }, + { + "epoch": 2.60256973424116, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7192094326019287, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7243956923484802, + "num_tokens": 613089819.0, + "step": 23699 + }, + { + "epoch": 2.6026795519437735, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6419777870178223, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7317947149276733, + "num_tokens": 613111268.0, + "step": 23700 + }, + { + "epoch": 2.602789369646387, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1941962242126465, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7364695072174072, + "num_tokens": 613138421.0, + "step": 23701 + }, + { + "epoch": 2.6028991873490006, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.47983455657959, + "learning_rate": 1e-06, + "loss": 0.9453, + "mean_token_accuracy": 0.7092166543006897, + "num_tokens": 613163254.0, + "step": 23702 + }, + { + "epoch": 2.6030090050516144, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1010165214538574, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7232656478881836, + "num_tokens": 613196287.0, + "step": 23703 + }, + { + "epoch": 2.603118822754228, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6009154319763184, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7217445969581604, + "num_tokens": 613220044.0, + "step": 23704 + }, + { + "epoch": 2.603228640456842, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.634459972381592, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7199469804763794, + "num_tokens": 613241739.0, + "step": 23705 + }, + { + "epoch": 2.603338458159455, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3695554733276367, + "learning_rate": 1e-06, + "loss": 0.9615, + "mean_token_accuracy": 0.7076386213302612, + "num_tokens": 613269674.0, + "step": 23706 + }, + { + "epoch": 2.603448275862069, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2780656814575195, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7274901866912842, + "num_tokens": 613299200.0, + "step": 23707 + }, + { + "epoch": 2.6035580935646827, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.224649429321289, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7484353184700012, + "num_tokens": 613326806.0, + "step": 23708 + }, + { + "epoch": 2.603667911267296, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0658562183380127, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7041900157928467, + "num_tokens": 613363218.0, + "step": 23709 + }, + { + "epoch": 2.60377772896991, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1640546321868896, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.7545010447502136, + "num_tokens": 613392091.0, + "step": 23710 + }, + { + "epoch": 2.6038875466725235, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.428870439529419, + "learning_rate": 1e-06, + "loss": 0.7742, + "mean_token_accuracy": 0.7522811889648438, + "num_tokens": 613415949.0, + "step": 23711 + }, + { + "epoch": 2.6039973643751373, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2720930576324463, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7211243510246277, + "num_tokens": 613442712.0, + "step": 23712 + }, + { + "epoch": 2.604107182077751, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.412777900695801, + "learning_rate": 1e-06, + "loss": 0.8774, + "mean_token_accuracy": 0.7230479121208191, + "num_tokens": 613468117.0, + "step": 23713 + }, + { + "epoch": 2.6042169997803644, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2683215141296387, + "learning_rate": 1e-06, + "loss": 0.945, + "mean_token_accuracy": 0.7076385617256165, + "num_tokens": 613497539.0, + "step": 23714 + }, + { + "epoch": 2.604326817482978, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3483614921569824, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7393540143966675, + "num_tokens": 613522873.0, + "step": 23715 + }, + { + "epoch": 2.604436635185592, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.367765426635742, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7038009166717529, + "num_tokens": 613549805.0, + "step": 23716 + }, + { + "epoch": 2.6045464528882056, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.31961989402771, + "learning_rate": 1e-06, + "loss": 0.934, + "mean_token_accuracy": 0.7181463837623596, + "num_tokens": 613576634.0, + "step": 23717 + }, + { + "epoch": 2.6046562705908194, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.542646884918213, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7258187532424927, + "num_tokens": 613600814.0, + "step": 23718 + }, + { + "epoch": 2.6047660882934327, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.361332893371582, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7290058732032776, + "num_tokens": 613626305.0, + "step": 23719 + }, + { + "epoch": 2.6048759059960465, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2729756832122803, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7391328811645508, + "num_tokens": 613654566.0, + "step": 23720 + }, + { + "epoch": 2.60498572369866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3508028984069824, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7428121566772461, + "num_tokens": 613678702.0, + "step": 23721 + }, + { + "epoch": 2.605095541401274, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.563873052597046, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.702478289604187, + "num_tokens": 613705791.0, + "step": 23722 + }, + { + "epoch": 2.6052053591038877, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.255030632019043, + "learning_rate": 1e-06, + "loss": 0.8913, + "mean_token_accuracy": 0.720928430557251, + "num_tokens": 613734905.0, + "step": 23723 + }, + { + "epoch": 2.605315176806501, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.488030433654785, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.7502096891403198, + "num_tokens": 613758211.0, + "step": 23724 + }, + { + "epoch": 2.605424994509115, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6971397399902344, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7234013080596924, + "num_tokens": 613778836.0, + "step": 23725 + }, + { + "epoch": 2.6055348122117286, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6328022480010986, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7274488210678101, + "num_tokens": 613801624.0, + "step": 23726 + }, + { + "epoch": 2.6056446299143423, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3655459880828857, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.728888988494873, + "num_tokens": 613829644.0, + "step": 23727 + }, + { + "epoch": 2.605754447616956, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.445568799972534, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7374963760375977, + "num_tokens": 613853936.0, + "step": 23728 + }, + { + "epoch": 2.6058642653195694, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6796162128448486, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7515753507614136, + "num_tokens": 613873765.0, + "step": 23729 + }, + { + "epoch": 2.605974083022183, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1602225303649902, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7269043922424316, + "num_tokens": 613901446.0, + "step": 23730 + }, + { + "epoch": 2.606083900724797, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3917534351348877, + "learning_rate": 1e-06, + "loss": 0.851, + "mean_token_accuracy": 0.7390952706336975, + "num_tokens": 613926484.0, + "step": 23731 + }, + { + "epoch": 2.6061937184274107, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.332073450088501, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7293490767478943, + "num_tokens": 613955178.0, + "step": 23732 + }, + { + "epoch": 2.6063035361300244, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.148731231689453, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7178629636764526, + "num_tokens": 613986262.0, + "step": 23733 + }, + { + "epoch": 2.6064133538326377, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.810978412628174, + "learning_rate": 1e-06, + "loss": 0.7484, + "mean_token_accuracy": 0.7569677233695984, + "num_tokens": 614004905.0, + "step": 23734 + }, + { + "epoch": 2.6065231715352515, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4835257530212402, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7388975024223328, + "num_tokens": 614028283.0, + "step": 23735 + }, + { + "epoch": 2.6066329892378652, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9192070960998535, + "learning_rate": 1e-06, + "loss": 0.8113, + "mean_token_accuracy": 0.7439614534378052, + "num_tokens": 614046571.0, + "step": 23736 + }, + { + "epoch": 2.6067428069404786, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7781083583831787, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7472869753837585, + "num_tokens": 614067079.0, + "step": 23737 + }, + { + "epoch": 2.6068526246430923, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.474644422531128, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7238377332687378, + "num_tokens": 614092766.0, + "step": 23738 + }, + { + "epoch": 2.606962442345706, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5701191425323486, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7367585301399231, + "num_tokens": 614115213.0, + "step": 23739 + }, + { + "epoch": 2.60707226004832, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7056760787963867, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.73973548412323, + "num_tokens": 614137069.0, + "step": 23740 + }, + { + "epoch": 2.6071820777509336, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7522644996643066, + "learning_rate": 1e-06, + "loss": 0.8003, + "mean_token_accuracy": 0.7561460137367249, + "num_tokens": 614160859.0, + "step": 23741 + }, + { + "epoch": 2.607291895453547, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.240969657897949, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7241055965423584, + "num_tokens": 614188900.0, + "step": 23742 + }, + { + "epoch": 2.6074017131561606, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.359734058380127, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7299154996871948, + "num_tokens": 614217035.0, + "step": 23743 + }, + { + "epoch": 2.6075115308587744, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5731658935546875, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.728584349155426, + "num_tokens": 614239744.0, + "step": 23744 + }, + { + "epoch": 2.607621348561388, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2977066040039062, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7351424694061279, + "num_tokens": 614266888.0, + "step": 23745 + }, + { + "epoch": 2.607731166264002, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.216049909591675, + "learning_rate": 1e-06, + "loss": 0.9267, + "mean_token_accuracy": 0.7125084400177002, + "num_tokens": 614295459.0, + "step": 23746 + }, + { + "epoch": 2.6078409839666152, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.395608901977539, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7423731088638306, + "num_tokens": 614319657.0, + "step": 23747 + }, + { + "epoch": 2.607950801669229, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4872488975524902, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7321025133132935, + "num_tokens": 614343494.0, + "step": 23748 + }, + { + "epoch": 2.6080606193718427, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2568769454956055, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7312803268432617, + "num_tokens": 614371395.0, + "step": 23749 + }, + { + "epoch": 2.6081704370744565, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3851728439331055, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7346745729446411, + "num_tokens": 614398280.0, + "step": 23750 + }, + { + "epoch": 2.6082802547770703, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4668374061584473, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7373283505439758, + "num_tokens": 614423501.0, + "step": 23751 + }, + { + "epoch": 2.6083900724796836, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.813617706298828, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.749483585357666, + "num_tokens": 614443447.0, + "step": 23752 + }, + { + "epoch": 2.6084998901822973, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.440455198287964, + "learning_rate": 1e-06, + "loss": 0.9208, + "mean_token_accuracy": 0.7126412391662598, + "num_tokens": 614470168.0, + "step": 23753 + }, + { + "epoch": 2.608609707884911, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.221909761428833, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7330853939056396, + "num_tokens": 614496569.0, + "step": 23754 + }, + { + "epoch": 2.608719525587525, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.406404972076416, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7206617593765259, + "num_tokens": 614522567.0, + "step": 23755 + }, + { + "epoch": 2.6088293432901386, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.400902509689331, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7248292565345764, + "num_tokens": 614547755.0, + "step": 23756 + }, + { + "epoch": 2.608939160992752, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.562408924102783, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7202292680740356, + "num_tokens": 614574353.0, + "step": 23757 + }, + { + "epoch": 2.6090489786953657, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.375680923461914, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7366551160812378, + "num_tokens": 614600494.0, + "step": 23758 + }, + { + "epoch": 2.6091587963979794, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.563070297241211, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.728379487991333, + "num_tokens": 614624797.0, + "step": 23759 + }, + { + "epoch": 2.6092686141005927, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2271270751953125, + "learning_rate": 1e-06, + "loss": 0.9613, + "mean_token_accuracy": 0.6993733048439026, + "num_tokens": 614655913.0, + "step": 23760 + }, + { + "epoch": 2.609378431803207, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5372557640075684, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.7516262531280518, + "num_tokens": 614679388.0, + "step": 23761 + }, + { + "epoch": 2.6094882495058203, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1768124103546143, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7415612936019897, + "num_tokens": 614709117.0, + "step": 23762 + }, + { + "epoch": 2.609598067208434, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8962228298187256, + "learning_rate": 1e-06, + "loss": 0.7548, + "mean_token_accuracy": 0.7526218891143799, + "num_tokens": 614727989.0, + "step": 23763 + }, + { + "epoch": 2.6097078849110478, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5291810035705566, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7431228160858154, + "num_tokens": 614750503.0, + "step": 23764 + }, + { + "epoch": 2.609817702613661, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.546170473098755, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7313411831855774, + "num_tokens": 614772555.0, + "step": 23765 + }, + { + "epoch": 2.609927520316275, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.410146713256836, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.715898334980011, + "num_tokens": 614798477.0, + "step": 23766 + }, + { + "epoch": 2.6100373380188886, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.376769781112671, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7440249919891357, + "num_tokens": 614821899.0, + "step": 23767 + }, + { + "epoch": 2.6101471557215024, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.210330009460449, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7382446527481079, + "num_tokens": 614853929.0, + "step": 23768 + }, + { + "epoch": 2.610256973424116, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2036616802215576, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7513666152954102, + "num_tokens": 614880210.0, + "step": 23769 + }, + { + "epoch": 2.6103667911267294, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4783999919891357, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.7525472640991211, + "num_tokens": 614903738.0, + "step": 23770 + }, + { + "epoch": 2.610476608829343, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2074637413024902, + "learning_rate": 1e-06, + "loss": 0.7776, + "mean_token_accuracy": 0.7564687728881836, + "num_tokens": 614929478.0, + "step": 23771 + }, + { + "epoch": 2.610586426531957, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.273855447769165, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7246549129486084, + "num_tokens": 614955346.0, + "step": 23772 + }, + { + "epoch": 2.6106962442345707, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2228846549987793, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.738162100315094, + "num_tokens": 614982587.0, + "step": 23773 + }, + { + "epoch": 2.6108060619371845, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4615323543548584, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.720346987247467, + "num_tokens": 615006632.0, + "step": 23774 + }, + { + "epoch": 2.6109158796397978, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2442641258239746, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7327254414558411, + "num_tokens": 615035959.0, + "step": 23775 + }, + { + "epoch": 2.6110256973424115, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2192223072052, + "learning_rate": 1e-06, + "loss": 0.9521, + "mean_token_accuracy": 0.7134251594543457, + "num_tokens": 615064681.0, + "step": 23776 + }, + { + "epoch": 2.6111355150450253, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4017860889434814, + "learning_rate": 1e-06, + "loss": 0.8485, + "mean_token_accuracy": 0.7305305004119873, + "num_tokens": 615090058.0, + "step": 23777 + }, + { + "epoch": 2.611245332747639, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.368340253829956, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7423343658447266, + "num_tokens": 615115131.0, + "step": 23778 + }, + { + "epoch": 2.611355150450253, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4577438831329346, + "learning_rate": 1e-06, + "loss": 0.767, + "mean_token_accuracy": 0.7542557716369629, + "num_tokens": 615137660.0, + "step": 23779 + }, + { + "epoch": 2.611464968152866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.312896728515625, + "learning_rate": 1e-06, + "loss": 0.8573, + "mean_token_accuracy": 0.739611029624939, + "num_tokens": 615164764.0, + "step": 23780 + }, + { + "epoch": 2.61157478585548, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2742974758148193, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.735039472579956, + "num_tokens": 615193518.0, + "step": 23781 + }, + { + "epoch": 2.6116846035580936, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.580275058746338, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7502687573432922, + "num_tokens": 615214833.0, + "step": 23782 + }, + { + "epoch": 2.6117944212607074, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3334014415740967, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7364097833633423, + "num_tokens": 615240146.0, + "step": 23783 + }, + { + "epoch": 2.611904238963321, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.326730489730835, + "learning_rate": 1e-06, + "loss": 0.8298, + "mean_token_accuracy": 0.7356753945350647, + "num_tokens": 615266486.0, + "step": 23784 + }, + { + "epoch": 2.6120140566659344, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.243089199066162, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7314877510070801, + "num_tokens": 615295792.0, + "step": 23785 + }, + { + "epoch": 2.612123874368548, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.267282724380493, + "learning_rate": 1e-06, + "loss": 0.7073, + "mean_token_accuracy": 0.7643259763717651, + "num_tokens": 615322940.0, + "step": 23786 + }, + { + "epoch": 2.612233692071162, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5950815677642822, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7367478609085083, + "num_tokens": 615344774.0, + "step": 23787 + }, + { + "epoch": 2.6123435097737753, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6688320636749268, + "learning_rate": 1e-06, + "loss": 0.8126, + "mean_token_accuracy": 0.7441816926002502, + "num_tokens": 615365924.0, + "step": 23788 + }, + { + "epoch": 2.612453327476389, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4526822566986084, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7364515066146851, + "num_tokens": 615390288.0, + "step": 23789 + }, + { + "epoch": 2.612563145179003, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4375784397125244, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7544595003128052, + "num_tokens": 615413048.0, + "step": 23790 + }, + { + "epoch": 2.6126729628816165, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4468977451324463, + "learning_rate": 1e-06, + "loss": 0.7842, + "mean_token_accuracy": 0.7462790012359619, + "num_tokens": 615436740.0, + "step": 23791 + }, + { + "epoch": 2.6127827805842303, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.24497127532959, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7348729372024536, + "num_tokens": 615462034.0, + "step": 23792 + }, + { + "epoch": 2.6128925982868436, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.361659288406372, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7265313863754272, + "num_tokens": 615488700.0, + "step": 23793 + }, + { + "epoch": 2.6130024159894574, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.446071147918701, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7330907583236694, + "num_tokens": 615513606.0, + "step": 23794 + }, + { + "epoch": 2.613112233692071, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5512170791625977, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7285329103469849, + "num_tokens": 615535227.0, + "step": 23795 + }, + { + "epoch": 2.613222051394685, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2428717613220215, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7169013619422913, + "num_tokens": 615562831.0, + "step": 23796 + }, + { + "epoch": 2.6133318690972986, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2280707359313965, + "learning_rate": 1e-06, + "loss": 0.9565, + "mean_token_accuracy": 0.705863893032074, + "num_tokens": 615592648.0, + "step": 23797 + }, + { + "epoch": 2.613441686799912, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6039488315582275, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.7338125705718994, + "num_tokens": 615614363.0, + "step": 23798 + }, + { + "epoch": 2.6135515045025257, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1281754970550537, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7030080556869507, + "num_tokens": 615647265.0, + "step": 23799 + }, + { + "epoch": 2.6136613222051395, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.440403938293457, + "learning_rate": 1e-06, + "loss": 0.7599, + "mean_token_accuracy": 0.7496491074562073, + "num_tokens": 615671285.0, + "step": 23800 + }, + { + "epoch": 2.6137711399077532, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.127312421798706, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7149308323860168, + "num_tokens": 615704263.0, + "step": 23801 + }, + { + "epoch": 2.613880957610367, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4158854484558105, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7313786745071411, + "num_tokens": 615728820.0, + "step": 23802 + }, + { + "epoch": 2.6139907753129803, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1365222930908203, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.717536449432373, + "num_tokens": 615758594.0, + "step": 23803 + }, + { + "epoch": 2.614100593015594, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2287375926971436, + "learning_rate": 1e-06, + "loss": 0.9611, + "mean_token_accuracy": 0.707139253616333, + "num_tokens": 615789841.0, + "step": 23804 + }, + { + "epoch": 2.614210410718208, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5445871353149414, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7293100953102112, + "num_tokens": 615811739.0, + "step": 23805 + }, + { + "epoch": 2.6143202284208216, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.134608507156372, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7210889458656311, + "num_tokens": 615843106.0, + "step": 23806 + }, + { + "epoch": 2.6144300461234353, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2395994663238525, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.7406032085418701, + "num_tokens": 615870852.0, + "step": 23807 + }, + { + "epoch": 2.6145398638260486, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5196139812469482, + "learning_rate": 1e-06, + "loss": 0.7416, + "mean_token_accuracy": 0.759118378162384, + "num_tokens": 615893427.0, + "step": 23808 + }, + { + "epoch": 2.6146496815286624, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2133266925811768, + "learning_rate": 1e-06, + "loss": 0.9296, + "mean_token_accuracy": 0.7075522541999817, + "num_tokens": 615922853.0, + "step": 23809 + }, + { + "epoch": 2.614759499231276, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6470775604248047, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7290592789649963, + "num_tokens": 615944594.0, + "step": 23810 + }, + { + "epoch": 2.61486931693389, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5626893043518066, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7372214794158936, + "num_tokens": 615968025.0, + "step": 23811 + }, + { + "epoch": 2.6149791346365037, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4683635234832764, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7110726237297058, + "num_tokens": 615996177.0, + "step": 23812 + }, + { + "epoch": 2.615088952339117, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.298534631729126, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7304844856262207, + "num_tokens": 616023342.0, + "step": 23813 + }, + { + "epoch": 2.6151987700417307, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.078214168548584, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7293152809143066, + "num_tokens": 616052964.0, + "step": 23814 + }, + { + "epoch": 2.6153085877443445, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3908636569976807, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7225260734558105, + "num_tokens": 616078639.0, + "step": 23815 + }, + { + "epoch": 2.615418405446958, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0925536155700684, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7380965948104858, + "num_tokens": 616108430.0, + "step": 23816 + }, + { + "epoch": 2.6155282231495716, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.267350196838379, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7442200779914856, + "num_tokens": 616132651.0, + "step": 23817 + }, + { + "epoch": 2.6156380408521853, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.303349733352661, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7143144607543945, + "num_tokens": 616160477.0, + "step": 23818 + }, + { + "epoch": 2.615747858554799, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8444936275482178, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7405120730400085, + "num_tokens": 616181366.0, + "step": 23819 + }, + { + "epoch": 2.615857676257413, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6994736194610596, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7254642248153687, + "num_tokens": 616203339.0, + "step": 23820 + }, + { + "epoch": 2.615967493960026, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.461822032928467, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.744062066078186, + "num_tokens": 616226918.0, + "step": 23821 + }, + { + "epoch": 2.61607731166264, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4007105827331543, + "learning_rate": 1e-06, + "loss": 0.9359, + "mean_token_accuracy": 0.712408721446991, + "num_tokens": 616253500.0, + "step": 23822 + }, + { + "epoch": 2.6161871293652537, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5523955821990967, + "learning_rate": 1e-06, + "loss": 0.8016, + "mean_token_accuracy": 0.7528505921363831, + "num_tokens": 616275443.0, + "step": 23823 + }, + { + "epoch": 2.6162969470678674, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.494191884994507, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7313022017478943, + "num_tokens": 616298483.0, + "step": 23824 + }, + { + "epoch": 2.616406764770481, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.563110113143921, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7319416999816895, + "num_tokens": 616321924.0, + "step": 23825 + }, + { + "epoch": 2.6165165824730945, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.339736223220825, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7298861742019653, + "num_tokens": 616349063.0, + "step": 23826 + }, + { + "epoch": 2.6166264001757082, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2486658096313477, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7260086536407471, + "num_tokens": 616375379.0, + "step": 23827 + }, + { + "epoch": 2.616736217878322, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.357714891433716, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7230417132377625, + "num_tokens": 616401438.0, + "step": 23828 + }, + { + "epoch": 2.6168460355809358, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.234569549560547, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7303469181060791, + "num_tokens": 616430323.0, + "step": 23829 + }, + { + "epoch": 2.6169558532835495, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2905025482177734, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.7445440292358398, + "num_tokens": 616457365.0, + "step": 23830 + }, + { + "epoch": 2.617065670986163, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.379458427429199, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7379169464111328, + "num_tokens": 616484086.0, + "step": 23831 + }, + { + "epoch": 2.6171754886887766, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1016175746917725, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.7170560956001282, + "num_tokens": 616515260.0, + "step": 23832 + }, + { + "epoch": 2.6172853063913903, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0734267234802246, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.723556399345398, + "num_tokens": 616545524.0, + "step": 23833 + }, + { + "epoch": 2.617395124094004, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4659273624420166, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7342073917388916, + "num_tokens": 616570629.0, + "step": 23834 + }, + { + "epoch": 2.617504941796618, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1853644847869873, + "learning_rate": 1e-06, + "loss": 0.8678, + "mean_token_accuracy": 0.7283905744552612, + "num_tokens": 616600138.0, + "step": 23835 + }, + { + "epoch": 2.617614759499231, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7539572715759277, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7169286608695984, + "num_tokens": 616621489.0, + "step": 23836 + }, + { + "epoch": 2.617724577201845, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.77009916305542, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7407664656639099, + "num_tokens": 616642458.0, + "step": 23837 + }, + { + "epoch": 2.6178343949044587, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.402703285217285, + "learning_rate": 1e-06, + "loss": 0.8064, + "mean_token_accuracy": 0.7399274110794067, + "num_tokens": 616668748.0, + "step": 23838 + }, + { + "epoch": 2.617944212607072, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1135287284851074, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7312145233154297, + "num_tokens": 616700767.0, + "step": 23839 + }, + { + "epoch": 2.618054030309686, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3786444664001465, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7496740818023682, + "num_tokens": 616725784.0, + "step": 23840 + }, + { + "epoch": 2.6181638480122995, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4958276748657227, + "learning_rate": 1e-06, + "loss": 0.7791, + "mean_token_accuracy": 0.7517073154449463, + "num_tokens": 616749963.0, + "step": 23841 + }, + { + "epoch": 2.6182736657149133, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4494407176971436, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7336503267288208, + "num_tokens": 616773116.0, + "step": 23842 + }, + { + "epoch": 2.618383483417527, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3861191272735596, + "learning_rate": 1e-06, + "loss": 0.9551, + "mean_token_accuracy": 0.702298641204834, + "num_tokens": 616799662.0, + "step": 23843 + }, + { + "epoch": 2.6184933011201403, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1091606616973877, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7268695831298828, + "num_tokens": 616832181.0, + "step": 23844 + }, + { + "epoch": 2.618603118822754, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.484438896179199, + "learning_rate": 1e-06, + "loss": 0.7085, + "mean_token_accuracy": 0.7673560380935669, + "num_tokens": 616856032.0, + "step": 23845 + }, + { + "epoch": 2.618712936525368, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.381486415863037, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7241693735122681, + "num_tokens": 616882922.0, + "step": 23846 + }, + { + "epoch": 2.6188227542279816, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.605966329574585, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.736234188079834, + "num_tokens": 616905003.0, + "step": 23847 + }, + { + "epoch": 2.6189325719305954, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2843856811523438, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7275165319442749, + "num_tokens": 616934182.0, + "step": 23848 + }, + { + "epoch": 2.6190423896332087, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3584864139556885, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7138736248016357, + "num_tokens": 616962792.0, + "step": 23849 + }, + { + "epoch": 2.6191522073358224, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1970508098602295, + "learning_rate": 1e-06, + "loss": 0.896, + "mean_token_accuracy": 0.7231061458587646, + "num_tokens": 616993715.0, + "step": 23850 + }, + { + "epoch": 2.619262025038436, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.611527681350708, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.739559531211853, + "num_tokens": 617016131.0, + "step": 23851 + }, + { + "epoch": 2.61937184274105, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0329933166503906, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7186923623085022, + "num_tokens": 617047681.0, + "step": 23852 + }, + { + "epoch": 2.6194816604436637, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.279555082321167, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7479594945907593, + "num_tokens": 617074011.0, + "step": 23853 + }, + { + "epoch": 2.619591478146277, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3497183322906494, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.742743968963623, + "num_tokens": 617098524.0, + "step": 23854 + }, + { + "epoch": 2.6197012958488908, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.215780258178711, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.725823163986206, + "num_tokens": 617125190.0, + "step": 23855 + }, + { + "epoch": 2.6198111135515045, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.493659734725952, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7453857660293579, + "num_tokens": 617149480.0, + "step": 23856 + }, + { + "epoch": 2.6199209312541183, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.370638847351074, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7307209968566895, + "num_tokens": 617175751.0, + "step": 23857 + }, + { + "epoch": 2.620030748956732, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.129645824432373, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7281199097633362, + "num_tokens": 617207187.0, + "step": 23858 + }, + { + "epoch": 2.6201405666593454, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4253859519958496, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.736693263053894, + "num_tokens": 617230627.0, + "step": 23859 + }, + { + "epoch": 2.620250384361959, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4574174880981445, + "learning_rate": 1e-06, + "loss": 0.9512, + "mean_token_accuracy": 0.6988041996955872, + "num_tokens": 617255474.0, + "step": 23860 + }, + { + "epoch": 2.620360202064573, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.374296188354492, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7362842559814453, + "num_tokens": 617283950.0, + "step": 23861 + }, + { + "epoch": 2.6204700197671866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2338650226593018, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7264820337295532, + "num_tokens": 617311662.0, + "step": 23862 + }, + { + "epoch": 2.6205798374698004, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3819642066955566, + "learning_rate": 1e-06, + "loss": 0.7554, + "mean_token_accuracy": 0.7468621730804443, + "num_tokens": 617337524.0, + "step": 23863 + }, + { + "epoch": 2.6206896551724137, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4920341968536377, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7305954694747925, + "num_tokens": 617362717.0, + "step": 23864 + }, + { + "epoch": 2.6207994728750275, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.304926872253418, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7254530787467957, + "num_tokens": 617390882.0, + "step": 23865 + }, + { + "epoch": 2.620909290577641, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.336235523223877, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7136498093605042, + "num_tokens": 617419244.0, + "step": 23866 + }, + { + "epoch": 2.6210191082802545, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.636291265487671, + "learning_rate": 1e-06, + "loss": 0.794, + "mean_token_accuracy": 0.7434596419334412, + "num_tokens": 617440822.0, + "step": 23867 + }, + { + "epoch": 2.6211289259828683, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1976802349090576, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7247411608695984, + "num_tokens": 617472444.0, + "step": 23868 + }, + { + "epoch": 2.621238743685482, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2536780834198, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.7414028644561768, + "num_tokens": 617498360.0, + "step": 23869 + }, + { + "epoch": 2.621348561388096, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.363077402114868, + "learning_rate": 1e-06, + "loss": 0.8238, + "mean_token_accuracy": 0.7332140207290649, + "num_tokens": 617524776.0, + "step": 23870 + }, + { + "epoch": 2.6214583790907096, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.097651720046997, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.714106023311615, + "num_tokens": 617559133.0, + "step": 23871 + }, + { + "epoch": 2.621568196793323, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0315253734588623, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7206168174743652, + "num_tokens": 617591498.0, + "step": 23872 + }, + { + "epoch": 2.6216780144959366, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2001030445098877, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7294678688049316, + "num_tokens": 617622204.0, + "step": 23873 + }, + { + "epoch": 2.6217878321985504, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6164705753326416, + "learning_rate": 1e-06, + "loss": 0.8837, + "mean_token_accuracy": 0.7231101393699646, + "num_tokens": 617645255.0, + "step": 23874 + }, + { + "epoch": 2.621897649901164, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.382096767425537, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7426835298538208, + "num_tokens": 617668651.0, + "step": 23875 + }, + { + "epoch": 2.622007467603778, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3930106163024902, + "learning_rate": 1e-06, + "loss": 0.9442, + "mean_token_accuracy": 0.7061626315116882, + "num_tokens": 617693808.0, + "step": 23876 + }, + { + "epoch": 2.622117285306391, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.611020803451538, + "learning_rate": 1e-06, + "loss": 0.7688, + "mean_token_accuracy": 0.7530168294906616, + "num_tokens": 617714789.0, + "step": 23877 + }, + { + "epoch": 2.622227103009005, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4802963733673096, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7309696674346924, + "num_tokens": 617740029.0, + "step": 23878 + }, + { + "epoch": 2.6223369207116187, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5535192489624023, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.745658814907074, + "num_tokens": 617763305.0, + "step": 23879 + }, + { + "epoch": 2.6224467384142325, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.346773386001587, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7545483112335205, + "num_tokens": 617788975.0, + "step": 23880 + }, + { + "epoch": 2.6225565561168462, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.110179901123047, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7274085283279419, + "num_tokens": 617821155.0, + "step": 23881 + }, + { + "epoch": 2.6226663738194596, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5707614421844482, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.754830539226532, + "num_tokens": 617843279.0, + "step": 23882 + }, + { + "epoch": 2.6227761915220733, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.243633985519409, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7199615240097046, + "num_tokens": 617871053.0, + "step": 23883 + }, + { + "epoch": 2.622886009224687, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0518596172332764, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7396310567855835, + "num_tokens": 617901241.0, + "step": 23884 + }, + { + "epoch": 2.622995826927301, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.435253381729126, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7139604091644287, + "num_tokens": 617925776.0, + "step": 23885 + }, + { + "epoch": 2.6231056446299146, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.536501169204712, + "learning_rate": 1e-06, + "loss": 0.7396, + "mean_token_accuracy": 0.7665029764175415, + "num_tokens": 617947516.0, + "step": 23886 + }, + { + "epoch": 2.623215462332528, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.895834445953369, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7445051670074463, + "num_tokens": 617966419.0, + "step": 23887 + }, + { + "epoch": 2.6233252800351416, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3816962242126465, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7350723743438721, + "num_tokens": 617991109.0, + "step": 23888 + }, + { + "epoch": 2.6234350977377554, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2517082691192627, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7229417562484741, + "num_tokens": 618020680.0, + "step": 23889 + }, + { + "epoch": 2.6235449154403687, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.504016876220703, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.750647783279419, + "num_tokens": 618043607.0, + "step": 23890 + }, + { + "epoch": 2.623654733142983, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.198456287384033, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7277250289916992, + "num_tokens": 618072111.0, + "step": 23891 + }, + { + "epoch": 2.6237645508455962, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.532209873199463, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7273149490356445, + "num_tokens": 618095676.0, + "step": 23892 + }, + { + "epoch": 2.62387436854821, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2312934398651123, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.7248429656028748, + "num_tokens": 618122311.0, + "step": 23893 + }, + { + "epoch": 2.6239841862508237, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4414429664611816, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7384535670280457, + "num_tokens": 618146648.0, + "step": 23894 + }, + { + "epoch": 2.624094003953437, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5925323963165283, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7502358555793762, + "num_tokens": 618167716.0, + "step": 23895 + }, + { + "epoch": 2.624203821656051, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.253258466720581, + "learning_rate": 1e-06, + "loss": 0.8289, + "mean_token_accuracy": 0.7349268198013306, + "num_tokens": 618194484.0, + "step": 23896 + }, + { + "epoch": 2.6243136393586646, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.550957202911377, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7414445877075195, + "num_tokens": 618215614.0, + "step": 23897 + }, + { + "epoch": 2.6244234570612783, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4856488704681396, + "learning_rate": 1e-06, + "loss": 0.9183, + "mean_token_accuracy": 0.7231142520904541, + "num_tokens": 618241722.0, + "step": 23898 + }, + { + "epoch": 2.624533274763892, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.250582456588745, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7263413667678833, + "num_tokens": 618270986.0, + "step": 23899 + }, + { + "epoch": 2.6246430924665054, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.40535569190979, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7357233762741089, + "num_tokens": 618298025.0, + "step": 23900 + }, + { + "epoch": 2.624752910169119, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6432430744171143, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7285917401313782, + "num_tokens": 618320589.0, + "step": 23901 + }, + { + "epoch": 2.624862727871733, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4327547550201416, + "learning_rate": 1e-06, + "loss": 0.7921, + "mean_token_accuracy": 0.7575602531433105, + "num_tokens": 618343548.0, + "step": 23902 + }, + { + "epoch": 2.6249725455743467, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5131657123565674, + "learning_rate": 1e-06, + "loss": 0.737, + "mean_token_accuracy": 0.7547425627708435, + "num_tokens": 618366871.0, + "step": 23903 + }, + { + "epoch": 2.6250823632769604, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1993906497955322, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7290372848510742, + "num_tokens": 618397442.0, + "step": 23904 + }, + { + "epoch": 2.6251921809795737, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1408846378326416, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7247911691665649, + "num_tokens": 618425474.0, + "step": 23905 + }, + { + "epoch": 2.6253019986821875, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2898244857788086, + "learning_rate": 1e-06, + "loss": 0.735, + "mean_token_accuracy": 0.7579464912414551, + "num_tokens": 618451059.0, + "step": 23906 + }, + { + "epoch": 2.6254118163848013, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1755993366241455, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7187466621398926, + "num_tokens": 618481342.0, + "step": 23907 + }, + { + "epoch": 2.625521634087415, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1299870014190674, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7364944219589233, + "num_tokens": 618512365.0, + "step": 23908 + }, + { + "epoch": 2.6256314517900288, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.23929762840271, + "learning_rate": 1e-06, + "loss": 0.8416, + "mean_token_accuracy": 0.736885666847229, + "num_tokens": 618540595.0, + "step": 23909 + }, + { + "epoch": 2.625741269492642, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6226677894592285, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7318744659423828, + "num_tokens": 618562278.0, + "step": 23910 + }, + { + "epoch": 2.625851087195256, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.362431764602661, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7400409579277039, + "num_tokens": 618587828.0, + "step": 23911 + }, + { + "epoch": 2.6259609048978696, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3028271198272705, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7188885807991028, + "num_tokens": 618615196.0, + "step": 23912 + }, + { + "epoch": 2.6260707226004834, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.135901689529419, + "learning_rate": 1e-06, + "loss": 0.7757, + "mean_token_accuracy": 0.7534605264663696, + "num_tokens": 618643927.0, + "step": 23913 + }, + { + "epoch": 2.626180540303097, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1033239364624023, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7120872735977173, + "num_tokens": 618676520.0, + "step": 23914 + }, + { + "epoch": 2.6262903580057104, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.493828535079956, + "learning_rate": 1e-06, + "loss": 0.8526, + "mean_token_accuracy": 0.7269307971000671, + "num_tokens": 618701042.0, + "step": 23915 + }, + { + "epoch": 2.626400175708324, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3306186199188232, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7422593832015991, + "num_tokens": 618727378.0, + "step": 23916 + }, + { + "epoch": 2.626509993410938, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2374885082244873, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.727219820022583, + "num_tokens": 618754896.0, + "step": 23917 + }, + { + "epoch": 2.6266198111135513, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.386549472808838, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7387115359306335, + "num_tokens": 618778671.0, + "step": 23918 + }, + { + "epoch": 2.626729628816165, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.30960750579834, + "learning_rate": 1e-06, + "loss": 0.9136, + "mean_token_accuracy": 0.7237430810928345, + "num_tokens": 618804593.0, + "step": 23919 + }, + { + "epoch": 2.6268394465187788, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4127273559570312, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7322561144828796, + "num_tokens": 618829558.0, + "step": 23920 + }, + { + "epoch": 2.6269492642213925, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.407595157623291, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7315303087234497, + "num_tokens": 618856096.0, + "step": 23921 + }, + { + "epoch": 2.6270590819240063, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.171909809112549, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7243669033050537, + "num_tokens": 618887366.0, + "step": 23922 + }, + { + "epoch": 2.6271688996266196, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.122553586959839, + "learning_rate": 1e-06, + "loss": 0.8899, + "mean_token_accuracy": 0.7231147885322571, + "num_tokens": 618918351.0, + "step": 23923 + }, + { + "epoch": 2.6272787173292333, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4867045879364014, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7259054183959961, + "num_tokens": 618944233.0, + "step": 23924 + }, + { + "epoch": 2.627388535031847, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4503328800201416, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7351976633071899, + "num_tokens": 618968428.0, + "step": 23925 + }, + { + "epoch": 2.627498352734461, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8129918575286865, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.746628999710083, + "num_tokens": 618989435.0, + "step": 23926 + }, + { + "epoch": 2.6276081704370746, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5579686164855957, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7263810634613037, + "num_tokens": 619011294.0, + "step": 23927 + }, + { + "epoch": 2.627717988139688, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6782164573669434, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7413011789321899, + "num_tokens": 619031707.0, + "step": 23928 + }, + { + "epoch": 2.6278278058423017, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3440048694610596, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7470080852508545, + "num_tokens": 619056051.0, + "step": 23929 + }, + { + "epoch": 2.6279376235449154, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.455751895904541, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7289604544639587, + "num_tokens": 619078179.0, + "step": 23930 + }, + { + "epoch": 2.628047441247529, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2216248512268066, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7297981381416321, + "num_tokens": 619107033.0, + "step": 23931 + }, + { + "epoch": 2.628157258950143, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.261392831802368, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.722645103931427, + "num_tokens": 619134208.0, + "step": 23932 + }, + { + "epoch": 2.6282670766527563, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3838343620300293, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7273731231689453, + "num_tokens": 619160739.0, + "step": 23933 + }, + { + "epoch": 2.62837689435537, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.482808828353882, + "learning_rate": 1e-06, + "loss": 0.8437, + "mean_token_accuracy": 0.7293746471405029, + "num_tokens": 619182600.0, + "step": 23934 + }, + { + "epoch": 2.628486712057984, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4579803943634033, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7223866581916809, + "num_tokens": 619207172.0, + "step": 23935 + }, + { + "epoch": 2.6285965297605975, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.300419330596924, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7399598956108093, + "num_tokens": 619234152.0, + "step": 23936 + }, + { + "epoch": 2.6287063474632113, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0885982513427734, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7275007367134094, + "num_tokens": 619264240.0, + "step": 23937 + }, + { + "epoch": 2.6288161651658246, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2171852588653564, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.728431224822998, + "num_tokens": 619292047.0, + "step": 23938 + }, + { + "epoch": 2.6289259828684384, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5851545333862305, + "learning_rate": 1e-06, + "loss": 0.7845, + "mean_token_accuracy": 0.7473977208137512, + "num_tokens": 619315084.0, + "step": 23939 + }, + { + "epoch": 2.629035800571052, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.942512273788452, + "learning_rate": 1e-06, + "loss": 0.868, + "mean_token_accuracy": 0.7285122871398926, + "num_tokens": 619335129.0, + "step": 23940 + }, + { + "epoch": 2.6291456182736654, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2695138454437256, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7254678010940552, + "num_tokens": 619362365.0, + "step": 23941 + }, + { + "epoch": 2.6292554359762796, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4695043563842773, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7245570421218872, + "num_tokens": 619385774.0, + "step": 23942 + }, + { + "epoch": 2.629365253678893, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3249683380126953, + "learning_rate": 1e-06, + "loss": 0.727, + "mean_token_accuracy": 0.7646430730819702, + "num_tokens": 619411958.0, + "step": 23943 + }, + { + "epoch": 2.6294750713815067, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1223156452178955, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7135021686553955, + "num_tokens": 619446957.0, + "step": 23944 + }, + { + "epoch": 2.6295848890841205, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4697163105010986, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7319433093070984, + "num_tokens": 619477566.0, + "step": 23945 + }, + { + "epoch": 2.629694706786734, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.510885238647461, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.741509199142456, + "num_tokens": 619501131.0, + "step": 23946 + }, + { + "epoch": 2.6298045244893475, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9845575094223022, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7284039855003357, + "num_tokens": 619535315.0, + "step": 23947 + }, + { + "epoch": 2.6299143421919613, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6193645000457764, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7376909255981445, + "num_tokens": 619556423.0, + "step": 23948 + }, + { + "epoch": 2.630024159894575, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0954627990722656, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7119293212890625, + "num_tokens": 619587941.0, + "step": 23949 + }, + { + "epoch": 2.630133977597189, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4074196815490723, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.736228346824646, + "num_tokens": 619612379.0, + "step": 23950 + }, + { + "epoch": 2.630243795299802, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.546339273452759, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7455794811248779, + "num_tokens": 619634800.0, + "step": 23951 + }, + { + "epoch": 2.630353613002416, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2629361152648926, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.733671247959137, + "num_tokens": 619661653.0, + "step": 23952 + }, + { + "epoch": 2.6304634307050296, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.262481212615967, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.7346762418746948, + "num_tokens": 619688426.0, + "step": 23953 + }, + { + "epoch": 2.6305732484076434, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5188066959381104, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7300462126731873, + "num_tokens": 619711954.0, + "step": 23954 + }, + { + "epoch": 2.630683066110257, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2368831634521484, + "learning_rate": 1e-06, + "loss": 0.9434, + "mean_token_accuracy": 0.7188527584075928, + "num_tokens": 619740104.0, + "step": 23955 + }, + { + "epoch": 2.6307928838128705, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4246487617492676, + "learning_rate": 1e-06, + "loss": 0.9176, + "mean_token_accuracy": 0.7290685176849365, + "num_tokens": 619763446.0, + "step": 23956 + }, + { + "epoch": 2.630902701515484, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3727190494537354, + "learning_rate": 1e-06, + "loss": 0.9101, + "mean_token_accuracy": 0.7140223979949951, + "num_tokens": 619789381.0, + "step": 23957 + }, + { + "epoch": 2.631012519218098, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.568052291870117, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7346148490905762, + "num_tokens": 619815071.0, + "step": 23958 + }, + { + "epoch": 2.6311223369207117, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.571460485458374, + "learning_rate": 1e-06, + "loss": 0.7833, + "mean_token_accuracy": 0.7446798086166382, + "num_tokens": 619836300.0, + "step": 23959 + }, + { + "epoch": 2.6312321546233255, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.332310676574707, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7170401811599731, + "num_tokens": 619864827.0, + "step": 23960 + }, + { + "epoch": 2.631341972325939, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.618609666824341, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7451317310333252, + "num_tokens": 619886514.0, + "step": 23961 + }, + { + "epoch": 2.6314517900285526, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3489065170288086, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7323527336120605, + "num_tokens": 619911601.0, + "step": 23962 + }, + { + "epoch": 2.6315616077311663, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.68776273727417, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.7378005385398865, + "num_tokens": 619933542.0, + "step": 23963 + }, + { + "epoch": 2.63167142543378, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2364859580993652, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.733462929725647, + "num_tokens": 619962953.0, + "step": 23964 + }, + { + "epoch": 2.631781243136394, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.336289405822754, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.7301210165023804, + "num_tokens": 619990685.0, + "step": 23965 + }, + { + "epoch": 2.631891060839007, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4402530193328857, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7295024991035461, + "num_tokens": 620016539.0, + "step": 23966 + }, + { + "epoch": 2.632000878541621, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0602023601531982, + "learning_rate": 1e-06, + "loss": 0.8937, + "mean_token_accuracy": 0.7180372476577759, + "num_tokens": 620048256.0, + "step": 23967 + }, + { + "epoch": 2.6321106962442347, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.626720905303955, + "learning_rate": 1e-06, + "loss": 0.8205, + "mean_token_accuracy": 0.7358094453811646, + "num_tokens": 620071741.0, + "step": 23968 + }, + { + "epoch": 2.632220513946848, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4208319187164307, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7291256189346313, + "num_tokens": 620096596.0, + "step": 23969 + }, + { + "epoch": 2.6323303316494617, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6393442153930664, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7339144945144653, + "num_tokens": 620118757.0, + "step": 23970 + }, + { + "epoch": 2.6324401493520755, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.183027744293213, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.6989963054656982, + "num_tokens": 620152284.0, + "step": 23971 + }, + { + "epoch": 2.6325499670546892, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7026736736297607, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7263062000274658, + "num_tokens": 620174262.0, + "step": 23972 + }, + { + "epoch": 2.632659784757303, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1424777507781982, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6889052987098694, + "num_tokens": 620205473.0, + "step": 23973 + }, + { + "epoch": 2.6327696024599163, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.305532455444336, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7163755893707275, + "num_tokens": 620231736.0, + "step": 23974 + }, + { + "epoch": 2.63287942016253, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.203434944152832, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7513652443885803, + "num_tokens": 620260991.0, + "step": 23975 + }, + { + "epoch": 2.632989237865144, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1992855072021484, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7311683297157288, + "num_tokens": 620288854.0, + "step": 23976 + }, + { + "epoch": 2.6330990555677576, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2699906826019287, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7103276252746582, + "num_tokens": 620317596.0, + "step": 23977 + }, + { + "epoch": 2.6332088732703713, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3487284183502197, + "learning_rate": 1e-06, + "loss": 0.7623, + "mean_token_accuracy": 0.7549050450325012, + "num_tokens": 620341641.0, + "step": 23978 + }, + { + "epoch": 2.6333186909729847, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.231588840484619, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7324649095535278, + "num_tokens": 620371222.0, + "step": 23979 + }, + { + "epoch": 2.6334285086755984, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.479853630065918, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7427606582641602, + "num_tokens": 620395234.0, + "step": 23980 + }, + { + "epoch": 2.633538326378212, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2104623317718506, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7334753274917603, + "num_tokens": 620423716.0, + "step": 23981 + }, + { + "epoch": 2.633648144080826, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6531848907470703, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7303865551948547, + "num_tokens": 620445412.0, + "step": 23982 + }, + { + "epoch": 2.6337579617834397, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5296173095703125, + "learning_rate": 1e-06, + "loss": 0.7871, + "mean_token_accuracy": 0.7509493827819824, + "num_tokens": 620468398.0, + "step": 23983 + }, + { + "epoch": 2.633867779486053, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.396237373352051, + "learning_rate": 1e-06, + "loss": 0.8545, + "mean_token_accuracy": 0.7316221594810486, + "num_tokens": 620492287.0, + "step": 23984 + }, + { + "epoch": 2.6339775971886668, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5267961025238037, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7402058243751526, + "num_tokens": 620514595.0, + "step": 23985 + }, + { + "epoch": 2.6340874148912805, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.215725898742676, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7332379817962646, + "num_tokens": 620543280.0, + "step": 23986 + }, + { + "epoch": 2.6341972325938943, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.604092597961426, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7319283485412598, + "num_tokens": 620564893.0, + "step": 23987 + }, + { + "epoch": 2.634307050296508, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.34448504447937, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7196689248085022, + "num_tokens": 620592556.0, + "step": 23988 + }, + { + "epoch": 2.6344168679991213, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.761599063873291, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7528423070907593, + "num_tokens": 620612564.0, + "step": 23989 + }, + { + "epoch": 2.634526685701735, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.349559783935547, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7461278438568115, + "num_tokens": 620637624.0, + "step": 23990 + }, + { + "epoch": 2.634636503404349, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4059085845947266, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7278200387954712, + "num_tokens": 620662855.0, + "step": 23991 + }, + { + "epoch": 2.6347463211069626, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.559877395629883, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7509406208992004, + "num_tokens": 620684595.0, + "step": 23992 + }, + { + "epoch": 2.6348561388095764, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3879010677337646, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7312669157981873, + "num_tokens": 620710732.0, + "step": 23993 + }, + { + "epoch": 2.6349659565121897, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1969642639160156, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7116044759750366, + "num_tokens": 620740712.0, + "step": 23994 + }, + { + "epoch": 2.6350757742148034, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.443037509918213, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.737837553024292, + "num_tokens": 620764298.0, + "step": 23995 + }, + { + "epoch": 2.635185591917417, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2964658737182617, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7226466536521912, + "num_tokens": 620790394.0, + "step": 23996 + }, + { + "epoch": 2.6352954096200305, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2810890674591064, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.712401270866394, + "num_tokens": 620817893.0, + "step": 23997 + }, + { + "epoch": 2.6354052273226443, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1355977058410645, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7473210096359253, + "num_tokens": 620844533.0, + "step": 23998 + }, + { + "epoch": 2.635515045025258, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7525794506073, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7257832288742065, + "num_tokens": 620867060.0, + "step": 23999 + }, + { + "epoch": 2.6356248627278718, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3381893634796143, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7301013469696045, + "num_tokens": 620893365.0, + "step": 24000 + }, + { + "epoch": 2.6357346804304855, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.24591064453125, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7431177496910095, + "num_tokens": 620920208.0, + "step": 24001 + }, + { + "epoch": 2.635844498133099, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6334598064422607, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.724281907081604, + "num_tokens": 620942974.0, + "step": 24002 + }, + { + "epoch": 2.6359543158357126, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.265002727508545, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7159988880157471, + "num_tokens": 620970769.0, + "step": 24003 + }, + { + "epoch": 2.6360641335383264, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1469767093658447, + "learning_rate": 1e-06, + "loss": 0.9539, + "mean_token_accuracy": 0.7056682109832764, + "num_tokens": 621003008.0, + "step": 24004 + }, + { + "epoch": 2.63617395124094, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.212437152862549, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7256301641464233, + "num_tokens": 621033307.0, + "step": 24005 + }, + { + "epoch": 2.636283768943554, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.179910898208618, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.729813814163208, + "num_tokens": 621063311.0, + "step": 24006 + }, + { + "epoch": 2.636393586646167, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2628836631774902, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7264906167984009, + "num_tokens": 621091081.0, + "step": 24007 + }, + { + "epoch": 2.636503404348781, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.851379156112671, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7174441814422607, + "num_tokens": 621111113.0, + "step": 24008 + }, + { + "epoch": 2.6366132220513947, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3540914058685303, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7279082536697388, + "num_tokens": 621137657.0, + "step": 24009 + }, + { + "epoch": 2.6367230397540085, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.653459310531616, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7483198046684265, + "num_tokens": 621157945.0, + "step": 24010 + }, + { + "epoch": 2.636832857456622, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7086775302886963, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7159042358398438, + "num_tokens": 621178900.0, + "step": 24011 + }, + { + "epoch": 2.6369426751592355, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.393710136413574, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7323157787322998, + "num_tokens": 621204561.0, + "step": 24012 + }, + { + "epoch": 2.6370524928618493, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.577450752258301, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.735599160194397, + "num_tokens": 621226032.0, + "step": 24013 + }, + { + "epoch": 2.637162310564463, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7464120388031006, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7258294820785522, + "num_tokens": 621245472.0, + "step": 24014 + }, + { + "epoch": 2.637272128267077, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5979511737823486, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7242566347122192, + "num_tokens": 621266922.0, + "step": 24015 + }, + { + "epoch": 2.6373819459696906, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.245492458343506, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7206317186355591, + "num_tokens": 621296209.0, + "step": 24016 + }, + { + "epoch": 2.637491763672304, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6710877418518066, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.718889594078064, + "num_tokens": 621318693.0, + "step": 24017 + }, + { + "epoch": 2.6376015813749176, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2610223293304443, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.7004252076148987, + "num_tokens": 621347752.0, + "step": 24018 + }, + { + "epoch": 2.6377113990775314, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3515889644622803, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.7577565908432007, + "num_tokens": 621373631.0, + "step": 24019 + }, + { + "epoch": 2.6378212167801447, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4674649238586426, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.7426304221153259, + "num_tokens": 621398262.0, + "step": 24020 + }, + { + "epoch": 2.637931034482759, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2949330806732178, + "learning_rate": 1e-06, + "loss": 0.7874, + "mean_token_accuracy": 0.7559057474136353, + "num_tokens": 621423360.0, + "step": 24021 + }, + { + "epoch": 2.638040852185372, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2420437335968018, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.733554482460022, + "num_tokens": 621451716.0, + "step": 24022 + }, + { + "epoch": 2.638150669887986, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5585083961486816, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7202598452568054, + "num_tokens": 621475198.0, + "step": 24023 + }, + { + "epoch": 2.6382604875905997, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.43601393699646, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7280049920082092, + "num_tokens": 621499649.0, + "step": 24024 + }, + { + "epoch": 2.638370305293213, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3375167846679688, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7395197153091431, + "num_tokens": 621524844.0, + "step": 24025 + }, + { + "epoch": 2.638480122995827, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1797947883605957, + "learning_rate": 1e-06, + "loss": 0.9529, + "mean_token_accuracy": 0.7054955959320068, + "num_tokens": 621556794.0, + "step": 24026 + }, + { + "epoch": 2.6385899406984406, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.771198034286499, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7304344177246094, + "num_tokens": 621576931.0, + "step": 24027 + }, + { + "epoch": 2.6386997584010543, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.21448016166687, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7259494066238403, + "num_tokens": 621603944.0, + "step": 24028 + }, + { + "epoch": 2.638809576103668, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.169511318206787, + "learning_rate": 1e-06, + "loss": 0.9347, + "mean_token_accuracy": 0.7138922214508057, + "num_tokens": 621635189.0, + "step": 24029 + }, + { + "epoch": 2.6389193938062814, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6962904930114746, + "learning_rate": 1e-06, + "loss": 0.7555, + "mean_token_accuracy": 0.75179123878479, + "num_tokens": 621656516.0, + "step": 24030 + }, + { + "epoch": 2.639029211508895, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.40566086769104, + "learning_rate": 1e-06, + "loss": 0.8914, + "mean_token_accuracy": 0.7293586134910583, + "num_tokens": 621681421.0, + "step": 24031 + }, + { + "epoch": 2.639139029211509, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4604458808898926, + "learning_rate": 1e-06, + "loss": 0.9354, + "mean_token_accuracy": 0.70299232006073, + "num_tokens": 621705867.0, + "step": 24032 + }, + { + "epoch": 2.6392488469141226, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.263287305831909, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7361429929733276, + "num_tokens": 621731823.0, + "step": 24033 + }, + { + "epoch": 2.6393586646167364, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8398377895355225, + "learning_rate": 1e-06, + "loss": 0.7581, + "mean_token_accuracy": 0.7499619126319885, + "num_tokens": 621751871.0, + "step": 24034 + }, + { + "epoch": 2.6394684823193497, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.28802227973938, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7333053946495056, + "num_tokens": 621778959.0, + "step": 24035 + }, + { + "epoch": 2.6395783000219635, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3180670738220215, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7201322317123413, + "num_tokens": 621806512.0, + "step": 24036 + }, + { + "epoch": 2.6396881177245772, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9719178676605225, + "learning_rate": 1e-06, + "loss": 0.7958, + "mean_token_accuracy": 0.7430630922317505, + "num_tokens": 621823949.0, + "step": 24037 + }, + { + "epoch": 2.639797935427191, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2367634773254395, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7426645755767822, + "num_tokens": 621851867.0, + "step": 24038 + }, + { + "epoch": 2.6399077531298047, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3496382236480713, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7367556095123291, + "num_tokens": 621877830.0, + "step": 24039 + }, + { + "epoch": 2.640017570832418, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3900301456451416, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.712889552116394, + "num_tokens": 621904075.0, + "step": 24040 + }, + { + "epoch": 2.640127388535032, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7443156242370605, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7478625774383545, + "num_tokens": 621923382.0, + "step": 24041 + }, + { + "epoch": 2.6402372062376456, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2542810440063477, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7295316457748413, + "num_tokens": 621953238.0, + "step": 24042 + }, + { + "epoch": 2.6403470239402593, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7915866374969482, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7446715831756592, + "num_tokens": 621972582.0, + "step": 24043 + }, + { + "epoch": 2.640456841642873, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4558444023132324, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7298036813735962, + "num_tokens": 621997752.0, + "step": 24044 + }, + { + "epoch": 2.6405666593454864, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.670163154602051, + "learning_rate": 1e-06, + "loss": 0.8276, + "mean_token_accuracy": 0.7424594163894653, + "num_tokens": 622020119.0, + "step": 24045 + }, + { + "epoch": 2.6406764770481, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.592616558074951, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7349718809127808, + "num_tokens": 622043496.0, + "step": 24046 + }, + { + "epoch": 2.640786294750714, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3025524616241455, + "learning_rate": 1e-06, + "loss": 0.8834, + "mean_token_accuracy": 0.7194176316261292, + "num_tokens": 622070300.0, + "step": 24047 + }, + { + "epoch": 2.6408961124533272, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.531689405441284, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7305612564086914, + "num_tokens": 622094411.0, + "step": 24048 + }, + { + "epoch": 2.641005930155941, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.061985731124878, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7084260582923889, + "num_tokens": 622128698.0, + "step": 24049 + }, + { + "epoch": 2.6411157478585547, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6288397312164307, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7328603267669678, + "num_tokens": 622148936.0, + "step": 24050 + }, + { + "epoch": 2.6412255655611685, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1892504692077637, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7359535098075867, + "num_tokens": 622176301.0, + "step": 24051 + }, + { + "epoch": 2.6413353832637823, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0932843685150146, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7130735516548157, + "num_tokens": 622208911.0, + "step": 24052 + }, + { + "epoch": 2.6414452009663956, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3576889038085938, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7142643332481384, + "num_tokens": 622236356.0, + "step": 24053 + }, + { + "epoch": 2.6415550186690093, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.305187940597534, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.728891909122467, + "num_tokens": 622262816.0, + "step": 24054 + }, + { + "epoch": 2.641664836371623, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.579904556274414, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7151479721069336, + "num_tokens": 622284963.0, + "step": 24055 + }, + { + "epoch": 2.641774654074237, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.706401824951172, + "learning_rate": 1e-06, + "loss": 0.7409, + "mean_token_accuracy": 0.7576544880867004, + "num_tokens": 622305749.0, + "step": 24056 + }, + { + "epoch": 2.6418844717768506, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.378628730773926, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7488024830818176, + "num_tokens": 622331647.0, + "step": 24057 + }, + { + "epoch": 2.641994289479464, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.484459400177002, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7235407829284668, + "num_tokens": 622354461.0, + "step": 24058 + }, + { + "epoch": 2.6421041071820777, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.452934503555298, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7244435548782349, + "num_tokens": 622378606.0, + "step": 24059 + }, + { + "epoch": 2.6422139248846914, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.319129705429077, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7321767210960388, + "num_tokens": 622404972.0, + "step": 24060 + }, + { + "epoch": 2.642323742587305, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1887595653533936, + "learning_rate": 1e-06, + "loss": 0.8999, + "mean_token_accuracy": 0.7157149314880371, + "num_tokens": 622437224.0, + "step": 24061 + }, + { + "epoch": 2.642433560289919, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.100839138031006, + "learning_rate": 1e-06, + "loss": 0.9745, + "mean_token_accuracy": 0.6915194988250732, + "num_tokens": 622470711.0, + "step": 24062 + }, + { + "epoch": 2.6425433779925322, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.469360828399658, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7446920871734619, + "num_tokens": 622495174.0, + "step": 24063 + }, + { + "epoch": 2.642653195695146, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.550696611404419, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7500208616256714, + "num_tokens": 622518425.0, + "step": 24064 + }, + { + "epoch": 2.6427630133977598, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.118981122970581, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7208017110824585, + "num_tokens": 622549915.0, + "step": 24065 + }, + { + "epoch": 2.6428728311003735, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6235671043395996, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7055200934410095, + "num_tokens": 622572444.0, + "step": 24066 + }, + { + "epoch": 2.6429826488029873, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7244935035705566, + "learning_rate": 1e-06, + "loss": 0.7649, + "mean_token_accuracy": 0.7496880292892456, + "num_tokens": 622593190.0, + "step": 24067 + }, + { + "epoch": 2.6430924665056006, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4470784664154053, + "learning_rate": 1e-06, + "loss": 0.8949, + "mean_token_accuracy": 0.7310084104537964, + "num_tokens": 622620392.0, + "step": 24068 + }, + { + "epoch": 2.6432022842082143, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3268356323242188, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.725374162197113, + "num_tokens": 622649585.0, + "step": 24069 + }, + { + "epoch": 2.643312101910828, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1949782371520996, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.739023745059967, + "num_tokens": 622676984.0, + "step": 24070 + }, + { + "epoch": 2.6434219196134414, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2025327682495117, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.721788763999939, + "num_tokens": 622705084.0, + "step": 24071 + }, + { + "epoch": 2.6435317373160556, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4087743759155273, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7246608734130859, + "num_tokens": 622732563.0, + "step": 24072 + }, + { + "epoch": 2.643641555018669, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5701406002044678, + "learning_rate": 1e-06, + "loss": 0.7924, + "mean_token_accuracy": 0.7457512021064758, + "num_tokens": 622757536.0, + "step": 24073 + }, + { + "epoch": 2.6437513727212827, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2505381107330322, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7281479835510254, + "num_tokens": 622787075.0, + "step": 24074 + }, + { + "epoch": 2.6438611904238964, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4702398777008057, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.7252689599990845, + "num_tokens": 622811946.0, + "step": 24075 + }, + { + "epoch": 2.6439710081265098, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.289198637008667, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7336069941520691, + "num_tokens": 622839809.0, + "step": 24076 + }, + { + "epoch": 2.6440808258291235, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7389047145843506, + "learning_rate": 1e-06, + "loss": 0.806, + "mean_token_accuracy": 0.7431610822677612, + "num_tokens": 622858369.0, + "step": 24077 + }, + { + "epoch": 2.6441906435317373, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.666386842727661, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7432674169540405, + "num_tokens": 622879353.0, + "step": 24078 + }, + { + "epoch": 2.644300461234351, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.543635845184326, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.714267373085022, + "num_tokens": 622906053.0, + "step": 24079 + }, + { + "epoch": 2.644410278936965, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2249834537506104, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7196913957595825, + "num_tokens": 622935393.0, + "step": 24080 + }, + { + "epoch": 2.644520096639578, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.334914207458496, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7221894860267639, + "num_tokens": 622961271.0, + "step": 24081 + }, + { + "epoch": 2.644629914342192, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2600059509277344, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.746360719203949, + "num_tokens": 622988400.0, + "step": 24082 + }, + { + "epoch": 2.6447397320448056, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5822253227233887, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7294251322746277, + "num_tokens": 623011022.0, + "step": 24083 + }, + { + "epoch": 2.6448495497474194, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.727841854095459, + "learning_rate": 1e-06, + "loss": 0.7394, + "mean_token_accuracy": 0.770042359828949, + "num_tokens": 623030796.0, + "step": 24084 + }, + { + "epoch": 2.644959367450033, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1566617488861084, + "learning_rate": 1e-06, + "loss": 0.9074, + "mean_token_accuracy": 0.7157810926437378, + "num_tokens": 623060043.0, + "step": 24085 + }, + { + "epoch": 2.6450691851526464, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.402261734008789, + "learning_rate": 1e-06, + "loss": 0.8225, + "mean_token_accuracy": 0.7480084896087646, + "num_tokens": 623082033.0, + "step": 24086 + }, + { + "epoch": 2.64517900285526, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3949697017669678, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.719382643699646, + "num_tokens": 623109445.0, + "step": 24087 + }, + { + "epoch": 2.645288820557874, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2613468170166016, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.716717004776001, + "num_tokens": 623138438.0, + "step": 24088 + }, + { + "epoch": 2.6453986382604877, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3908820152282715, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7405948638916016, + "num_tokens": 623162892.0, + "step": 24089 + }, + { + "epoch": 2.6455084559631015, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 3.030789613723755, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.743903636932373, + "num_tokens": 623181200.0, + "step": 24090 + }, + { + "epoch": 2.645618273665715, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.278071641921997, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7300277948379517, + "num_tokens": 623209572.0, + "step": 24091 + }, + { + "epoch": 2.6457280913683285, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5535340309143066, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7196454405784607, + "num_tokens": 623233462.0, + "step": 24092 + }, + { + "epoch": 2.6458379090709423, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2227206230163574, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7289032936096191, + "num_tokens": 623260152.0, + "step": 24093 + }, + { + "epoch": 2.645947726773556, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1426773071289062, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7218966484069824, + "num_tokens": 623290976.0, + "step": 24094 + }, + { + "epoch": 2.64605754447617, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5674166679382324, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7162994146347046, + "num_tokens": 623314911.0, + "step": 24095 + }, + { + "epoch": 2.646167362178783, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.303234100341797, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7346755266189575, + "num_tokens": 623342413.0, + "step": 24096 + }, + { + "epoch": 2.646277179881397, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2059531211853027, + "learning_rate": 1e-06, + "loss": 0.9081, + "mean_token_accuracy": 0.7147029042243958, + "num_tokens": 623371308.0, + "step": 24097 + }, + { + "epoch": 2.6463869975840106, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1464359760284424, + "learning_rate": 1e-06, + "loss": 0.7425, + "mean_token_accuracy": 0.7609768509864807, + "num_tokens": 623398913.0, + "step": 24098 + }, + { + "epoch": 2.646496815286624, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1513009071350098, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7139971256256104, + "num_tokens": 623430023.0, + "step": 24099 + }, + { + "epoch": 2.6466066329892377, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.107619285583496, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.717513918876648, + "num_tokens": 623461192.0, + "step": 24100 + }, + { + "epoch": 2.6467164506918515, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3726322650909424, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7418351173400879, + "num_tokens": 623487347.0, + "step": 24101 + }, + { + "epoch": 2.646826268394465, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.425239324569702, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7209898233413696, + "num_tokens": 623511378.0, + "step": 24102 + }, + { + "epoch": 2.646936086097079, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4093127250671387, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7228089570999146, + "num_tokens": 623535263.0, + "step": 24103 + }, + { + "epoch": 2.6470459037996923, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.362213134765625, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7192811965942383, + "num_tokens": 623562398.0, + "step": 24104 + }, + { + "epoch": 2.647155721502306, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.537646770477295, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.733672022819519, + "num_tokens": 623586222.0, + "step": 24105 + }, + { + "epoch": 2.64726553920492, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.66438889503479, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7269377112388611, + "num_tokens": 623608154.0, + "step": 24106 + }, + { + "epoch": 2.6473753569075336, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4675357341766357, + "learning_rate": 1e-06, + "loss": 0.948, + "mean_token_accuracy": 0.7057210803031921, + "num_tokens": 623634219.0, + "step": 24107 + }, + { + "epoch": 2.6474851746101473, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.413553476333618, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.724174976348877, + "num_tokens": 623661239.0, + "step": 24108 + }, + { + "epoch": 2.6475949923127606, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.216461658477783, + "learning_rate": 1e-06, + "loss": 0.7664, + "mean_token_accuracy": 0.7522971630096436, + "num_tokens": 623687698.0, + "step": 24109 + }, + { + "epoch": 2.6477048100153744, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4179511070251465, + "learning_rate": 1e-06, + "loss": 0.8157, + "mean_token_accuracy": 0.735079824924469, + "num_tokens": 623712239.0, + "step": 24110 + }, + { + "epoch": 2.647814627717988, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.575897693634033, + "learning_rate": 1e-06, + "loss": 0.7676, + "mean_token_accuracy": 0.746666431427002, + "num_tokens": 623734006.0, + "step": 24111 + }, + { + "epoch": 2.647924445420602, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3650100231170654, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7266438603401184, + "num_tokens": 623761280.0, + "step": 24112 + }, + { + "epoch": 2.6480342631232157, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2782368659973145, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7191810011863708, + "num_tokens": 623788872.0, + "step": 24113 + }, + { + "epoch": 2.648144080825829, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3776590824127197, + "learning_rate": 1e-06, + "loss": 0.9199, + "mean_token_accuracy": 0.709426760673523, + "num_tokens": 623815823.0, + "step": 24114 + }, + { + "epoch": 2.6482538985284427, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.44008469581604, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7263413667678833, + "num_tokens": 623841142.0, + "step": 24115 + }, + { + "epoch": 2.6483637162310565, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.294464111328125, + "learning_rate": 1e-06, + "loss": 0.9757, + "mean_token_accuracy": 0.6952219009399414, + "num_tokens": 623868966.0, + "step": 24116 + }, + { + "epoch": 2.6484735339336702, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.559704542160034, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7348629236221313, + "num_tokens": 623889833.0, + "step": 24117 + }, + { + "epoch": 2.648583351636284, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3823530673980713, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7173782587051392, + "num_tokens": 623916392.0, + "step": 24118 + }, + { + "epoch": 2.6486931693388973, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3930699825286865, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7284508943557739, + "num_tokens": 623941110.0, + "step": 24119 + }, + { + "epoch": 2.648802987041511, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.354710817337036, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7414911985397339, + "num_tokens": 623965608.0, + "step": 24120 + }, + { + "epoch": 2.648912804744125, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.417898416519165, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7323299050331116, + "num_tokens": 623992636.0, + "step": 24121 + }, + { + "epoch": 2.649022622446738, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3859448432922363, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7455546855926514, + "num_tokens": 624016363.0, + "step": 24122 + }, + { + "epoch": 2.6491324401493523, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2412772178649902, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7459802627563477, + "num_tokens": 624043771.0, + "step": 24123 + }, + { + "epoch": 2.6492422578519657, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1838700771331787, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.723315417766571, + "num_tokens": 624072386.0, + "step": 24124 + }, + { + "epoch": 2.6493520755545794, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.382782459259033, + "learning_rate": 1e-06, + "loss": 0.8917, + "mean_token_accuracy": 0.7302605509757996, + "num_tokens": 624097076.0, + "step": 24125 + }, + { + "epoch": 2.649461893257193, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4007413387298584, + "learning_rate": 1e-06, + "loss": 0.7619, + "mean_token_accuracy": 0.7496553659439087, + "num_tokens": 624120131.0, + "step": 24126 + }, + { + "epoch": 2.6495717109598065, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.285731315612793, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.742438793182373, + "num_tokens": 624148254.0, + "step": 24127 + }, + { + "epoch": 2.6496815286624202, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.212400436401367, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.714084267616272, + "num_tokens": 624176985.0, + "step": 24128 + }, + { + "epoch": 2.649791346365034, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.447101354598999, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7350500822067261, + "num_tokens": 624202640.0, + "step": 24129 + }, + { + "epoch": 2.6499011640676478, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5011565685272217, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7470526695251465, + "num_tokens": 624225318.0, + "step": 24130 + }, + { + "epoch": 2.6500109817702615, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5127041339874268, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7354397177696228, + "num_tokens": 624249215.0, + "step": 24131 + }, + { + "epoch": 2.650120799472875, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5672497749328613, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.7346483469009399, + "num_tokens": 624271947.0, + "step": 24132 + }, + { + "epoch": 2.6502306171754886, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.410595178604126, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7384040951728821, + "num_tokens": 624294819.0, + "step": 24133 + }, + { + "epoch": 2.6503404348781023, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.525770425796509, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7315435409545898, + "num_tokens": 624319693.0, + "step": 24134 + }, + { + "epoch": 2.650450252580716, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0975306034088135, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.729324221611023, + "num_tokens": 624350989.0, + "step": 24135 + }, + { + "epoch": 2.65056007028333, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1517202854156494, + "learning_rate": 1e-06, + "loss": 0.7656, + "mean_token_accuracy": 0.7568522095680237, + "num_tokens": 624380414.0, + "step": 24136 + }, + { + "epoch": 2.650669887985943, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4002976417541504, + "learning_rate": 1e-06, + "loss": 0.8714, + "mean_token_accuracy": 0.7219226360321045, + "num_tokens": 624406399.0, + "step": 24137 + }, + { + "epoch": 2.650779705688557, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2910165786743164, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.7106490731239319, + "num_tokens": 624433901.0, + "step": 24138 + }, + { + "epoch": 2.6508895233911707, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3606314659118652, + "learning_rate": 1e-06, + "loss": 0.8239, + "mean_token_accuracy": 0.7405809164047241, + "num_tokens": 624458660.0, + "step": 24139 + }, + { + "epoch": 2.6509993410937844, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5635616779327393, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7357615828514099, + "num_tokens": 624481456.0, + "step": 24140 + }, + { + "epoch": 2.651109158796398, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5874924659729004, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7334370613098145, + "num_tokens": 624504676.0, + "step": 24141 + }, + { + "epoch": 2.6512189764990115, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2029078006744385, + "learning_rate": 1e-06, + "loss": 1.0073, + "mean_token_accuracy": 0.6962963342666626, + "num_tokens": 624535033.0, + "step": 24142 + }, + { + "epoch": 2.6513287942016253, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.57047963142395, + "learning_rate": 1e-06, + "loss": 0.788, + "mean_token_accuracy": 0.7430803179740906, + "num_tokens": 624557535.0, + "step": 24143 + }, + { + "epoch": 2.651438611904239, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2874372005462646, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.731013298034668, + "num_tokens": 624583930.0, + "step": 24144 + }, + { + "epoch": 2.6515484296068528, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.599717855453491, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7302906513214111, + "num_tokens": 624607216.0, + "step": 24145 + }, + { + "epoch": 2.6516582473094665, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3935389518737793, + "learning_rate": 1e-06, + "loss": 0.7989, + "mean_token_accuracy": 0.7455379962921143, + "num_tokens": 624629725.0, + "step": 24146 + }, + { + "epoch": 2.65176806501208, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.350142002105713, + "learning_rate": 1e-06, + "loss": 0.8566, + "mean_token_accuracy": 0.7280128002166748, + "num_tokens": 624654995.0, + "step": 24147 + }, + { + "epoch": 2.6518778827146936, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5226986408233643, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7509815096855164, + "num_tokens": 624677595.0, + "step": 24148 + }, + { + "epoch": 2.6519877004173074, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3863160610198975, + "learning_rate": 1e-06, + "loss": 0.8958, + "mean_token_accuracy": 0.7181543111801147, + "num_tokens": 624705954.0, + "step": 24149 + }, + { + "epoch": 2.6520975181199207, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2710824012756348, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7172311544418335, + "num_tokens": 624735000.0, + "step": 24150 + }, + { + "epoch": 2.6522073358225344, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7246105670928955, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7222616672515869, + "num_tokens": 624756511.0, + "step": 24151 + }, + { + "epoch": 2.652317153525148, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2100627422332764, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7133558988571167, + "num_tokens": 624784308.0, + "step": 24152 + }, + { + "epoch": 2.652426971227762, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.473665714263916, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7291172742843628, + "num_tokens": 624809534.0, + "step": 24153 + }, + { + "epoch": 2.6525367889303757, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3653903007507324, + "learning_rate": 1e-06, + "loss": 0.7983, + "mean_token_accuracy": 0.7442224025726318, + "num_tokens": 624834936.0, + "step": 24154 + }, + { + "epoch": 2.652646606632989, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1861391067504883, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.7147090435028076, + "num_tokens": 624864487.0, + "step": 24155 + }, + { + "epoch": 2.6527564243356028, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.472369432449341, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7120540142059326, + "num_tokens": 624889573.0, + "step": 24156 + }, + { + "epoch": 2.6528662420382165, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2537155151367188, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7092326879501343, + "num_tokens": 624918369.0, + "step": 24157 + }, + { + "epoch": 2.6529760597408303, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3937511444091797, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7365914583206177, + "num_tokens": 624944122.0, + "step": 24158 + }, + { + "epoch": 2.653085877443444, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6077094078063965, + "learning_rate": 1e-06, + "loss": 0.7352, + "mean_token_accuracy": 0.7593063712120056, + "num_tokens": 624964762.0, + "step": 24159 + }, + { + "epoch": 2.6531956951460574, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4264187812805176, + "learning_rate": 1e-06, + "loss": 0.9125, + "mean_token_accuracy": 0.7183436155319214, + "num_tokens": 624989315.0, + "step": 24160 + }, + { + "epoch": 2.653305512848671, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.423994541168213, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7106859683990479, + "num_tokens": 625016020.0, + "step": 24161 + }, + { + "epoch": 2.653415330551285, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0615322589874268, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7319760322570801, + "num_tokens": 625046600.0, + "step": 24162 + }, + { + "epoch": 2.6535251482538986, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.334645986557007, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7283344268798828, + "num_tokens": 625073179.0, + "step": 24163 + }, + { + "epoch": 2.6536349659565124, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2026078701019287, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7451514005661011, + "num_tokens": 625100817.0, + "step": 24164 + }, + { + "epoch": 2.6537447836591257, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.103302001953125, + "learning_rate": 1e-06, + "loss": 0.9513, + "mean_token_accuracy": 0.7077186107635498, + "num_tokens": 625132091.0, + "step": 24165 + }, + { + "epoch": 2.6538546013617395, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.568892240524292, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7307289838790894, + "num_tokens": 625156293.0, + "step": 24166 + }, + { + "epoch": 2.653964419064353, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2652523517608643, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7102399468421936, + "num_tokens": 625185119.0, + "step": 24167 + }, + { + "epoch": 2.654074236766967, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2724668979644775, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7201855182647705, + "num_tokens": 625216115.0, + "step": 24168 + }, + { + "epoch": 2.6541840544695807, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3504996299743652, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.729342520236969, + "num_tokens": 625243558.0, + "step": 24169 + }, + { + "epoch": 2.654293872172194, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.504366397857666, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7372658252716064, + "num_tokens": 625266309.0, + "step": 24170 + }, + { + "epoch": 2.654403689874808, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1760053634643555, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7121235728263855, + "num_tokens": 625295082.0, + "step": 24171 + }, + { + "epoch": 2.6545135075774215, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5364840030670166, + "learning_rate": 1e-06, + "loss": 0.7677, + "mean_token_accuracy": 0.7513856887817383, + "num_tokens": 625319200.0, + "step": 24172 + }, + { + "epoch": 2.6546233252800353, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.397334575653076, + "learning_rate": 1e-06, + "loss": 0.785, + "mean_token_accuracy": 0.754886269569397, + "num_tokens": 625344812.0, + "step": 24173 + }, + { + "epoch": 2.654733142982649, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2675554752349854, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.7327063083648682, + "num_tokens": 625371736.0, + "step": 24174 + }, + { + "epoch": 2.6548429606852624, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2533226013183594, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.7378259897232056, + "num_tokens": 625400843.0, + "step": 24175 + }, + { + "epoch": 2.654952778387876, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8233017921447754, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7249051332473755, + "num_tokens": 625421941.0, + "step": 24176 + }, + { + "epoch": 2.65506259609049, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7411744594573975, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7351319789886475, + "num_tokens": 625443964.0, + "step": 24177 + }, + { + "epoch": 2.655172413793103, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.736015796661377, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.7449936866760254, + "num_tokens": 625463927.0, + "step": 24178 + }, + { + "epoch": 2.655282231495717, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1825644969940186, + "learning_rate": 1e-06, + "loss": 0.8288, + "mean_token_accuracy": 0.739205002784729, + "num_tokens": 625492348.0, + "step": 24179 + }, + { + "epoch": 2.6553920491983307, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6213467121124268, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7383432388305664, + "num_tokens": 625513604.0, + "step": 24180 + }, + { + "epoch": 2.6555018669009445, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6297590732574463, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7341026067733765, + "num_tokens": 625534927.0, + "step": 24181 + }, + { + "epoch": 2.6556116846035582, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.514392614364624, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7483274936676025, + "num_tokens": 625558477.0, + "step": 24182 + }, + { + "epoch": 2.6557215023061715, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.236593246459961, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7364359498023987, + "num_tokens": 625585865.0, + "step": 24183 + }, + { + "epoch": 2.6558313200087853, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3617103099823, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7257392406463623, + "num_tokens": 625610298.0, + "step": 24184 + }, + { + "epoch": 2.655941137711399, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.375098943710327, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7365583181381226, + "num_tokens": 625635271.0, + "step": 24185 + }, + { + "epoch": 2.656050955414013, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.230370283126831, + "learning_rate": 1e-06, + "loss": 0.9515, + "mean_token_accuracy": 0.7024505138397217, + "num_tokens": 625667778.0, + "step": 24186 + }, + { + "epoch": 2.6561607731166266, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.295922040939331, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7244841456413269, + "num_tokens": 625693015.0, + "step": 24187 + }, + { + "epoch": 2.65627059081924, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.370645046234131, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7181161046028137, + "num_tokens": 625726287.0, + "step": 24188 + }, + { + "epoch": 2.6563804085218536, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.721179962158203, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7326027154922485, + "num_tokens": 625747179.0, + "step": 24189 + }, + { + "epoch": 2.6564902262244674, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3979058265686035, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7227824926376343, + "num_tokens": 625772030.0, + "step": 24190 + }, + { + "epoch": 2.656600043927081, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.318530321121216, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7247267961502075, + "num_tokens": 625800036.0, + "step": 24191 + }, + { + "epoch": 2.656709861629695, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.239290475845337, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7229897379875183, + "num_tokens": 625828947.0, + "step": 24192 + }, + { + "epoch": 2.6568196793323082, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.366936445236206, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7206407189369202, + "num_tokens": 625855482.0, + "step": 24193 + }, + { + "epoch": 2.656929497034922, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2856433391571045, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.7459705471992493, + "num_tokens": 625881288.0, + "step": 24194 + }, + { + "epoch": 2.6570393147375357, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4669101238250732, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7177467346191406, + "num_tokens": 625905574.0, + "step": 24195 + }, + { + "epoch": 2.6571491324401495, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.147883892059326, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7301499843597412, + "num_tokens": 625934805.0, + "step": 24196 + }, + { + "epoch": 2.6572589501427633, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.636815309524536, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7397019863128662, + "num_tokens": 625957195.0, + "step": 24197 + }, + { + "epoch": 2.6573687678453766, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.627242088317871, + "learning_rate": 1e-06, + "loss": 0.6849, + "mean_token_accuracy": 0.7728540897369385, + "num_tokens": 625976035.0, + "step": 24198 + }, + { + "epoch": 2.6574785855479903, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6439871788024902, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7226556539535522, + "num_tokens": 625999073.0, + "step": 24199 + }, + { + "epoch": 2.657588403250604, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.290093421936035, + "learning_rate": 1e-06, + "loss": 0.8504, + "mean_token_accuracy": 0.7289020419120789, + "num_tokens": 626026380.0, + "step": 24200 + }, + { + "epoch": 2.6576982209532174, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.391277551651001, + "learning_rate": 1e-06, + "loss": 0.8977, + "mean_token_accuracy": 0.7215807437896729, + "num_tokens": 626052800.0, + "step": 24201 + }, + { + "epoch": 2.6578080386558316, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.486032724380493, + "learning_rate": 1e-06, + "loss": 0.8262, + "mean_token_accuracy": 0.7284565567970276, + "num_tokens": 626076458.0, + "step": 24202 + }, + { + "epoch": 2.657917856358445, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.629415512084961, + "learning_rate": 1e-06, + "loss": 0.7984, + "mean_token_accuracy": 0.7466626167297363, + "num_tokens": 626097548.0, + "step": 24203 + }, + { + "epoch": 2.6580276740610587, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.473426342010498, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7426762580871582, + "num_tokens": 626121697.0, + "step": 24204 + }, + { + "epoch": 2.6581374917636724, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5566208362579346, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7380751371383667, + "num_tokens": 626143533.0, + "step": 24205 + }, + { + "epoch": 2.6582473094662857, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3061509132385254, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7186055779457092, + "num_tokens": 626169981.0, + "step": 24206 + }, + { + "epoch": 2.6583571271688995, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3298184871673584, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7240067720413208, + "num_tokens": 626197217.0, + "step": 24207 + }, + { + "epoch": 2.6584669448715132, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0208370685577393, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7338516116142273, + "num_tokens": 626229495.0, + "step": 24208 + }, + { + "epoch": 2.658576762574127, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1381819248199463, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7245621681213379, + "num_tokens": 626257022.0, + "step": 24209 + }, + { + "epoch": 2.6586865802767408, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4798715114593506, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7206757664680481, + "num_tokens": 626280771.0, + "step": 24210 + }, + { + "epoch": 2.658796397979354, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4837288856506348, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7214437127113342, + "num_tokens": 626305088.0, + "step": 24211 + }, + { + "epoch": 2.658906215681968, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7452073097229004, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7196897864341736, + "num_tokens": 626327391.0, + "step": 24212 + }, + { + "epoch": 2.6590160333845816, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.325129985809326, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7398113012313843, + "num_tokens": 626354456.0, + "step": 24213 + }, + { + "epoch": 2.6591258510871953, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2284584045410156, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7422910332679749, + "num_tokens": 626380975.0, + "step": 24214 + }, + { + "epoch": 2.659235668789809, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2513270378112793, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.748999834060669, + "num_tokens": 626408294.0, + "step": 24215 + }, + { + "epoch": 2.6593454864924224, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2669925689697266, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7353946566581726, + "num_tokens": 626436301.0, + "step": 24216 + }, + { + "epoch": 2.659455304195036, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3066282272338867, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7222979068756104, + "num_tokens": 626461838.0, + "step": 24217 + }, + { + "epoch": 2.65956512189765, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.118945837020874, + "learning_rate": 1e-06, + "loss": 0.8897, + "mean_token_accuracy": 0.7183138132095337, + "num_tokens": 626491759.0, + "step": 24218 + }, + { + "epoch": 2.6596749396002637, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.399348020553589, + "learning_rate": 1e-06, + "loss": 0.8303, + "mean_token_accuracy": 0.7364408373832703, + "num_tokens": 626516598.0, + "step": 24219 + }, + { + "epoch": 2.6597847573028774, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4696366786956787, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7129366397857666, + "num_tokens": 626542760.0, + "step": 24220 + }, + { + "epoch": 2.6598945750054908, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3495068550109863, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7361609935760498, + "num_tokens": 626568063.0, + "step": 24221 + }, + { + "epoch": 2.6600043927081045, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9300612211227417, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7355329990386963, + "num_tokens": 626602420.0, + "step": 24222 + }, + { + "epoch": 2.6601142104107183, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.608689546585083, + "learning_rate": 1e-06, + "loss": 0.7499, + "mean_token_accuracy": 0.7533811926841736, + "num_tokens": 626622888.0, + "step": 24223 + }, + { + "epoch": 2.660224028113332, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9739506244659424, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7423044443130493, + "num_tokens": 626655392.0, + "step": 24224 + }, + { + "epoch": 2.660333845815946, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0700626373291016, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7185387015342712, + "num_tokens": 626687978.0, + "step": 24225 + }, + { + "epoch": 2.660443663518559, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4170761108398438, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7325046062469482, + "num_tokens": 626714453.0, + "step": 24226 + }, + { + "epoch": 2.660553481221173, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6119487285614014, + "learning_rate": 1e-06, + "loss": 0.8477, + "mean_token_accuracy": 0.7354556322097778, + "num_tokens": 626738129.0, + "step": 24227 + }, + { + "epoch": 2.6606632989237866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.512333869934082, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7334668636322021, + "num_tokens": 626762084.0, + "step": 24228 + }, + { + "epoch": 2.6607731166264, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8446240425109863, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7307894229888916, + "num_tokens": 626786533.0, + "step": 24229 + }, + { + "epoch": 2.6608829343290137, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.618629217147827, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7359741926193237, + "num_tokens": 626811012.0, + "step": 24230 + }, + { + "epoch": 2.6609927520316274, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.364682912826538, + "learning_rate": 1e-06, + "loss": 0.7659, + "mean_token_accuracy": 0.7547298669815063, + "num_tokens": 626835039.0, + "step": 24231 + }, + { + "epoch": 2.661102569734241, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3548455238342285, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7312784194946289, + "num_tokens": 626859162.0, + "step": 24232 + }, + { + "epoch": 2.661212387436855, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2175493240356445, + "learning_rate": 1e-06, + "loss": 0.9293, + "mean_token_accuracy": 0.7095108032226562, + "num_tokens": 626890271.0, + "step": 24233 + }, + { + "epoch": 2.6613222051394683, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.510176420211792, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7253779172897339, + "num_tokens": 626916816.0, + "step": 24234 + }, + { + "epoch": 2.661432022842082, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0866212844848633, + "learning_rate": 1e-06, + "loss": 0.9263, + "mean_token_accuracy": 0.7135704755783081, + "num_tokens": 626948544.0, + "step": 24235 + }, + { + "epoch": 2.661541840544696, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6695499420166016, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7187323570251465, + "num_tokens": 626969038.0, + "step": 24236 + }, + { + "epoch": 2.6616516582473095, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.106186866760254, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7219709157943726, + "num_tokens": 627001751.0, + "step": 24237 + }, + { + "epoch": 2.6617614759499233, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.460665464401245, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7288630604743958, + "num_tokens": 627025947.0, + "step": 24238 + }, + { + "epoch": 2.6618712936525366, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.74391770362854, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7375046014785767, + "num_tokens": 627045642.0, + "step": 24239 + }, + { + "epoch": 2.6619811113551504, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.199087381362915, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7273597121238708, + "num_tokens": 627072345.0, + "step": 24240 + }, + { + "epoch": 2.662090929057764, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2893857955932617, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7373603582382202, + "num_tokens": 627100585.0, + "step": 24241 + }, + { + "epoch": 2.662200746760378, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.321643590927124, + "learning_rate": 1e-06, + "loss": 0.9339, + "mean_token_accuracy": 0.7158876657485962, + "num_tokens": 627126318.0, + "step": 24242 + }, + { + "epoch": 2.6623105644629916, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.269818067550659, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7150185108184814, + "num_tokens": 627153684.0, + "step": 24243 + }, + { + "epoch": 2.662420382165605, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.24072003364563, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7280245423316956, + "num_tokens": 627182507.0, + "step": 24244 + }, + { + "epoch": 2.6625301998682187, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.22745418548584, + "learning_rate": 1e-06, + "loss": 0.7058, + "mean_token_accuracy": 0.7720133066177368, + "num_tokens": 627210042.0, + "step": 24245 + }, + { + "epoch": 2.6626400175708325, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2467117309570312, + "learning_rate": 1e-06, + "loss": 0.7422, + "mean_token_accuracy": 0.7609213590621948, + "num_tokens": 627235725.0, + "step": 24246 + }, + { + "epoch": 2.662749835273446, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1999740600585938, + "learning_rate": 1e-06, + "loss": 0.8896, + "mean_token_accuracy": 0.7188352346420288, + "num_tokens": 627264374.0, + "step": 24247 + }, + { + "epoch": 2.66285965297606, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1665093898773193, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7156809568405151, + "num_tokens": 627292718.0, + "step": 24248 + }, + { + "epoch": 2.6629694706786733, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.187767505645752, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7043790221214294, + "num_tokens": 627323275.0, + "step": 24249 + }, + { + "epoch": 2.663079288381287, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.533891201019287, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7466920614242554, + "num_tokens": 627347055.0, + "step": 24250 + }, + { + "epoch": 2.663189106083901, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2664644718170166, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7185141444206238, + "num_tokens": 627375292.0, + "step": 24251 + }, + { + "epoch": 2.663298923786514, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.301912546157837, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7107514142990112, + "num_tokens": 627404543.0, + "step": 24252 + }, + { + "epoch": 2.6634087414891283, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1624839305877686, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7210420370101929, + "num_tokens": 627435946.0, + "step": 24253 + }, + { + "epoch": 2.6635185591917416, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.34138822555542, + "learning_rate": 1e-06, + "loss": 0.8831, + "mean_token_accuracy": 0.7202466726303101, + "num_tokens": 627463374.0, + "step": 24254 + }, + { + "epoch": 2.6636283768943554, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6114046573638916, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7337382435798645, + "num_tokens": 627486138.0, + "step": 24255 + }, + { + "epoch": 2.663738194596969, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.295621156692505, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7314852476119995, + "num_tokens": 627512833.0, + "step": 24256 + }, + { + "epoch": 2.6638480122995825, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.041935443878174, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7164753079414368, + "num_tokens": 627545399.0, + "step": 24257 + }, + { + "epoch": 2.663957830002196, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2691650390625, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7214868664741516, + "num_tokens": 627571910.0, + "step": 24258 + }, + { + "epoch": 2.66406764770481, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.363210439682007, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7217596173286438, + "num_tokens": 627600017.0, + "step": 24259 + }, + { + "epoch": 2.6641774654074237, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1739845275878906, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7267066836357117, + "num_tokens": 627630488.0, + "step": 24260 + }, + { + "epoch": 2.6642872831100375, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.487743854522705, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7352216243743896, + "num_tokens": 627654978.0, + "step": 24261 + }, + { + "epoch": 2.664397100812651, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.663137197494507, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7232999205589294, + "num_tokens": 627676512.0, + "step": 24262 + }, + { + "epoch": 2.6645069185152646, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.464360475540161, + "learning_rate": 1e-06, + "loss": 0.9332, + "mean_token_accuracy": 0.709581732749939, + "num_tokens": 627702865.0, + "step": 24263 + }, + { + "epoch": 2.6646167362178783, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4018900394439697, + "learning_rate": 1e-06, + "loss": 0.8327, + "mean_token_accuracy": 0.737328290939331, + "num_tokens": 627729759.0, + "step": 24264 + }, + { + "epoch": 2.664726553920492, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0940279960632324, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7155271172523499, + "num_tokens": 627761297.0, + "step": 24265 + }, + { + "epoch": 2.664836371623106, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.573843002319336, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.7406156063079834, + "num_tokens": 627783809.0, + "step": 24266 + }, + { + "epoch": 2.664946189325719, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.465155839920044, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7396484613418579, + "num_tokens": 627808469.0, + "step": 24267 + }, + { + "epoch": 2.665056007028333, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3356199264526367, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7465792894363403, + "num_tokens": 627832696.0, + "step": 24268 + }, + { + "epoch": 2.6651658247309467, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.425506591796875, + "learning_rate": 1e-06, + "loss": 0.7685, + "mean_token_accuracy": 0.7470577359199524, + "num_tokens": 627857509.0, + "step": 24269 + }, + { + "epoch": 2.6652756424335604, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2764875888824463, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.7539867162704468, + "num_tokens": 627883020.0, + "step": 24270 + }, + { + "epoch": 2.665385460136174, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.386484146118164, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7288845777511597, + "num_tokens": 627907355.0, + "step": 24271 + }, + { + "epoch": 2.6654952778387875, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.490429162979126, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7334877252578735, + "num_tokens": 627934318.0, + "step": 24272 + }, + { + "epoch": 2.6656050955414012, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.353870153427124, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.740706205368042, + "num_tokens": 627962342.0, + "step": 24273 + }, + { + "epoch": 2.665714913244015, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4164438247680664, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.709155797958374, + "num_tokens": 627989770.0, + "step": 24274 + }, + { + "epoch": 2.6658247309466288, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6359660625457764, + "learning_rate": 1e-06, + "loss": 0.7178, + "mean_token_accuracy": 0.7613648176193237, + "num_tokens": 628009019.0, + "step": 24275 + }, + { + "epoch": 2.6659345486492425, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.035351276397705, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7021695375442505, + "num_tokens": 628044962.0, + "step": 24276 + }, + { + "epoch": 2.666044366351856, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4350342750549316, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7434119582176208, + "num_tokens": 628069155.0, + "step": 24277 + }, + { + "epoch": 2.6661541840544696, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4598915576934814, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7261893153190613, + "num_tokens": 628094640.0, + "step": 24278 + }, + { + "epoch": 2.6662640017570833, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4166462421417236, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.720733642578125, + "num_tokens": 628118746.0, + "step": 24279 + }, + { + "epoch": 2.6663738194596966, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2383673191070557, + "learning_rate": 1e-06, + "loss": 0.7404, + "mean_token_accuracy": 0.7610321640968323, + "num_tokens": 628143374.0, + "step": 24280 + }, + { + "epoch": 2.6664836371623104, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0763797760009766, + "learning_rate": 1e-06, + "loss": 0.7968, + "mean_token_accuracy": 0.7436776161193848, + "num_tokens": 628175194.0, + "step": 24281 + }, + { + "epoch": 2.666593454864924, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.425229072570801, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7392438650131226, + "num_tokens": 628199375.0, + "step": 24282 + }, + { + "epoch": 2.666703272567538, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3758914470672607, + "learning_rate": 1e-06, + "loss": 0.8231, + "mean_token_accuracy": 0.7386219501495361, + "num_tokens": 628224584.0, + "step": 24283 + }, + { + "epoch": 2.6668130902701517, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5649056434631348, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.723694920539856, + "num_tokens": 628246848.0, + "step": 24284 + }, + { + "epoch": 2.666922907972765, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3496339321136475, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7030308246612549, + "num_tokens": 628274470.0, + "step": 24285 + }, + { + "epoch": 2.6670327256753787, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3791470527648926, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7495086193084717, + "num_tokens": 628299899.0, + "step": 24286 + }, + { + "epoch": 2.6671425433779925, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4298689365386963, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7144134044647217, + "num_tokens": 628323819.0, + "step": 24287 + }, + { + "epoch": 2.6672523610806063, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4989733695983887, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.736921489238739, + "num_tokens": 628346219.0, + "step": 24288 + }, + { + "epoch": 2.66736217878322, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.384493589401245, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7322227954864502, + "num_tokens": 628370534.0, + "step": 24289 + }, + { + "epoch": 2.6674719964858333, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3120477199554443, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7214662432670593, + "num_tokens": 628397247.0, + "step": 24290 + }, + { + "epoch": 2.667581814188447, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2086479663848877, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7343577146530151, + "num_tokens": 628424134.0, + "step": 24291 + }, + { + "epoch": 2.667691631891061, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2764861583709717, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7515050172805786, + "num_tokens": 628451886.0, + "step": 24292 + }, + { + "epoch": 2.6678014495936746, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.381382942199707, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7369861006736755, + "num_tokens": 628475401.0, + "step": 24293 + }, + { + "epoch": 2.6679112672962884, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.579497814178467, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7314547300338745, + "num_tokens": 628496463.0, + "step": 24294 + }, + { + "epoch": 2.6680210849989017, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6934008598327637, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.743615448474884, + "num_tokens": 628518543.0, + "step": 24295 + }, + { + "epoch": 2.6681309027015154, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2981159687042236, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7266879081726074, + "num_tokens": 628546183.0, + "step": 24296 + }, + { + "epoch": 2.668240720404129, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.469895124435425, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7351542711257935, + "num_tokens": 628569383.0, + "step": 24297 + }, + { + "epoch": 2.668350538106743, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.423224925994873, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7433286905288696, + "num_tokens": 628592940.0, + "step": 24298 + }, + { + "epoch": 2.6684603558093567, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3195605278015137, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7248023152351379, + "num_tokens": 628620055.0, + "step": 24299 + }, + { + "epoch": 2.66857017351197, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2994930744171143, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7360390424728394, + "num_tokens": 628645968.0, + "step": 24300 + }, + { + "epoch": 2.6686799912145838, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5153565406799316, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7283909320831299, + "num_tokens": 628667921.0, + "step": 24301 + }, + { + "epoch": 2.6687898089171975, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.438901901245117, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7409676313400269, + "num_tokens": 628693110.0, + "step": 24302 + }, + { + "epoch": 2.668899626619811, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.479444742202759, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.721967875957489, + "num_tokens": 628716271.0, + "step": 24303 + }, + { + "epoch": 2.669009444322425, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.35347056388855, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7255734801292419, + "num_tokens": 628741671.0, + "step": 24304 + }, + { + "epoch": 2.6691192620250384, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3354616165161133, + "learning_rate": 1e-06, + "loss": 0.9388, + "mean_token_accuracy": 0.7123804688453674, + "num_tokens": 628768746.0, + "step": 24305 + }, + { + "epoch": 2.669229079727652, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.505936622619629, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7518786191940308, + "num_tokens": 628792673.0, + "step": 24306 + }, + { + "epoch": 2.669338897430266, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3162922859191895, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7203696370124817, + "num_tokens": 628817965.0, + "step": 24307 + }, + { + "epoch": 2.669448715132879, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0200889110565186, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7053737640380859, + "num_tokens": 628852368.0, + "step": 24308 + }, + { + "epoch": 2.669558532835493, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6497952938079834, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7351081967353821, + "num_tokens": 628876487.0, + "step": 24309 + }, + { + "epoch": 2.6696683505381067, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2175421714782715, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7291408777236938, + "num_tokens": 628904968.0, + "step": 24310 + }, + { + "epoch": 2.6697781682407205, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.068756580352783, + "learning_rate": 1e-06, + "loss": 0.9606, + "mean_token_accuracy": 0.7051876783370972, + "num_tokens": 628936609.0, + "step": 24311 + }, + { + "epoch": 2.669887985943334, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.611398220062256, + "learning_rate": 1e-06, + "loss": 0.7153, + "mean_token_accuracy": 0.7622358798980713, + "num_tokens": 628956272.0, + "step": 24312 + }, + { + "epoch": 2.6699978036459475, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.540966749191284, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7358191013336182, + "num_tokens": 628978794.0, + "step": 24313 + }, + { + "epoch": 2.6701076213485613, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5375568866729736, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7221161127090454, + "num_tokens": 629004138.0, + "step": 24314 + }, + { + "epoch": 2.670217439051175, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4085965156555176, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.7270008325576782, + "num_tokens": 629030168.0, + "step": 24315 + }, + { + "epoch": 2.670327256753789, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.311126947402954, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7282572984695435, + "num_tokens": 629057981.0, + "step": 24316 + }, + { + "epoch": 2.6704370744564025, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1994595527648926, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7379164099693298, + "num_tokens": 629086282.0, + "step": 24317 + }, + { + "epoch": 2.670546892159016, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4677834510803223, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7390165328979492, + "num_tokens": 629111573.0, + "step": 24318 + }, + { + "epoch": 2.6706567098616296, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5508174896240234, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7338415384292603, + "num_tokens": 629134667.0, + "step": 24319 + }, + { + "epoch": 2.6707665275642434, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0989127159118652, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.7180908918380737, + "num_tokens": 629169421.0, + "step": 24320 + }, + { + "epoch": 2.670876345266857, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.669389009475708, + "learning_rate": 1e-06, + "loss": 0.8071, + "mean_token_accuracy": 0.7443325519561768, + "num_tokens": 629190802.0, + "step": 24321 + }, + { + "epoch": 2.670986162969471, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.51235294342041, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7266520261764526, + "num_tokens": 629214455.0, + "step": 24322 + }, + { + "epoch": 2.671095980672084, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.344461441040039, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.7406103014945984, + "num_tokens": 629239574.0, + "step": 24323 + }, + { + "epoch": 2.671205798374698, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.27341890335083, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7133763432502747, + "num_tokens": 629266362.0, + "step": 24324 + }, + { + "epoch": 2.6713156160773117, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3964219093322754, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7344004511833191, + "num_tokens": 629290907.0, + "step": 24325 + }, + { + "epoch": 2.6714254337799255, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.352848768234253, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.734424352645874, + "num_tokens": 629317945.0, + "step": 24326 + }, + { + "epoch": 2.6715352514825392, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.558772563934326, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7349789142608643, + "num_tokens": 629340176.0, + "step": 24327 + }, + { + "epoch": 2.6716450691851525, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3575599193573, + "learning_rate": 1e-06, + "loss": 0.9576, + "mean_token_accuracy": 0.6986883282661438, + "num_tokens": 629367833.0, + "step": 24328 + }, + { + "epoch": 2.6717548868877663, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.34765887260437, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7225615978240967, + "num_tokens": 629395065.0, + "step": 24329 + }, + { + "epoch": 2.67186470459038, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.424163579940796, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7349563837051392, + "num_tokens": 629420042.0, + "step": 24330 + }, + { + "epoch": 2.6719745222929934, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9037201404571533, + "learning_rate": 1e-06, + "loss": 0.7363, + "mean_token_accuracy": 0.759782075881958, + "num_tokens": 629438858.0, + "step": 24331 + }, + { + "epoch": 2.672084339995607, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.34458589553833, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7378786206245422, + "num_tokens": 629463482.0, + "step": 24332 + }, + { + "epoch": 2.672194157698221, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.411569356918335, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7237133979797363, + "num_tokens": 629489477.0, + "step": 24333 + }, + { + "epoch": 2.6723039754008346, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1395339965820312, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7481702566146851, + "num_tokens": 629516528.0, + "step": 24334 + }, + { + "epoch": 2.6724137931034484, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.529052734375, + "learning_rate": 1e-06, + "loss": 0.7539, + "mean_token_accuracy": 0.752166748046875, + "num_tokens": 629538327.0, + "step": 24335 + }, + { + "epoch": 2.6725236108060617, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.79902720451355, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7401324510574341, + "num_tokens": 629558109.0, + "step": 24336 + }, + { + "epoch": 2.6726334285086755, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5199782848358154, + "learning_rate": 1e-06, + "loss": 0.7795, + "mean_token_accuracy": 0.7509223222732544, + "num_tokens": 629581221.0, + "step": 24337 + }, + { + "epoch": 2.6727432462112892, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.271838665008545, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7201079726219177, + "num_tokens": 629610989.0, + "step": 24338 + }, + { + "epoch": 2.672853063913903, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2946629524230957, + "learning_rate": 1e-06, + "loss": 0.9452, + "mean_token_accuracy": 0.6985210180282593, + "num_tokens": 629639140.0, + "step": 24339 + }, + { + "epoch": 2.6729628816165167, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8840889930725098, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7452457547187805, + "num_tokens": 629658917.0, + "step": 24340 + }, + { + "epoch": 2.67307269931913, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.535057783126831, + "learning_rate": 1e-06, + "loss": 0.8166, + "mean_token_accuracy": 0.7417166233062744, + "num_tokens": 629683062.0, + "step": 24341 + }, + { + "epoch": 2.673182517021744, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1924855709075928, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7064357995986938, + "num_tokens": 629717786.0, + "step": 24342 + }, + { + "epoch": 2.6732923347243576, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.634232759475708, + "learning_rate": 1e-06, + "loss": 0.7764, + "mean_token_accuracy": 0.7434741854667664, + "num_tokens": 629738750.0, + "step": 24343 + }, + { + "epoch": 2.6734021524269713, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7164816856384277, + "learning_rate": 1e-06, + "loss": 0.9, + "mean_token_accuracy": 0.7200295329093933, + "num_tokens": 629759552.0, + "step": 24344 + }, + { + "epoch": 2.673511970129585, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.345109701156616, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7438755035400391, + "num_tokens": 629787628.0, + "step": 24345 + }, + { + "epoch": 2.6736217878321984, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3282363414764404, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7369437217712402, + "num_tokens": 629812867.0, + "step": 24346 + }, + { + "epoch": 2.673731605534812, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.350003719329834, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7235757112503052, + "num_tokens": 629839150.0, + "step": 24347 + }, + { + "epoch": 2.673841423237426, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.369472026824951, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7274283170700073, + "num_tokens": 629863973.0, + "step": 24348 + }, + { + "epoch": 2.6739512409400397, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.459306001663208, + "learning_rate": 1e-06, + "loss": 0.8438, + "mean_token_accuracy": 0.7297755479812622, + "num_tokens": 629888021.0, + "step": 24349 + }, + { + "epoch": 2.6740610586426534, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0969271659851074, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7100629210472107, + "num_tokens": 629920231.0, + "step": 24350 + }, + { + "epoch": 2.6741708763452667, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2403945922851562, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7277376651763916, + "num_tokens": 629947745.0, + "step": 24351 + }, + { + "epoch": 2.6742806940478805, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.633784532546997, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7323020696640015, + "num_tokens": 629969679.0, + "step": 24352 + }, + { + "epoch": 2.6743905117504942, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2722575664520264, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7108510732650757, + "num_tokens": 630001784.0, + "step": 24353 + }, + { + "epoch": 2.674500329453108, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.236339807510376, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.714468240737915, + "num_tokens": 630033023.0, + "step": 24354 + }, + { + "epoch": 2.6746101471557218, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.635018825531006, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7237025499343872, + "num_tokens": 630059274.0, + "step": 24355 + }, + { + "epoch": 2.674719964858335, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.442445755004883, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7208170294761658, + "num_tokens": 630085439.0, + "step": 24356 + }, + { + "epoch": 2.674829782560949, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3482916355133057, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7162725329399109, + "num_tokens": 630111706.0, + "step": 24357 + }, + { + "epoch": 2.6749396002635626, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2307839393615723, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.727215051651001, + "num_tokens": 630140800.0, + "step": 24358 + }, + { + "epoch": 2.675049417966176, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.176685094833374, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7201603055000305, + "num_tokens": 630170430.0, + "step": 24359 + }, + { + "epoch": 2.6751592356687897, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4856159687042236, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7430367469787598, + "num_tokens": 630192434.0, + "step": 24360 + }, + { + "epoch": 2.6752690533714034, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3990588188171387, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.73808354139328, + "num_tokens": 630216439.0, + "step": 24361 + }, + { + "epoch": 2.675378871074017, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.225551128387451, + "learning_rate": 1e-06, + "loss": 0.9128, + "mean_token_accuracy": 0.7140928506851196, + "num_tokens": 630244907.0, + "step": 24362 + }, + { + "epoch": 2.675488688776631, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2224719524383545, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7308365106582642, + "num_tokens": 630273273.0, + "step": 24363 + }, + { + "epoch": 2.6755985064792442, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3287932872772217, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.728044867515564, + "num_tokens": 630299721.0, + "step": 24364 + }, + { + "epoch": 2.675708324181858, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3872716426849365, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.723232626914978, + "num_tokens": 630325033.0, + "step": 24365 + }, + { + "epoch": 2.6758181418844718, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3835010528564453, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.723670482635498, + "num_tokens": 630351318.0, + "step": 24366 + }, + { + "epoch": 2.6759279595870855, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2978672981262207, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7235596179962158, + "num_tokens": 630378922.0, + "step": 24367 + }, + { + "epoch": 2.6760377772896993, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1511282920837402, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7163190841674805, + "num_tokens": 630408757.0, + "step": 24368 + }, + { + "epoch": 2.6761475949923126, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3770341873168945, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.743933916091919, + "num_tokens": 630435441.0, + "step": 24369 + }, + { + "epoch": 2.6762574126949263, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3225514888763428, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7147467136383057, + "num_tokens": 630462981.0, + "step": 24370 + }, + { + "epoch": 2.67636723039754, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.589838981628418, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7504757642745972, + "num_tokens": 630484003.0, + "step": 24371 + }, + { + "epoch": 2.676477048100154, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.431629180908203, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7279208302497864, + "num_tokens": 630510191.0, + "step": 24372 + }, + { + "epoch": 2.6765868658027676, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.59800124168396, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7362162470817566, + "num_tokens": 630532773.0, + "step": 24373 + }, + { + "epoch": 2.676696683505381, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.402172565460205, + "learning_rate": 1e-06, + "loss": 0.7861, + "mean_token_accuracy": 0.7430056929588318, + "num_tokens": 630557174.0, + "step": 24374 + }, + { + "epoch": 2.6768065012079947, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4775795936584473, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7423741221427917, + "num_tokens": 630580489.0, + "step": 24375 + }, + { + "epoch": 2.6769163189106084, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.221343755722046, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7287330031394958, + "num_tokens": 630609870.0, + "step": 24376 + }, + { + "epoch": 2.677026136613222, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.622056722640991, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7310183644294739, + "num_tokens": 630630976.0, + "step": 24377 + }, + { + "epoch": 2.677135954315836, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.720655918121338, + "learning_rate": 1e-06, + "loss": 0.7754, + "mean_token_accuracy": 0.7461950182914734, + "num_tokens": 630651152.0, + "step": 24378 + }, + { + "epoch": 2.6772457720184493, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2683486938476562, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.710913896560669, + "num_tokens": 630679285.0, + "step": 24379 + }, + { + "epoch": 2.677355589721063, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2762703895568848, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7161144018173218, + "num_tokens": 630706399.0, + "step": 24380 + }, + { + "epoch": 2.677465407423677, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2313756942749023, + "learning_rate": 1e-06, + "loss": 0.7104, + "mean_token_accuracy": 0.7667919993400574, + "num_tokens": 630732261.0, + "step": 24381 + }, + { + "epoch": 2.67757522512629, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4555726051330566, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7427099943161011, + "num_tokens": 630755814.0, + "step": 24382 + }, + { + "epoch": 2.6776850428289043, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3179914951324463, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7206756472587585, + "num_tokens": 630782783.0, + "step": 24383 + }, + { + "epoch": 2.6777948605315176, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.045285940170288, + "learning_rate": 1e-06, + "loss": 0.9133, + "mean_token_accuracy": 0.7097401022911072, + "num_tokens": 630815627.0, + "step": 24384 + }, + { + "epoch": 2.6779046782341314, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 3.040170907974243, + "learning_rate": 1e-06, + "loss": 0.7684, + "mean_token_accuracy": 0.7521579265594482, + "num_tokens": 630832948.0, + "step": 24385 + }, + { + "epoch": 2.678014495936745, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.323406934738159, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7375266551971436, + "num_tokens": 630860164.0, + "step": 24386 + }, + { + "epoch": 2.6781243136393584, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.306614637374878, + "learning_rate": 1e-06, + "loss": 0.8836, + "mean_token_accuracy": 0.7262129187583923, + "num_tokens": 630889955.0, + "step": 24387 + }, + { + "epoch": 2.678234131341972, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2798709869384766, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7177239656448364, + "num_tokens": 630917747.0, + "step": 24388 + }, + { + "epoch": 2.678343949044586, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2319459915161133, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7195874452590942, + "num_tokens": 630945446.0, + "step": 24389 + }, + { + "epoch": 2.6784537667471997, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.335843801498413, + "learning_rate": 1e-06, + "loss": 0.9405, + "mean_token_accuracy": 0.7070621252059937, + "num_tokens": 630973759.0, + "step": 24390 + }, + { + "epoch": 2.6785635844498135, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.31300687789917, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.720091700553894, + "num_tokens": 631001578.0, + "step": 24391 + }, + { + "epoch": 2.6786734021524268, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.052234649658203, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7291423082351685, + "num_tokens": 631032517.0, + "step": 24392 + }, + { + "epoch": 2.6787832198550405, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6661856174468994, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7297523021697998, + "num_tokens": 631054209.0, + "step": 24393 + }, + { + "epoch": 2.6788930375576543, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9236373901367188, + "learning_rate": 1e-06, + "loss": 0.7668, + "mean_token_accuracy": 0.7533356547355652, + "num_tokens": 631073210.0, + "step": 24394 + }, + { + "epoch": 2.679002855260268, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3064932823181152, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7350909113883972, + "num_tokens": 631099917.0, + "step": 24395 + }, + { + "epoch": 2.679112672962882, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3520240783691406, + "learning_rate": 1e-06, + "loss": 0.9656, + "mean_token_accuracy": 0.7017039060592651, + "num_tokens": 631128902.0, + "step": 24396 + }, + { + "epoch": 2.679222490665495, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.230532169342041, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7083622217178345, + "num_tokens": 631157803.0, + "step": 24397 + }, + { + "epoch": 2.679332308368109, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.714447259902954, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7393278479576111, + "num_tokens": 631177743.0, + "step": 24398 + }, + { + "epoch": 2.6794421260707226, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3882100582122803, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7451761960983276, + "num_tokens": 631201585.0, + "step": 24399 + }, + { + "epoch": 2.6795519437733364, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9704618453979492, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7398947477340698, + "num_tokens": 631235262.0, + "step": 24400 + }, + { + "epoch": 2.67966176147595, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.617619514465332, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7234400510787964, + "num_tokens": 631257057.0, + "step": 24401 + }, + { + "epoch": 2.6797715791785635, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0413999557495117, + "learning_rate": 1e-06, + "loss": 0.9991, + "mean_token_accuracy": 0.7141075730323792, + "num_tokens": 631287863.0, + "step": 24402 + }, + { + "epoch": 2.679881396881177, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4218978881835938, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7253744602203369, + "num_tokens": 631312985.0, + "step": 24403 + }, + { + "epoch": 2.679991214583791, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.544538974761963, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7314255833625793, + "num_tokens": 631337562.0, + "step": 24404 + }, + { + "epoch": 2.6801010322864047, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.765547513961792, + "learning_rate": 1e-06, + "loss": 0.7742, + "mean_token_accuracy": 0.7455146908760071, + "num_tokens": 631357719.0, + "step": 24405 + }, + { + "epoch": 2.6802108499890185, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0919110774993896, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7283951640129089, + "num_tokens": 631389130.0, + "step": 24406 + }, + { + "epoch": 2.680320667691632, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.644418239593506, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7371827960014343, + "num_tokens": 631410423.0, + "step": 24407 + }, + { + "epoch": 2.6804304853942456, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.220270872116089, + "learning_rate": 1e-06, + "loss": 0.7734, + "mean_token_accuracy": 0.7501400709152222, + "num_tokens": 631437968.0, + "step": 24408 + }, + { + "epoch": 2.6805403030968593, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2386457920074463, + "learning_rate": 1e-06, + "loss": 0.7277, + "mean_token_accuracy": 0.7611589431762695, + "num_tokens": 631463678.0, + "step": 24409 + }, + { + "epoch": 2.6806501207994726, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1729767322540283, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7355951070785522, + "num_tokens": 631494062.0, + "step": 24410 + }, + { + "epoch": 2.6807599385020864, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4223575592041016, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7355284094810486, + "num_tokens": 631519307.0, + "step": 24411 + }, + { + "epoch": 2.6808697562047, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.561687707901001, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7424442768096924, + "num_tokens": 631542890.0, + "step": 24412 + }, + { + "epoch": 2.680979573907314, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.08807635307312, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.7021470665931702, + "num_tokens": 631577584.0, + "step": 24413 + }, + { + "epoch": 2.6810893916099277, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3697617053985596, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7203383445739746, + "num_tokens": 631602628.0, + "step": 24414 + }, + { + "epoch": 2.681199209312541, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5104057788848877, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7423789501190186, + "num_tokens": 631625565.0, + "step": 24415 + }, + { + "epoch": 2.6813090270151547, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.558128833770752, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7395267486572266, + "num_tokens": 631648466.0, + "step": 24416 + }, + { + "epoch": 2.6814188447177685, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3758015632629395, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7231596112251282, + "num_tokens": 631672970.0, + "step": 24417 + }, + { + "epoch": 2.6815286624203822, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2865731716156006, + "learning_rate": 1e-06, + "loss": 0.8796, + "mean_token_accuracy": 0.7178646922111511, + "num_tokens": 631700133.0, + "step": 24418 + }, + { + "epoch": 2.681638480122996, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3642749786376953, + "learning_rate": 1e-06, + "loss": 0.9404, + "mean_token_accuracy": 0.7053071856498718, + "num_tokens": 631726860.0, + "step": 24419 + }, + { + "epoch": 2.6817482978256093, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.132636785507202, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7260633707046509, + "num_tokens": 631754956.0, + "step": 24420 + }, + { + "epoch": 2.681858115528223, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.549511194229126, + "learning_rate": 1e-06, + "loss": 0.7355, + "mean_token_accuracy": 0.758399486541748, + "num_tokens": 631775773.0, + "step": 24421 + }, + { + "epoch": 2.681967933230837, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.530402898788452, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7382464408874512, + "num_tokens": 631798551.0, + "step": 24422 + }, + { + "epoch": 2.6820777509334506, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.246067523956299, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7295990586280823, + "num_tokens": 631826646.0, + "step": 24423 + }, + { + "epoch": 2.6821875686360643, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.956605076789856, + "learning_rate": 1e-06, + "loss": 0.7705, + "mean_token_accuracy": 0.7583061456680298, + "num_tokens": 631858987.0, + "step": 24424 + }, + { + "epoch": 2.6822973863386776, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.436249256134033, + "learning_rate": 1e-06, + "loss": 0.8627, + "mean_token_accuracy": 0.7269124984741211, + "num_tokens": 631882069.0, + "step": 24425 + }, + { + "epoch": 2.6824072040412914, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.217907190322876, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7289551496505737, + "num_tokens": 631911029.0, + "step": 24426 + }, + { + "epoch": 2.682517021743905, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4258649349212646, + "learning_rate": 1e-06, + "loss": 0.9737, + "mean_token_accuracy": 0.7011747360229492, + "num_tokens": 631937513.0, + "step": 24427 + }, + { + "epoch": 2.682626839446519, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1946604251861572, + "learning_rate": 1e-06, + "loss": 0.7674, + "mean_token_accuracy": 0.7465658187866211, + "num_tokens": 631964628.0, + "step": 24428 + }, + { + "epoch": 2.6827366571491327, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6825180053710938, + "learning_rate": 1e-06, + "loss": 0.8626, + "mean_token_accuracy": 0.7268593907356262, + "num_tokens": 631987694.0, + "step": 24429 + }, + { + "epoch": 2.682846474851746, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3210670948028564, + "learning_rate": 1e-06, + "loss": 0.9209, + "mean_token_accuracy": 0.7148937582969666, + "num_tokens": 632018211.0, + "step": 24430 + }, + { + "epoch": 2.6829562925543597, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2507712841033936, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.7214736938476562, + "num_tokens": 632047901.0, + "step": 24431 + }, + { + "epoch": 2.6830661102569735, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6279776096343994, + "learning_rate": 1e-06, + "loss": 0.7335, + "mean_token_accuracy": 0.7587010860443115, + "num_tokens": 632068329.0, + "step": 24432 + }, + { + "epoch": 2.683175927959587, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.395125150680542, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7232065796852112, + "num_tokens": 632094069.0, + "step": 24433 + }, + { + "epoch": 2.683285745662201, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1670563220977783, + "learning_rate": 1e-06, + "loss": 0.853, + "mean_token_accuracy": 0.7287235260009766, + "num_tokens": 632123887.0, + "step": 24434 + }, + { + "epoch": 2.6833955633648143, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.288684844970703, + "learning_rate": 1e-06, + "loss": 0.9969, + "mean_token_accuracy": 0.6974248886108398, + "num_tokens": 632151927.0, + "step": 24435 + }, + { + "epoch": 2.683505381067428, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7479472160339355, + "learning_rate": 1e-06, + "loss": 0.8167, + "mean_token_accuracy": 0.7423275709152222, + "num_tokens": 632172273.0, + "step": 24436 + }, + { + "epoch": 2.683615198770042, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.338911533355713, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7241921424865723, + "num_tokens": 632200268.0, + "step": 24437 + }, + { + "epoch": 2.683725016472655, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1308836936950684, + "learning_rate": 1e-06, + "loss": 0.9099, + "mean_token_accuracy": 0.7133634090423584, + "num_tokens": 632232815.0, + "step": 24438 + }, + { + "epoch": 2.683834834175269, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.417522430419922, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7310904264450073, + "num_tokens": 632256901.0, + "step": 24439 + }, + { + "epoch": 2.6839446518778827, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.32517671585083, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7296760678291321, + "num_tokens": 632282897.0, + "step": 24440 + }, + { + "epoch": 2.6840544695804964, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5410783290863037, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7333036661148071, + "num_tokens": 632307063.0, + "step": 24441 + }, + { + "epoch": 2.68416428728311, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2007877826690674, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7326710820198059, + "num_tokens": 632336231.0, + "step": 24442 + }, + { + "epoch": 2.6842741049857235, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1976516246795654, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7370913028717041, + "num_tokens": 632365393.0, + "step": 24443 + }, + { + "epoch": 2.6843839226883373, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3603477478027344, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7045061588287354, + "num_tokens": 632392173.0, + "step": 24444 + }, + { + "epoch": 2.684493740390951, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.371208429336548, + "learning_rate": 1e-06, + "loss": 0.9173, + "mean_token_accuracy": 0.7162042856216431, + "num_tokens": 632419194.0, + "step": 24445 + }, + { + "epoch": 2.6846035580935648, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0424435138702393, + "learning_rate": 1e-06, + "loss": 0.8501, + "mean_token_accuracy": 0.7277466654777527, + "num_tokens": 632452581.0, + "step": 24446 + }, + { + "epoch": 2.6847133757961785, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0482492446899414, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7076996564865112, + "num_tokens": 632485790.0, + "step": 24447 + }, + { + "epoch": 2.684823193498792, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.744419574737549, + "learning_rate": 1e-06, + "loss": 0.7671, + "mean_token_accuracy": 0.7526993751525879, + "num_tokens": 632504942.0, + "step": 24448 + }, + { + "epoch": 2.6849330112014056, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.343831777572632, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7483220100402832, + "num_tokens": 632528163.0, + "step": 24449 + }, + { + "epoch": 2.6850428289040194, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.229659080505371, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7196263074874878, + "num_tokens": 632556386.0, + "step": 24450 + }, + { + "epoch": 2.685152646606633, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6610605716705322, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.736039400100708, + "num_tokens": 632578355.0, + "step": 24451 + }, + { + "epoch": 2.685262464309247, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.372615337371826, + "learning_rate": 1e-06, + "loss": 0.752, + "mean_token_accuracy": 0.7571916580200195, + "num_tokens": 632603086.0, + "step": 24452 + }, + { + "epoch": 2.68537228201186, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3150486946105957, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7333462238311768, + "num_tokens": 632631197.0, + "step": 24453 + }, + { + "epoch": 2.685482099714474, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.224470853805542, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7215962409973145, + "num_tokens": 632659157.0, + "step": 24454 + }, + { + "epoch": 2.6855919174170877, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4238455295562744, + "learning_rate": 1e-06, + "loss": 0.9193, + "mean_token_accuracy": 0.7103745937347412, + "num_tokens": 632684876.0, + "step": 24455 + }, + { + "epoch": 2.6857017351197014, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7960453033447266, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7491145133972168, + "num_tokens": 632704324.0, + "step": 24456 + }, + { + "epoch": 2.685811552822315, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.181706666946411, + "learning_rate": 1e-06, + "loss": 0.8612, + "mean_token_accuracy": 0.7355367541313171, + "num_tokens": 632732603.0, + "step": 24457 + }, + { + "epoch": 2.6859213705249285, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0526556968688965, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.7227270603179932, + "num_tokens": 632766272.0, + "step": 24458 + }, + { + "epoch": 2.6860311882275423, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.294771194458008, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7282930016517639, + "num_tokens": 632793284.0, + "step": 24459 + }, + { + "epoch": 2.686141005930156, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2300376892089844, + "learning_rate": 1e-06, + "loss": 0.9823, + "mean_token_accuracy": 0.6998435854911804, + "num_tokens": 632822643.0, + "step": 24460 + }, + { + "epoch": 2.6862508236327693, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.366408109664917, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7226372361183167, + "num_tokens": 632848903.0, + "step": 24461 + }, + { + "epoch": 2.686360641335383, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5054681301116943, + "learning_rate": 1e-06, + "loss": 0.7886, + "mean_token_accuracy": 0.7517192959785461, + "num_tokens": 632871012.0, + "step": 24462 + }, + { + "epoch": 2.686470459037997, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5635764598846436, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7429918050765991, + "num_tokens": 632893130.0, + "step": 24463 + }, + { + "epoch": 2.6865802767406106, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.210984230041504, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7135276198387146, + "num_tokens": 632921317.0, + "step": 24464 + }, + { + "epoch": 2.6866900944432244, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2782723903656006, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.7365908622741699, + "num_tokens": 632949586.0, + "step": 24465 + }, + { + "epoch": 2.6867999121458377, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.477977752685547, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7101579308509827, + "num_tokens": 632975352.0, + "step": 24466 + }, + { + "epoch": 2.6869097298484514, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1343274116516113, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7490513920783997, + "num_tokens": 633004096.0, + "step": 24467 + }, + { + "epoch": 2.687019547551065, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.057448148727417, + "learning_rate": 1e-06, + "loss": 0.9377, + "mean_token_accuracy": 0.7094861268997192, + "num_tokens": 633035934.0, + "step": 24468 + }, + { + "epoch": 2.687129365253679, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.088961124420166, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7389257550239563, + "num_tokens": 633066967.0, + "step": 24469 + }, + { + "epoch": 2.6872391829562927, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4754505157470703, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7267218828201294, + "num_tokens": 633091102.0, + "step": 24470 + }, + { + "epoch": 2.687349000658906, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.319415807723999, + "learning_rate": 1e-06, + "loss": 0.7661, + "mean_token_accuracy": 0.7500088214874268, + "num_tokens": 633115293.0, + "step": 24471 + }, + { + "epoch": 2.68745881836152, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5510175228118896, + "learning_rate": 1e-06, + "loss": 0.7969, + "mean_token_accuracy": 0.7395778894424438, + "num_tokens": 633138373.0, + "step": 24472 + }, + { + "epoch": 2.6875686360641335, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.41540789604187, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7382105588912964, + "num_tokens": 633162004.0, + "step": 24473 + }, + { + "epoch": 2.6876784537667473, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.415201425552368, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7384270429611206, + "num_tokens": 633185919.0, + "step": 24474 + }, + { + "epoch": 2.687788271469361, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.251941442489624, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7290680408477783, + "num_tokens": 633214269.0, + "step": 24475 + }, + { + "epoch": 2.6878980891719744, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2641148567199707, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7234313488006592, + "num_tokens": 633242678.0, + "step": 24476 + }, + { + "epoch": 2.688007906874588, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5846052169799805, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7177184224128723, + "num_tokens": 633264820.0, + "step": 24477 + }, + { + "epoch": 2.688117724577202, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9266972541809082, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7148287892341614, + "num_tokens": 633298596.0, + "step": 24478 + }, + { + "epoch": 2.6882275422798156, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.231938362121582, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7409411072731018, + "num_tokens": 633325792.0, + "step": 24479 + }, + { + "epoch": 2.6883373599824294, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5711376667022705, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.738102912902832, + "num_tokens": 633348353.0, + "step": 24480 + }, + { + "epoch": 2.6884471776850427, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.472412109375, + "learning_rate": 1e-06, + "loss": 0.8171, + "mean_token_accuracy": 0.7421079277992249, + "num_tokens": 633371318.0, + "step": 24481 + }, + { + "epoch": 2.6885569953876565, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.100208282470703, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7285506129264832, + "num_tokens": 633400630.0, + "step": 24482 + }, + { + "epoch": 2.6886668130902702, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3581626415252686, + "learning_rate": 1e-06, + "loss": 0.7658, + "mean_token_accuracy": 0.7563364505767822, + "num_tokens": 633425392.0, + "step": 24483 + }, + { + "epoch": 2.6887766307928835, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.533196210861206, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7309940457344055, + "num_tokens": 633450601.0, + "step": 24484 + }, + { + "epoch": 2.6888864484954977, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.393507242202759, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7364459037780762, + "num_tokens": 633475397.0, + "step": 24485 + }, + { + "epoch": 2.688996266198111, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.410557270050049, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7303159236907959, + "num_tokens": 633500874.0, + "step": 24486 + }, + { + "epoch": 2.689106083900725, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2117011547088623, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7137347459793091, + "num_tokens": 633530432.0, + "step": 24487 + }, + { + "epoch": 2.6892159016033386, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.333465337753296, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7395395040512085, + "num_tokens": 633557693.0, + "step": 24488 + }, + { + "epoch": 2.689325719305952, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.271430253982544, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7402383089065552, + "num_tokens": 633584872.0, + "step": 24489 + }, + { + "epoch": 2.6894355370085656, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.76141619682312, + "learning_rate": 1e-06, + "loss": 0.7658, + "mean_token_accuracy": 0.7533084154129028, + "num_tokens": 633605096.0, + "step": 24490 + }, + { + "epoch": 2.6895453547111794, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4964993000030518, + "learning_rate": 1e-06, + "loss": 0.8061, + "mean_token_accuracy": 0.7403382658958435, + "num_tokens": 633628223.0, + "step": 24491 + }, + { + "epoch": 2.689655172413793, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.361013412475586, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7367871999740601, + "num_tokens": 633654837.0, + "step": 24492 + }, + { + "epoch": 2.689764990116407, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5287835597991943, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7256067991256714, + "num_tokens": 633678396.0, + "step": 24493 + }, + { + "epoch": 2.68987480781902, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2833986282348633, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7196076512336731, + "num_tokens": 633709747.0, + "step": 24494 + }, + { + "epoch": 2.689984625521634, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.328380584716797, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7034422159194946, + "num_tokens": 633738224.0, + "step": 24495 + }, + { + "epoch": 2.6900944432242477, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.235507011413574, + "learning_rate": 1e-06, + "loss": 0.9476, + "mean_token_accuracy": 0.7031725645065308, + "num_tokens": 633765948.0, + "step": 24496 + }, + { + "epoch": 2.6902042609268615, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.542177200317383, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.741898775100708, + "num_tokens": 633789191.0, + "step": 24497 + }, + { + "epoch": 2.6903140786294752, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9150002002716064, + "learning_rate": 1e-06, + "loss": 0.7116, + "mean_token_accuracy": 0.7650128602981567, + "num_tokens": 633806814.0, + "step": 24498 + }, + { + "epoch": 2.6904238963320886, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.260418653488159, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7352083325386047, + "num_tokens": 633836908.0, + "step": 24499 + }, + { + "epoch": 2.6905337140347023, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.531515121459961, + "learning_rate": 1e-06, + "loss": 0.7692, + "mean_token_accuracy": 0.7555301189422607, + "num_tokens": 633859780.0, + "step": 24500 + }, + { + "epoch": 2.690643531737316, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.376044511795044, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7062571048736572, + "num_tokens": 633885859.0, + "step": 24501 + }, + { + "epoch": 2.69075334943993, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.525125503540039, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7320624589920044, + "num_tokens": 633909121.0, + "step": 24502 + }, + { + "epoch": 2.6908631671425436, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.241140604019165, + "learning_rate": 1e-06, + "loss": 0.9538, + "mean_token_accuracy": 0.7003454566001892, + "num_tokens": 633938167.0, + "step": 24503 + }, + { + "epoch": 2.690972984845157, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.202773332595825, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.7455710768699646, + "num_tokens": 633965109.0, + "step": 24504 + }, + { + "epoch": 2.6910828025477707, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3829925060272217, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7339526414871216, + "num_tokens": 633988185.0, + "step": 24505 + }, + { + "epoch": 2.6911926202503844, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5365219116210938, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7159189581871033, + "num_tokens": 634012442.0, + "step": 24506 + }, + { + "epoch": 2.691302437952998, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4351584911346436, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7404710650444031, + "num_tokens": 634036849.0, + "step": 24507 + }, + { + "epoch": 2.691412255655612, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4941558837890625, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7183250784873962, + "num_tokens": 634061913.0, + "step": 24508 + }, + { + "epoch": 2.6915220733582252, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.336327075958252, + "learning_rate": 1e-06, + "loss": 0.9313, + "mean_token_accuracy": 0.7060570120811462, + "num_tokens": 634088505.0, + "step": 24509 + }, + { + "epoch": 2.691631891060839, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1552493572235107, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.724635660648346, + "num_tokens": 634119812.0, + "step": 24510 + }, + { + "epoch": 2.6917417087634528, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.358461618423462, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.736011266708374, + "num_tokens": 634144470.0, + "step": 24511 + }, + { + "epoch": 2.691851526466066, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.015578269958496, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7208476066589355, + "num_tokens": 634178419.0, + "step": 24512 + }, + { + "epoch": 2.69196134416868, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.397475004196167, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7207124829292297, + "num_tokens": 634204048.0, + "step": 24513 + }, + { + "epoch": 2.6920711618712936, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.654574394226074, + "learning_rate": 1e-06, + "loss": 0.7927, + "mean_token_accuracy": 0.7547681331634521, + "num_tokens": 634224837.0, + "step": 24514 + }, + { + "epoch": 2.6921809795739073, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.273277759552002, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.7254266738891602, + "num_tokens": 634251640.0, + "step": 24515 + }, + { + "epoch": 2.692290797276521, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1675028800964355, + "learning_rate": 1e-06, + "loss": 1.0142, + "mean_token_accuracy": 0.6853678226470947, + "num_tokens": 634285192.0, + "step": 24516 + }, + { + "epoch": 2.6924006149791344, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2196803092956543, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7302455306053162, + "num_tokens": 634313160.0, + "step": 24517 + }, + { + "epoch": 2.692510432681748, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6352148056030273, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7597952485084534, + "num_tokens": 634334500.0, + "step": 24518 + }, + { + "epoch": 2.692620250384362, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3229081630706787, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7253590822219849, + "num_tokens": 634361753.0, + "step": 24519 + }, + { + "epoch": 2.6927300680869757, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.343663215637207, + "learning_rate": 1e-06, + "loss": 0.7913, + "mean_token_accuracy": 0.742351770401001, + "num_tokens": 634387044.0, + "step": 24520 + }, + { + "epoch": 2.6928398857895894, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1966097354888916, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7146416902542114, + "num_tokens": 634416653.0, + "step": 24521 + }, + { + "epoch": 2.6929497034922028, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.566019058227539, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7413911819458008, + "num_tokens": 634440434.0, + "step": 24522 + }, + { + "epoch": 2.6930595211948165, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.258826732635498, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7251222133636475, + "num_tokens": 634466287.0, + "step": 24523 + }, + { + "epoch": 2.6931693388974303, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3742318153381348, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7437025308609009, + "num_tokens": 634491980.0, + "step": 24524 + }, + { + "epoch": 2.693279156600044, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4350595474243164, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7390441298484802, + "num_tokens": 634515161.0, + "step": 24525 + }, + { + "epoch": 2.693388974302658, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6680896282196045, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7305111289024353, + "num_tokens": 634536852.0, + "step": 24526 + }, + { + "epoch": 2.693498792005271, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4135963916778564, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7406259775161743, + "num_tokens": 634561649.0, + "step": 24527 + }, + { + "epoch": 2.693608609707885, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2354626655578613, + "learning_rate": 1e-06, + "loss": 0.7976, + "mean_token_accuracy": 0.7468428015708923, + "num_tokens": 634586384.0, + "step": 24528 + }, + { + "epoch": 2.6937184274104986, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1204166412353516, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7286397218704224, + "num_tokens": 634617927.0, + "step": 24529 + }, + { + "epoch": 2.6938282451131124, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.277754545211792, + "learning_rate": 1e-06, + "loss": 0.8115, + "mean_token_accuracy": 0.7414101958274841, + "num_tokens": 634644921.0, + "step": 24530 + }, + { + "epoch": 2.693938062815726, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.747828245162964, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7532027959823608, + "num_tokens": 634665892.0, + "step": 24531 + }, + { + "epoch": 2.6940478805183394, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1696555614471436, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7358865737915039, + "num_tokens": 634696236.0, + "step": 24532 + }, + { + "epoch": 2.694157698220953, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.202449083328247, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7405858039855957, + "num_tokens": 634725664.0, + "step": 24533 + }, + { + "epoch": 2.694267515923567, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.485572338104248, + "learning_rate": 1e-06, + "loss": 0.8256, + "mean_token_accuracy": 0.7367878556251526, + "num_tokens": 634749180.0, + "step": 24534 + }, + { + "epoch": 2.6943773336261807, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6519854068756104, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7374483346939087, + "num_tokens": 634769158.0, + "step": 24535 + }, + { + "epoch": 2.6944871513287945, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.497145414352417, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7319310307502747, + "num_tokens": 634791663.0, + "step": 24536 + }, + { + "epoch": 2.6945969690314078, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4125540256500244, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7336433529853821, + "num_tokens": 634817045.0, + "step": 24537 + }, + { + "epoch": 2.6947067867340215, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7318546772003174, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7353324294090271, + "num_tokens": 634838364.0, + "step": 24538 + }, + { + "epoch": 2.6948166044366353, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2557930946350098, + "learning_rate": 1e-06, + "loss": 0.9399, + "mean_token_accuracy": 0.7125696539878845, + "num_tokens": 634867696.0, + "step": 24539 + }, + { + "epoch": 2.6949264221392486, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7077767848968506, + "learning_rate": 1e-06, + "loss": 0.8758, + "mean_token_accuracy": 0.7273682355880737, + "num_tokens": 634887761.0, + "step": 24540 + }, + { + "epoch": 2.6950362398418624, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5704550743103027, + "learning_rate": 1e-06, + "loss": 0.85, + "mean_token_accuracy": 0.7362322807312012, + "num_tokens": 634910623.0, + "step": 24541 + }, + { + "epoch": 2.695146057544476, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5373311042785645, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7292547821998596, + "num_tokens": 634933794.0, + "step": 24542 + }, + { + "epoch": 2.69525587524709, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2551040649414062, + "learning_rate": 1e-06, + "loss": 0.9639, + "mean_token_accuracy": 0.7037938833236694, + "num_tokens": 634963445.0, + "step": 24543 + }, + { + "epoch": 2.6953656929497036, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.546269178390503, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7429027557373047, + "num_tokens": 634984866.0, + "step": 24544 + }, + { + "epoch": 2.695475510652317, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.320373058319092, + "learning_rate": 1e-06, + "loss": 0.8287, + "mean_token_accuracy": 0.7381733655929565, + "num_tokens": 635012109.0, + "step": 24545 + }, + { + "epoch": 2.6955853283549307, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2713308334350586, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7197182774543762, + "num_tokens": 635040466.0, + "step": 24546 + }, + { + "epoch": 2.6956951460575445, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5010759830474854, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7374775409698486, + "num_tokens": 635064335.0, + "step": 24547 + }, + { + "epoch": 2.695804963760158, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.386157274246216, + "learning_rate": 1e-06, + "loss": 0.9063, + "mean_token_accuracy": 0.7209873199462891, + "num_tokens": 635091295.0, + "step": 24548 + }, + { + "epoch": 2.695914781462772, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.193418264389038, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7153969407081604, + "num_tokens": 635122405.0, + "step": 24549 + }, + { + "epoch": 2.6960245991653853, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7406156063079834, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.756123960018158, + "num_tokens": 635143309.0, + "step": 24550 + }, + { + "epoch": 2.696134416867999, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.277456521987915, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7363853454589844, + "num_tokens": 635172218.0, + "step": 24551 + }, + { + "epoch": 2.696244234570613, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.043712854385376, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.7065817713737488, + "num_tokens": 635206805.0, + "step": 24552 + }, + { + "epoch": 2.6963540522732266, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3291876316070557, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.728151798248291, + "num_tokens": 635233987.0, + "step": 24553 + }, + { + "epoch": 2.6964638699758403, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5744519233703613, + "learning_rate": 1e-06, + "loss": 0.9249, + "mean_token_accuracy": 0.7097231149673462, + "num_tokens": 635257790.0, + "step": 24554 + }, + { + "epoch": 2.6965736876784536, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.555971622467041, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7385433912277222, + "num_tokens": 635280021.0, + "step": 24555 + }, + { + "epoch": 2.6966835053810674, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.227449655532837, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7146085500717163, + "num_tokens": 635309456.0, + "step": 24556 + }, + { + "epoch": 2.696793323083681, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.500126838684082, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7359317541122437, + "num_tokens": 635331597.0, + "step": 24557 + }, + { + "epoch": 2.696903140786295, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.913458824157715, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7352566123008728, + "num_tokens": 635352324.0, + "step": 24558 + }, + { + "epoch": 2.6970129584889087, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.561335563659668, + "learning_rate": 1e-06, + "loss": 0.7808, + "mean_token_accuracy": 0.74360191822052, + "num_tokens": 635373940.0, + "step": 24559 + }, + { + "epoch": 2.697122776191522, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0174460411071777, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7222471237182617, + "num_tokens": 635407931.0, + "step": 24560 + }, + { + "epoch": 2.6972325938941357, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1427552700042725, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.725051999092102, + "num_tokens": 635437344.0, + "step": 24561 + }, + { + "epoch": 2.6973424115967495, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.488232135772705, + "learning_rate": 1e-06, + "loss": 0.8951, + "mean_token_accuracy": 0.7187087535858154, + "num_tokens": 635462421.0, + "step": 24562 + }, + { + "epoch": 2.697452229299363, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.80578875541687, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7456908226013184, + "num_tokens": 635481273.0, + "step": 24563 + }, + { + "epoch": 2.697562047001977, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.447503089904785, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7250111103057861, + "num_tokens": 635505387.0, + "step": 24564 + }, + { + "epoch": 2.6976718647045903, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3027470111846924, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7113644480705261, + "num_tokens": 635533811.0, + "step": 24565 + }, + { + "epoch": 2.697781682407204, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2815942764282227, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7322163581848145, + "num_tokens": 635561271.0, + "step": 24566 + }, + { + "epoch": 2.697891500109818, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.439027786254883, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7216817140579224, + "num_tokens": 635584208.0, + "step": 24567 + }, + { + "epoch": 2.698001317812431, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2287416458129883, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.754609227180481, + "num_tokens": 635611111.0, + "step": 24568 + }, + { + "epoch": 2.698111135515045, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2588813304901123, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7370002269744873, + "num_tokens": 635638897.0, + "step": 24569 + }, + { + "epoch": 2.6982209532176586, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1055715084075928, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7321134805679321, + "num_tokens": 635670072.0, + "step": 24570 + }, + { + "epoch": 2.6983307709202724, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8329453468322754, + "learning_rate": 1e-06, + "loss": 0.7659, + "mean_token_accuracy": 0.7496199607849121, + "num_tokens": 635688613.0, + "step": 24571 + }, + { + "epoch": 2.698440588622886, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5403130054473877, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7420803308486938, + "num_tokens": 635711296.0, + "step": 24572 + }, + { + "epoch": 2.6985504063254995, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.195709466934204, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7320960164070129, + "num_tokens": 635741297.0, + "step": 24573 + }, + { + "epoch": 2.6986602240281132, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0762219429016113, + "learning_rate": 1e-06, + "loss": 0.9531, + "mean_token_accuracy": 0.718828558921814, + "num_tokens": 635776310.0, + "step": 24574 + }, + { + "epoch": 2.698770041730727, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3766002655029297, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7276800870895386, + "num_tokens": 635799526.0, + "step": 24575 + }, + { + "epoch": 2.6988798594333407, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3607964515686035, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7190501689910889, + "num_tokens": 635827342.0, + "step": 24576 + }, + { + "epoch": 2.6989896771359545, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3607711791992188, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7316207885742188, + "num_tokens": 635853904.0, + "step": 24577 + }, + { + "epoch": 2.699099494838568, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3665313720703125, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7287318706512451, + "num_tokens": 635882300.0, + "step": 24578 + }, + { + "epoch": 2.6992093125411816, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.202091693878174, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.743148684501648, + "num_tokens": 635911264.0, + "step": 24579 + }, + { + "epoch": 2.6993191302437953, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2223284244537354, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7348917126655579, + "num_tokens": 635937409.0, + "step": 24580 + }, + { + "epoch": 2.699428947946409, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3392841815948486, + "learning_rate": 1e-06, + "loss": 0.8953, + "mean_token_accuracy": 0.7246543169021606, + "num_tokens": 635962176.0, + "step": 24581 + }, + { + "epoch": 2.699538765649023, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4974558353424072, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7399351596832275, + "num_tokens": 635985408.0, + "step": 24582 + }, + { + "epoch": 2.699648583351636, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5268990993499756, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7501511573791504, + "num_tokens": 636008460.0, + "step": 24583 + }, + { + "epoch": 2.69975840105425, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.440328359603882, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7179217338562012, + "num_tokens": 636033600.0, + "step": 24584 + }, + { + "epoch": 2.6998682187568637, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.104010820388794, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7036035060882568, + "num_tokens": 636064531.0, + "step": 24585 + }, + { + "epoch": 2.6999780364594774, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.454862117767334, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.737297773361206, + "num_tokens": 636088199.0, + "step": 24586 + }, + { + "epoch": 2.700087854162091, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3635549545288086, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7367410659790039, + "num_tokens": 636114960.0, + "step": 24587 + }, + { + "epoch": 2.7001976718647045, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.356217861175537, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7109805345535278, + "num_tokens": 636142518.0, + "step": 24588 + }, + { + "epoch": 2.7003074895673183, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.223904609680176, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.7283211946487427, + "num_tokens": 636170558.0, + "step": 24589 + }, + { + "epoch": 2.700417307269932, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.430241584777832, + "learning_rate": 1e-06, + "loss": 0.8013, + "mean_token_accuracy": 0.7384501099586487, + "num_tokens": 636195178.0, + "step": 24590 + }, + { + "epoch": 2.7005271249725453, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.574974775314331, + "learning_rate": 1e-06, + "loss": 0.9519, + "mean_token_accuracy": 0.7025030851364136, + "num_tokens": 636219507.0, + "step": 24591 + }, + { + "epoch": 2.700636942675159, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2266182899475098, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7148646116256714, + "num_tokens": 636249262.0, + "step": 24592 + }, + { + "epoch": 2.700746760377773, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.177196502685547, + "learning_rate": 1e-06, + "loss": 1.0055, + "mean_token_accuracy": 0.6951534748077393, + "num_tokens": 636282270.0, + "step": 24593 + }, + { + "epoch": 2.7008565780803866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3493847846984863, + "learning_rate": 1e-06, + "loss": 0.9253, + "mean_token_accuracy": 0.7095667123794556, + "num_tokens": 636309437.0, + "step": 24594 + }, + { + "epoch": 2.7009663957830004, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5952048301696777, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7472760677337646, + "num_tokens": 636334045.0, + "step": 24595 + }, + { + "epoch": 2.7010762134856137, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5114662647247314, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7286913394927979, + "num_tokens": 636356655.0, + "step": 24596 + }, + { + "epoch": 2.7011860311882274, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3710906505584717, + "learning_rate": 1e-06, + "loss": 0.7739, + "mean_token_accuracy": 0.7534575462341309, + "num_tokens": 636380594.0, + "step": 24597 + }, + { + "epoch": 2.701295848890841, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.241804361343384, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7262264490127563, + "num_tokens": 636409274.0, + "step": 24598 + }, + { + "epoch": 2.701405666593455, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5277771949768066, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7434588670730591, + "num_tokens": 636431881.0, + "step": 24599 + }, + { + "epoch": 2.7015154842960687, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.552011728286743, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.718999981880188, + "num_tokens": 636456702.0, + "step": 24600 + }, + { + "epoch": 2.701625301998682, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4957492351531982, + "learning_rate": 1e-06, + "loss": 0.7368, + "mean_token_accuracy": 0.764002799987793, + "num_tokens": 636477902.0, + "step": 24601 + }, + { + "epoch": 2.7017351197012958, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3090622425079346, + "learning_rate": 1e-06, + "loss": 0.8559, + "mean_token_accuracy": 0.7283141016960144, + "num_tokens": 636506103.0, + "step": 24602 + }, + { + "epoch": 2.7018449374039095, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.18979549407959, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7353358268737793, + "num_tokens": 636534483.0, + "step": 24603 + }, + { + "epoch": 2.7019547551065233, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.33888840675354, + "learning_rate": 1e-06, + "loss": 0.8381, + "mean_token_accuracy": 0.7326328158378601, + "num_tokens": 636560499.0, + "step": 24604 + }, + { + "epoch": 2.702064572809137, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.362576961517334, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7294368743896484, + "num_tokens": 636587021.0, + "step": 24605 + }, + { + "epoch": 2.7021743905117503, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.35074520111084, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7341795563697815, + "num_tokens": 636611808.0, + "step": 24606 + }, + { + "epoch": 2.702284208214364, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.133646011352539, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7259193658828735, + "num_tokens": 636642235.0, + "step": 24607 + }, + { + "epoch": 2.702394025916978, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3456640243530273, + "learning_rate": 1e-06, + "loss": 0.9206, + "mean_token_accuracy": 0.7236326932907104, + "num_tokens": 636668339.0, + "step": 24608 + }, + { + "epoch": 2.7025038436195916, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.404646158218384, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.722508430480957, + "num_tokens": 636694031.0, + "step": 24609 + }, + { + "epoch": 2.7026136613222054, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2728874683380127, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7365759015083313, + "num_tokens": 636721196.0, + "step": 24610 + }, + { + "epoch": 2.7027234790248187, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4762113094329834, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7220327854156494, + "num_tokens": 636744077.0, + "step": 24611 + }, + { + "epoch": 2.7028332967274324, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3300042152404785, + "learning_rate": 1e-06, + "loss": 0.9032, + "mean_token_accuracy": 0.7264959812164307, + "num_tokens": 636771047.0, + "step": 24612 + }, + { + "epoch": 2.702943114430046, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5815303325653076, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7452350854873657, + "num_tokens": 636797984.0, + "step": 24613 + }, + { + "epoch": 2.7030529321326595, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.564380645751953, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7414802312850952, + "num_tokens": 636820040.0, + "step": 24614 + }, + { + "epoch": 2.7031627498352737, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.288728952407837, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7519351243972778, + "num_tokens": 636847213.0, + "step": 24615 + }, + { + "epoch": 2.703272567537887, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3761112689971924, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7344293594360352, + "num_tokens": 636873425.0, + "step": 24616 + }, + { + "epoch": 2.703382385240501, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4942238330841064, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7363519668579102, + "num_tokens": 636895224.0, + "step": 24617 + }, + { + "epoch": 2.7034922029431145, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1466057300567627, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7092373371124268, + "num_tokens": 636925505.0, + "step": 24618 + }, + { + "epoch": 2.703602020645728, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3000848293304443, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.732957124710083, + "num_tokens": 636950985.0, + "step": 24619 + }, + { + "epoch": 2.7037118383483416, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6060588359832764, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7368621826171875, + "num_tokens": 636972902.0, + "step": 24620 + }, + { + "epoch": 2.7038216560509554, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.479680061340332, + "learning_rate": 1e-06, + "loss": 0.8821, + "mean_token_accuracy": 0.7290413975715637, + "num_tokens": 636995027.0, + "step": 24621 + }, + { + "epoch": 2.703931473753569, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.215630292892456, + "learning_rate": 1e-06, + "loss": 0.9341, + "mean_token_accuracy": 0.7118934392929077, + "num_tokens": 637025443.0, + "step": 24622 + }, + { + "epoch": 2.704041291456183, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3692452907562256, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7345952391624451, + "num_tokens": 637053295.0, + "step": 24623 + }, + { + "epoch": 2.704151109158796, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3655312061309814, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.7354822158813477, + "num_tokens": 637078599.0, + "step": 24624 + }, + { + "epoch": 2.70426092686141, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0823471546173096, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7227450609207153, + "num_tokens": 637109623.0, + "step": 24625 + }, + { + "epoch": 2.7043707445640237, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.415525197982788, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7285202145576477, + "num_tokens": 637132960.0, + "step": 24626 + }, + { + "epoch": 2.7044805622666375, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3140063285827637, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7412451505661011, + "num_tokens": 637159563.0, + "step": 24627 + }, + { + "epoch": 2.7045903799692512, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1125760078430176, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7440499663352966, + "num_tokens": 637189795.0, + "step": 24628 + }, + { + "epoch": 2.7047001976718645, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2409348487854004, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7163902521133423, + "num_tokens": 637217854.0, + "step": 24629 + }, + { + "epoch": 2.7048100153744783, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.246901273727417, + "learning_rate": 1e-06, + "loss": 0.7279, + "mean_token_accuracy": 0.7590171098709106, + "num_tokens": 637246068.0, + "step": 24630 + }, + { + "epoch": 2.704919833077092, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2479264736175537, + "learning_rate": 1e-06, + "loss": 0.907, + "mean_token_accuracy": 0.7174955010414124, + "num_tokens": 637275907.0, + "step": 24631 + }, + { + "epoch": 2.705029650779706, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5898518562316895, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7295066714286804, + "num_tokens": 637298231.0, + "step": 24632 + }, + { + "epoch": 2.7051394684823196, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3179802894592285, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7436383962631226, + "num_tokens": 637324818.0, + "step": 24633 + }, + { + "epoch": 2.705249286184933, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2342636585235596, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7407087087631226, + "num_tokens": 637351693.0, + "step": 24634 + }, + { + "epoch": 2.7053591038875466, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.400895833969116, + "learning_rate": 1e-06, + "loss": 0.7916, + "mean_token_accuracy": 0.7513678073883057, + "num_tokens": 637376108.0, + "step": 24635 + }, + { + "epoch": 2.7054689215901604, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.548046827316284, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7289530634880066, + "num_tokens": 637402645.0, + "step": 24636 + }, + { + "epoch": 2.705578739292774, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.392230987548828, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.7163246870040894, + "num_tokens": 637427056.0, + "step": 24637 + }, + { + "epoch": 2.705688556995388, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1752235889434814, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7218295931816101, + "num_tokens": 637457793.0, + "step": 24638 + }, + { + "epoch": 2.705798374698001, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.351923704147339, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7460455894470215, + "num_tokens": 637481962.0, + "step": 24639 + }, + { + "epoch": 2.705908192400615, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.741217613220215, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7494505047798157, + "num_tokens": 637502153.0, + "step": 24640 + }, + { + "epoch": 2.7060180101032287, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2833614349365234, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7158390879631042, + "num_tokens": 637529072.0, + "step": 24641 + }, + { + "epoch": 2.706127827805842, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7747795581817627, + "learning_rate": 1e-06, + "loss": 0.7725, + "mean_token_accuracy": 0.7530202865600586, + "num_tokens": 637549332.0, + "step": 24642 + }, + { + "epoch": 2.706237645508456, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6989970207214355, + "learning_rate": 1e-06, + "loss": 0.7676, + "mean_token_accuracy": 0.7566288709640503, + "num_tokens": 637571205.0, + "step": 24643 + }, + { + "epoch": 2.7063474632110696, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.353670358657837, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7216286063194275, + "num_tokens": 637599018.0, + "step": 24644 + }, + { + "epoch": 2.7064572809136833, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4097516536712646, + "learning_rate": 1e-06, + "loss": 0.9219, + "mean_token_accuracy": 0.7142748832702637, + "num_tokens": 637626618.0, + "step": 24645 + }, + { + "epoch": 2.706567098616297, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1699464321136475, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7253172993659973, + "num_tokens": 637655494.0, + "step": 24646 + }, + { + "epoch": 2.7066769163189104, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3802483081817627, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7292084097862244, + "num_tokens": 637680457.0, + "step": 24647 + }, + { + "epoch": 2.706786734021524, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2817447185516357, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7320384383201599, + "num_tokens": 637708205.0, + "step": 24648 + }, + { + "epoch": 2.706896551724138, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 3.144602060317993, + "learning_rate": 1e-06, + "loss": 0.7686, + "mean_token_accuracy": 0.7504491806030273, + "num_tokens": 637725835.0, + "step": 24649 + }, + { + "epoch": 2.7070063694267517, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0884034633636475, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7291129231452942, + "num_tokens": 637757318.0, + "step": 24650 + }, + { + "epoch": 2.7071161871293654, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1796317100524902, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7240942120552063, + "num_tokens": 637786045.0, + "step": 24651 + }, + { + "epoch": 2.7072260048319787, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3214879035949707, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7338600158691406, + "num_tokens": 637814344.0, + "step": 24652 + }, + { + "epoch": 2.7073358225345925, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9497703313827515, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7174724340438843, + "num_tokens": 637851542.0, + "step": 24653 + }, + { + "epoch": 2.7074456402372062, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.291818141937256, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.6965442895889282, + "num_tokens": 637880353.0, + "step": 24654 + }, + { + "epoch": 2.70755545793982, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.342883586883545, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.725757360458374, + "num_tokens": 637905596.0, + "step": 24655 + }, + { + "epoch": 2.7076652756424338, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.422858238220215, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7263998985290527, + "num_tokens": 637930958.0, + "step": 24656 + }, + { + "epoch": 2.707775093345047, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5849201679229736, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.7425658702850342, + "num_tokens": 637952486.0, + "step": 24657 + }, + { + "epoch": 2.707884911047661, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.363375425338745, + "learning_rate": 1e-06, + "loss": 0.9625, + "mean_token_accuracy": 0.7093040347099304, + "num_tokens": 637980785.0, + "step": 24658 + }, + { + "epoch": 2.7079947287502746, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.145292043685913, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7207233905792236, + "num_tokens": 638012640.0, + "step": 24659 + }, + { + "epoch": 2.7081045464528883, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6202497482299805, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7478834986686707, + "num_tokens": 638035306.0, + "step": 24660 + }, + { + "epoch": 2.708214364155502, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.266432523727417, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7285258769989014, + "num_tokens": 638064193.0, + "step": 24661 + }, + { + "epoch": 2.7083241818581154, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4339263439178467, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7479957938194275, + "num_tokens": 638088972.0, + "step": 24662 + }, + { + "epoch": 2.708433999560729, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6947169303894043, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7398704290390015, + "num_tokens": 638110260.0, + "step": 24663 + }, + { + "epoch": 2.708543817263343, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.6186304092407227, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7379290461540222, + "num_tokens": 638132932.0, + "step": 24664 + }, + { + "epoch": 2.7086536349659562, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.403036117553711, + "learning_rate": 1e-06, + "loss": 0.9791, + "mean_token_accuracy": 0.7054903507232666, + "num_tokens": 638159454.0, + "step": 24665 + }, + { + "epoch": 2.7087634526685704, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.402179479598999, + "learning_rate": 1e-06, + "loss": 0.8197, + "mean_token_accuracy": 0.7389533519744873, + "num_tokens": 638183545.0, + "step": 24666 + }, + { + "epoch": 2.7088732703711838, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.054421901702881, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7303181290626526, + "num_tokens": 638215187.0, + "step": 24667 + }, + { + "epoch": 2.7089830880737975, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4660542011260986, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7287050485610962, + "num_tokens": 638240218.0, + "step": 24668 + }, + { + "epoch": 2.7090929057764113, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.091907501220703, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7306507229804993, + "num_tokens": 638271344.0, + "step": 24669 + }, + { + "epoch": 2.7092027234790246, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.004512071609497, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7264448404312134, + "num_tokens": 638303577.0, + "step": 24670 + }, + { + "epoch": 2.7093125411816383, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.241454601287842, + "learning_rate": 1e-06, + "loss": 0.9203, + "mean_token_accuracy": 0.7138913869857788, + "num_tokens": 638330973.0, + "step": 24671 + }, + { + "epoch": 2.709422358884252, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.222198724746704, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7323986887931824, + "num_tokens": 638357139.0, + "step": 24672 + }, + { + "epoch": 2.709532176586866, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 3.163606882095337, + "learning_rate": 1e-06, + "loss": 0.8281, + "mean_token_accuracy": 0.7380682826042175, + "num_tokens": 638373815.0, + "step": 24673 + }, + { + "epoch": 2.7096419942894796, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5133166313171387, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7058452367782593, + "num_tokens": 638399938.0, + "step": 24674 + }, + { + "epoch": 2.709751811992093, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.499359130859375, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7426696419715881, + "num_tokens": 638422808.0, + "step": 24675 + }, + { + "epoch": 2.7098616296947067, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1183886528015137, + "learning_rate": 1e-06, + "loss": 0.7619, + "mean_token_accuracy": 0.7563468813896179, + "num_tokens": 638450788.0, + "step": 24676 + }, + { + "epoch": 2.7099714473973204, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.349079132080078, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7304528951644897, + "num_tokens": 638478890.0, + "step": 24677 + }, + { + "epoch": 2.710081265099934, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1503353118896484, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7366750836372375, + "num_tokens": 638510335.0, + "step": 24678 + }, + { + "epoch": 2.710191082802548, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.477921724319458, + "learning_rate": 1e-06, + "loss": 0.8475, + "mean_token_accuracy": 0.7320513725280762, + "num_tokens": 638533854.0, + "step": 24679 + }, + { + "epoch": 2.7103009005051613, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3142480850219727, + "learning_rate": 1e-06, + "loss": 0.9182, + "mean_token_accuracy": 0.7114912271499634, + "num_tokens": 638561507.0, + "step": 24680 + }, + { + "epoch": 2.710410718207775, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.424504280090332, + "learning_rate": 1e-06, + "loss": 0.8398, + "mean_token_accuracy": 0.7362139821052551, + "num_tokens": 638586013.0, + "step": 24681 + }, + { + "epoch": 2.7105205359103888, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3481979370117188, + "learning_rate": 1e-06, + "loss": 0.7683, + "mean_token_accuracy": 0.7518531680107117, + "num_tokens": 638609740.0, + "step": 24682 + }, + { + "epoch": 2.7106303536130025, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4046847820281982, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7220615744590759, + "num_tokens": 638636714.0, + "step": 24683 + }, + { + "epoch": 2.7107401713156163, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3922245502471924, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7441604137420654, + "num_tokens": 638661166.0, + "step": 24684 + }, + { + "epoch": 2.7108499890182296, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1546177864074707, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.7201381325721741, + "num_tokens": 638689317.0, + "step": 24685 + }, + { + "epoch": 2.7109598067208434, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4272608757019043, + "learning_rate": 1e-06, + "loss": 0.9262, + "mean_token_accuracy": 0.7103430032730103, + "num_tokens": 638714501.0, + "step": 24686 + }, + { + "epoch": 2.711069624423457, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.491844415664673, + "learning_rate": 1e-06, + "loss": 0.8954, + "mean_token_accuracy": 0.7165550589561462, + "num_tokens": 638739101.0, + "step": 24687 + }, + { + "epoch": 2.711179442126071, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3344550132751465, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.716113805770874, + "num_tokens": 638766889.0, + "step": 24688 + }, + { + "epoch": 2.7112892598286846, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5812830924987793, + "learning_rate": 1e-06, + "loss": 0.8216, + "mean_token_accuracy": 0.7340693473815918, + "num_tokens": 638789471.0, + "step": 24689 + }, + { + "epoch": 2.711399077531298, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.615222215652466, + "learning_rate": 1e-06, + "loss": 0.8139, + "mean_token_accuracy": 0.7359635233879089, + "num_tokens": 638810942.0, + "step": 24690 + }, + { + "epoch": 2.7115088952339117, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.111589193344116, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7315313816070557, + "num_tokens": 638843824.0, + "step": 24691 + }, + { + "epoch": 2.7116187129365255, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0103461742401123, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7293202877044678, + "num_tokens": 638874574.0, + "step": 24692 + }, + { + "epoch": 2.7117285306391388, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2159624099731445, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7280152440071106, + "num_tokens": 638903284.0, + "step": 24693 + }, + { + "epoch": 2.7118383483417525, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5215044021606445, + "learning_rate": 1e-06, + "loss": 0.9037, + "mean_token_accuracy": 0.7183462977409363, + "num_tokens": 638928789.0, + "step": 24694 + }, + { + "epoch": 2.7119481660443663, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1114089488983154, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7086423635482788, + "num_tokens": 638960128.0, + "step": 24695 + }, + { + "epoch": 2.71205798374698, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.251364231109619, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7331717014312744, + "num_tokens": 638986557.0, + "step": 24696 + }, + { + "epoch": 2.712167801449594, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0300180912017822, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7141133546829224, + "num_tokens": 639019260.0, + "step": 24697 + }, + { + "epoch": 2.712277619152207, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.026034355163574, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7250460982322693, + "num_tokens": 639051229.0, + "step": 24698 + }, + { + "epoch": 2.712387436854821, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2356953620910645, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7264413833618164, + "num_tokens": 639080680.0, + "step": 24699 + }, + { + "epoch": 2.7124972545574346, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0620956420898438, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.740364134311676, + "num_tokens": 639112770.0, + "step": 24700 + }, + { + "epoch": 2.7126070722600484, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.8222718238830566, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.734175980091095, + "num_tokens": 639132757.0, + "step": 24701 + }, + { + "epoch": 2.712716889962662, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4582486152648926, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7474043369293213, + "num_tokens": 639155224.0, + "step": 24702 + }, + { + "epoch": 2.7128267076652754, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3500888347625732, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7284752130508423, + "num_tokens": 639180807.0, + "step": 24703 + }, + { + "epoch": 2.712936525367889, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.792724132537842, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7244904637336731, + "num_tokens": 639200014.0, + "step": 24704 + }, + { + "epoch": 2.713046343070503, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7425696849823, + "learning_rate": 1e-06, + "loss": 0.7438, + "mean_token_accuracy": 0.7577232122421265, + "num_tokens": 639219260.0, + "step": 24705 + }, + { + "epoch": 2.7131561607731167, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1753475666046143, + "learning_rate": 1e-06, + "loss": 0.9028, + "mean_token_accuracy": 0.7218772768974304, + "num_tokens": 639250650.0, + "step": 24706 + }, + { + "epoch": 2.7132659784757305, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4103915691375732, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7383613586425781, + "num_tokens": 639275641.0, + "step": 24707 + }, + { + "epoch": 2.713375796178344, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5336976051330566, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7345089316368103, + "num_tokens": 639297757.0, + "step": 24708 + }, + { + "epoch": 2.7134856138809575, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2155075073242188, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.730984091758728, + "num_tokens": 639325714.0, + "step": 24709 + }, + { + "epoch": 2.7135954315835713, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.64133620262146, + "learning_rate": 1e-06, + "loss": 0.7831, + "mean_token_accuracy": 0.7419283390045166, + "num_tokens": 639345003.0, + "step": 24710 + }, + { + "epoch": 2.713705249286185, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3867831230163574, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.7314665913581848, + "num_tokens": 639369979.0, + "step": 24711 + }, + { + "epoch": 2.713815066988799, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2747275829315186, + "learning_rate": 1e-06, + "loss": 0.9047, + "mean_token_accuracy": 0.7268499732017517, + "num_tokens": 639398053.0, + "step": 24712 + }, + { + "epoch": 2.713924884691412, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.198620557785034, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7129814624786377, + "num_tokens": 639428209.0, + "step": 24713 + }, + { + "epoch": 2.714034702394026, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1764352321624756, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7396271228790283, + "num_tokens": 639454434.0, + "step": 24714 + }, + { + "epoch": 2.7141445200966396, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4050967693328857, + "learning_rate": 1e-06, + "loss": 0.7189, + "mean_token_accuracy": 0.7645260095596313, + "num_tokens": 639476779.0, + "step": 24715 + }, + { + "epoch": 2.7142543377992534, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3096206188201904, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7382931709289551, + "num_tokens": 639503392.0, + "step": 24716 + }, + { + "epoch": 2.714364155501867, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.365178346633911, + "learning_rate": 1e-06, + "loss": 0.7558, + "mean_token_accuracy": 0.7584096193313599, + "num_tokens": 639527342.0, + "step": 24717 + }, + { + "epoch": 2.7144739732044805, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5028786659240723, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7417139410972595, + "num_tokens": 639550694.0, + "step": 24718 + }, + { + "epoch": 2.7145837909070942, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.9962728023529053, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7299481630325317, + "num_tokens": 639585039.0, + "step": 24719 + }, + { + "epoch": 2.714693608609708, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.616131544113159, + "learning_rate": 1e-06, + "loss": 0.7961, + "mean_token_accuracy": 0.7456715106964111, + "num_tokens": 639607137.0, + "step": 24720 + }, + { + "epoch": 2.7148034263123213, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2225143909454346, + "learning_rate": 1e-06, + "loss": 0.8554, + "mean_token_accuracy": 0.7314232587814331, + "num_tokens": 639635501.0, + "step": 24721 + }, + { + "epoch": 2.714913244014935, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2643120288848877, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7176351547241211, + "num_tokens": 639665461.0, + "step": 24722 + }, + { + "epoch": 2.715023061717549, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.272005319595337, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7153232097625732, + "num_tokens": 639693491.0, + "step": 24723 + }, + { + "epoch": 2.7151328794201626, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.666247844696045, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7338207960128784, + "num_tokens": 639715777.0, + "step": 24724 + }, + { + "epoch": 2.7152426971227763, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3625941276550293, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7303791046142578, + "num_tokens": 639741348.0, + "step": 24725 + }, + { + "epoch": 2.7153525148253896, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.4931113719940186, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7342545986175537, + "num_tokens": 639764709.0, + "step": 24726 + }, + { + "epoch": 2.7154623325280034, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.145519733428955, + "learning_rate": 1e-06, + "loss": 0.8201, + "mean_token_accuracy": 0.7451132535934448, + "num_tokens": 639793037.0, + "step": 24727 + }, + { + "epoch": 2.715572150230617, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3120503425598145, + "learning_rate": 1e-06, + "loss": 0.9264, + "mean_token_accuracy": 0.7145832777023315, + "num_tokens": 639820215.0, + "step": 24728 + }, + { + "epoch": 2.715681967933231, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5439655780792236, + "learning_rate": 1e-06, + "loss": 0.8942, + "mean_token_accuracy": 0.7256634831428528, + "num_tokens": 639842429.0, + "step": 24729 + }, + { + "epoch": 2.7157917856358447, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.280137062072754, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7264991998672485, + "num_tokens": 639873513.0, + "step": 24730 + }, + { + "epoch": 2.715901603338458, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.37253999710083, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7124786376953125, + "num_tokens": 639899481.0, + "step": 24731 + }, + { + "epoch": 2.7160114210410717, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.405059576034546, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7294777631759644, + "num_tokens": 639924799.0, + "step": 24732 + }, + { + "epoch": 2.7161212387436855, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.272305727005005, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7274844646453857, + "num_tokens": 639951117.0, + "step": 24733 + }, + { + "epoch": 2.7162310564462993, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3293392658233643, + "learning_rate": 1e-06, + "loss": 0.8271, + "mean_token_accuracy": 0.734919548034668, + "num_tokens": 639974759.0, + "step": 24734 + }, + { + "epoch": 2.716340874148913, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.707789659500122, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.7245413064956665, + "num_tokens": 639997730.0, + "step": 24735 + }, + { + "epoch": 2.7164506918515263, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5753672122955322, + "learning_rate": 1e-06, + "loss": 0.8203, + "mean_token_accuracy": 0.73714280128479, + "num_tokens": 640020029.0, + "step": 24736 + }, + { + "epoch": 2.71656050955414, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.325021266937256, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7350565195083618, + "num_tokens": 640048122.0, + "step": 24737 + }, + { + "epoch": 2.716670327256754, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.169407606124878, + "learning_rate": 1e-06, + "loss": 0.9127, + "mean_token_accuracy": 0.7108908295631409, + "num_tokens": 640077578.0, + "step": 24738 + }, + { + "epoch": 2.7167801449593676, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.351200819015503, + "learning_rate": 1e-06, + "loss": 0.7219, + "mean_token_accuracy": 0.7637498378753662, + "num_tokens": 640102825.0, + "step": 24739 + }, + { + "epoch": 2.7168899626619814, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2905056476593018, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.739370584487915, + "num_tokens": 640132048.0, + "step": 24740 + }, + { + "epoch": 2.7169997803645947, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.643712043762207, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7315114736557007, + "num_tokens": 640156823.0, + "step": 24741 + }, + { + "epoch": 2.7171095980672084, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.518087863922119, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7233077883720398, + "num_tokens": 640180505.0, + "step": 24742 + }, + { + "epoch": 2.717219415769822, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7165393829345703, + "learning_rate": 1e-06, + "loss": 0.7432, + "mean_token_accuracy": 0.7581827640533447, + "num_tokens": 640199882.0, + "step": 24743 + }, + { + "epoch": 2.7173292334724355, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3172991275787354, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7399392127990723, + "num_tokens": 640228809.0, + "step": 24744 + }, + { + "epoch": 2.7174390511750497, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3730599880218506, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.716719925403595, + "num_tokens": 640256072.0, + "step": 24745 + }, + { + "epoch": 2.717548868877663, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.305514335632324, + "learning_rate": 1e-06, + "loss": 0.8696, + "mean_token_accuracy": 0.7495748996734619, + "num_tokens": 640282519.0, + "step": 24746 + }, + { + "epoch": 2.7176586865802768, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3754327297210693, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7179707288742065, + "num_tokens": 640307450.0, + "step": 24747 + }, + { + "epoch": 2.7177685042828905, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.507070779800415, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7354162931442261, + "num_tokens": 640329891.0, + "step": 24748 + }, + { + "epoch": 2.717878321985504, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.0234434604644775, + "learning_rate": 1e-06, + "loss": 0.9752, + "mean_token_accuracy": 0.7017194628715515, + "num_tokens": 640364993.0, + "step": 24749 + }, + { + "epoch": 2.7179881396881176, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2810568809509277, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7312695980072021, + "num_tokens": 640392942.0, + "step": 24750 + }, + { + "epoch": 2.7180979573907313, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.277383327484131, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7235311269760132, + "num_tokens": 640419676.0, + "step": 24751 + }, + { + "epoch": 2.718207775093345, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.221513032913208, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7397938370704651, + "num_tokens": 640448091.0, + "step": 24752 + }, + { + "epoch": 2.718317592795959, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3119442462921143, + "learning_rate": 1e-06, + "loss": 0.9216, + "mean_token_accuracy": 0.7169054746627808, + "num_tokens": 640477778.0, + "step": 24753 + }, + { + "epoch": 2.718427410498572, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.448395013809204, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7311520576477051, + "num_tokens": 640500398.0, + "step": 24754 + }, + { + "epoch": 2.718537228201186, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2473409175872803, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7433093190193176, + "num_tokens": 640528757.0, + "step": 24755 + }, + { + "epoch": 2.7186470459037997, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3626863956451416, + "learning_rate": 1e-06, + "loss": 0.936, + "mean_token_accuracy": 0.7066399455070496, + "num_tokens": 640554774.0, + "step": 24756 + }, + { + "epoch": 2.7187568636064134, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.251856565475464, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7227606773376465, + "num_tokens": 640582467.0, + "step": 24757 + }, + { + "epoch": 2.718866681309027, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.174825668334961, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7214810848236084, + "num_tokens": 640614279.0, + "step": 24758 + }, + { + "epoch": 2.7189764990116405, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2122230529785156, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7231765985488892, + "num_tokens": 640642335.0, + "step": 24759 + }, + { + "epoch": 2.7190863167142543, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.320399522781372, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7096638679504395, + "num_tokens": 640669322.0, + "step": 24760 + }, + { + "epoch": 2.719196134416868, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.9257588386535645, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7313530445098877, + "num_tokens": 640687288.0, + "step": 24761 + }, + { + "epoch": 2.719305952119482, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.5066444873809814, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7288989424705505, + "num_tokens": 640710975.0, + "step": 24762 + }, + { + "epoch": 2.7194157698220955, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 1.905647873878479, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7273598909378052, + "num_tokens": 640747982.0, + "step": 24763 + }, + { + "epoch": 2.719525587524709, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3561904430389404, + "learning_rate": 1e-06, + "loss": 0.798, + "mean_token_accuracy": 0.7545922994613647, + "num_tokens": 640775056.0, + "step": 24764 + }, + { + "epoch": 2.7196354052273226, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.273812770843506, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.7195870876312256, + "num_tokens": 640802794.0, + "step": 24765 + }, + { + "epoch": 2.7197452229299364, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3703773021698, + "learning_rate": 1e-06, + "loss": 0.9123, + "mean_token_accuracy": 0.7188737392425537, + "num_tokens": 640828970.0, + "step": 24766 + }, + { + "epoch": 2.71985504063255, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.340580940246582, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7360047101974487, + "num_tokens": 640856830.0, + "step": 24767 + }, + { + "epoch": 2.719964858335164, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.1731576919555664, + "learning_rate": 1e-06, + "loss": 0.9438, + "mean_token_accuracy": 0.709875762462616, + "num_tokens": 640888052.0, + "step": 24768 + }, + { + "epoch": 2.720074676037777, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.379476547241211, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.745150625705719, + "num_tokens": 640912422.0, + "step": 24769 + }, + { + "epoch": 2.720184493740391, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.7078146934509277, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.735424280166626, + "num_tokens": 640933340.0, + "step": 24770 + }, + { + "epoch": 2.7202943114430047, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.699012517929077, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.753525972366333, + "num_tokens": 640953858.0, + "step": 24771 + }, + { + "epoch": 2.720404129145618, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.368227005004883, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.733319103717804, + "num_tokens": 640979591.0, + "step": 24772 + }, + { + "epoch": 2.720513946848232, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.229962110519409, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7292518019676208, + "num_tokens": 641005981.0, + "step": 24773 + }, + { + "epoch": 2.7206237645508455, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.471031665802002, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7442123889923096, + "num_tokens": 641030596.0, + "step": 24774 + }, + { + "epoch": 2.7207335822534593, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3409018516540527, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7070852518081665, + "num_tokens": 641057299.0, + "step": 24775 + }, + { + "epoch": 2.720843399956073, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.514911651611328, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7298526167869568, + "num_tokens": 641083248.0, + "step": 24776 + }, + { + "epoch": 2.7209532176586864, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.386695146560669, + "learning_rate": 1e-06, + "loss": 0.9475, + "mean_token_accuracy": 0.7041655778884888, + "num_tokens": 641110826.0, + "step": 24777 + }, + { + "epoch": 2.7210630353613, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.357015371322632, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7471200227737427, + "num_tokens": 641138748.0, + "step": 24778 + }, + { + "epoch": 2.721172853063914, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.307642698287964, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7184854745864868, + "num_tokens": 641167288.0, + "step": 24779 + }, + { + "epoch": 2.7212826707665276, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.2111456394195557, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7237638831138611, + "num_tokens": 641196385.0, + "step": 24780 + }, + { + "epoch": 2.7213924884691414, + "ewc_loss": 1.5735626220703125e-05, + "grad_norm": 2.3927128314971924, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.72297203540802, + "num_tokens": 641222366.0, + "step": 24781 + }, + { + "epoch": 2.7215023061717547, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.04819655418396, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7246832847595215, + "num_tokens": 641255955.0, + "step": 24782 + }, + { + "epoch": 2.7216121238743685, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.496809244155884, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7170484066009521, + "num_tokens": 641280874.0, + "step": 24783 + }, + { + "epoch": 2.721721941576982, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5131723880767822, + "learning_rate": 1e-06, + "loss": 0.8116, + "mean_token_accuracy": 0.7385138273239136, + "num_tokens": 641303592.0, + "step": 24784 + }, + { + "epoch": 2.721831759279596, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.042994976043701, + "learning_rate": 1e-06, + "loss": 0.9669, + "mean_token_accuracy": 0.6972705721855164, + "num_tokens": 641336893.0, + "step": 24785 + }, + { + "epoch": 2.7219415769822097, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.48443603515625, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7375200986862183, + "num_tokens": 641359312.0, + "step": 24786 + }, + { + "epoch": 2.722051394684823, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4971859455108643, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7499881982803345, + "num_tokens": 641383693.0, + "step": 24787 + }, + { + "epoch": 2.722161212387437, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5063135623931885, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7336451411247253, + "num_tokens": 641407736.0, + "step": 24788 + }, + { + "epoch": 2.7222710300900506, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.327800989151001, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7037920951843262, + "num_tokens": 641433221.0, + "step": 24789 + }, + { + "epoch": 2.7223808477926643, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.530590295791626, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7276163697242737, + "num_tokens": 641456588.0, + "step": 24790 + }, + { + "epoch": 2.722490665495278, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0316073894500732, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.712549090385437, + "num_tokens": 641491270.0, + "step": 24791 + }, + { + "epoch": 2.7226004831978914, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1093857288360596, + "learning_rate": 1e-06, + "loss": 0.9059, + "mean_token_accuracy": 0.720861554145813, + "num_tokens": 641524657.0, + "step": 24792 + }, + { + "epoch": 2.722710300900505, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.295057773590088, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7119478583335876, + "num_tokens": 641551827.0, + "step": 24793 + }, + { + "epoch": 2.722820118603119, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.878596544265747, + "learning_rate": 1e-06, + "loss": 0.7749, + "mean_token_accuracy": 0.7578040361404419, + "num_tokens": 641570975.0, + "step": 24794 + }, + { + "epoch": 2.722929936305732, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.654454469680786, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7379464507102966, + "num_tokens": 641591883.0, + "step": 24795 + }, + { + "epoch": 2.7230397540083464, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.480207681655884, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7301248908042908, + "num_tokens": 641615483.0, + "step": 24796 + }, + { + "epoch": 2.7231495717109597, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3268160820007324, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7514714598655701, + "num_tokens": 641640463.0, + "step": 24797 + }, + { + "epoch": 2.7232593894135735, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.242645263671875, + "learning_rate": 1e-06, + "loss": 0.8766, + "mean_token_accuracy": 0.718482255935669, + "num_tokens": 641669341.0, + "step": 24798 + }, + { + "epoch": 2.7233692071161872, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4460244178771973, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7109360694885254, + "num_tokens": 641695953.0, + "step": 24799 + }, + { + "epoch": 2.7234790248188006, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.500929355621338, + "learning_rate": 1e-06, + "loss": 0.8541, + "mean_token_accuracy": 0.7277604341506958, + "num_tokens": 641720964.0, + "step": 24800 + }, + { + "epoch": 2.7235888425214143, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1101083755493164, + "learning_rate": 1e-06, + "loss": 0.9145, + "mean_token_accuracy": 0.7177333831787109, + "num_tokens": 641752724.0, + "step": 24801 + }, + { + "epoch": 2.723698660224028, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.616665840148926, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7517511248588562, + "num_tokens": 641774184.0, + "step": 24802 + }, + { + "epoch": 2.723808477926642, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2372303009033203, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7194854021072388, + "num_tokens": 641802550.0, + "step": 24803 + }, + { + "epoch": 2.7239182956292556, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.673861265182495, + "learning_rate": 1e-06, + "loss": 0.745, + "mean_token_accuracy": 0.7568542957305908, + "num_tokens": 641822945.0, + "step": 24804 + }, + { + "epoch": 2.724028113331869, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.371401786804199, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7461990714073181, + "num_tokens": 641848120.0, + "step": 24805 + }, + { + "epoch": 2.7241379310344827, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2781341075897217, + "learning_rate": 1e-06, + "loss": 0.8028, + "mean_token_accuracy": 0.7479733824729919, + "num_tokens": 641874104.0, + "step": 24806 + }, + { + "epoch": 2.7242477487370964, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.422055244445801, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7395399212837219, + "num_tokens": 641898951.0, + "step": 24807 + }, + { + "epoch": 2.72435756643971, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5053963661193848, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7468662261962891, + "num_tokens": 641921362.0, + "step": 24808 + }, + { + "epoch": 2.724467384142324, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 3.026719570159912, + "learning_rate": 1e-06, + "loss": 0.7991, + "mean_token_accuracy": 0.7450743317604065, + "num_tokens": 641938341.0, + "step": 24809 + }, + { + "epoch": 2.7245772018449372, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.056062936782837, + "learning_rate": 1e-06, + "loss": 0.9164, + "mean_token_accuracy": 0.7099096179008484, + "num_tokens": 641970748.0, + "step": 24810 + }, + { + "epoch": 2.724687019547551, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2119975090026855, + "learning_rate": 1e-06, + "loss": 0.9178, + "mean_token_accuracy": 0.7097673416137695, + "num_tokens": 642001075.0, + "step": 24811 + }, + { + "epoch": 2.7247968372501647, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.328659772872925, + "learning_rate": 1e-06, + "loss": 0.7873, + "mean_token_accuracy": 0.7502753734588623, + "num_tokens": 642026261.0, + "step": 24812 + }, + { + "epoch": 2.7249066549527785, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.250305414199829, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7183589935302734, + "num_tokens": 642055981.0, + "step": 24813 + }, + { + "epoch": 2.7250164726553923, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2863292694091797, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7074260711669922, + "num_tokens": 642082195.0, + "step": 24814 + }, + { + "epoch": 2.7251262903580056, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1723997592926025, + "learning_rate": 1e-06, + "loss": 0.8701, + "mean_token_accuracy": 0.7277867794036865, + "num_tokens": 642114405.0, + "step": 24815 + }, + { + "epoch": 2.7252361080606193, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3267159461975098, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7202876210212708, + "num_tokens": 642139251.0, + "step": 24816 + }, + { + "epoch": 2.725345925763233, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.527243137359619, + "learning_rate": 1e-06, + "loss": 0.8138, + "mean_token_accuracy": 0.7413182258605957, + "num_tokens": 642163813.0, + "step": 24817 + }, + { + "epoch": 2.725455743465847, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2632510662078857, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7366670370101929, + "num_tokens": 642189645.0, + "step": 24818 + }, + { + "epoch": 2.7255655611684606, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1821084022521973, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7339158058166504, + "num_tokens": 642219610.0, + "step": 24819 + }, + { + "epoch": 2.725675378871074, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6293041706085205, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7381178140640259, + "num_tokens": 642240897.0, + "step": 24820 + }, + { + "epoch": 2.7257851965736877, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2829885482788086, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7236326336860657, + "num_tokens": 642270152.0, + "step": 24821 + }, + { + "epoch": 2.7258950142763014, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.204116106033325, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.7299126386642456, + "num_tokens": 642303524.0, + "step": 24822 + }, + { + "epoch": 2.7260048319789147, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5370118618011475, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7277158498764038, + "num_tokens": 642326904.0, + "step": 24823 + }, + { + "epoch": 2.7261146496815285, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4918556213378906, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7284147143363953, + "num_tokens": 642352381.0, + "step": 24824 + }, + { + "epoch": 2.7262244673841423, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3693506717681885, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7290056943893433, + "num_tokens": 642378034.0, + "step": 24825 + }, + { + "epoch": 2.726334285086756, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.426804304122925, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7407398223876953, + "num_tokens": 642403044.0, + "step": 24826 + }, + { + "epoch": 2.7264441027893698, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.462303876876831, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7233676910400391, + "num_tokens": 642426545.0, + "step": 24827 + }, + { + "epoch": 2.726553920491983, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.314344882965088, + "learning_rate": 1e-06, + "loss": 0.7703, + "mean_token_accuracy": 0.7548658847808838, + "num_tokens": 642452117.0, + "step": 24828 + }, + { + "epoch": 2.726663738194597, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.483811378479004, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7185373306274414, + "num_tokens": 642478164.0, + "step": 24829 + }, + { + "epoch": 2.7267735558972106, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1923818588256836, + "learning_rate": 1e-06, + "loss": 0.9548, + "mean_token_accuracy": 0.7073726654052734, + "num_tokens": 642506938.0, + "step": 24830 + }, + { + "epoch": 2.7268833735998244, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.094015121459961, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.712204098701477, + "num_tokens": 642542506.0, + "step": 24831 + }, + { + "epoch": 2.726993191302438, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.580031394958496, + "learning_rate": 1e-06, + "loss": 0.8297, + "mean_token_accuracy": 0.7378846406936646, + "num_tokens": 642563542.0, + "step": 24832 + }, + { + "epoch": 2.7271030090050514, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.088707208633423, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7264895439147949, + "num_tokens": 642595227.0, + "step": 24833 + }, + { + "epoch": 2.727212826707665, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.911264419555664, + "learning_rate": 1e-06, + "loss": 0.7628, + "mean_token_accuracy": 0.748336672782898, + "num_tokens": 642613641.0, + "step": 24834 + }, + { + "epoch": 2.727322644410279, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2212953567504883, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7244325280189514, + "num_tokens": 642640041.0, + "step": 24835 + }, + { + "epoch": 2.7274324621128927, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.735631227493286, + "learning_rate": 1e-06, + "loss": 0.741, + "mean_token_accuracy": 0.7630082964897156, + "num_tokens": 642658862.0, + "step": 24836 + }, + { + "epoch": 2.7275422798155065, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4682071208953857, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7382918000221252, + "num_tokens": 642683284.0, + "step": 24837 + }, + { + "epoch": 2.7276520975181198, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1133158206939697, + "learning_rate": 1e-06, + "loss": 0.8048, + "mean_token_accuracy": 0.744008481502533, + "num_tokens": 642711859.0, + "step": 24838 + }, + { + "epoch": 2.7277619152207335, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.005021572113037, + "learning_rate": 1e-06, + "loss": 0.8959, + "mean_token_accuracy": 0.7175322771072388, + "num_tokens": 642749230.0, + "step": 24839 + }, + { + "epoch": 2.7278717329233473, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3556864261627197, + "learning_rate": 1e-06, + "loss": 0.904, + "mean_token_accuracy": 0.7173733711242676, + "num_tokens": 642775552.0, + "step": 24840 + }, + { + "epoch": 2.727981550625961, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.556077480316162, + "learning_rate": 1e-06, + "loss": 0.7856, + "mean_token_accuracy": 0.7539838552474976, + "num_tokens": 642797175.0, + "step": 24841 + }, + { + "epoch": 2.728091368328575, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3783159255981445, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7282878756523132, + "num_tokens": 642823312.0, + "step": 24842 + }, + { + "epoch": 2.728201186031188, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.287243604660034, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7248870730400085, + "num_tokens": 642851144.0, + "step": 24843 + }, + { + "epoch": 2.728311003733802, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3376219272613525, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7377797365188599, + "num_tokens": 642878449.0, + "step": 24844 + }, + { + "epoch": 2.7284208214364156, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4873740673065186, + "learning_rate": 1e-06, + "loss": 0.764, + "mean_token_accuracy": 0.7546665668487549, + "num_tokens": 642900043.0, + "step": 24845 + }, + { + "epoch": 2.728530639139029, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0629067420959473, + "learning_rate": 1e-06, + "loss": 0.9321, + "mean_token_accuracy": 0.7073684930801392, + "num_tokens": 642933975.0, + "step": 24846 + }, + { + "epoch": 2.728640456841643, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.553635835647583, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.7376159429550171, + "num_tokens": 642957037.0, + "step": 24847 + }, + { + "epoch": 2.7287502745442564, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.270307779312134, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7386106252670288, + "num_tokens": 642983748.0, + "step": 24848 + }, + { + "epoch": 2.72886009224687, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0268096923828125, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7333444356918335, + "num_tokens": 643015345.0, + "step": 24849 + }, + { + "epoch": 2.728969909949484, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6540255546569824, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7290286421775818, + "num_tokens": 643037919.0, + "step": 24850 + }, + { + "epoch": 2.7290797276520973, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.157654285430908, + "learning_rate": 1e-06, + "loss": 0.8263, + "mean_token_accuracy": 0.7430790662765503, + "num_tokens": 643066006.0, + "step": 24851 + }, + { + "epoch": 2.729189545354711, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2298083305358887, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7477630376815796, + "num_tokens": 643092663.0, + "step": 24852 + }, + { + "epoch": 2.729299363057325, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.224259853363037, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7282085418701172, + "num_tokens": 643122114.0, + "step": 24853 + }, + { + "epoch": 2.7294091807599385, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5216684341430664, + "learning_rate": 1e-06, + "loss": 0.8319, + "mean_token_accuracy": 0.7407110333442688, + "num_tokens": 643143373.0, + "step": 24854 + }, + { + "epoch": 2.7295189984625523, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.324833869934082, + "learning_rate": 1e-06, + "loss": 0.7836, + "mean_token_accuracy": 0.7497963309288025, + "num_tokens": 643166868.0, + "step": 24855 + }, + { + "epoch": 2.7296288161651656, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3721470832824707, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7222427129745483, + "num_tokens": 643195898.0, + "step": 24856 + }, + { + "epoch": 2.7297386338677794, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0968475341796875, + "learning_rate": 1e-06, + "loss": 0.9223, + "mean_token_accuracy": 0.7067883610725403, + "num_tokens": 643228559.0, + "step": 24857 + }, + { + "epoch": 2.729848451570393, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.307422161102295, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7503275275230408, + "num_tokens": 643254623.0, + "step": 24858 + }, + { + "epoch": 2.729958269273007, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.356691837310791, + "learning_rate": 1e-06, + "loss": 0.8066, + "mean_token_accuracy": 0.7454441785812378, + "num_tokens": 643280531.0, + "step": 24859 + }, + { + "epoch": 2.7300680869756206, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.709747791290283, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7396916747093201, + "num_tokens": 643300152.0, + "step": 24860 + }, + { + "epoch": 2.730177904678234, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.214370012283325, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7322825193405151, + "num_tokens": 643327754.0, + "step": 24861 + }, + { + "epoch": 2.7302877223808477, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5234954357147217, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7397149801254272, + "num_tokens": 643352108.0, + "step": 24862 + }, + { + "epoch": 2.7303975400834615, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.193286180496216, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7286335229873657, + "num_tokens": 643380193.0, + "step": 24863 + }, + { + "epoch": 2.7305073577860752, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2098145484924316, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7139166593551636, + "num_tokens": 643409522.0, + "step": 24864 + }, + { + "epoch": 2.730617175488689, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.237399101257324, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7347333431243896, + "num_tokens": 643437342.0, + "step": 24865 + }, + { + "epoch": 2.7307269931913023, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.35400390625, + "learning_rate": 1e-06, + "loss": 0.8592, + "mean_token_accuracy": 0.7288382053375244, + "num_tokens": 643463951.0, + "step": 24866 + }, + { + "epoch": 2.730836810893916, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.405687093734741, + "learning_rate": 1e-06, + "loss": 0.7687, + "mean_token_accuracy": 0.7536643147468567, + "num_tokens": 643487187.0, + "step": 24867 + }, + { + "epoch": 2.73094662859653, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.587364435195923, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7246092557907104, + "num_tokens": 643508957.0, + "step": 24868 + }, + { + "epoch": 2.7310564462991436, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.354236364364624, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7305059432983398, + "num_tokens": 643534542.0, + "step": 24869 + }, + { + "epoch": 2.7311662640017573, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3718721866607666, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7347721457481384, + "num_tokens": 643560555.0, + "step": 24870 + }, + { + "epoch": 2.7312760817043706, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.009443759918213, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7246049046516418, + "num_tokens": 643594435.0, + "step": 24871 + }, + { + "epoch": 2.7313858994069844, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.193277359008789, + "learning_rate": 1e-06, + "loss": 0.9266, + "mean_token_accuracy": 0.710963249206543, + "num_tokens": 643623464.0, + "step": 24872 + }, + { + "epoch": 2.731495717109598, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2936506271362305, + "learning_rate": 1e-06, + "loss": 0.8549, + "mean_token_accuracy": 0.7272568941116333, + "num_tokens": 643651590.0, + "step": 24873 + }, + { + "epoch": 2.7316055348122115, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.523531436920166, + "learning_rate": 1e-06, + "loss": 0.8639, + "mean_token_accuracy": 0.7332562804222107, + "num_tokens": 643679139.0, + "step": 24874 + }, + { + "epoch": 2.7317153525148252, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4707367420196533, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.721880316734314, + "num_tokens": 643704449.0, + "step": 24875 + }, + { + "epoch": 2.731825170217439, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5648460388183594, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.73934006690979, + "num_tokens": 643727376.0, + "step": 24876 + }, + { + "epoch": 2.7319349879200527, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4697229862213135, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7262840867042542, + "num_tokens": 643751934.0, + "step": 24877 + }, + { + "epoch": 2.7320448056226665, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.335510015487671, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7319803833961487, + "num_tokens": 643778941.0, + "step": 24878 + }, + { + "epoch": 2.73215462332528, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.366290330886841, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7323093414306641, + "num_tokens": 643805336.0, + "step": 24879 + }, + { + "epoch": 2.7322644410278936, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1654045581817627, + "learning_rate": 1e-06, + "loss": 0.9609, + "mean_token_accuracy": 0.702312171459198, + "num_tokens": 643834946.0, + "step": 24880 + }, + { + "epoch": 2.7323742587305073, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.142179489135742, + "learning_rate": 1e-06, + "loss": 0.8747, + "mean_token_accuracy": 0.723203182220459, + "num_tokens": 643864791.0, + "step": 24881 + }, + { + "epoch": 2.732484076433121, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1960504055023193, + "learning_rate": 1e-06, + "loss": 0.9096, + "mean_token_accuracy": 0.7208651304244995, + "num_tokens": 643896644.0, + "step": 24882 + }, + { + "epoch": 2.732593894135735, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.363064765930176, + "learning_rate": 1e-06, + "loss": 0.8675, + "mean_token_accuracy": 0.7404983043670654, + "num_tokens": 643924896.0, + "step": 24883 + }, + { + "epoch": 2.732703711838348, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.145698308944702, + "learning_rate": 1e-06, + "loss": 0.8552, + "mean_token_accuracy": 0.742260217666626, + "num_tokens": 643955376.0, + "step": 24884 + }, + { + "epoch": 2.732813529540962, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1162798404693604, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7449302673339844, + "num_tokens": 643984504.0, + "step": 24885 + }, + { + "epoch": 2.7329233472435757, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.425555467605591, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7280392646789551, + "num_tokens": 644010345.0, + "step": 24886 + }, + { + "epoch": 2.7330331649461894, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.458979368209839, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7346764802932739, + "num_tokens": 644034153.0, + "step": 24887 + }, + { + "epoch": 2.733142982648803, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3290271759033203, + "learning_rate": 1e-06, + "loss": 0.8127, + "mean_token_accuracy": 0.7341626286506653, + "num_tokens": 644059787.0, + "step": 24888 + }, + { + "epoch": 2.7332528003514165, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4905974864959717, + "learning_rate": 1e-06, + "loss": 0.8056, + "mean_token_accuracy": 0.7466416954994202, + "num_tokens": 644082650.0, + "step": 24889 + }, + { + "epoch": 2.7333626180540302, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4546658992767334, + "learning_rate": 1e-06, + "loss": 0.787, + "mean_token_accuracy": 0.7489143013954163, + "num_tokens": 644106614.0, + "step": 24890 + }, + { + "epoch": 2.733472435756644, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.502739906311035, + "learning_rate": 1e-06, + "loss": 0.8867, + "mean_token_accuracy": 0.7230914831161499, + "num_tokens": 644130761.0, + "step": 24891 + }, + { + "epoch": 2.7335822534592578, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5104143619537354, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7474961280822754, + "num_tokens": 644154838.0, + "step": 24892 + }, + { + "epoch": 2.7336920711618715, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4968156814575195, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7432472109794617, + "num_tokens": 644179150.0, + "step": 24893 + }, + { + "epoch": 2.733801888864485, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.211977243423462, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7408068180084229, + "num_tokens": 644206990.0, + "step": 24894 + }, + { + "epoch": 2.7339117065670986, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.48359751701355, + "learning_rate": 1e-06, + "loss": 0.9423, + "mean_token_accuracy": 0.7126247882843018, + "num_tokens": 644233162.0, + "step": 24895 + }, + { + "epoch": 2.7340215242697123, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5036563873291016, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7097350358963013, + "num_tokens": 644256968.0, + "step": 24896 + }, + { + "epoch": 2.734131341972326, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2776296138763428, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7248569130897522, + "num_tokens": 644284581.0, + "step": 24897 + }, + { + "epoch": 2.73424115967494, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5909805297851562, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.733173131942749, + "num_tokens": 644308226.0, + "step": 24898 + }, + { + "epoch": 2.734350977377553, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1578316688537598, + "learning_rate": 1e-06, + "loss": 0.9309, + "mean_token_accuracy": 0.7150278091430664, + "num_tokens": 644340434.0, + "step": 24899 + }, + { + "epoch": 2.734460795080167, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.238206386566162, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.7350113987922668, + "num_tokens": 644371815.0, + "step": 24900 + }, + { + "epoch": 2.7345706127827807, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3328113555908203, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7297995090484619, + "num_tokens": 644396875.0, + "step": 24901 + }, + { + "epoch": 2.734680430485394, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.283413887023926, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7249099612236023, + "num_tokens": 644423665.0, + "step": 24902 + }, + { + "epoch": 2.7347902481880078, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.13912034034729, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7167055606842041, + "num_tokens": 644455080.0, + "step": 24903 + }, + { + "epoch": 2.7349000658906215, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.48317289352417, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7468153238296509, + "num_tokens": 644478301.0, + "step": 24904 + }, + { + "epoch": 2.7350098835932353, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5940752029418945, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7400785684585571, + "num_tokens": 644500203.0, + "step": 24905 + }, + { + "epoch": 2.735119701295849, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2776856422424316, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7211695909500122, + "num_tokens": 644528204.0, + "step": 24906 + }, + { + "epoch": 2.7352295189984623, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.760894536972046, + "learning_rate": 1e-06, + "loss": 0.7737, + "mean_token_accuracy": 0.7508081793785095, + "num_tokens": 644547240.0, + "step": 24907 + }, + { + "epoch": 2.735339336701076, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.552751064300537, + "learning_rate": 1e-06, + "loss": 0.794, + "mean_token_accuracy": 0.741564154624939, + "num_tokens": 644570855.0, + "step": 24908 + }, + { + "epoch": 2.73544915440369, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.446648359298706, + "learning_rate": 1e-06, + "loss": 0.9304, + "mean_token_accuracy": 0.7111760377883911, + "num_tokens": 644595127.0, + "step": 24909 + }, + { + "epoch": 2.7355589721063036, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1342499256134033, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7345086336135864, + "num_tokens": 644625702.0, + "step": 24910 + }, + { + "epoch": 2.7356687898089174, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6077938079833984, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7179428935050964, + "num_tokens": 644649169.0, + "step": 24911 + }, + { + "epoch": 2.7357786075115307, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.715425729751587, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7477411031723022, + "num_tokens": 644670190.0, + "step": 24912 + }, + { + "epoch": 2.7358884252141444, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.342637777328491, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7292299270629883, + "num_tokens": 644694420.0, + "step": 24913 + }, + { + "epoch": 2.735998242916758, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4950125217437744, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7239751815795898, + "num_tokens": 644717165.0, + "step": 24914 + }, + { + "epoch": 2.736108060619372, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2386021614074707, + "learning_rate": 1e-06, + "loss": 0.8191, + "mean_token_accuracy": 0.7423512935638428, + "num_tokens": 644745186.0, + "step": 24915 + }, + { + "epoch": 2.7362178783219857, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2872958183288574, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7192535400390625, + "num_tokens": 644773772.0, + "step": 24916 + }, + { + "epoch": 2.736327696024599, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4152321815490723, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7188270688056946, + "num_tokens": 644800001.0, + "step": 24917 + }, + { + "epoch": 2.736437513727213, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3951897621154785, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7237526178359985, + "num_tokens": 644825060.0, + "step": 24918 + }, + { + "epoch": 2.7365473314298265, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.489753484725952, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7351175546646118, + "num_tokens": 644848268.0, + "step": 24919 + }, + { + "epoch": 2.7366571491324403, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.690640687942505, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7405024766921997, + "num_tokens": 644868479.0, + "step": 24920 + }, + { + "epoch": 2.736766966835054, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.848032832145691, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7279118299484253, + "num_tokens": 644904886.0, + "step": 24921 + }, + { + "epoch": 2.7368767845376674, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.992722749710083, + "learning_rate": 1e-06, + "loss": 0.9214, + "mean_token_accuracy": 0.7119637727737427, + "num_tokens": 644942507.0, + "step": 24922 + }, + { + "epoch": 2.736986602240281, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7209832668304443, + "learning_rate": 1e-06, + "loss": 0.7431, + "mean_token_accuracy": 0.7607342600822449, + "num_tokens": 644960996.0, + "step": 24923 + }, + { + "epoch": 2.737096419942895, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2188682556152344, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7339529991149902, + "num_tokens": 644991209.0, + "step": 24924 + }, + { + "epoch": 2.737206237645508, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2254960536956787, + "learning_rate": 1e-06, + "loss": 0.9586, + "mean_token_accuracy": 0.7130460143089294, + "num_tokens": 645021253.0, + "step": 24925 + }, + { + "epoch": 2.7373160553481224, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4481914043426514, + "learning_rate": 1e-06, + "loss": 0.7926, + "mean_token_accuracy": 0.7486454844474792, + "num_tokens": 645045743.0, + "step": 24926 + }, + { + "epoch": 2.7374258730507357, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.44000506401062, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7301032543182373, + "num_tokens": 645070358.0, + "step": 24927 + }, + { + "epoch": 2.7375356907533495, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.486773729324341, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.7495317459106445, + "num_tokens": 645093569.0, + "step": 24928 + }, + { + "epoch": 2.737645508455963, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2073452472686768, + "learning_rate": 1e-06, + "loss": 0.8858, + "mean_token_accuracy": 0.7190844416618347, + "num_tokens": 645123189.0, + "step": 24929 + }, + { + "epoch": 2.7377553261585765, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3345112800598145, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7418384552001953, + "num_tokens": 645148340.0, + "step": 24930 + }, + { + "epoch": 2.7378651438611903, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6427500247955322, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7521442770957947, + "num_tokens": 645170493.0, + "step": 24931 + }, + { + "epoch": 2.737974961563804, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.437830924987793, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7205784916877747, + "num_tokens": 645196556.0, + "step": 24932 + }, + { + "epoch": 2.738084779266418, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.376906394958496, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7272416353225708, + "num_tokens": 645222041.0, + "step": 24933 + }, + { + "epoch": 2.7381945969690316, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5182089805603027, + "learning_rate": 1e-06, + "loss": 0.759, + "mean_token_accuracy": 0.7509632706642151, + "num_tokens": 645243118.0, + "step": 24934 + }, + { + "epoch": 2.738304414671645, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 3.032546281814575, + "learning_rate": 1e-06, + "loss": 0.8085, + "mean_token_accuracy": 0.7365089654922485, + "num_tokens": 645261831.0, + "step": 24935 + }, + { + "epoch": 2.7384142323742586, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3480021953582764, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7144324779510498, + "num_tokens": 645289555.0, + "step": 24936 + }, + { + "epoch": 2.7385240500768724, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.605733871459961, + "learning_rate": 1e-06, + "loss": 0.7881, + "mean_token_accuracy": 0.7438393831253052, + "num_tokens": 645311314.0, + "step": 24937 + }, + { + "epoch": 2.738633867779486, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3114171028137207, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7134194374084473, + "num_tokens": 645341909.0, + "step": 24938 + }, + { + "epoch": 2.7387436854821, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.591304302215576, + "learning_rate": 1e-06, + "loss": 0.9151, + "mean_token_accuracy": 0.7168125510215759, + "num_tokens": 645366844.0, + "step": 24939 + }, + { + "epoch": 2.738853503184713, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6245882511138916, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7260711193084717, + "num_tokens": 645389199.0, + "step": 24940 + }, + { + "epoch": 2.738963320887327, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1151416301727295, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.72528076171875, + "num_tokens": 645423038.0, + "step": 24941 + }, + { + "epoch": 2.7390731385899407, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.226627826690674, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7463164925575256, + "num_tokens": 645451731.0, + "step": 24942 + }, + { + "epoch": 2.7391829562925545, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2485406398773193, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7301877737045288, + "num_tokens": 645478334.0, + "step": 24943 + }, + { + "epoch": 2.7392927739951682, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.426440715789795, + "learning_rate": 1e-06, + "loss": 0.8599, + "mean_token_accuracy": 0.7294892072677612, + "num_tokens": 645502011.0, + "step": 24944 + }, + { + "epoch": 2.7394025916977816, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4154114723205566, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7296769618988037, + "num_tokens": 645526537.0, + "step": 24945 + }, + { + "epoch": 2.7395124094003953, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.126279592514038, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.764085054397583, + "num_tokens": 645555576.0, + "step": 24946 + }, + { + "epoch": 2.739622227103009, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.599379777908325, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7362058162689209, + "num_tokens": 645579414.0, + "step": 24947 + }, + { + "epoch": 2.739732044805623, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6840476989746094, + "learning_rate": 1e-06, + "loss": 0.7746, + "mean_token_accuracy": 0.7443327307701111, + "num_tokens": 645600487.0, + "step": 24948 + }, + { + "epoch": 2.7398418625082366, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2488820552825928, + "learning_rate": 1e-06, + "loss": 0.8665, + "mean_token_accuracy": 0.7251266837120056, + "num_tokens": 645627786.0, + "step": 24949 + }, + { + "epoch": 2.73995168021085, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.610933303833008, + "learning_rate": 1e-06, + "loss": 0.8121, + "mean_token_accuracy": 0.7419484853744507, + "num_tokens": 645650574.0, + "step": 24950 + }, + { + "epoch": 2.7400614979134637, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.057610273361206, + "learning_rate": 1e-06, + "loss": 0.9429, + "mean_token_accuracy": 0.7057051062583923, + "num_tokens": 645687087.0, + "step": 24951 + }, + { + "epoch": 2.7401713156160774, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.417551040649414, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.737598180770874, + "num_tokens": 645710254.0, + "step": 24952 + }, + { + "epoch": 2.7402811333186907, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.169393301010132, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7208300828933716, + "num_tokens": 645740072.0, + "step": 24953 + }, + { + "epoch": 2.7403909510213045, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.9410955905914307, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.742165744304657, + "num_tokens": 645758717.0, + "step": 24954 + }, + { + "epoch": 2.7405007687239182, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3115625381469727, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7194513082504272, + "num_tokens": 645783376.0, + "step": 24955 + }, + { + "epoch": 2.740610586426532, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4331228733062744, + "learning_rate": 1e-06, + "loss": 0.9056, + "mean_token_accuracy": 0.7204594612121582, + "num_tokens": 645810540.0, + "step": 24956 + }, + { + "epoch": 2.7407204041291457, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4205944538116455, + "learning_rate": 1e-06, + "loss": 0.8973, + "mean_token_accuracy": 0.7288251519203186, + "num_tokens": 645836400.0, + "step": 24957 + }, + { + "epoch": 2.740830221831759, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1401429176330566, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7126786112785339, + "num_tokens": 645865217.0, + "step": 24958 + }, + { + "epoch": 2.740940039534373, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0713188648223877, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7162960767745972, + "num_tokens": 645897124.0, + "step": 24959 + }, + { + "epoch": 2.7410498572369866, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.56038761138916, + "learning_rate": 1e-06, + "loss": 0.7465, + "mean_token_accuracy": 0.7583973407745361, + "num_tokens": 645919008.0, + "step": 24960 + }, + { + "epoch": 2.7411596749396003, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5226082801818848, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7281287908554077, + "num_tokens": 645944119.0, + "step": 24961 + }, + { + "epoch": 2.741269492642214, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.162435531616211, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.7119473218917847, + "num_tokens": 645976846.0, + "step": 24962 + }, + { + "epoch": 2.7413793103448274, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4657883644104004, + "learning_rate": 1e-06, + "loss": 0.8233, + "mean_token_accuracy": 0.7408207654953003, + "num_tokens": 646000486.0, + "step": 24963 + }, + { + "epoch": 2.741489128047441, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3454623222351074, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7104669809341431, + "num_tokens": 646028394.0, + "step": 24964 + }, + { + "epoch": 2.741598945750055, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1143717765808105, + "learning_rate": 1e-06, + "loss": 0.9276, + "mean_token_accuracy": 0.7164372205734253, + "num_tokens": 646058019.0, + "step": 24965 + }, + { + "epoch": 2.7417087634526687, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.481330394744873, + "learning_rate": 1e-06, + "loss": 0.9186, + "mean_token_accuracy": 0.7194902300834656, + "num_tokens": 646082838.0, + "step": 24966 + }, + { + "epoch": 2.7418185811552824, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1906638145446777, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7415732145309448, + "num_tokens": 646112665.0, + "step": 24967 + }, + { + "epoch": 2.7419283988578957, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1154398918151855, + "learning_rate": 1e-06, + "loss": 0.89, + "mean_token_accuracy": 0.7187942862510681, + "num_tokens": 646143820.0, + "step": 24968 + }, + { + "epoch": 2.7420382165605095, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.307210922241211, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7191973924636841, + "num_tokens": 646170636.0, + "step": 24969 + }, + { + "epoch": 2.7421480342631233, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6006181240081787, + "learning_rate": 1e-06, + "loss": 0.883, + "mean_token_accuracy": 0.7220185399055481, + "num_tokens": 646192243.0, + "step": 24970 + }, + { + "epoch": 2.742257851965737, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.160060167312622, + "learning_rate": 1e-06, + "loss": 0.9415, + "mean_token_accuracy": 0.7100280523300171, + "num_tokens": 646224106.0, + "step": 24971 + }, + { + "epoch": 2.7423676696683508, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2601265907287598, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7188245058059692, + "num_tokens": 646251003.0, + "step": 24972 + }, + { + "epoch": 2.742477487370964, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.298189878463745, + "learning_rate": 1e-06, + "loss": 0.9227, + "mean_token_accuracy": 0.7109919786453247, + "num_tokens": 646278161.0, + "step": 24973 + }, + { + "epoch": 2.742587305073578, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.560075283050537, + "learning_rate": 1e-06, + "loss": 0.7363, + "mean_token_accuracy": 0.7582552433013916, + "num_tokens": 646300023.0, + "step": 24974 + }, + { + "epoch": 2.7426971227761916, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1725783348083496, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7173442840576172, + "num_tokens": 646330269.0, + "step": 24975 + }, + { + "epoch": 2.742806940478805, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2402257919311523, + "learning_rate": 1e-06, + "loss": 0.9058, + "mean_token_accuracy": 0.7167041301727295, + "num_tokens": 646359177.0, + "step": 24976 + }, + { + "epoch": 2.742916758181419, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.745652675628662, + "learning_rate": 1e-06, + "loss": 0.832, + "mean_token_accuracy": 0.7375220656394958, + "num_tokens": 646378780.0, + "step": 24977 + }, + { + "epoch": 2.7430265758840324, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2494001388549805, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7386505603790283, + "num_tokens": 646408380.0, + "step": 24978 + }, + { + "epoch": 2.743136393586646, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2492146492004395, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.7275101542472839, + "num_tokens": 646436526.0, + "step": 24979 + }, + { + "epoch": 2.74324621128926, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6815216541290283, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7482459545135498, + "num_tokens": 646457486.0, + "step": 24980 + }, + { + "epoch": 2.7433560289918733, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6304588317871094, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7356607913970947, + "num_tokens": 646478215.0, + "step": 24981 + }, + { + "epoch": 2.743465846694487, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3329520225524902, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.731619119644165, + "num_tokens": 646502529.0, + "step": 24982 + }, + { + "epoch": 2.7435756643971008, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.214179277420044, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.7208670377731323, + "num_tokens": 646532780.0, + "step": 24983 + }, + { + "epoch": 2.7436854820997145, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.522162437438965, + "learning_rate": 1e-06, + "loss": 0.8079, + "mean_token_accuracy": 0.7459148168563843, + "num_tokens": 646555900.0, + "step": 24984 + }, + { + "epoch": 2.7437952998023283, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.9622021913528442, + "learning_rate": 1e-06, + "loss": 0.9492, + "mean_token_accuracy": 0.7050454616546631, + "num_tokens": 646590148.0, + "step": 24985 + }, + { + "epoch": 2.7439051175049416, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4445457458496094, + "learning_rate": 1e-06, + "loss": 0.7889, + "mean_token_accuracy": 0.7498247623443604, + "num_tokens": 646615531.0, + "step": 24986 + }, + { + "epoch": 2.7440149352075554, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.353337049484253, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7427443861961365, + "num_tokens": 646640800.0, + "step": 24987 + }, + { + "epoch": 2.744124752910169, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6068930625915527, + "learning_rate": 1e-06, + "loss": 0.8329, + "mean_token_accuracy": 0.731158971786499, + "num_tokens": 646664533.0, + "step": 24988 + }, + { + "epoch": 2.744234570612783, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.482342004776001, + "learning_rate": 1e-06, + "loss": 0.78, + "mean_token_accuracy": 0.7477964162826538, + "num_tokens": 646687570.0, + "step": 24989 + }, + { + "epoch": 2.7443443883153966, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3518784046173096, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7173566818237305, + "num_tokens": 646713050.0, + "step": 24990 + }, + { + "epoch": 2.74445420601801, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4974682331085205, + "learning_rate": 1e-06, + "loss": 0.7872, + "mean_token_accuracy": 0.7466062307357788, + "num_tokens": 646736710.0, + "step": 24991 + }, + { + "epoch": 2.7445640237206237, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.505202054977417, + "learning_rate": 1e-06, + "loss": 0.834, + "mean_token_accuracy": 0.7372452020645142, + "num_tokens": 646760872.0, + "step": 24992 + }, + { + "epoch": 2.7446738414232374, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4638431072235107, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.7513138651847839, + "num_tokens": 646783717.0, + "step": 24993 + }, + { + "epoch": 2.744783659125851, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4067330360412598, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7215348482131958, + "num_tokens": 646810412.0, + "step": 24994 + }, + { + "epoch": 2.744893476828465, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6693973541259766, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7266016006469727, + "num_tokens": 646833771.0, + "step": 24995 + }, + { + "epoch": 2.7450032945310783, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4042654037475586, + "learning_rate": 1e-06, + "loss": 0.8427, + "mean_token_accuracy": 0.7366331219673157, + "num_tokens": 646858231.0, + "step": 24996 + }, + { + "epoch": 2.745113112233692, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.741405963897705, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7324813008308411, + "num_tokens": 646879355.0, + "step": 24997 + }, + { + "epoch": 2.745222929936306, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2051634788513184, + "learning_rate": 1e-06, + "loss": 0.9711, + "mean_token_accuracy": 0.6999967694282532, + "num_tokens": 646910891.0, + "step": 24998 + }, + { + "epoch": 2.7453327476389195, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2655749320983887, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7423528432846069, + "num_tokens": 646938572.0, + "step": 24999 + }, + { + "epoch": 2.7454425653415333, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2714931964874268, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.742103636264801, + "num_tokens": 646964803.0, + "step": 25000 + }, + { + "epoch": 2.7455523830441466, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.079284906387329, + "learning_rate": 1e-06, + "loss": 0.889, + "mean_token_accuracy": 0.7267599105834961, + "num_tokens": 646997019.0, + "step": 25001 + }, + { + "epoch": 2.7456622007467604, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4903564453125, + "learning_rate": 1e-06, + "loss": 0.7986, + "mean_token_accuracy": 0.7469280958175659, + "num_tokens": 647020212.0, + "step": 25002 + }, + { + "epoch": 2.745772018449374, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1739511489868164, + "learning_rate": 1e-06, + "loss": 0.9132, + "mean_token_accuracy": 0.7170488238334656, + "num_tokens": 647049845.0, + "step": 25003 + }, + { + "epoch": 2.7458818361519874, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2979164123535156, + "learning_rate": 1e-06, + "loss": 0.9643, + "mean_token_accuracy": 0.7132928371429443, + "num_tokens": 647078205.0, + "step": 25004 + }, + { + "epoch": 2.745991653854601, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.993180274963379, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7176154255867004, + "num_tokens": 647113634.0, + "step": 25005 + }, + { + "epoch": 2.746101471557215, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6489017009735107, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7304449677467346, + "num_tokens": 647136875.0, + "step": 25006 + }, + { + "epoch": 2.7462112892598287, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.325669288635254, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7345545291900635, + "num_tokens": 647162932.0, + "step": 25007 + }, + { + "epoch": 2.7463211069624425, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2582499980926514, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7156845927238464, + "num_tokens": 647190846.0, + "step": 25008 + }, + { + "epoch": 2.746430924665056, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.221848487854004, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7258216142654419, + "num_tokens": 647219098.0, + "step": 25009 + }, + { + "epoch": 2.7465407423676695, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4582247734069824, + "learning_rate": 1e-06, + "loss": 0.8332, + "mean_token_accuracy": 0.725949764251709, + "num_tokens": 647243920.0, + "step": 25010 + }, + { + "epoch": 2.7466505600702833, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.52729868888855, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.7322920560836792, + "num_tokens": 647268324.0, + "step": 25011 + }, + { + "epoch": 2.746760377772897, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4747509956359863, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7344428896903992, + "num_tokens": 647291575.0, + "step": 25012 + }, + { + "epoch": 2.746870195475511, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4203310012817383, + "learning_rate": 1e-06, + "loss": 0.7548, + "mean_token_accuracy": 0.7557694911956787, + "num_tokens": 647314862.0, + "step": 25013 + }, + { + "epoch": 2.746980013178124, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3634450435638428, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7272240519523621, + "num_tokens": 647338860.0, + "step": 25014 + }, + { + "epoch": 2.747089830880738, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.162757635116577, + "learning_rate": 1e-06, + "loss": 0.8763, + "mean_token_accuracy": 0.7290060520172119, + "num_tokens": 647367882.0, + "step": 25015 + }, + { + "epoch": 2.7471996485833516, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3206818103790283, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7124893665313721, + "num_tokens": 647394841.0, + "step": 25016 + }, + { + "epoch": 2.7473094662859654, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.338300943374634, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7327070236206055, + "num_tokens": 647419836.0, + "step": 25017 + }, + { + "epoch": 2.747419283988579, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4504904747009277, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7372386455535889, + "num_tokens": 647445445.0, + "step": 25018 + }, + { + "epoch": 2.7475291016911925, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1546618938446045, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7386432886123657, + "num_tokens": 647476707.0, + "step": 25019 + }, + { + "epoch": 2.7476389193938062, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 3.2076644897460938, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7458381652832031, + "num_tokens": 647493810.0, + "step": 25020 + }, + { + "epoch": 2.74774873709642, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.24873423576355, + "learning_rate": 1e-06, + "loss": 0.8, + "mean_token_accuracy": 0.7466550469398499, + "num_tokens": 647521427.0, + "step": 25021 + }, + { + "epoch": 2.7478585547990337, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.235661745071411, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.72796231508255, + "num_tokens": 647549213.0, + "step": 25022 + }, + { + "epoch": 2.7479683725016475, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0795552730560303, + "learning_rate": 1e-06, + "loss": 0.9884, + "mean_token_accuracy": 0.6926294565200806, + "num_tokens": 647581008.0, + "step": 25023 + }, + { + "epoch": 2.748078190204261, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3479340076446533, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7404998540878296, + "num_tokens": 647605790.0, + "step": 25024 + }, + { + "epoch": 2.7481880079068746, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.266422986984253, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7341550588607788, + "num_tokens": 647631816.0, + "step": 25025 + }, + { + "epoch": 2.7482978256094883, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.475895404815674, + "learning_rate": 1e-06, + "loss": 0.7999, + "mean_token_accuracy": 0.7452372312545776, + "num_tokens": 647656111.0, + "step": 25026 + }, + { + "epoch": 2.7484076433121016, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.639521837234497, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7349201440811157, + "num_tokens": 647677258.0, + "step": 25027 + }, + { + "epoch": 2.748517461014716, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.813753843307495, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7351324558258057, + "num_tokens": 647697016.0, + "step": 25028 + }, + { + "epoch": 2.748627278717329, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.378960371017456, + "learning_rate": 1e-06, + "loss": 0.8704, + "mean_token_accuracy": 0.7253621816635132, + "num_tokens": 647723095.0, + "step": 25029 + }, + { + "epoch": 2.748737096419943, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4122672080993652, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7211463451385498, + "num_tokens": 647747672.0, + "step": 25030 + }, + { + "epoch": 2.7488469141225567, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4076168537139893, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7304798364639282, + "num_tokens": 647774007.0, + "step": 25031 + }, + { + "epoch": 2.74895673182517, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2736005783081055, + "learning_rate": 1e-06, + "loss": 0.9195, + "mean_token_accuracy": 0.7222568392753601, + "num_tokens": 647804272.0, + "step": 25032 + }, + { + "epoch": 2.7490665495277837, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5288283824920654, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7416304349899292, + "num_tokens": 647827428.0, + "step": 25033 + }, + { + "epoch": 2.7491763672303975, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6439621448516846, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.734779417514801, + "num_tokens": 647849150.0, + "step": 25034 + }, + { + "epoch": 2.7492861849330112, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2777099609375, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7474410533905029, + "num_tokens": 647876113.0, + "step": 25035 + }, + { + "epoch": 2.749396002635625, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2772538661956787, + "learning_rate": 1e-06, + "loss": 0.8984, + "mean_token_accuracy": 0.7194147109985352, + "num_tokens": 647903846.0, + "step": 25036 + }, + { + "epoch": 2.7495058203382383, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.9989463090896606, + "learning_rate": 1e-06, + "loss": 0.9464, + "mean_token_accuracy": 0.7067093849182129, + "num_tokens": 647938368.0, + "step": 25037 + }, + { + "epoch": 2.749615638040852, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.074333906173706, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7205950021743774, + "num_tokens": 647970228.0, + "step": 25038 + }, + { + "epoch": 2.749725455743466, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.171140670776367, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.722015917301178, + "num_tokens": 647999975.0, + "step": 25039 + }, + { + "epoch": 2.7498352734460796, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.444301128387451, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7336561679840088, + "num_tokens": 648025289.0, + "step": 25040 + }, + { + "epoch": 2.7499450911486933, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2009224891662598, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7338262796401978, + "num_tokens": 648056141.0, + "step": 25041 + }, + { + "epoch": 2.7500549088513067, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1353955268859863, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7330600619316101, + "num_tokens": 648087984.0, + "step": 25042 + }, + { + "epoch": 2.7501647265539204, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.12626576423645, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7426751852035522, + "num_tokens": 648118362.0, + "step": 25043 + }, + { + "epoch": 2.750274544256534, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.316622495651245, + "learning_rate": 1e-06, + "loss": 0.7563, + "mean_token_accuracy": 0.7576324939727783, + "num_tokens": 648142816.0, + "step": 25044 + }, + { + "epoch": 2.750384361959148, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2305166721343994, + "learning_rate": 1e-06, + "loss": 0.7643, + "mean_token_accuracy": 0.7536170482635498, + "num_tokens": 648170513.0, + "step": 25045 + }, + { + "epoch": 2.7504941796617617, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3436925411224365, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7380615472793579, + "num_tokens": 648195166.0, + "step": 25046 + }, + { + "epoch": 2.750603997364375, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2859928607940674, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.728145956993103, + "num_tokens": 648222767.0, + "step": 25047 + }, + { + "epoch": 2.7507138150669888, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.025296926498413, + "learning_rate": 1e-06, + "loss": 0.8811, + "mean_token_accuracy": 0.7269959449768066, + "num_tokens": 648255722.0, + "step": 25048 + }, + { + "epoch": 2.7508236327696025, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4505598545074463, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7462665438652039, + "num_tokens": 648280043.0, + "step": 25049 + }, + { + "epoch": 2.7509334504722163, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2917401790618896, + "learning_rate": 1e-06, + "loss": 0.9401, + "mean_token_accuracy": 0.7060353755950928, + "num_tokens": 648307814.0, + "step": 25050 + }, + { + "epoch": 2.75104326817483, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2811882495880127, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7173072099685669, + "num_tokens": 648336217.0, + "step": 25051 + }, + { + "epoch": 2.7511530858774433, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2393784523010254, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7473098039627075, + "num_tokens": 648362643.0, + "step": 25052 + }, + { + "epoch": 2.751262903580057, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2925145626068115, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7179005146026611, + "num_tokens": 648389974.0, + "step": 25053 + }, + { + "epoch": 2.751372721282671, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.246695041656494, + "learning_rate": 1e-06, + "loss": 0.9349, + "mean_token_accuracy": 0.7097388505935669, + "num_tokens": 648418638.0, + "step": 25054 + }, + { + "epoch": 2.751482538985284, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5472445487976074, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7348113059997559, + "num_tokens": 648440188.0, + "step": 25055 + }, + { + "epoch": 2.7515923566878984, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.455493211746216, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7523733973503113, + "num_tokens": 648463041.0, + "step": 25056 + }, + { + "epoch": 2.7517021743905117, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.345728874206543, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7448515892028809, + "num_tokens": 648489706.0, + "step": 25057 + }, + { + "epoch": 2.7518119920931254, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.454618215560913, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7258659601211548, + "num_tokens": 648513222.0, + "step": 25058 + }, + { + "epoch": 2.751921809795739, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.514709711074829, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7444772720336914, + "num_tokens": 648538625.0, + "step": 25059 + }, + { + "epoch": 2.7520316274983525, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4782986640930176, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7351794242858887, + "num_tokens": 648562289.0, + "step": 25060 + }, + { + "epoch": 2.7521414452009663, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0726852416992188, + "learning_rate": 1e-06, + "loss": 0.9258, + "mean_token_accuracy": 0.7154099345207214, + "num_tokens": 648595217.0, + "step": 25061 + }, + { + "epoch": 2.75225126290358, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4122421741485596, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7298855185508728, + "num_tokens": 648620006.0, + "step": 25062 + }, + { + "epoch": 2.7523610806061938, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.203947067260742, + "learning_rate": 1e-06, + "loss": 0.9703, + "mean_token_accuracy": 0.6934396028518677, + "num_tokens": 648648475.0, + "step": 25063 + }, + { + "epoch": 2.7524708983088075, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.404829978942871, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7330154180526733, + "num_tokens": 648672852.0, + "step": 25064 + }, + { + "epoch": 2.752580716011421, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4029040336608887, + "learning_rate": 1e-06, + "loss": 0.7781, + "mean_token_accuracy": 0.7567130327224731, + "num_tokens": 648698109.0, + "step": 25065 + }, + { + "epoch": 2.7526905337140346, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2573490142822266, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7253217101097107, + "num_tokens": 648725701.0, + "step": 25066 + }, + { + "epoch": 2.7528003514166484, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4255285263061523, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7577319145202637, + "num_tokens": 648749283.0, + "step": 25067 + }, + { + "epoch": 2.752910169119262, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.431673288345337, + "learning_rate": 1e-06, + "loss": 0.7586, + "mean_token_accuracy": 0.7615964412689209, + "num_tokens": 648771991.0, + "step": 25068 + }, + { + "epoch": 2.753019986821876, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.434758424758911, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7430002689361572, + "num_tokens": 648795677.0, + "step": 25069 + }, + { + "epoch": 2.753129804524489, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.510265588760376, + "learning_rate": 1e-06, + "loss": 0.854, + "mean_token_accuracy": 0.7333971261978149, + "num_tokens": 648818931.0, + "step": 25070 + }, + { + "epoch": 2.753239622227103, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3468987941741943, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7152073383331299, + "num_tokens": 648847331.0, + "step": 25071 + }, + { + "epoch": 2.7533494399297167, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6618447303771973, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.723091721534729, + "num_tokens": 648868817.0, + "step": 25072 + }, + { + "epoch": 2.7534592576323305, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3630990982055664, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7227579355239868, + "num_tokens": 648895614.0, + "step": 25073 + }, + { + "epoch": 2.753569075334944, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3191330432891846, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7367300987243652, + "num_tokens": 648921781.0, + "step": 25074 + }, + { + "epoch": 2.7536788930375575, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.196685791015625, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7119666337966919, + "num_tokens": 648952615.0, + "step": 25075 + }, + { + "epoch": 2.7537887107401713, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3654286861419678, + "learning_rate": 1e-06, + "loss": 0.8186, + "mean_token_accuracy": 0.7417553663253784, + "num_tokens": 648976681.0, + "step": 25076 + }, + { + "epoch": 2.753898528442785, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4864561557769775, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7383531928062439, + "num_tokens": 649001451.0, + "step": 25077 + }, + { + "epoch": 2.754008346145399, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.218977689743042, + "learning_rate": 1e-06, + "loss": 0.9083, + "mean_token_accuracy": 0.7101090550422668, + "num_tokens": 649030563.0, + "step": 25078 + }, + { + "epoch": 2.7541181638480126, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6342854499816895, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7296954989433289, + "num_tokens": 649052981.0, + "step": 25079 + }, + { + "epoch": 2.754227981550626, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2277841567993164, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7315608263015747, + "num_tokens": 649081091.0, + "step": 25080 + }, + { + "epoch": 2.7543377992532396, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4211082458496094, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7337457537651062, + "num_tokens": 649105045.0, + "step": 25081 + }, + { + "epoch": 2.7544476169558534, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.220130205154419, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7385571599006653, + "num_tokens": 649133027.0, + "step": 25082 + }, + { + "epoch": 2.7545574346584667, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.787324905395508, + "learning_rate": 1e-06, + "loss": 0.7852, + "mean_token_accuracy": 0.7452313303947449, + "num_tokens": 649151761.0, + "step": 25083 + }, + { + "epoch": 2.7546672523610805, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0706045627593994, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7256913781166077, + "num_tokens": 649183703.0, + "step": 25084 + }, + { + "epoch": 2.754777070063694, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4246082305908203, + "learning_rate": 1e-06, + "loss": 0.8895, + "mean_token_accuracy": 0.7238287925720215, + "num_tokens": 649209334.0, + "step": 25085 + }, + { + "epoch": 2.754886887766308, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3102760314941406, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.750399112701416, + "num_tokens": 649235028.0, + "step": 25086 + }, + { + "epoch": 2.7549967054689217, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4657480716705322, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7328891754150391, + "num_tokens": 649258176.0, + "step": 25087 + }, + { + "epoch": 2.755106523171535, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4946842193603516, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7353448867797852, + "num_tokens": 649281539.0, + "step": 25088 + }, + { + "epoch": 2.755216340874149, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.060500383377075, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.7003772854804993, + "num_tokens": 649312403.0, + "step": 25089 + }, + { + "epoch": 2.7553261585767626, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2675564289093018, + "learning_rate": 1e-06, + "loss": 0.9286, + "mean_token_accuracy": 0.7113670110702515, + "num_tokens": 649341352.0, + "step": 25090 + }, + { + "epoch": 2.7554359762793763, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.583489179611206, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7385785579681396, + "num_tokens": 649363178.0, + "step": 25091 + }, + { + "epoch": 2.75554579398199, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.8819832801818848, + "learning_rate": 1e-06, + "loss": 0.7612, + "mean_token_accuracy": 0.7534067630767822, + "num_tokens": 649381050.0, + "step": 25092 + }, + { + "epoch": 2.7556556116846034, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4709548950195312, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7355573773384094, + "num_tokens": 649405317.0, + "step": 25093 + }, + { + "epoch": 2.755765429387217, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.369758129119873, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7368852496147156, + "num_tokens": 649430816.0, + "step": 25094 + }, + { + "epoch": 2.755875247089831, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3662383556365967, + "learning_rate": 1e-06, + "loss": 0.7694, + "mean_token_accuracy": 0.7560372352600098, + "num_tokens": 649456318.0, + "step": 25095 + }, + { + "epoch": 2.7559850647924446, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5053904056549072, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7331634759902954, + "num_tokens": 649480182.0, + "step": 25096 + }, + { + "epoch": 2.7560948824950584, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5545432567596436, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7380211353302002, + "num_tokens": 649504133.0, + "step": 25097 + }, + { + "epoch": 2.7562047001976717, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2043306827545166, + "learning_rate": 1e-06, + "loss": 0.9007, + "mean_token_accuracy": 0.727185070514679, + "num_tokens": 649533342.0, + "step": 25098 + }, + { + "epoch": 2.7563145179002855, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.10186505317688, + "learning_rate": 1e-06, + "loss": 0.9217, + "mean_token_accuracy": 0.7092772722244263, + "num_tokens": 649567140.0, + "step": 25099 + }, + { + "epoch": 2.7564243356028992, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3256869316101074, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.736434817314148, + "num_tokens": 649591643.0, + "step": 25100 + }, + { + "epoch": 2.756534153305513, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0695111751556396, + "learning_rate": 1e-06, + "loss": 0.925, + "mean_token_accuracy": 0.7134522199630737, + "num_tokens": 649624483.0, + "step": 25101 + }, + { + "epoch": 2.7566439710081267, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.293473720550537, + "learning_rate": 1e-06, + "loss": 0.7041, + "mean_token_accuracy": 0.7725965976715088, + "num_tokens": 649650361.0, + "step": 25102 + }, + { + "epoch": 2.75675378871074, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.651979446411133, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.7212284207344055, + "num_tokens": 649670929.0, + "step": 25103 + }, + { + "epoch": 2.756863606413354, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 1.993815302848816, + "learning_rate": 1e-06, + "loss": 0.8799, + "mean_token_accuracy": 0.72948157787323, + "num_tokens": 649704762.0, + "step": 25104 + }, + { + "epoch": 2.7569734241159676, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5209808349609375, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7115931510925293, + "num_tokens": 649728888.0, + "step": 25105 + }, + { + "epoch": 2.757083241818581, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2311007976531982, + "learning_rate": 1e-06, + "loss": 0.9662, + "mean_token_accuracy": 0.7000206708908081, + "num_tokens": 649756537.0, + "step": 25106 + }, + { + "epoch": 2.757193059521195, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3568453788757324, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.730820894241333, + "num_tokens": 649784537.0, + "step": 25107 + }, + { + "epoch": 2.7573028772238084, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.273716449737549, + "learning_rate": 1e-06, + "loss": 0.9414, + "mean_token_accuracy": 0.7061070799827576, + "num_tokens": 649812101.0, + "step": 25108 + }, + { + "epoch": 2.757412694926422, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0639405250549316, + "learning_rate": 1e-06, + "loss": 0.8042, + "mean_token_accuracy": 0.7426650524139404, + "num_tokens": 649840318.0, + "step": 25109 + }, + { + "epoch": 2.757522512629036, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3772270679473877, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7416907548904419, + "num_tokens": 649865001.0, + "step": 25110 + }, + { + "epoch": 2.7576323303316492, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4630801677703857, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7492257356643677, + "num_tokens": 649889271.0, + "step": 25111 + }, + { + "epoch": 2.757742148034263, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3033664226531982, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7220237851142883, + "num_tokens": 649916816.0, + "step": 25112 + }, + { + "epoch": 2.7578519657368767, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.458484172821045, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7141777276992798, + "num_tokens": 649941801.0, + "step": 25113 + }, + { + "epoch": 2.7579617834394905, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.547456979751587, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7332817912101746, + "num_tokens": 649964154.0, + "step": 25114 + }, + { + "epoch": 2.7580716011421043, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.156489849090576, + "learning_rate": 1e-06, + "loss": 0.8985, + "mean_token_accuracy": 0.7229579091072083, + "num_tokens": 649994431.0, + "step": 25115 + }, + { + "epoch": 2.7581814188447176, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.201612710952759, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.73814857006073, + "num_tokens": 650023704.0, + "step": 25116 + }, + { + "epoch": 2.7582912365473313, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5654489994049072, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7293184995651245, + "num_tokens": 650046884.0, + "step": 25117 + }, + { + "epoch": 2.758401054249945, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3441061973571777, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7496857643127441, + "num_tokens": 650071453.0, + "step": 25118 + }, + { + "epoch": 2.758510871952559, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2835886478424072, + "learning_rate": 1e-06, + "loss": 0.7889, + "mean_token_accuracy": 0.7491053342819214, + "num_tokens": 650096453.0, + "step": 25119 + }, + { + "epoch": 2.7586206896551726, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.488175392150879, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7295438051223755, + "num_tokens": 650120391.0, + "step": 25120 + }, + { + "epoch": 2.758730507357786, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.495016574859619, + "learning_rate": 1e-06, + "loss": 0.8609, + "mean_token_accuracy": 0.72374027967453, + "num_tokens": 650144385.0, + "step": 25121 + }, + { + "epoch": 2.7588403250603997, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.441927194595337, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7125139236450195, + "num_tokens": 650169422.0, + "step": 25122 + }, + { + "epoch": 2.7589501427630134, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.345121383666992, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.7257562875747681, + "num_tokens": 650194483.0, + "step": 25123 + }, + { + "epoch": 2.759059960465627, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4456534385681152, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.7239885926246643, + "num_tokens": 650220021.0, + "step": 25124 + }, + { + "epoch": 2.759169778168241, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2692453861236572, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7286726832389832, + "num_tokens": 650246245.0, + "step": 25125 + }, + { + "epoch": 2.7592795958708543, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.377929449081421, + "learning_rate": 1e-06, + "loss": 0.8968, + "mean_token_accuracy": 0.7198632955551147, + "num_tokens": 650272698.0, + "step": 25126 + }, + { + "epoch": 2.759389413573468, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3810980319976807, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7370784282684326, + "num_tokens": 650297928.0, + "step": 25127 + }, + { + "epoch": 2.7594992312760818, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.465121030807495, + "learning_rate": 1e-06, + "loss": 0.894, + "mean_token_accuracy": 0.7227482199668884, + "num_tokens": 650323221.0, + "step": 25128 + }, + { + "epoch": 2.7596090489786955, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.096691608428955, + "learning_rate": 1e-06, + "loss": 0.8029, + "mean_token_accuracy": 0.7430688738822937, + "num_tokens": 650354418.0, + "step": 25129 + }, + { + "epoch": 2.7597188666813093, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4448790550231934, + "learning_rate": 1e-06, + "loss": 0.7347, + "mean_token_accuracy": 0.7564724683761597, + "num_tokens": 650376365.0, + "step": 25130 + }, + { + "epoch": 2.7598286843839226, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.436532735824585, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7149389386177063, + "num_tokens": 650400045.0, + "step": 25131 + }, + { + "epoch": 2.7599385020865363, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.9633047580718994, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7385013103485107, + "num_tokens": 650418615.0, + "step": 25132 + }, + { + "epoch": 2.76004831978915, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.231207847595215, + "learning_rate": 1e-06, + "loss": 0.7743, + "mean_token_accuracy": 0.7529650926589966, + "num_tokens": 650443488.0, + "step": 25133 + }, + { + "epoch": 2.7601581374917634, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.230228900909424, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7232639789581299, + "num_tokens": 650473019.0, + "step": 25134 + }, + { + "epoch": 2.760267955194377, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.275908946990967, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7408983111381531, + "num_tokens": 650499245.0, + "step": 25135 + }, + { + "epoch": 2.760377772896991, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4742655754089355, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7329533100128174, + "num_tokens": 650522784.0, + "step": 25136 + }, + { + "epoch": 2.7604875905996047, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.596618413925171, + "learning_rate": 1e-06, + "loss": 0.8125, + "mean_token_accuracy": 0.7362143993377686, + "num_tokens": 650543810.0, + "step": 25137 + }, + { + "epoch": 2.7605974083022184, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4973392486572266, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7432951927185059, + "num_tokens": 650565417.0, + "step": 25138 + }, + { + "epoch": 2.7607072260048318, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1546454429626465, + "learning_rate": 1e-06, + "loss": 0.8607, + "mean_token_accuracy": 0.7339307069778442, + "num_tokens": 650596977.0, + "step": 25139 + }, + { + "epoch": 2.7608170437074455, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.45438814163208, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.721463680267334, + "num_tokens": 650621860.0, + "step": 25140 + }, + { + "epoch": 2.7609268614100593, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.418377637863159, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7282547354698181, + "num_tokens": 650646498.0, + "step": 25141 + }, + { + "epoch": 2.761036679112673, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.8395907878875732, + "learning_rate": 1e-06, + "loss": 0.7696, + "mean_token_accuracy": 0.7530816793441772, + "num_tokens": 650664077.0, + "step": 25142 + }, + { + "epoch": 2.761146496815287, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7175161838531494, + "learning_rate": 1e-06, + "loss": 0.8961, + "mean_token_accuracy": 0.7245747447013855, + "num_tokens": 650686151.0, + "step": 25143 + }, + { + "epoch": 2.7612563145179, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.358316659927368, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7056663036346436, + "num_tokens": 650714203.0, + "step": 25144 + }, + { + "epoch": 2.761366132220514, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7107179164886475, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7510380744934082, + "num_tokens": 650734054.0, + "step": 25145 + }, + { + "epoch": 2.7614759499231276, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4260880947113037, + "learning_rate": 1e-06, + "loss": 0.9108, + "mean_token_accuracy": 0.7112606167793274, + "num_tokens": 650761726.0, + "step": 25146 + }, + { + "epoch": 2.7615857676257414, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.562882900238037, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7355443835258484, + "num_tokens": 650784206.0, + "step": 25147 + }, + { + "epoch": 2.761695585328355, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5262017250061035, + "learning_rate": 1e-06, + "loss": 0.8588, + "mean_token_accuracy": 0.7223618030548096, + "num_tokens": 650807736.0, + "step": 25148 + }, + { + "epoch": 2.7618054030309684, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4381520748138428, + "learning_rate": 1e-06, + "loss": 0.809, + "mean_token_accuracy": 0.7381291389465332, + "num_tokens": 650830542.0, + "step": 25149 + }, + { + "epoch": 2.761915220733582, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2351224422454834, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7380837798118591, + "num_tokens": 650859543.0, + "step": 25150 + }, + { + "epoch": 2.762025038436196, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.501215934753418, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7439956068992615, + "num_tokens": 650881841.0, + "step": 25151 + }, + { + "epoch": 2.7621348561388097, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4637269973754883, + "learning_rate": 1e-06, + "loss": 0.7782, + "mean_token_accuracy": 0.7432546615600586, + "num_tokens": 650906005.0, + "step": 25152 + }, + { + "epoch": 2.7622446738414235, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2003259658813477, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7335731983184814, + "num_tokens": 650933378.0, + "step": 25153 + }, + { + "epoch": 2.762354491544037, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.410421848297119, + "learning_rate": 1e-06, + "loss": 0.9045, + "mean_token_accuracy": 0.7147300839424133, + "num_tokens": 650960849.0, + "step": 25154 + }, + { + "epoch": 2.7624643092466505, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.021463632583618, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7278024554252625, + "num_tokens": 650990245.0, + "step": 25155 + }, + { + "epoch": 2.7625741269492643, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.8487792015075684, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7426950931549072, + "num_tokens": 651007513.0, + "step": 25156 + }, + { + "epoch": 2.7626839446518776, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4426627159118652, + "learning_rate": 1e-06, + "loss": 0.8585, + "mean_token_accuracy": 0.7352675199508667, + "num_tokens": 651033184.0, + "step": 25157 + }, + { + "epoch": 2.762793762354492, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2308921813964844, + "learning_rate": 1e-06, + "loss": 0.8476, + "mean_token_accuracy": 0.7287395596504211, + "num_tokens": 651061072.0, + "step": 25158 + }, + { + "epoch": 2.762903580057105, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.472339391708374, + "learning_rate": 1e-06, + "loss": 0.9622, + "mean_token_accuracy": 0.7013990879058838, + "num_tokens": 651086402.0, + "step": 25159 + }, + { + "epoch": 2.763013397759719, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.398367166519165, + "learning_rate": 1e-06, + "loss": 0.8724, + "mean_token_accuracy": 0.7259078025817871, + "num_tokens": 651110355.0, + "step": 25160 + }, + { + "epoch": 2.7631232154623326, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4462246894836426, + "learning_rate": 1e-06, + "loss": 0.6759, + "mean_token_accuracy": 0.7862434387207031, + "num_tokens": 651132937.0, + "step": 25161 + }, + { + "epoch": 2.763233033164946, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2845206260681152, + "learning_rate": 1e-06, + "loss": 0.8901, + "mean_token_accuracy": 0.7308433651924133, + "num_tokens": 651159254.0, + "step": 25162 + }, + { + "epoch": 2.7633428508675597, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.417698860168457, + "learning_rate": 1e-06, + "loss": 0.8936, + "mean_token_accuracy": 0.7196728587150574, + "num_tokens": 651183771.0, + "step": 25163 + }, + { + "epoch": 2.7634526685701735, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3556277751922607, + "learning_rate": 1e-06, + "loss": 0.8893, + "mean_token_accuracy": 0.7321350574493408, + "num_tokens": 651210647.0, + "step": 25164 + }, + { + "epoch": 2.763562486272787, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5945708751678467, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7293510437011719, + "num_tokens": 651232450.0, + "step": 25165 + }, + { + "epoch": 2.763672303975401, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2840261459350586, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7127439975738525, + "num_tokens": 651262363.0, + "step": 25166 + }, + { + "epoch": 2.7637821216780143, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1937179565429688, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7241217494010925, + "num_tokens": 651294423.0, + "step": 25167 + }, + { + "epoch": 2.763891939380628, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.452983856201172, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7219961881637573, + "num_tokens": 651320244.0, + "step": 25168 + }, + { + "epoch": 2.764001757083242, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.430337905883789, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7304834127426147, + "num_tokens": 651345228.0, + "step": 25169 + }, + { + "epoch": 2.7641115747858556, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3620359897613525, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7404660582542419, + "num_tokens": 651370252.0, + "step": 25170 + }, + { + "epoch": 2.7642213924884693, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1176648139953613, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7379401922225952, + "num_tokens": 651399536.0, + "step": 25171 + }, + { + "epoch": 2.7643312101910826, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3206701278686523, + "learning_rate": 1e-06, + "loss": 0.9288, + "mean_token_accuracy": 0.7115530371665955, + "num_tokens": 651427551.0, + "step": 25172 + }, + { + "epoch": 2.7644410278936964, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2755870819091797, + "learning_rate": 1e-06, + "loss": 0.9097, + "mean_token_accuracy": 0.7232922315597534, + "num_tokens": 651456709.0, + "step": 25173 + }, + { + "epoch": 2.76455084559631, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3240435123443604, + "learning_rate": 1e-06, + "loss": 0.8692, + "mean_token_accuracy": 0.7300360202789307, + "num_tokens": 651483211.0, + "step": 25174 + }, + { + "epoch": 2.764660663298924, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.785128116607666, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7460125684738159, + "num_tokens": 651503020.0, + "step": 25175 + }, + { + "epoch": 2.7647704810015377, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.383476972579956, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7292668223381042, + "num_tokens": 651527231.0, + "step": 25176 + }, + { + "epoch": 2.764880298704151, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2897017002105713, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7460090517997742, + "num_tokens": 651554015.0, + "step": 25177 + }, + { + "epoch": 2.7649901164067647, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3648314476013184, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7334241271018982, + "num_tokens": 651580250.0, + "step": 25178 + }, + { + "epoch": 2.7650999341093785, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.202822208404541, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.736840546131134, + "num_tokens": 651606831.0, + "step": 25179 + }, + { + "epoch": 2.7652097518119922, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.443347454071045, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7334654331207275, + "num_tokens": 651629974.0, + "step": 25180 + }, + { + "epoch": 2.765319569514606, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.228275775909424, + "learning_rate": 1e-06, + "loss": 0.8561, + "mean_token_accuracy": 0.7325387001037598, + "num_tokens": 651657057.0, + "step": 25181 + }, + { + "epoch": 2.7654293872172193, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3566277027130127, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7458739280700684, + "num_tokens": 651682385.0, + "step": 25182 + }, + { + "epoch": 2.765539204919833, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4041810035705566, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7265172004699707, + "num_tokens": 651706546.0, + "step": 25183 + }, + { + "epoch": 2.765649022622447, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.115023612976074, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.7479717135429382, + "num_tokens": 651734169.0, + "step": 25184 + }, + { + "epoch": 2.76575884032506, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.335928201675415, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7343394756317139, + "num_tokens": 651760196.0, + "step": 25185 + }, + { + "epoch": 2.765868658027674, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.173121452331543, + "learning_rate": 1e-06, + "loss": 0.9518, + "mean_token_accuracy": 0.7174030542373657, + "num_tokens": 651791156.0, + "step": 25186 + }, + { + "epoch": 2.7659784757302877, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7861762046813965, + "learning_rate": 1e-06, + "loss": 0.7836, + "mean_token_accuracy": 0.7507011890411377, + "num_tokens": 651809139.0, + "step": 25187 + }, + { + "epoch": 2.7660882934329014, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.627587080001831, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.744194507598877, + "num_tokens": 651829919.0, + "step": 25188 + }, + { + "epoch": 2.766198111135515, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.286693572998047, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7140003442764282, + "num_tokens": 651859312.0, + "step": 25189 + }, + { + "epoch": 2.7663079288381285, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1583216190338135, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.7289513349533081, + "num_tokens": 651889055.0, + "step": 25190 + }, + { + "epoch": 2.7664177465407422, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7870078086853027, + "learning_rate": 1e-06, + "loss": 0.6951, + "mean_token_accuracy": 0.7678454518318176, + "num_tokens": 651906880.0, + "step": 25191 + }, + { + "epoch": 2.766527564243356, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4740729331970215, + "learning_rate": 1e-06, + "loss": 0.8022, + "mean_token_accuracy": 0.7434250712394714, + "num_tokens": 651929096.0, + "step": 25192 + }, + { + "epoch": 2.7666373819459698, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.701397657394409, + "learning_rate": 1e-06, + "loss": 0.9067, + "mean_token_accuracy": 0.7154375314712524, + "num_tokens": 651951758.0, + "step": 25193 + }, + { + "epoch": 2.7667471996485835, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3284549713134766, + "learning_rate": 1e-06, + "loss": 0.8647, + "mean_token_accuracy": 0.7261266708374023, + "num_tokens": 651976696.0, + "step": 25194 + }, + { + "epoch": 2.766857017351197, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.221827983856201, + "learning_rate": 1e-06, + "loss": 0.8023, + "mean_token_accuracy": 0.7444530725479126, + "num_tokens": 652004851.0, + "step": 25195 + }, + { + "epoch": 2.7669668350538106, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5811285972595215, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7293252944946289, + "num_tokens": 652027235.0, + "step": 25196 + }, + { + "epoch": 2.7670766527564243, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2624282836914062, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7245269417762756, + "num_tokens": 652053645.0, + "step": 25197 + }, + { + "epoch": 2.767186470459038, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.285041570663452, + "learning_rate": 1e-06, + "loss": 0.909, + "mean_token_accuracy": 0.7215540409088135, + "num_tokens": 652080875.0, + "step": 25198 + }, + { + "epoch": 2.767296288161652, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.399420976638794, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7200782299041748, + "num_tokens": 652105946.0, + "step": 25199 + }, + { + "epoch": 2.767406105864265, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.671685218811035, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7324874997138977, + "num_tokens": 652132432.0, + "step": 25200 + }, + { + "epoch": 2.767515923566879, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.0155270099639893, + "learning_rate": 1e-06, + "loss": 0.9471, + "mean_token_accuracy": 0.70271235704422, + "num_tokens": 652168571.0, + "step": 25201 + }, + { + "epoch": 2.7676257412694927, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.4822680950164795, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7362489700317383, + "num_tokens": 652192561.0, + "step": 25202 + }, + { + "epoch": 2.7677355589721064, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.524336338043213, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7301617860794067, + "num_tokens": 652216199.0, + "step": 25203 + }, + { + "epoch": 2.76784537667472, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.578782796859741, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7176670432090759, + "num_tokens": 652240440.0, + "step": 25204 + }, + { + "epoch": 2.7679551943773335, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.544997453689575, + "learning_rate": 1e-06, + "loss": 0.9043, + "mean_token_accuracy": 0.7248323559761047, + "num_tokens": 652264582.0, + "step": 25205 + }, + { + "epoch": 2.7680650120799473, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2732486724853516, + "learning_rate": 1e-06, + "loss": 0.7857, + "mean_token_accuracy": 0.7520795464515686, + "num_tokens": 652291585.0, + "step": 25206 + }, + { + "epoch": 2.768174829782561, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.1727805137634277, + "learning_rate": 1e-06, + "loss": 0.9508, + "mean_token_accuracy": 0.7086778879165649, + "num_tokens": 652323196.0, + "step": 25207 + }, + { + "epoch": 2.7682846474851748, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.176271915435791, + "learning_rate": 1e-06, + "loss": 0.9424, + "mean_token_accuracy": 0.7021019458770752, + "num_tokens": 652355899.0, + "step": 25208 + }, + { + "epoch": 2.7683944651877885, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.388956069946289, + "learning_rate": 1e-06, + "loss": 0.9274, + "mean_token_accuracy": 0.704542875289917, + "num_tokens": 652381810.0, + "step": 25209 + }, + { + "epoch": 2.768504282890402, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5390686988830566, + "learning_rate": 1e-06, + "loss": 0.7851, + "mean_token_accuracy": 0.7506474256515503, + "num_tokens": 652403246.0, + "step": 25210 + }, + { + "epoch": 2.7686141005930156, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.438915729522705, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7161450386047363, + "num_tokens": 652427886.0, + "step": 25211 + }, + { + "epoch": 2.7687239182956294, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3151679039001465, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7364493608474731, + "num_tokens": 652454926.0, + "step": 25212 + }, + { + "epoch": 2.7688337359982427, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3726818561553955, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7461602091789246, + "num_tokens": 652481219.0, + "step": 25213 + }, + { + "epoch": 2.7689435537008564, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.2303059101104736, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7186930775642395, + "num_tokens": 652507642.0, + "step": 25214 + }, + { + "epoch": 2.76905337140347, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5385000705718994, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.7554540038108826, + "num_tokens": 652530668.0, + "step": 25215 + }, + { + "epoch": 2.769163189106084, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6631276607513428, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7386187314987183, + "num_tokens": 652552515.0, + "step": 25216 + }, + { + "epoch": 2.7692730068086977, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.6400463581085205, + "learning_rate": 1e-06, + "loss": 0.7701, + "mean_token_accuracy": 0.7521974444389343, + "num_tokens": 652572944.0, + "step": 25217 + }, + { + "epoch": 2.769382824511311, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.163496255874634, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7240067720413208, + "num_tokens": 652601639.0, + "step": 25218 + }, + { + "epoch": 2.7694926422139248, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3703150749206543, + "learning_rate": 1e-06, + "loss": 0.9139, + "mean_token_accuracy": 0.7162758111953735, + "num_tokens": 652626592.0, + "step": 25219 + }, + { + "epoch": 2.7696024599165385, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.3143463134765625, + "learning_rate": 1e-06, + "loss": 0.9472, + "mean_token_accuracy": 0.7018197774887085, + "num_tokens": 652655322.0, + "step": 25220 + }, + { + "epoch": 2.7697122776191523, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.5625383853912354, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7461166977882385, + "num_tokens": 652679234.0, + "step": 25221 + }, + { + "epoch": 2.769822095321766, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.7533693313598633, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7364179491996765, + "num_tokens": 652699212.0, + "step": 25222 + }, + { + "epoch": 2.7699319130243794, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.307781457901001, + "learning_rate": 1e-06, + "loss": 0.8306, + "mean_token_accuracy": 0.7383389472961426, + "num_tokens": 652726972.0, + "step": 25223 + }, + { + "epoch": 2.770041730726993, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.486212730407715, + "learning_rate": 1e-06, + "loss": 0.8063, + "mean_token_accuracy": 0.7509043216705322, + "num_tokens": 652748924.0, + "step": 25224 + }, + { + "epoch": 2.770151548429607, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5397789478302, + "learning_rate": 1e-06, + "loss": 0.8905, + "mean_token_accuracy": 0.7281683087348938, + "num_tokens": 652772025.0, + "step": 25225 + }, + { + "epoch": 2.7702613661322206, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.358607530593872, + "learning_rate": 1e-06, + "loss": 0.7867, + "mean_token_accuracy": 0.7482679486274719, + "num_tokens": 652798692.0, + "step": 25226 + }, + { + "epoch": 2.7703711838348344, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1402194499969482, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7256702780723572, + "num_tokens": 652830187.0, + "step": 25227 + }, + { + "epoch": 2.7704810015374477, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.452375650405884, + "learning_rate": 1e-06, + "loss": 0.8002, + "mean_token_accuracy": 0.7389233112335205, + "num_tokens": 652853772.0, + "step": 25228 + }, + { + "epoch": 2.7705908192400615, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.478111505508423, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7286389470100403, + "num_tokens": 652879406.0, + "step": 25229 + }, + { + "epoch": 2.770700636942675, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.427436351776123, + "learning_rate": 1e-06, + "loss": 0.7572, + "mean_token_accuracy": 0.7485721707344055, + "num_tokens": 652904322.0, + "step": 25230 + }, + { + "epoch": 2.770810454645289, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1971774101257324, + "learning_rate": 1e-06, + "loss": 0.8519, + "mean_token_accuracy": 0.7341643571853638, + "num_tokens": 652932569.0, + "step": 25231 + }, + { + "epoch": 2.7709202723479027, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.346320152282715, + "learning_rate": 1e-06, + "loss": 0.8771, + "mean_token_accuracy": 0.7274099588394165, + "num_tokens": 652957935.0, + "step": 25232 + }, + { + "epoch": 2.771030090050516, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.334934711456299, + "learning_rate": 1e-06, + "loss": 0.8555, + "mean_token_accuracy": 0.7231759428977966, + "num_tokens": 652983410.0, + "step": 25233 + }, + { + "epoch": 2.77113990775313, + "ewc_loss": 1.5854835510253906e-05, + "grad_norm": 2.28194260597229, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7203619480133057, + "num_tokens": 653009975.0, + "step": 25234 + }, + { + "epoch": 2.7712497254557436, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2025301456451416, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.745613694190979, + "num_tokens": 653036431.0, + "step": 25235 + }, + { + "epoch": 2.771359543158357, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5495197772979736, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7465446591377258, + "num_tokens": 653058546.0, + "step": 25236 + }, + { + "epoch": 2.771469360860971, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.306818723678589, + "learning_rate": 1e-06, + "loss": 0.7447, + "mean_token_accuracy": 0.7582814693450928, + "num_tokens": 653083335.0, + "step": 25237 + }, + { + "epoch": 2.7715791785635844, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2405028343200684, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7433757185935974, + "num_tokens": 653110027.0, + "step": 25238 + }, + { + "epoch": 2.771688996266198, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.33821964263916, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7337111234664917, + "num_tokens": 653136265.0, + "step": 25239 + }, + { + "epoch": 2.771798813968812, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.484071969985962, + "learning_rate": 1e-06, + "loss": 0.855, + "mean_token_accuracy": 0.7269793152809143, + "num_tokens": 653159842.0, + "step": 25240 + }, + { + "epoch": 2.771908631671425, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4231293201446533, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7310112714767456, + "num_tokens": 653186815.0, + "step": 25241 + }, + { + "epoch": 2.772018449374039, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4666361808776855, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7205374240875244, + "num_tokens": 653211266.0, + "step": 25242 + }, + { + "epoch": 2.7721282670766527, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2531158924102783, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.724197268486023, + "num_tokens": 653240510.0, + "step": 25243 + }, + { + "epoch": 2.7722380847792665, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.659278631210327, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.716140627861023, + "num_tokens": 653261176.0, + "step": 25244 + }, + { + "epoch": 2.7723479024818802, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.205657482147217, + "learning_rate": 1e-06, + "loss": 0.9272, + "mean_token_accuracy": 0.7084863185882568, + "num_tokens": 653290730.0, + "step": 25245 + }, + { + "epoch": 2.7724577201844935, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4168922901153564, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7181735634803772, + "num_tokens": 653316082.0, + "step": 25246 + }, + { + "epoch": 2.7725675378871073, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3396763801574707, + "learning_rate": 1e-06, + "loss": 0.7967, + "mean_token_accuracy": 0.744072675704956, + "num_tokens": 653341875.0, + "step": 25247 + }, + { + "epoch": 2.772677355589721, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.447570562362671, + "learning_rate": 1e-06, + "loss": 0.7635, + "mean_token_accuracy": 0.7521747946739197, + "num_tokens": 653366545.0, + "step": 25248 + }, + { + "epoch": 2.772787173292335, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.472893238067627, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7252828478813171, + "num_tokens": 653391297.0, + "step": 25249 + }, + { + "epoch": 2.7728969909949486, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.46696138381958, + "learning_rate": 1e-06, + "loss": 0.8751, + "mean_token_accuracy": 0.7386940717697144, + "num_tokens": 653414622.0, + "step": 25250 + }, + { + "epoch": 2.773006808697562, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.446241617202759, + "learning_rate": 1e-06, + "loss": 0.7767, + "mean_token_accuracy": 0.7525621652603149, + "num_tokens": 653438389.0, + "step": 25251 + }, + { + "epoch": 2.7731166264001756, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.446357011795044, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7249698638916016, + "num_tokens": 653462007.0, + "step": 25252 + }, + { + "epoch": 2.7732264441027894, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.139613628387451, + "learning_rate": 1e-06, + "loss": 0.8713, + "mean_token_accuracy": 0.7225372791290283, + "num_tokens": 653491402.0, + "step": 25253 + }, + { + "epoch": 2.773336261805403, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.253835678100586, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7111153602600098, + "num_tokens": 653521194.0, + "step": 25254 + }, + { + "epoch": 2.773446079508017, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4851584434509277, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7207201719284058, + "num_tokens": 653546587.0, + "step": 25255 + }, + { + "epoch": 2.7735558972106302, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3694000244140625, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.7249749898910522, + "num_tokens": 653572844.0, + "step": 25256 + }, + { + "epoch": 2.773665714913244, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0776610374450684, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7176414728164673, + "num_tokens": 653604279.0, + "step": 25257 + }, + { + "epoch": 2.7737755326158577, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2366597652435303, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7205463647842407, + "num_tokens": 653630499.0, + "step": 25258 + }, + { + "epoch": 2.7738853503184715, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4735310077667236, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7335273027420044, + "num_tokens": 653652860.0, + "step": 25259 + }, + { + "epoch": 2.7739951680210853, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.460947036743164, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7236720323562622, + "num_tokens": 653678067.0, + "step": 25260 + }, + { + "epoch": 2.7741049857236986, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.280907392501831, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7277216911315918, + "num_tokens": 653705602.0, + "step": 25261 + }, + { + "epoch": 2.7742148034263123, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5663859844207764, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.754486620426178, + "num_tokens": 653728381.0, + "step": 25262 + }, + { + "epoch": 2.774324621128926, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5023367404937744, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7270211577415466, + "num_tokens": 653752289.0, + "step": 25263 + }, + { + "epoch": 2.7744344388315394, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4823694229125977, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7426187992095947, + "num_tokens": 653776572.0, + "step": 25264 + }, + { + "epoch": 2.774544256534153, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4000203609466553, + "learning_rate": 1e-06, + "loss": 0.84, + "mean_token_accuracy": 0.739717960357666, + "num_tokens": 653801743.0, + "step": 25265 + }, + { + "epoch": 2.774654074236767, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3904643058776855, + "learning_rate": 1e-06, + "loss": 0.9015, + "mean_token_accuracy": 0.7170055508613586, + "num_tokens": 653828938.0, + "step": 25266 + }, + { + "epoch": 2.7747638919393807, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3036727905273438, + "learning_rate": 1e-06, + "loss": 0.8685, + "mean_token_accuracy": 0.7312754392623901, + "num_tokens": 653855618.0, + "step": 25267 + }, + { + "epoch": 2.7748737096419944, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5126988887786865, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7464014291763306, + "num_tokens": 653878047.0, + "step": 25268 + }, + { + "epoch": 2.7749835273446077, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.067817449569702, + "learning_rate": 1e-06, + "loss": 0.957, + "mean_token_accuracy": 0.7051888704299927, + "num_tokens": 653910661.0, + "step": 25269 + }, + { + "epoch": 2.7750933450472215, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7155332565307617, + "learning_rate": 1e-06, + "loss": 0.7985, + "mean_token_accuracy": 0.7440785765647888, + "num_tokens": 653931833.0, + "step": 25270 + }, + { + "epoch": 2.7752031627498353, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3732903003692627, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.72368323802948, + "num_tokens": 653956526.0, + "step": 25271 + }, + { + "epoch": 2.775312980452449, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.374727725982666, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7235570549964905, + "num_tokens": 653981615.0, + "step": 25272 + }, + { + "epoch": 2.7754227981550628, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.22713041305542, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7394497990608215, + "num_tokens": 654009960.0, + "step": 25273 + }, + { + "epoch": 2.775532615857676, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.569622755050659, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7324867844581604, + "num_tokens": 654032444.0, + "step": 25274 + }, + { + "epoch": 2.77564243356029, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.8948416709899902, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7322667837142944, + "num_tokens": 654051166.0, + "step": 25275 + }, + { + "epoch": 2.7757522512629036, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.439560651779175, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7145558595657349, + "num_tokens": 654076902.0, + "step": 25276 + }, + { + "epoch": 2.7758620689655173, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.6223268508911133, + "learning_rate": 1e-06, + "loss": 0.7703, + "mean_token_accuracy": 0.7584472894668579, + "num_tokens": 654096938.0, + "step": 25277 + }, + { + "epoch": 2.775971886668131, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2928595542907715, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7345466613769531, + "num_tokens": 654122767.0, + "step": 25278 + }, + { + "epoch": 2.7760817043707444, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.563419818878174, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7286781072616577, + "num_tokens": 654145677.0, + "step": 25279 + }, + { + "epoch": 2.776191522073358, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3679046630859375, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7342828512191772, + "num_tokens": 654169535.0, + "step": 25280 + }, + { + "epoch": 2.776301339775972, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3891255855560303, + "learning_rate": 1e-06, + "loss": 0.7463, + "mean_token_accuracy": 0.7624130845069885, + "num_tokens": 654192822.0, + "step": 25281 + }, + { + "epoch": 2.7764111574785857, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0241401195526123, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7235215902328491, + "num_tokens": 654225419.0, + "step": 25282 + }, + { + "epoch": 2.7765209751811994, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3430826663970947, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7336481809616089, + "num_tokens": 654249585.0, + "step": 25283 + }, + { + "epoch": 2.7766307928838128, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.461513042449951, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7307571768760681, + "num_tokens": 654273537.0, + "step": 25284 + }, + { + "epoch": 2.7767406105864265, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1642062664031982, + "learning_rate": 1e-06, + "loss": 0.8797, + "mean_token_accuracy": 0.7228364944458008, + "num_tokens": 654303510.0, + "step": 25285 + }, + { + "epoch": 2.7768504282890403, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1070547103881836, + "learning_rate": 1e-06, + "loss": 0.9228, + "mean_token_accuracy": 0.7130413055419922, + "num_tokens": 654333327.0, + "step": 25286 + }, + { + "epoch": 2.7769602459916536, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7011353969573975, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7336094975471497, + "num_tokens": 654353996.0, + "step": 25287 + }, + { + "epoch": 2.777070063694268, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7243010997772217, + "learning_rate": 1e-06, + "loss": 0.7778, + "mean_token_accuracy": 0.7503799796104431, + "num_tokens": 654374873.0, + "step": 25288 + }, + { + "epoch": 2.777179881396881, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5274460315704346, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7446756362915039, + "num_tokens": 654397303.0, + "step": 25289 + }, + { + "epoch": 2.777289699099495, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4675328731536865, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7171988487243652, + "num_tokens": 654421715.0, + "step": 25290 + }, + { + "epoch": 2.7773995168021086, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3626372814178467, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7230426669120789, + "num_tokens": 654446136.0, + "step": 25291 + }, + { + "epoch": 2.777509334504722, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4459269046783447, + "learning_rate": 1e-06, + "loss": 0.7319, + "mean_token_accuracy": 0.75947105884552, + "num_tokens": 654469152.0, + "step": 25292 + }, + { + "epoch": 2.7776191522073357, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.23930025100708, + "learning_rate": 1e-06, + "loss": 0.7668, + "mean_token_accuracy": 0.7647441029548645, + "num_tokens": 654495377.0, + "step": 25293 + }, + { + "epoch": 2.7777289699099494, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5996527671813965, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7297177910804749, + "num_tokens": 654516281.0, + "step": 25294 + }, + { + "epoch": 2.777838787612563, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4800212383270264, + "learning_rate": 1e-06, + "loss": 0.827, + "mean_token_accuracy": 0.7381606101989746, + "num_tokens": 654540311.0, + "step": 25295 + }, + { + "epoch": 2.777948605315177, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5881187915802, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.732621967792511, + "num_tokens": 654563102.0, + "step": 25296 + }, + { + "epoch": 2.7780584230177903, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2397594451904297, + "learning_rate": 1e-06, + "loss": 0.9222, + "mean_token_accuracy": 0.7167702317237854, + "num_tokens": 654592510.0, + "step": 25297 + }, + { + "epoch": 2.778168240720404, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4339942932128906, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7401700615882874, + "num_tokens": 654615826.0, + "step": 25298 + }, + { + "epoch": 2.778278058423018, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5024659633636475, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.723458468914032, + "num_tokens": 654641014.0, + "step": 25299 + }, + { + "epoch": 2.7783878761256315, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.237694263458252, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7428093552589417, + "num_tokens": 654669484.0, + "step": 25300 + }, + { + "epoch": 2.7784976938282453, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.507953405380249, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7290730476379395, + "num_tokens": 654691943.0, + "step": 25301 + }, + { + "epoch": 2.7786075115308586, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1056878566741943, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.739662766456604, + "num_tokens": 654724098.0, + "step": 25302 + }, + { + "epoch": 2.7787173292334724, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.330176830291748, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7264509201049805, + "num_tokens": 654751649.0, + "step": 25303 + }, + { + "epoch": 2.778827146936086, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7030444145202637, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7437431216239929, + "num_tokens": 654774086.0, + "step": 25304 + }, + { + "epoch": 2.7789369646387, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.305612564086914, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7122149467468262, + "num_tokens": 654800126.0, + "step": 25305 + }, + { + "epoch": 2.7790467823413136, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.263442039489746, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7200878262519836, + "num_tokens": 654828206.0, + "step": 25306 + }, + { + "epoch": 2.779156600043927, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.69230055809021, + "learning_rate": 1e-06, + "loss": 0.6986, + "mean_token_accuracy": 0.7744464874267578, + "num_tokens": 654847164.0, + "step": 25307 + }, + { + "epoch": 2.7792664177465407, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.6197075843811035, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.7409900426864624, + "num_tokens": 654869336.0, + "step": 25308 + }, + { + "epoch": 2.7793762354491545, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1667582988739014, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7406545877456665, + "num_tokens": 654897779.0, + "step": 25309 + }, + { + "epoch": 2.779486053151768, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.736048936843872, + "learning_rate": 1e-06, + "loss": 0.824, + "mean_token_accuracy": 0.7346186637878418, + "num_tokens": 654917220.0, + "step": 25310 + }, + { + "epoch": 2.779595870854382, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.68157958984375, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7279797196388245, + "num_tokens": 654939692.0, + "step": 25311 + }, + { + "epoch": 2.7797056885569953, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.35375714302063, + "learning_rate": 1e-06, + "loss": 0.9367, + "mean_token_accuracy": 0.7053499817848206, + "num_tokens": 654965842.0, + "step": 25312 + }, + { + "epoch": 2.779815506259609, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.308445692062378, + "learning_rate": 1e-06, + "loss": 0.9079, + "mean_token_accuracy": 0.7168137431144714, + "num_tokens": 654993830.0, + "step": 25313 + }, + { + "epoch": 2.779925323962223, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4849345684051514, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.737407922744751, + "num_tokens": 655015958.0, + "step": 25314 + }, + { + "epoch": 2.780035141664836, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2280166149139404, + "learning_rate": 1e-06, + "loss": 0.7865, + "mean_token_accuracy": 0.7494574785232544, + "num_tokens": 655044444.0, + "step": 25315 + }, + { + "epoch": 2.78014495936745, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2553091049194336, + "learning_rate": 1e-06, + "loss": 0.8505, + "mean_token_accuracy": 0.7401317358016968, + "num_tokens": 655071755.0, + "step": 25316 + }, + { + "epoch": 2.7802547770700636, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.297887086868286, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7023220062255859, + "num_tokens": 655099814.0, + "step": 25317 + }, + { + "epoch": 2.7803645947726774, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.24647855758667, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.744464635848999, + "num_tokens": 655129052.0, + "step": 25318 + }, + { + "epoch": 2.780474412475291, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3690004348754883, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.748485803604126, + "num_tokens": 655153983.0, + "step": 25319 + }, + { + "epoch": 2.7805842301779045, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3372180461883545, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7250200510025024, + "num_tokens": 655180614.0, + "step": 25320 + }, + { + "epoch": 2.780694047880518, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2327046394348145, + "learning_rate": 1e-06, + "loss": 0.9114, + "mean_token_accuracy": 0.7185125350952148, + "num_tokens": 655209629.0, + "step": 25321 + }, + { + "epoch": 2.780803865583132, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3970704078674316, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.7434930801391602, + "num_tokens": 655234356.0, + "step": 25322 + }, + { + "epoch": 2.7809136832857457, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2384650707244873, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7295981049537659, + "num_tokens": 655263774.0, + "step": 25323 + }, + { + "epoch": 2.7810235009883595, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3296427726745605, + "learning_rate": 1e-06, + "loss": 0.837, + "mean_token_accuracy": 0.7368702292442322, + "num_tokens": 655290867.0, + "step": 25324 + }, + { + "epoch": 2.781133318690973, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.211092948913574, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7186766862869263, + "num_tokens": 655319273.0, + "step": 25325 + }, + { + "epoch": 2.7812431363935866, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.323807716369629, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7483733296394348, + "num_tokens": 655344099.0, + "step": 25326 + }, + { + "epoch": 2.7813529540962003, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.127368450164795, + "learning_rate": 1e-06, + "loss": 0.9327, + "mean_token_accuracy": 0.7069670557975769, + "num_tokens": 655376303.0, + "step": 25327 + }, + { + "epoch": 2.781462771798814, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.365121364593506, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7369428873062134, + "num_tokens": 655404148.0, + "step": 25328 + }, + { + "epoch": 2.781572589501428, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.6181857585906982, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.72386634349823, + "num_tokens": 655425556.0, + "step": 25329 + }, + { + "epoch": 2.781682407204041, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.460341453552246, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.7547181844711304, + "num_tokens": 655448683.0, + "step": 25330 + }, + { + "epoch": 2.781792224906655, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.246387481689453, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7362908124923706, + "num_tokens": 655475223.0, + "step": 25331 + }, + { + "epoch": 2.7819020426092687, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.154724359512329, + "learning_rate": 1e-06, + "loss": 0.8105, + "mean_token_accuracy": 0.7400616407394409, + "num_tokens": 655503829.0, + "step": 25332 + }, + { + "epoch": 2.7820118603118824, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3042328357696533, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7315755486488342, + "num_tokens": 655530470.0, + "step": 25333 + }, + { + "epoch": 2.782121678014496, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5984389781951904, + "learning_rate": 1e-06, + "loss": 0.8248, + "mean_token_accuracy": 0.7388873100280762, + "num_tokens": 655553107.0, + "step": 25334 + }, + { + "epoch": 2.7822314957171095, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.321830987930298, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7341419458389282, + "num_tokens": 655579577.0, + "step": 25335 + }, + { + "epoch": 2.7823413134197232, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4166953563690186, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7463555932044983, + "num_tokens": 655604063.0, + "step": 25336 + }, + { + "epoch": 2.782451131122337, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.414220094680786, + "learning_rate": 1e-06, + "loss": 0.8065, + "mean_token_accuracy": 0.7464863657951355, + "num_tokens": 655629811.0, + "step": 25337 + }, + { + "epoch": 2.7825609488249503, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.376178503036499, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7127934694290161, + "num_tokens": 655658096.0, + "step": 25338 + }, + { + "epoch": 2.7826707665275645, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.363464593887329, + "learning_rate": 1e-06, + "loss": 0.8847, + "mean_token_accuracy": 0.7219419479370117, + "num_tokens": 655686339.0, + "step": 25339 + }, + { + "epoch": 2.782780584230178, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.568009614944458, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7180919647216797, + "num_tokens": 655710433.0, + "step": 25340 + }, + { + "epoch": 2.7828904019327916, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2644503116607666, + "learning_rate": 1e-06, + "loss": 0.8723, + "mean_token_accuracy": 0.7275308966636658, + "num_tokens": 655738374.0, + "step": 25341 + }, + { + "epoch": 2.7830002196354053, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4439857006073, + "learning_rate": 1e-06, + "loss": 0.8382, + "mean_token_accuracy": 0.730694055557251, + "num_tokens": 655763612.0, + "step": 25342 + }, + { + "epoch": 2.7831100373380186, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2045481204986572, + "learning_rate": 1e-06, + "loss": 0.876, + "mean_token_accuracy": 0.7182507514953613, + "num_tokens": 655794128.0, + "step": 25343 + }, + { + "epoch": 2.7832198550406324, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5427439212799072, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7386574149131775, + "num_tokens": 655817587.0, + "step": 25344 + }, + { + "epoch": 2.783329672743246, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.601933717727661, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7342641353607178, + "num_tokens": 655840349.0, + "step": 25345 + }, + { + "epoch": 2.78343949044586, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4650492668151855, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7405768632888794, + "num_tokens": 655863924.0, + "step": 25346 + }, + { + "epoch": 2.7835493081484737, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.6170551776885986, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7348048686981201, + "num_tokens": 655886611.0, + "step": 25347 + }, + { + "epoch": 2.783659125851087, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.308121919631958, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7346634864807129, + "num_tokens": 655914869.0, + "step": 25348 + }, + { + "epoch": 2.7837689435537007, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.297647476196289, + "learning_rate": 1e-06, + "loss": 0.8032, + "mean_token_accuracy": 0.7428689002990723, + "num_tokens": 655941675.0, + "step": 25349 + }, + { + "epoch": 2.7838787612563145, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.1682651042938232, + "learning_rate": 1e-06, + "loss": 0.9661, + "mean_token_accuracy": 0.7085320949554443, + "num_tokens": 655971798.0, + "step": 25350 + }, + { + "epoch": 2.7839885789589283, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.379697322845459, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7521851062774658, + "num_tokens": 655995942.0, + "step": 25351 + }, + { + "epoch": 2.784098396661542, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.0292768478393555, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7109347581863403, + "num_tokens": 656030686.0, + "step": 25352 + }, + { + "epoch": 2.7842082143641553, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2665903568267822, + "learning_rate": 1e-06, + "loss": 0.8518, + "mean_token_accuracy": 0.7288830280303955, + "num_tokens": 656060557.0, + "step": 25353 + }, + { + "epoch": 2.784318032066769, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3067173957824707, + "learning_rate": 1e-06, + "loss": 0.8037, + "mean_token_accuracy": 0.7497938871383667, + "num_tokens": 656086004.0, + "step": 25354 + }, + { + "epoch": 2.784427849769383, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3916208744049072, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7211195230484009, + "num_tokens": 656111463.0, + "step": 25355 + }, + { + "epoch": 2.7845376674719966, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5555503368377686, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7314157485961914, + "num_tokens": 656136462.0, + "step": 25356 + }, + { + "epoch": 2.7846474851746104, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4373764991760254, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7336657047271729, + "num_tokens": 656160109.0, + "step": 25357 + }, + { + "epoch": 2.7847573028772237, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3465232849121094, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7268857955932617, + "num_tokens": 656184727.0, + "step": 25358 + }, + { + "epoch": 2.7848671205798374, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.8172006607055664, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7379902601242065, + "num_tokens": 656204424.0, + "step": 25359 + }, + { + "epoch": 2.784976938282451, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.3031554222106934, + "learning_rate": 1e-06, + "loss": 0.8807, + "mean_token_accuracy": 0.7228039503097534, + "num_tokens": 656231785.0, + "step": 25360 + }, + { + "epoch": 2.785086755985065, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5263803005218506, + "learning_rate": 1e-06, + "loss": 0.8392, + "mean_token_accuracy": 0.7304977774620056, + "num_tokens": 656255960.0, + "step": 25361 + }, + { + "epoch": 2.7851965736876787, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 1.993076205253601, + "learning_rate": 1e-06, + "loss": 0.9251, + "mean_token_accuracy": 0.7042067050933838, + "num_tokens": 656289877.0, + "step": 25362 + }, + { + "epoch": 2.785306391390292, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5957283973693848, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7204014658927917, + "num_tokens": 656313410.0, + "step": 25363 + }, + { + "epoch": 2.7854162090929058, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4217326641082764, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7142549753189087, + "num_tokens": 656338139.0, + "step": 25364 + }, + { + "epoch": 2.7855260267955195, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.383077383041382, + "learning_rate": 1e-06, + "loss": 0.874, + "mean_token_accuracy": 0.7195910215377808, + "num_tokens": 656365724.0, + "step": 25365 + }, + { + "epoch": 2.785635844498133, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2832605838775635, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7318479418754578, + "num_tokens": 656397010.0, + "step": 25366 + }, + { + "epoch": 2.7857456622007466, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2277638912200928, + "learning_rate": 1e-06, + "loss": 0.9657, + "mean_token_accuracy": 0.7018907070159912, + "num_tokens": 656426045.0, + "step": 25367 + }, + { + "epoch": 2.7858554799033604, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7266359329223633, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7498917579650879, + "num_tokens": 656445907.0, + "step": 25368 + }, + { + "epoch": 2.785965297605974, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.36739444732666, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.721280574798584, + "num_tokens": 656471183.0, + "step": 25369 + }, + { + "epoch": 2.786075115308588, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.240370988845825, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7284505367279053, + "num_tokens": 656499016.0, + "step": 25370 + }, + { + "epoch": 2.786184933011201, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.535679578781128, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7280252575874329, + "num_tokens": 656523225.0, + "step": 25371 + }, + { + "epoch": 2.786294750713815, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4457879066467285, + "learning_rate": 1e-06, + "loss": 0.9933, + "mean_token_accuracy": 0.70964115858078, + "num_tokens": 656548239.0, + "step": 25372 + }, + { + "epoch": 2.7864045684164287, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2521114349365234, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7357404828071594, + "num_tokens": 656574403.0, + "step": 25373 + }, + { + "epoch": 2.7865143861190425, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.502974033355713, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7464971542358398, + "num_tokens": 656597325.0, + "step": 25374 + }, + { + "epoch": 2.786624203821656, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2009358406066895, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7293106913566589, + "num_tokens": 656624767.0, + "step": 25375 + }, + { + "epoch": 2.7867340215242695, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4358184337615967, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7314759492874146, + "num_tokens": 656650865.0, + "step": 25376 + }, + { + "epoch": 2.7868438392268833, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.2489662170410156, + "learning_rate": 1e-06, + "loss": 0.8761, + "mean_token_accuracy": 0.7257534265518188, + "num_tokens": 656679434.0, + "step": 25377 + }, + { + "epoch": 2.786953656929497, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.094693422317505, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7395360469818115, + "num_tokens": 656707461.0, + "step": 25378 + }, + { + "epoch": 2.787063474632111, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.4918861389160156, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7373394966125488, + "num_tokens": 656731067.0, + "step": 25379 + }, + { + "epoch": 2.7871732923347246, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.429856300354004, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7069936394691467, + "num_tokens": 656756016.0, + "step": 25380 + }, + { + "epoch": 2.787283110037338, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.6451363563537598, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7480164766311646, + "num_tokens": 656776780.0, + "step": 25381 + }, + { + "epoch": 2.7873929277399516, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.386202812194824, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7240854501724243, + "num_tokens": 656802819.0, + "step": 25382 + }, + { + "epoch": 2.7875027454425654, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5947425365448, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7332290410995483, + "num_tokens": 656825854.0, + "step": 25383 + }, + { + "epoch": 2.787612563145179, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.7810161113739014, + "learning_rate": 1e-06, + "loss": 0.7939, + "mean_token_accuracy": 0.7463438510894775, + "num_tokens": 656846330.0, + "step": 25384 + }, + { + "epoch": 2.787722380847793, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.5763895511627197, + "learning_rate": 1e-06, + "loss": 0.813, + "mean_token_accuracy": 0.7427920699119568, + "num_tokens": 656870508.0, + "step": 25385 + }, + { + "epoch": 2.787832198550406, + "ewc_loss": 1.5974044799804688e-05, + "grad_norm": 2.478515148162842, + "learning_rate": 1e-06, + "loss": 0.7762, + "mean_token_accuracy": 0.7500420212745667, + "num_tokens": 656894746.0, + "step": 25386 + }, + { + "epoch": 2.78794201625302, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.164114236831665, + "learning_rate": 1e-06, + "loss": 0.818, + "mean_token_accuracy": 0.7410850524902344, + "num_tokens": 656924634.0, + "step": 25387 + }, + { + "epoch": 2.7880518339556337, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.066596031188965, + "learning_rate": 1e-06, + "loss": 0.8406, + "mean_token_accuracy": 0.7407207489013672, + "num_tokens": 656954425.0, + "step": 25388 + }, + { + "epoch": 2.7881616516582475, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.364427089691162, + "learning_rate": 1e-06, + "loss": 0.8668, + "mean_token_accuracy": 0.7234320640563965, + "num_tokens": 656981932.0, + "step": 25389 + }, + { + "epoch": 2.7882714693608612, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2362658977508545, + "learning_rate": 1e-06, + "loss": 0.9918, + "mean_token_accuracy": 0.6960991621017456, + "num_tokens": 657014208.0, + "step": 25390 + }, + { + "epoch": 2.7883812870634745, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.042745351791382, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.7263224124908447, + "num_tokens": 657043290.0, + "step": 25391 + }, + { + "epoch": 2.7884911047660883, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.647019624710083, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7438016533851624, + "num_tokens": 657064427.0, + "step": 25392 + }, + { + "epoch": 2.788600922468702, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3774054050445557, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7328759431838989, + "num_tokens": 657089835.0, + "step": 25393 + }, + { + "epoch": 2.7887107401713154, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2551939487457275, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.731525719165802, + "num_tokens": 657117918.0, + "step": 25394 + }, + { + "epoch": 2.788820557873929, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1437485218048096, + "learning_rate": 1e-06, + "loss": 0.9566, + "mean_token_accuracy": 0.699751079082489, + "num_tokens": 657150714.0, + "step": 25395 + }, + { + "epoch": 2.788930375576543, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.369842052459717, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7315267324447632, + "num_tokens": 657175174.0, + "step": 25396 + }, + { + "epoch": 2.7890401932791566, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4110331535339355, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7198280096054077, + "num_tokens": 657202885.0, + "step": 25397 + }, + { + "epoch": 2.7891500109817704, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1609160900115967, + "learning_rate": 1e-06, + "loss": 0.8828, + "mean_token_accuracy": 0.7270996570587158, + "num_tokens": 657231845.0, + "step": 25398 + }, + { + "epoch": 2.7892598286843837, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1738715171813965, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7295311689376831, + "num_tokens": 657261026.0, + "step": 25399 + }, + { + "epoch": 2.7893696463869975, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.206137180328369, + "learning_rate": 1e-06, + "loss": 0.9213, + "mean_token_accuracy": 0.7172409892082214, + "num_tokens": 657291041.0, + "step": 25400 + }, + { + "epoch": 2.7894794640896112, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.202874183654785, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.731594443321228, + "num_tokens": 657319078.0, + "step": 25401 + }, + { + "epoch": 2.789589281792225, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.7783913612365723, + "learning_rate": 1e-06, + "loss": 0.7276, + "mean_token_accuracy": 0.7627758979797363, + "num_tokens": 657338584.0, + "step": 25402 + }, + { + "epoch": 2.7896990994948387, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.604548454284668, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.7248581647872925, + "num_tokens": 657361728.0, + "step": 25403 + }, + { + "epoch": 2.789808917197452, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.549931764602661, + "learning_rate": 1e-06, + "loss": 0.7893, + "mean_token_accuracy": 0.7412634491920471, + "num_tokens": 657384977.0, + "step": 25404 + }, + { + "epoch": 2.789918734900066, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1286308765411377, + "learning_rate": 1e-06, + "loss": 0.9116, + "mean_token_accuracy": 0.7130241394042969, + "num_tokens": 657415883.0, + "step": 25405 + }, + { + "epoch": 2.7900285526026796, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5079519748687744, + "learning_rate": 1e-06, + "loss": 0.7256, + "mean_token_accuracy": 0.768007755279541, + "num_tokens": 657438089.0, + "step": 25406 + }, + { + "epoch": 2.7901383703052933, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1420974731445312, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7198542356491089, + "num_tokens": 657467505.0, + "step": 25407 + }, + { + "epoch": 2.790248188007907, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2420144081115723, + "learning_rate": 1e-06, + "loss": 0.9402, + "mean_token_accuracy": 0.7131698131561279, + "num_tokens": 657496368.0, + "step": 25408 + }, + { + "epoch": 2.7903580057105204, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.07084059715271, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7302145957946777, + "num_tokens": 657529212.0, + "step": 25409 + }, + { + "epoch": 2.790467823413134, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2301104068756104, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.7332844734191895, + "num_tokens": 657556847.0, + "step": 25410 + }, + { + "epoch": 2.790577641115748, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.183166027069092, + "learning_rate": 1e-06, + "loss": 0.9279, + "mean_token_accuracy": 0.713046669960022, + "num_tokens": 657587255.0, + "step": 25411 + }, + { + "epoch": 2.7906874588183617, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.692983865737915, + "learning_rate": 1e-06, + "loss": 0.7558, + "mean_token_accuracy": 0.756580114364624, + "num_tokens": 657607807.0, + "step": 25412 + }, + { + "epoch": 2.7907972765209754, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.303666114807129, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7564324736595154, + "num_tokens": 657632519.0, + "step": 25413 + }, + { + "epoch": 2.7909070942235887, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2892420291900635, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7011531591415405, + "num_tokens": 657660615.0, + "step": 25414 + }, + { + "epoch": 2.7910169119262025, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.31969952583313, + "learning_rate": 1e-06, + "loss": 0.91, + "mean_token_accuracy": 0.712591290473938, + "num_tokens": 657688699.0, + "step": 25415 + }, + { + "epoch": 2.7911267296288162, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.260631799697876, + "learning_rate": 1e-06, + "loss": 0.8312, + "mean_token_accuracy": 0.7367423176765442, + "num_tokens": 657717203.0, + "step": 25416 + }, + { + "epoch": 2.7912365473314296, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3674228191375732, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7280862331390381, + "num_tokens": 657743949.0, + "step": 25417 + }, + { + "epoch": 2.7913463650340438, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.103917121887207, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7228235602378845, + "num_tokens": 657774713.0, + "step": 25418 + }, + { + "epoch": 2.791456182736657, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4105608463287354, + "learning_rate": 1e-06, + "loss": 0.8005, + "mean_token_accuracy": 0.7556835412979126, + "num_tokens": 657799743.0, + "step": 25419 + }, + { + "epoch": 2.791566000439271, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1371209621429443, + "learning_rate": 1e-06, + "loss": 0.7748, + "mean_token_accuracy": 0.7449350357055664, + "num_tokens": 657829342.0, + "step": 25420 + }, + { + "epoch": 2.7916758181418846, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.301521062850952, + "learning_rate": 1e-06, + "loss": 0.7492, + "mean_token_accuracy": 0.7601827383041382, + "num_tokens": 657854566.0, + "step": 25421 + }, + { + "epoch": 2.791785635844498, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5046839714050293, + "learning_rate": 1e-06, + "loss": 0.8052, + "mean_token_accuracy": 0.7439155578613281, + "num_tokens": 657879366.0, + "step": 25422 + }, + { + "epoch": 2.7918954535471117, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3373756408691406, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.732528567314148, + "num_tokens": 657905414.0, + "step": 25423 + }, + { + "epoch": 2.7920052712497254, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3461344242095947, + "learning_rate": 1e-06, + "loss": 0.891, + "mean_token_accuracy": 0.7319289445877075, + "num_tokens": 657931013.0, + "step": 25424 + }, + { + "epoch": 2.792115088952339, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.259115219116211, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.7084304094314575, + "num_tokens": 657959293.0, + "step": 25425 + }, + { + "epoch": 2.792224906654953, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2483935356140137, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.71253502368927, + "num_tokens": 657985708.0, + "step": 25426 + }, + { + "epoch": 2.7923347243575662, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3027844429016113, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7297677397727966, + "num_tokens": 658013079.0, + "step": 25427 + }, + { + "epoch": 2.79244454206018, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1976237297058105, + "learning_rate": 1e-06, + "loss": 0.9044, + "mean_token_accuracy": 0.7182911038398743, + "num_tokens": 658042567.0, + "step": 25428 + }, + { + "epoch": 2.7925543597627938, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1623833179473877, + "learning_rate": 1e-06, + "loss": 0.8345, + "mean_token_accuracy": 0.7349146604537964, + "num_tokens": 658072626.0, + "step": 25429 + }, + { + "epoch": 2.7926641774654075, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1059398651123047, + "learning_rate": 1e-06, + "loss": 0.9018, + "mean_token_accuracy": 0.7153053283691406, + "num_tokens": 658102681.0, + "step": 25430 + }, + { + "epoch": 2.7927739951680213, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.6142349243164062, + "learning_rate": 1e-06, + "loss": 0.7614, + "mean_token_accuracy": 0.7523783445358276, + "num_tokens": 658125645.0, + "step": 25431 + }, + { + "epoch": 2.7928838128706346, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.521606922149658, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7391164898872375, + "num_tokens": 658148908.0, + "step": 25432 + }, + { + "epoch": 2.7929936305732483, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4727470874786377, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7230290770530701, + "num_tokens": 658171824.0, + "step": 25433 + }, + { + "epoch": 2.793103448275862, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.36348032951355, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7443437576293945, + "num_tokens": 658196709.0, + "step": 25434 + }, + { + "epoch": 2.793213265978476, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.826171398162842, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7533737421035767, + "num_tokens": 658216649.0, + "step": 25435 + }, + { + "epoch": 2.7933230836810896, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.398982524871826, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7497944831848145, + "num_tokens": 658241009.0, + "step": 25436 + }, + { + "epoch": 2.793432901383703, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4006173610687256, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7232059240341187, + "num_tokens": 658267432.0, + "step": 25437 + }, + { + "epoch": 2.7935427190863167, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.324721097946167, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7314801216125488, + "num_tokens": 658294948.0, + "step": 25438 + }, + { + "epoch": 2.7936525367889304, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4346375465393066, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7485932111740112, + "num_tokens": 658319781.0, + "step": 25439 + }, + { + "epoch": 2.793762354491544, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2927918434143066, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7235908508300781, + "num_tokens": 658347955.0, + "step": 25440 + }, + { + "epoch": 2.793872172194158, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.61224102973938, + "learning_rate": 1e-06, + "loss": 0.7496, + "mean_token_accuracy": 0.7581782937049866, + "num_tokens": 658370227.0, + "step": 25441 + }, + { + "epoch": 2.7939819898967713, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2364120483398438, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7335865497589111, + "num_tokens": 658399478.0, + "step": 25442 + }, + { + "epoch": 2.794091807599385, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.7339625358581543, + "learning_rate": 1e-06, + "loss": 0.7982, + "mean_token_accuracy": 0.7357418537139893, + "num_tokens": 658419334.0, + "step": 25443 + }, + { + "epoch": 2.794201625301999, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2122902870178223, + "learning_rate": 1e-06, + "loss": 0.8264, + "mean_token_accuracy": 0.7454077005386353, + "num_tokens": 658446821.0, + "step": 25444 + }, + { + "epoch": 2.794311443004612, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.308192729949951, + "learning_rate": 1e-06, + "loss": 0.7506, + "mean_token_accuracy": 0.7598128318786621, + "num_tokens": 658472890.0, + "step": 25445 + }, + { + "epoch": 2.794421260707226, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.158773183822632, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7264703512191772, + "num_tokens": 658501453.0, + "step": 25446 + }, + { + "epoch": 2.7945310784098396, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.379500389099121, + "learning_rate": 1e-06, + "loss": 0.8026, + "mean_token_accuracy": 0.7491104006767273, + "num_tokens": 658525789.0, + "step": 25447 + }, + { + "epoch": 2.7946408961124534, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.597698450088501, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7350151538848877, + "num_tokens": 658547999.0, + "step": 25448 + }, + { + "epoch": 2.794750713815067, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.7361702919006348, + "learning_rate": 1e-06, + "loss": 0.7425, + "mean_token_accuracy": 0.7542396783828735, + "num_tokens": 658567804.0, + "step": 25449 + }, + { + "epoch": 2.7948605315176804, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2669565677642822, + "learning_rate": 1e-06, + "loss": 0.8523, + "mean_token_accuracy": 0.7258629202842712, + "num_tokens": 658595068.0, + "step": 25450 + }, + { + "epoch": 2.794970349220294, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.535146951675415, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7301720380783081, + "num_tokens": 658621032.0, + "step": 25451 + }, + { + "epoch": 2.795080166922908, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5345358848571777, + "learning_rate": 1e-06, + "loss": 0.8091, + "mean_token_accuracy": 0.7385651469230652, + "num_tokens": 658641357.0, + "step": 25452 + }, + { + "epoch": 2.7951899846255217, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.245300054550171, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7095842361450195, + "num_tokens": 658671439.0, + "step": 25453 + }, + { + "epoch": 2.7952998023281355, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4047460556030273, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7332615852355957, + "num_tokens": 658696053.0, + "step": 25454 + }, + { + "epoch": 2.7954096200307488, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3398733139038086, + "learning_rate": 1e-06, + "loss": 0.8284, + "mean_token_accuracy": 0.7454126477241516, + "num_tokens": 658721063.0, + "step": 25455 + }, + { + "epoch": 2.7955194377333625, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3422763347625732, + "learning_rate": 1e-06, + "loss": 0.7914, + "mean_token_accuracy": 0.7509905099868774, + "num_tokens": 658746630.0, + "step": 25456 + }, + { + "epoch": 2.7956292554359763, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.314703941345215, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7394767999649048, + "num_tokens": 658771703.0, + "step": 25457 + }, + { + "epoch": 2.79573907313859, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1663100719451904, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7386044263839722, + "num_tokens": 658800801.0, + "step": 25458 + }, + { + "epoch": 2.795848890841204, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.100438117980957, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7235312461853027, + "num_tokens": 658832820.0, + "step": 25459 + }, + { + "epoch": 2.795958708543817, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.0717499256134033, + "learning_rate": 1e-06, + "loss": 0.9509, + "mean_token_accuracy": 0.7005367875099182, + "num_tokens": 658868258.0, + "step": 25460 + }, + { + "epoch": 2.796068526246431, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4532406330108643, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7312933206558228, + "num_tokens": 658891424.0, + "step": 25461 + }, + { + "epoch": 2.7961783439490446, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2133374214172363, + "learning_rate": 1e-06, + "loss": 0.929, + "mean_token_accuracy": 0.7074543237686157, + "num_tokens": 658919905.0, + "step": 25462 + }, + { + "epoch": 2.7962881616516584, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4050025939941406, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7313326597213745, + "num_tokens": 658945164.0, + "step": 25463 + }, + { + "epoch": 2.796397979354272, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.430839776992798, + "learning_rate": 1e-06, + "loss": 0.7424, + "mean_token_accuracy": 0.754778265953064, + "num_tokens": 658967595.0, + "step": 25464 + }, + { + "epoch": 2.7965077970568855, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.078500270843506, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7321211099624634, + "num_tokens": 658999509.0, + "step": 25465 + }, + { + "epoch": 2.796617614759499, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2706735134124756, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7270210981369019, + "num_tokens": 659028209.0, + "step": 25466 + }, + { + "epoch": 2.796727432462113, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.283090829849243, + "learning_rate": 1e-06, + "loss": 0.821, + "mean_token_accuracy": 0.7322912216186523, + "num_tokens": 659054733.0, + "step": 25467 + }, + { + "epoch": 2.7968372501647263, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2592227458953857, + "learning_rate": 1e-06, + "loss": 0.9663, + "mean_token_accuracy": 0.6946583390235901, + "num_tokens": 659086566.0, + "step": 25468 + }, + { + "epoch": 2.7969470678673405, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.325840473175049, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7162012457847595, + "num_tokens": 659113076.0, + "step": 25469 + }, + { + "epoch": 2.797056885569954, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2649688720703125, + "learning_rate": 1e-06, + "loss": 0.779, + "mean_token_accuracy": 0.7463902235031128, + "num_tokens": 659140510.0, + "step": 25470 + }, + { + "epoch": 2.7971667032725676, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1918938159942627, + "learning_rate": 1e-06, + "loss": 0.8603, + "mean_token_accuracy": 0.7300593852996826, + "num_tokens": 659171065.0, + "step": 25471 + }, + { + "epoch": 2.7972765209751813, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.330164909362793, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7154703140258789, + "num_tokens": 659198950.0, + "step": 25472 + }, + { + "epoch": 2.7973863386777946, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1800284385681152, + "learning_rate": 1e-06, + "loss": 0.9155, + "mean_token_accuracy": 0.7143568992614746, + "num_tokens": 659228794.0, + "step": 25473 + }, + { + "epoch": 2.7974961563804084, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.578744411468506, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.7259178161621094, + "num_tokens": 659250785.0, + "step": 25474 + }, + { + "epoch": 2.797605974083022, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.526914119720459, + "learning_rate": 1e-06, + "loss": 0.704, + "mean_token_accuracy": 0.7681125998497009, + "num_tokens": 659273650.0, + "step": 25475 + }, + { + "epoch": 2.797715791785636, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3588650226593018, + "learning_rate": 1e-06, + "loss": 0.757, + "mean_token_accuracy": 0.7527104020118713, + "num_tokens": 659297636.0, + "step": 25476 + }, + { + "epoch": 2.7978256094882497, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.428730010986328, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7304988503456116, + "num_tokens": 659325961.0, + "step": 25477 + }, + { + "epoch": 2.797935427190863, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2131948471069336, + "learning_rate": 1e-06, + "loss": 0.922, + "mean_token_accuracy": 0.7134498357772827, + "num_tokens": 659355236.0, + "step": 25478 + }, + { + "epoch": 2.7980452448934767, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.38156795501709, + "learning_rate": 1e-06, + "loss": 0.7998, + "mean_token_accuracy": 0.747269868850708, + "num_tokens": 659380549.0, + "step": 25479 + }, + { + "epoch": 2.7981550625960905, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.382938861846924, + "learning_rate": 1e-06, + "loss": 0.7608, + "mean_token_accuracy": 0.7541805505752563, + "num_tokens": 659404170.0, + "step": 25480 + }, + { + "epoch": 2.7982648802987042, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2560484409332275, + "learning_rate": 1e-06, + "loss": 0.7717, + "mean_token_accuracy": 0.7574585676193237, + "num_tokens": 659429358.0, + "step": 25481 + }, + { + "epoch": 2.798374698001318, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4906728267669678, + "learning_rate": 1e-06, + "loss": 0.8221, + "mean_token_accuracy": 0.7355649471282959, + "num_tokens": 659452045.0, + "step": 25482 + }, + { + "epoch": 2.7984845157039313, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.087782859802246, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7253122329711914, + "num_tokens": 659482183.0, + "step": 25483 + }, + { + "epoch": 2.798594333406545, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4212191104888916, + "learning_rate": 1e-06, + "loss": 0.8783, + "mean_token_accuracy": 0.722892165184021, + "num_tokens": 659506878.0, + "step": 25484 + }, + { + "epoch": 2.798704151109159, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.290160655975342, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7272031903266907, + "num_tokens": 659534891.0, + "step": 25485 + }, + { + "epoch": 2.7988139688117726, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3066916465759277, + "learning_rate": 1e-06, + "loss": 0.8185, + "mean_token_accuracy": 0.7379109263420105, + "num_tokens": 659563039.0, + "step": 25486 + }, + { + "epoch": 2.7989237865143863, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4064621925354004, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7184547185897827, + "num_tokens": 659587458.0, + "step": 25487 + }, + { + "epoch": 2.7990336042169996, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.457991361618042, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7380248308181763, + "num_tokens": 659610684.0, + "step": 25488 + }, + { + "epoch": 2.7991434219196134, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5331125259399414, + "learning_rate": 1e-06, + "loss": 0.7608, + "mean_token_accuracy": 0.7553030252456665, + "num_tokens": 659631616.0, + "step": 25489 + }, + { + "epoch": 2.799253239622227, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.333296537399292, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7222431898117065, + "num_tokens": 659659811.0, + "step": 25490 + }, + { + "epoch": 2.799363057324841, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.6232006549835205, + "learning_rate": 1e-06, + "loss": 0.7736, + "mean_token_accuracy": 0.7509067058563232, + "num_tokens": 659681866.0, + "step": 25491 + }, + { + "epoch": 2.7994728750274547, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.25740647315979, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7273648381233215, + "num_tokens": 659708529.0, + "step": 25492 + }, + { + "epoch": 2.799582692730068, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3219923973083496, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7342575788497925, + "num_tokens": 659737239.0, + "step": 25493 + }, + { + "epoch": 2.7996925104326817, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.6368868350982666, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7427369356155396, + "num_tokens": 659758319.0, + "step": 25494 + }, + { + "epoch": 2.7998023281352955, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.488443374633789, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7295851707458496, + "num_tokens": 659781992.0, + "step": 25495 + }, + { + "epoch": 2.799912145837909, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.681126594543457, + "learning_rate": 1e-06, + "loss": 0.8457, + "mean_token_accuracy": 0.7296565771102905, + "num_tokens": 659803204.0, + "step": 25496 + }, + { + "epoch": 2.8000219635405226, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.457926034927368, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7293559312820435, + "num_tokens": 659827306.0, + "step": 25497 + }, + { + "epoch": 2.8001317812431363, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.6508145332336426, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7355440855026245, + "num_tokens": 659847819.0, + "step": 25498 + }, + { + "epoch": 2.80024159894575, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5178048610687256, + "learning_rate": 1e-06, + "loss": 0.8989, + "mean_token_accuracy": 0.7241474390029907, + "num_tokens": 659871600.0, + "step": 25499 + }, + { + "epoch": 2.800351416648364, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2023675441741943, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7396095395088196, + "num_tokens": 659900805.0, + "step": 25500 + }, + { + "epoch": 2.800461234350977, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1840245723724365, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7254867553710938, + "num_tokens": 659929084.0, + "step": 25501 + }, + { + "epoch": 2.800571052053591, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1397557258605957, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7169272303581238, + "num_tokens": 659961327.0, + "step": 25502 + }, + { + "epoch": 2.8006808697562047, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.483525276184082, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7105932235717773, + "num_tokens": 659986817.0, + "step": 25503 + }, + { + "epoch": 2.8007906874588184, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.7243974208831787, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7213197946548462, + "num_tokens": 660008085.0, + "step": 25504 + }, + { + "epoch": 2.800900505161432, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.3541407585144043, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7452359199523926, + "num_tokens": 660034970.0, + "step": 25505 + }, + { + "epoch": 2.8010103228640455, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.653935432434082, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7443447113037109, + "num_tokens": 660056318.0, + "step": 25506 + }, + { + "epoch": 2.8011201405666593, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4698474407196045, + "learning_rate": 1e-06, + "loss": 0.7761, + "mean_token_accuracy": 0.7505037784576416, + "num_tokens": 660079209.0, + "step": 25507 + }, + { + "epoch": 2.801229958269273, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2785072326660156, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7093487977981567, + "num_tokens": 660107530.0, + "step": 25508 + }, + { + "epoch": 2.8013397759718868, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.5017292499542236, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7397016882896423, + "num_tokens": 660131087.0, + "step": 25509 + }, + { + "epoch": 2.8014495936745005, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.1959478855133057, + "learning_rate": 1e-06, + "loss": 0.958, + "mean_token_accuracy": 0.7147440910339355, + "num_tokens": 660162306.0, + "step": 25510 + }, + { + "epoch": 2.801559411377114, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4313459396362305, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7367068529129028, + "num_tokens": 660186485.0, + "step": 25511 + }, + { + "epoch": 2.8016692290797276, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2967047691345215, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7276768684387207, + "num_tokens": 660210493.0, + "step": 25512 + }, + { + "epoch": 2.8017790467823414, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.4405622482299805, + "learning_rate": 1e-06, + "loss": 0.8979, + "mean_token_accuracy": 0.7174168825149536, + "num_tokens": 660233809.0, + "step": 25513 + }, + { + "epoch": 2.801888864484955, + "ewc_loss": 1.609325408935547e-05, + "grad_norm": 2.2387447357177734, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7324272990226746, + "num_tokens": 660262261.0, + "step": 25514 + }, + { + "epoch": 2.801998682187569, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.541897773742676, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7425963878631592, + "num_tokens": 660285282.0, + "step": 25515 + }, + { + "epoch": 2.802108499890182, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3783328533172607, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7379401326179504, + "num_tokens": 660312026.0, + "step": 25516 + }, + { + "epoch": 2.802218317592796, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5332772731781006, + "learning_rate": 1e-06, + "loss": 0.8455, + "mean_token_accuracy": 0.7263572216033936, + "num_tokens": 660336797.0, + "step": 25517 + }, + { + "epoch": 2.8023281352954097, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3853917121887207, + "learning_rate": 1e-06, + "loss": 0.8331, + "mean_token_accuracy": 0.7346949577331543, + "num_tokens": 660363648.0, + "step": 25518 + }, + { + "epoch": 2.802437952998023, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.576866626739502, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7319961786270142, + "num_tokens": 660388352.0, + "step": 25519 + }, + { + "epoch": 2.802547770700637, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.501856803894043, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.729526698589325, + "num_tokens": 660412677.0, + "step": 25520 + }, + { + "epoch": 2.8026575884032505, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2075884342193604, + "learning_rate": 1e-06, + "loss": 0.9564, + "mean_token_accuracy": 0.7060315012931824, + "num_tokens": 660442384.0, + "step": 25521 + }, + { + "epoch": 2.8027674061058643, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.140674114227295, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7465271353721619, + "num_tokens": 660471277.0, + "step": 25522 + }, + { + "epoch": 2.802877223808478, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.269613265991211, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7303922176361084, + "num_tokens": 660499776.0, + "step": 25523 + }, + { + "epoch": 2.8029870415110913, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.165822982788086, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7217049598693848, + "num_tokens": 660529542.0, + "step": 25524 + }, + { + "epoch": 2.803096859213705, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.327997922897339, + "learning_rate": 1e-06, + "loss": 0.8596, + "mean_token_accuracy": 0.725588321685791, + "num_tokens": 660557255.0, + "step": 25525 + }, + { + "epoch": 2.803206676916319, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.628762722015381, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.7417820692062378, + "num_tokens": 660577448.0, + "step": 25526 + }, + { + "epoch": 2.8033164946189326, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0438356399536133, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7221611738204956, + "num_tokens": 660611802.0, + "step": 25527 + }, + { + "epoch": 2.8034263123215464, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.7893364429473877, + "learning_rate": 1e-06, + "loss": 0.7278, + "mean_token_accuracy": 0.7614141702651978, + "num_tokens": 660632030.0, + "step": 25528 + }, + { + "epoch": 2.8035361300241597, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3803067207336426, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7169198989868164, + "num_tokens": 660656917.0, + "step": 25529 + }, + { + "epoch": 2.8036459477267734, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6685891151428223, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7534306049346924, + "num_tokens": 660677591.0, + "step": 25530 + }, + { + "epoch": 2.803755765429387, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.29784893989563, + "learning_rate": 1e-06, + "loss": 0.7814, + "mean_token_accuracy": 0.7480393648147583, + "num_tokens": 660704066.0, + "step": 25531 + }, + { + "epoch": 2.803865583132001, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.032470464706421, + "learning_rate": 1e-06, + "loss": 0.913, + "mean_token_accuracy": 0.7193014621734619, + "num_tokens": 660736158.0, + "step": 25532 + }, + { + "epoch": 2.8039754008346147, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.21030330657959, + "learning_rate": 1e-06, + "loss": 0.8574, + "mean_token_accuracy": 0.739270031452179, + "num_tokens": 660764931.0, + "step": 25533 + }, + { + "epoch": 2.804085218537228, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3523600101470947, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.7207346558570862, + "num_tokens": 660791878.0, + "step": 25534 + }, + { + "epoch": 2.804195036239842, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3694846630096436, + "learning_rate": 1e-06, + "loss": 0.808, + "mean_token_accuracy": 0.7416025400161743, + "num_tokens": 660818303.0, + "step": 25535 + }, + { + "epoch": 2.8043048539424555, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1807148456573486, + "learning_rate": 1e-06, + "loss": 0.8631, + "mean_token_accuracy": 0.7260972261428833, + "num_tokens": 660847686.0, + "step": 25536 + }, + { + "epoch": 2.8044146716450693, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.305990695953369, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7287136316299438, + "num_tokens": 660874618.0, + "step": 25537 + }, + { + "epoch": 2.804524489347683, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6204488277435303, + "learning_rate": 1e-06, + "loss": 0.7731, + "mean_token_accuracy": 0.7520994544029236, + "num_tokens": 660895920.0, + "step": 25538 + }, + { + "epoch": 2.8046343070502964, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6350691318511963, + "learning_rate": 1e-06, + "loss": 0.7959, + "mean_token_accuracy": 0.7362786531448364, + "num_tokens": 660918930.0, + "step": 25539 + }, + { + "epoch": 2.80474412475291, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2497589588165283, + "learning_rate": 1e-06, + "loss": 0.783, + "mean_token_accuracy": 0.7546278238296509, + "num_tokens": 660946257.0, + "step": 25540 + }, + { + "epoch": 2.804853942455524, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4457640647888184, + "learning_rate": 1e-06, + "loss": 0.902, + "mean_token_accuracy": 0.7212575078010559, + "num_tokens": 660970851.0, + "step": 25541 + }, + { + "epoch": 2.8049637601581376, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.170222282409668, + "learning_rate": 1e-06, + "loss": 0.9719, + "mean_token_accuracy": 0.7011820673942566, + "num_tokens": 661001842.0, + "step": 25542 + }, + { + "epoch": 2.8050735778607514, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1463966369628906, + "learning_rate": 1e-06, + "loss": 0.872, + "mean_token_accuracy": 0.7230838537216187, + "num_tokens": 661033093.0, + "step": 25543 + }, + { + "epoch": 2.8051833955633647, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5819969177246094, + "learning_rate": 1e-06, + "loss": 0.8317, + "mean_token_accuracy": 0.7414048910140991, + "num_tokens": 661055272.0, + "step": 25544 + }, + { + "epoch": 2.8052932132659785, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3447751998901367, + "learning_rate": 1e-06, + "loss": 0.8396, + "mean_token_accuracy": 0.7335003614425659, + "num_tokens": 661080432.0, + "step": 25545 + }, + { + "epoch": 2.8054030309685922, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4742074012756348, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7423954010009766, + "num_tokens": 661105508.0, + "step": 25546 + }, + { + "epoch": 2.8055128486712055, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2215003967285156, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.7283312082290649, + "num_tokens": 661135946.0, + "step": 25547 + }, + { + "epoch": 2.8056226663738193, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5273430347442627, + "learning_rate": 1e-06, + "loss": 0.8646, + "mean_token_accuracy": 0.7291048765182495, + "num_tokens": 661159454.0, + "step": 25548 + }, + { + "epoch": 2.805732484076433, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.459840774536133, + "learning_rate": 1e-06, + "loss": 0.7514, + "mean_token_accuracy": 0.7656041383743286, + "num_tokens": 661182410.0, + "step": 25549 + }, + { + "epoch": 2.805842301779047, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.891655206680298, + "learning_rate": 1e-06, + "loss": 0.8204, + "mean_token_accuracy": 0.7454370260238647, + "num_tokens": 661201229.0, + "step": 25550 + }, + { + "epoch": 2.8059521194816606, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.284705877304077, + "learning_rate": 1e-06, + "loss": 0.9386, + "mean_token_accuracy": 0.7116859555244446, + "num_tokens": 661229079.0, + "step": 25551 + }, + { + "epoch": 2.806061937184274, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.469763994216919, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.735124945640564, + "num_tokens": 661252597.0, + "step": 25552 + }, + { + "epoch": 2.8061717548868876, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.323572874069214, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7075143456459045, + "num_tokens": 661280899.0, + "step": 25553 + }, + { + "epoch": 2.8062815725895014, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.318094253540039, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7313740253448486, + "num_tokens": 661309258.0, + "step": 25554 + }, + { + "epoch": 2.806391390292115, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.456296920776367, + "learning_rate": 1e-06, + "loss": 0.9103, + "mean_token_accuracy": 0.7125616669654846, + "num_tokens": 661334630.0, + "step": 25555 + }, + { + "epoch": 2.806501207994729, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3133533000946045, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7373478412628174, + "num_tokens": 661361697.0, + "step": 25556 + }, + { + "epoch": 2.806611025697342, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4236130714416504, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7351338267326355, + "num_tokens": 661387248.0, + "step": 25557 + }, + { + "epoch": 2.806720843399956, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3943891525268555, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7262319326400757, + "num_tokens": 661412647.0, + "step": 25558 + }, + { + "epoch": 2.8068306611025697, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3127455711364746, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7279454469680786, + "num_tokens": 661438448.0, + "step": 25559 + }, + { + "epoch": 2.8069404788051835, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.450745105743408, + "learning_rate": 1e-06, + "loss": 0.8302, + "mean_token_accuracy": 0.7358707189559937, + "num_tokens": 661463863.0, + "step": 25560 + }, + { + "epoch": 2.8070502965077972, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2084622383117676, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7244718074798584, + "num_tokens": 661492452.0, + "step": 25561 + }, + { + "epoch": 2.8071601142104106, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.639169692993164, + "learning_rate": 1e-06, + "loss": 0.8237, + "mean_token_accuracy": 0.7410681247711182, + "num_tokens": 661514392.0, + "step": 25562 + }, + { + "epoch": 2.8072699319130243, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2094759941101074, + "learning_rate": 1e-06, + "loss": 0.7938, + "mean_token_accuracy": 0.7438747882843018, + "num_tokens": 661540750.0, + "step": 25563 + }, + { + "epoch": 2.807379749615638, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.415663003921509, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7128036618232727, + "num_tokens": 661568038.0, + "step": 25564 + }, + { + "epoch": 2.807489567318252, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.255640745162964, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7367203235626221, + "num_tokens": 661594930.0, + "step": 25565 + }, + { + "epoch": 2.8075993850208656, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1978931427001953, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7204098701477051, + "num_tokens": 661623558.0, + "step": 25566 + }, + { + "epoch": 2.807709202723479, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2119996547698975, + "learning_rate": 1e-06, + "loss": 0.8278, + "mean_token_accuracy": 0.73186194896698, + "num_tokens": 661652206.0, + "step": 25567 + }, + { + "epoch": 2.8078190204260927, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1491472721099854, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.71910160779953, + "num_tokens": 661683089.0, + "step": 25568 + }, + { + "epoch": 2.8079288381287064, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4033262729644775, + "learning_rate": 1e-06, + "loss": 0.7663, + "mean_token_accuracy": 0.7538060545921326, + "num_tokens": 661707272.0, + "step": 25569 + }, + { + "epoch": 2.80803865583132, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4880008697509766, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7427282333374023, + "num_tokens": 661730598.0, + "step": 25570 + }, + { + "epoch": 2.808148473533934, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5209896564483643, + "learning_rate": 1e-06, + "loss": 0.7949, + "mean_token_accuracy": 0.7466349601745605, + "num_tokens": 661753754.0, + "step": 25571 + }, + { + "epoch": 2.8082582912365472, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3851637840270996, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.738034188747406, + "num_tokens": 661779277.0, + "step": 25572 + }, + { + "epoch": 2.808368108939161, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.492222785949707, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7396552562713623, + "num_tokens": 661804152.0, + "step": 25573 + }, + { + "epoch": 2.8084779266417748, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.574002742767334, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7194287180900574, + "num_tokens": 661826690.0, + "step": 25574 + }, + { + "epoch": 2.808587744344388, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.151851177215576, + "learning_rate": 1e-06, + "loss": 0.9376, + "mean_token_accuracy": 0.7178682088851929, + "num_tokens": 661858275.0, + "step": 25575 + }, + { + "epoch": 2.808697562047002, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4155733585357666, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7202879190444946, + "num_tokens": 661885693.0, + "step": 25576 + }, + { + "epoch": 2.8088073797496156, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4044220447540283, + "learning_rate": 1e-06, + "loss": 0.8681, + "mean_token_accuracy": 0.7266196012496948, + "num_tokens": 661910365.0, + "step": 25577 + }, + { + "epoch": 2.8089171974522293, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.688729763031006, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.7479673027992249, + "num_tokens": 661930340.0, + "step": 25578 + }, + { + "epoch": 2.809027015154843, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.691551446914673, + "learning_rate": 1e-06, + "loss": 0.8377, + "mean_token_accuracy": 0.7339229583740234, + "num_tokens": 661950143.0, + "step": 25579 + }, + { + "epoch": 2.8091368328574564, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.70929217338562, + "learning_rate": 1e-06, + "loss": 0.7877, + "mean_token_accuracy": 0.7474260330200195, + "num_tokens": 661972186.0, + "step": 25580 + }, + { + "epoch": 2.80924665056007, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.263349771499634, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7384678721427917, + "num_tokens": 661998612.0, + "step": 25581 + }, + { + "epoch": 2.809356468262684, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3195183277130127, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7368987798690796, + "num_tokens": 662027212.0, + "step": 25582 + }, + { + "epoch": 2.8094662859652977, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.496264696121216, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7399066686630249, + "num_tokens": 662052504.0, + "step": 25583 + }, + { + "epoch": 2.8095761036679114, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4671974182128906, + "learning_rate": 1e-06, + "loss": 0.7662, + "mean_token_accuracy": 0.7515538334846497, + "num_tokens": 662075237.0, + "step": 25584 + }, + { + "epoch": 2.8096859213705248, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.362190008163452, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7249054908752441, + "num_tokens": 662100237.0, + "step": 25585 + }, + { + "epoch": 2.8097957390731385, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.332552671432495, + "learning_rate": 1e-06, + "loss": 0.7691, + "mean_token_accuracy": 0.7580549716949463, + "num_tokens": 662125680.0, + "step": 25586 + }, + { + "epoch": 2.8099055567757523, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.732949733734131, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7269501090049744, + "num_tokens": 662146244.0, + "step": 25587 + }, + { + "epoch": 2.810015374478366, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1891591548919678, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7359028458595276, + "num_tokens": 662177330.0, + "step": 25588 + }, + { + "epoch": 2.81012519218098, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.692031145095825, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.7539445161819458, + "num_tokens": 662197208.0, + "step": 25589 + }, + { + "epoch": 2.810235009883593, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7509801387786865, + "learning_rate": 1e-06, + "loss": 0.7119, + "mean_token_accuracy": 0.7672209739685059, + "num_tokens": 662216227.0, + "step": 25590 + }, + { + "epoch": 2.810344827586207, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4341888427734375, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.7238433957099915, + "num_tokens": 662241684.0, + "step": 25591 + }, + { + "epoch": 2.8104546452888206, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.72800874710083, + "learning_rate": 1e-06, + "loss": 0.8134, + "mean_token_accuracy": 0.7401981353759766, + "num_tokens": 662262461.0, + "step": 25592 + }, + { + "epoch": 2.8105644629914344, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3624320030212402, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7398064732551575, + "num_tokens": 662287571.0, + "step": 25593 + }, + { + "epoch": 2.810674280694048, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.1745898723602295, + "learning_rate": 1e-06, + "loss": 0.8689, + "mean_token_accuracy": 0.7278367280960083, + "num_tokens": 662317521.0, + "step": 25594 + }, + { + "epoch": 2.8107840983966614, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.204024314880371, + "learning_rate": 1e-06, + "loss": 0.9552, + "mean_token_accuracy": 0.7099912762641907, + "num_tokens": 662346423.0, + "step": 25595 + }, + { + "epoch": 2.810893916099275, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.679877758026123, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7259112000465393, + "num_tokens": 662367305.0, + "step": 25596 + }, + { + "epoch": 2.811003733801889, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.364093065261841, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7320255041122437, + "num_tokens": 662392026.0, + "step": 25597 + }, + { + "epoch": 2.8111135515045023, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5083651542663574, + "learning_rate": 1e-06, + "loss": 0.7796, + "mean_token_accuracy": 0.7478110194206238, + "num_tokens": 662417973.0, + "step": 25598 + }, + { + "epoch": 2.8112233692071165, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.0972819328308105, + "learning_rate": 1e-06, + "loss": 0.8367, + "mean_token_accuracy": 0.734325110912323, + "num_tokens": 662449961.0, + "step": 25599 + }, + { + "epoch": 2.8113331869097298, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3678648471832275, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7276392579078674, + "num_tokens": 662474543.0, + "step": 25600 + }, + { + "epoch": 2.8114430046123435, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3228464126586914, + "learning_rate": 1e-06, + "loss": 0.9177, + "mean_token_accuracy": 0.7143465876579285, + "num_tokens": 662503803.0, + "step": 25601 + }, + { + "epoch": 2.8115528223149573, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.017388343811035, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7349005341529846, + "num_tokens": 662538210.0, + "step": 25602 + }, + { + "epoch": 2.8116626400175706, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.5873475074768066, + "learning_rate": 1e-06, + "loss": 0.7402, + "mean_token_accuracy": 0.7594212293624878, + "num_tokens": 662559687.0, + "step": 25603 + }, + { + "epoch": 2.8117724577201844, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.4001753330230713, + "learning_rate": 1e-06, + "loss": 0.7706, + "mean_token_accuracy": 0.7496618032455444, + "num_tokens": 662583448.0, + "step": 25604 + }, + { + "epoch": 2.811882275422798, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.3925583362579346, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7256278991699219, + "num_tokens": 662608705.0, + "step": 25605 + }, + { + "epoch": 2.811992093125412, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.6303467750549316, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7108672857284546, + "num_tokens": 662630715.0, + "step": 25606 + }, + { + "epoch": 2.8121019108280256, + "ewc_loss": 1.621246337890625e-05, + "grad_norm": 2.2915091514587402, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.7208261489868164, + "num_tokens": 662657962.0, + "step": 25607 + }, + { + "epoch": 2.812211728530639, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.414525032043457, + "learning_rate": 1e-06, + "loss": 0.8682, + "mean_token_accuracy": 0.7351964712142944, + "num_tokens": 662684548.0, + "step": 25608 + }, + { + "epoch": 2.8123215462332527, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.376434087753296, + "learning_rate": 1e-06, + "loss": 0.7951, + "mean_token_accuracy": 0.7443232536315918, + "num_tokens": 662710245.0, + "step": 25609 + }, + { + "epoch": 2.8124313639358665, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4721014499664307, + "learning_rate": 1e-06, + "loss": 0.83, + "mean_token_accuracy": 0.7378284335136414, + "num_tokens": 662735937.0, + "step": 25610 + }, + { + "epoch": 2.81254118163848, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.73378586769104, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7371330261230469, + "num_tokens": 662755845.0, + "step": 25611 + }, + { + "epoch": 2.812650999341094, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.466021776199341, + "learning_rate": 1e-06, + "loss": 0.8841, + "mean_token_accuracy": 0.7258588671684265, + "num_tokens": 662781038.0, + "step": 25612 + }, + { + "epoch": 2.8127608170437073, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.514486789703369, + "learning_rate": 1e-06, + "loss": 0.8866, + "mean_token_accuracy": 0.7318433523178101, + "num_tokens": 662805811.0, + "step": 25613 + }, + { + "epoch": 2.812870634746321, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.630743980407715, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7480267286300659, + "num_tokens": 662826943.0, + "step": 25614 + }, + { + "epoch": 2.812980452448935, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1953084468841553, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7256218194961548, + "num_tokens": 662854053.0, + "step": 25615 + }, + { + "epoch": 2.8130902701515486, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.272233486175537, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7180647850036621, + "num_tokens": 662882986.0, + "step": 25616 + }, + { + "epoch": 2.8132000878541623, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2715914249420166, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.732734203338623, + "num_tokens": 662913476.0, + "step": 25617 + }, + { + "epoch": 2.8133099055567756, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3621344566345215, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.74123215675354, + "num_tokens": 662940969.0, + "step": 25618 + }, + { + "epoch": 2.8134197232593894, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1438963413238525, + "learning_rate": 1e-06, + "loss": 0.7648, + "mean_token_accuracy": 0.7486591935157776, + "num_tokens": 662969481.0, + "step": 25619 + }, + { + "epoch": 2.813529540962003, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.733916997909546, + "learning_rate": 1e-06, + "loss": 0.8422, + "mean_token_accuracy": 0.7233633995056152, + "num_tokens": 662990900.0, + "step": 25620 + }, + { + "epoch": 2.813639358664617, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0823774337768555, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.6988880634307861, + "num_tokens": 663025348.0, + "step": 25621 + }, + { + "epoch": 2.8137491763672307, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1710739135742188, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7281073331832886, + "num_tokens": 663055529.0, + "step": 25622 + }, + { + "epoch": 2.813858994069844, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.308039665222168, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7435791492462158, + "num_tokens": 663082171.0, + "step": 25623 + }, + { + "epoch": 2.8139688117724577, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7019944190979004, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7444726228713989, + "num_tokens": 663102817.0, + "step": 25624 + }, + { + "epoch": 2.8140786294750715, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.399527072906494, + "learning_rate": 1e-06, + "loss": 0.8872, + "mean_token_accuracy": 0.7226244211196899, + "num_tokens": 663126762.0, + "step": 25625 + }, + { + "epoch": 2.814188447177685, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5798609256744385, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7346396446228027, + "num_tokens": 663149265.0, + "step": 25626 + }, + { + "epoch": 2.8142982648802986, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.628580093383789, + "learning_rate": 1e-06, + "loss": 0.7935, + "mean_token_accuracy": 0.7477298974990845, + "num_tokens": 663171640.0, + "step": 25627 + }, + { + "epoch": 2.8144080825829123, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0186352729797363, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.742347240447998, + "num_tokens": 663205859.0, + "step": 25628 + }, + { + "epoch": 2.814517900285526, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4604196548461914, + "learning_rate": 1e-06, + "loss": 0.7891, + "mean_token_accuracy": 0.7529672980308533, + "num_tokens": 663229362.0, + "step": 25629 + }, + { + "epoch": 2.81462771798814, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.49019718170166, + "learning_rate": 1e-06, + "loss": 0.7849, + "mean_token_accuracy": 0.7516612410545349, + "num_tokens": 663251463.0, + "step": 25630 + }, + { + "epoch": 2.814737535690753, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4935739040374756, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7258865833282471, + "num_tokens": 663275502.0, + "step": 25631 + }, + { + "epoch": 2.814847353393367, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.159512996673584, + "learning_rate": 1e-06, + "loss": 0.8695, + "mean_token_accuracy": 0.7273772358894348, + "num_tokens": 663305986.0, + "step": 25632 + }, + { + "epoch": 2.8149571710959806, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3801870346069336, + "learning_rate": 1e-06, + "loss": 0.7694, + "mean_token_accuracy": 0.7551497220993042, + "num_tokens": 663330428.0, + "step": 25633 + }, + { + "epoch": 2.8150669887985944, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3917582035064697, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7307968735694885, + "num_tokens": 663356398.0, + "step": 25634 + }, + { + "epoch": 2.815176806501208, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7348082065582275, + "learning_rate": 1e-06, + "loss": 0.8339, + "mean_token_accuracy": 0.7328031063079834, + "num_tokens": 663376647.0, + "step": 25635 + }, + { + "epoch": 2.8152866242038215, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5704305171966553, + "learning_rate": 1e-06, + "loss": 0.8184, + "mean_token_accuracy": 0.741439700126648, + "num_tokens": 663399079.0, + "step": 25636 + }, + { + "epoch": 2.8153964419064352, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.366750717163086, + "learning_rate": 1e-06, + "loss": 0.8412, + "mean_token_accuracy": 0.7422176599502563, + "num_tokens": 663424783.0, + "step": 25637 + }, + { + "epoch": 2.815506259609049, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0863912105560303, + "learning_rate": 1e-06, + "loss": 0.867, + "mean_token_accuracy": 0.7225279808044434, + "num_tokens": 663456579.0, + "step": 25638 + }, + { + "epoch": 2.8156160773116627, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.404514789581299, + "learning_rate": 1e-06, + "loss": 0.881, + "mean_token_accuracy": 0.7217074632644653, + "num_tokens": 663481581.0, + "step": 25639 + }, + { + "epoch": 2.8157258950142765, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.488603115081787, + "learning_rate": 1e-06, + "loss": 0.8229, + "mean_token_accuracy": 0.7363150119781494, + "num_tokens": 663504593.0, + "step": 25640 + }, + { + "epoch": 2.81583571271689, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3660037517547607, + "learning_rate": 1e-06, + "loss": 0.8364, + "mean_token_accuracy": 0.7415636777877808, + "num_tokens": 663529251.0, + "step": 25641 + }, + { + "epoch": 2.8159455304195036, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.417295217514038, + "learning_rate": 1e-06, + "loss": 0.7799, + "mean_token_accuracy": 0.7523046731948853, + "num_tokens": 663553956.0, + "step": 25642 + }, + { + "epoch": 2.8160553481221173, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5294265747070312, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.7415686845779419, + "num_tokens": 663575866.0, + "step": 25643 + }, + { + "epoch": 2.816165165824731, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0496740341186523, + "learning_rate": 1e-06, + "loss": 0.8733, + "mean_token_accuracy": 0.7229539155960083, + "num_tokens": 663609903.0, + "step": 25644 + }, + { + "epoch": 2.816274983527345, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2940046787261963, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7304285764694214, + "num_tokens": 663637331.0, + "step": 25645 + }, + { + "epoch": 2.816384801229958, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.306353807449341, + "learning_rate": 1e-06, + "loss": 0.8269, + "mean_token_accuracy": 0.7401106357574463, + "num_tokens": 663667723.0, + "step": 25646 + }, + { + "epoch": 2.816494618932572, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.430569887161255, + "learning_rate": 1e-06, + "loss": 0.8838, + "mean_token_accuracy": 0.7241173386573792, + "num_tokens": 663691961.0, + "step": 25647 + }, + { + "epoch": 2.8166044366351857, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.8042166233062744, + "learning_rate": 1e-06, + "loss": 0.7599, + "mean_token_accuracy": 0.75836181640625, + "num_tokens": 663712363.0, + "step": 25648 + }, + { + "epoch": 2.816714254337799, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.056447982788086, + "learning_rate": 1e-06, + "loss": 0.9027, + "mean_token_accuracy": 0.717947781085968, + "num_tokens": 663744232.0, + "step": 25649 + }, + { + "epoch": 2.816824072040413, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.283284902572632, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7296624779701233, + "num_tokens": 663771402.0, + "step": 25650 + }, + { + "epoch": 2.8169338897430265, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.280341386795044, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7316811084747314, + "num_tokens": 663798841.0, + "step": 25651 + }, + { + "epoch": 2.8170437074456403, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.280522346496582, + "learning_rate": 1e-06, + "loss": 0.8546, + "mean_token_accuracy": 0.7243180871009827, + "num_tokens": 663825270.0, + "step": 25652 + }, + { + "epoch": 2.817153525148254, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.040435552597046, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7157257199287415, + "num_tokens": 663856581.0, + "step": 25653 + }, + { + "epoch": 2.8172633428508673, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2868919372558594, + "learning_rate": 1e-06, + "loss": 0.7737, + "mean_token_accuracy": 0.7548926472663879, + "num_tokens": 663882578.0, + "step": 25654 + }, + { + "epoch": 2.817373160553481, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2995784282684326, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7260169982910156, + "num_tokens": 663908507.0, + "step": 25655 + }, + { + "epoch": 2.817482978256095, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.343012571334839, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7438088059425354, + "num_tokens": 663934012.0, + "step": 25656 + }, + { + "epoch": 2.8175927959587086, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 1.9358649253845215, + "learning_rate": 1e-06, + "loss": 0.8787, + "mean_token_accuracy": 0.7194116711616516, + "num_tokens": 663971194.0, + "step": 25657 + }, + { + "epoch": 2.8177026136613224, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3648223876953125, + "learning_rate": 1e-06, + "loss": 0.7904, + "mean_token_accuracy": 0.7447125911712646, + "num_tokens": 663996352.0, + "step": 25658 + }, + { + "epoch": 2.8178124313639357, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6625707149505615, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.724916934967041, + "num_tokens": 664016776.0, + "step": 25659 + }, + { + "epoch": 2.8179222490665494, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3425936698913574, + "learning_rate": 1e-06, + "loss": 0.7662, + "mean_token_accuracy": 0.7564562559127808, + "num_tokens": 664042536.0, + "step": 25660 + }, + { + "epoch": 2.818032066769163, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2933568954467773, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.7552853226661682, + "num_tokens": 664070228.0, + "step": 25661 + }, + { + "epoch": 2.818141884471777, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4952967166900635, + "learning_rate": 1e-06, + "loss": 0.7973, + "mean_token_accuracy": 0.7445557713508606, + "num_tokens": 664093475.0, + "step": 25662 + }, + { + "epoch": 2.8182517021743907, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6280765533447266, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7328765392303467, + "num_tokens": 664114673.0, + "step": 25663 + }, + { + "epoch": 2.818361519877004, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4450855255126953, + "learning_rate": 1e-06, + "loss": 0.841, + "mean_token_accuracy": 0.7333680391311646, + "num_tokens": 664139000.0, + "step": 25664 + }, + { + "epoch": 2.8184713375796178, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2392776012420654, + "learning_rate": 1e-06, + "loss": 0.9064, + "mean_token_accuracy": 0.7203332781791687, + "num_tokens": 664169452.0, + "step": 25665 + }, + { + "epoch": 2.8185811552822315, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.592115640640259, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7419848442077637, + "num_tokens": 664190664.0, + "step": 25666 + }, + { + "epoch": 2.8186909729848453, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4639077186584473, + "learning_rate": 1e-06, + "loss": 0.8795, + "mean_token_accuracy": 0.731498122215271, + "num_tokens": 664214414.0, + "step": 25667 + }, + { + "epoch": 2.818800790687459, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.185046672821045, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7324806451797485, + "num_tokens": 664243671.0, + "step": 25668 + }, + { + "epoch": 2.8189106083900723, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.243257999420166, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7204824090003967, + "num_tokens": 664272682.0, + "step": 25669 + }, + { + "epoch": 2.819020426092686, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4093775749206543, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.733546257019043, + "num_tokens": 664297167.0, + "step": 25670 + }, + { + "epoch": 2.8191302437953, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1796531677246094, + "learning_rate": 1e-06, + "loss": 0.9117, + "mean_token_accuracy": 0.7200043201446533, + "num_tokens": 664327436.0, + "step": 25671 + }, + { + "epoch": 2.8192400614979136, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.314535617828369, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7305192947387695, + "num_tokens": 664355794.0, + "step": 25672 + }, + { + "epoch": 2.8193498792005274, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2386608123779297, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7078282237052917, + "num_tokens": 664385123.0, + "step": 25673 + }, + { + "epoch": 2.8194596969031407, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 3.3588016033172607, + "learning_rate": 1e-06, + "loss": 0.707, + "mean_token_accuracy": 0.7687808871269226, + "num_tokens": 664399166.0, + "step": 25674 + }, + { + "epoch": 2.8195695146057544, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2835147380828857, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7210528254508972, + "num_tokens": 664426734.0, + "step": 25675 + }, + { + "epoch": 2.819679332308368, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4579949378967285, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7504785060882568, + "num_tokens": 664449383.0, + "step": 25676 + }, + { + "epoch": 2.8197891500109815, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.190302848815918, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7370380163192749, + "num_tokens": 664478943.0, + "step": 25677 + }, + { + "epoch": 2.8198989677135953, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.501514196395874, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7253156900405884, + "num_tokens": 664504458.0, + "step": 25678 + }, + { + "epoch": 2.820008785416209, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.080718755722046, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7197975516319275, + "num_tokens": 664535890.0, + "step": 25679 + }, + { + "epoch": 2.820118603118823, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.565228223800659, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7393442392349243, + "num_tokens": 664558345.0, + "step": 25680 + }, + { + "epoch": 2.8202284208214365, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.531104326248169, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7362404465675354, + "num_tokens": 664580299.0, + "step": 25681 + }, + { + "epoch": 2.82033823852405, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.8619370460510254, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.740676999092102, + "num_tokens": 664599077.0, + "step": 25682 + }, + { + "epoch": 2.8204480562266636, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4923813343048096, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7331353425979614, + "num_tokens": 664624121.0, + "step": 25683 + }, + { + "epoch": 2.8205578739292774, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 1.973595380783081, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7304761409759521, + "num_tokens": 664660951.0, + "step": 25684 + }, + { + "epoch": 2.820667691631891, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5324344635009766, + "learning_rate": 1e-06, + "loss": 0.7716, + "mean_token_accuracy": 0.7559947967529297, + "num_tokens": 664684410.0, + "step": 25685 + }, + { + "epoch": 2.820777509334505, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5556325912475586, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7337363362312317, + "num_tokens": 664706851.0, + "step": 25686 + }, + { + "epoch": 2.820887327037118, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.458696126937866, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7289915680885315, + "num_tokens": 664731863.0, + "step": 25687 + }, + { + "epoch": 2.820997144739732, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.714845657348633, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7381911873817444, + "num_tokens": 664752228.0, + "step": 25688 + }, + { + "epoch": 2.8211069624423457, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5228796005249023, + "learning_rate": 1e-06, + "loss": 0.7154, + "mean_token_accuracy": 0.7624504566192627, + "num_tokens": 664773933.0, + "step": 25689 + }, + { + "epoch": 2.8212167801449595, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1888530254364014, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7301260828971863, + "num_tokens": 664802609.0, + "step": 25690 + }, + { + "epoch": 2.8213265978475732, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.880312204360962, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7302387356758118, + "num_tokens": 664821200.0, + "step": 25691 + }, + { + "epoch": 2.8214364155501865, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.339221239089966, + "learning_rate": 1e-06, + "loss": 0.888, + "mean_token_accuracy": 0.7188431620597839, + "num_tokens": 664848426.0, + "step": 25692 + }, + { + "epoch": 2.8215462332528003, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.246290922164917, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7339469194412231, + "num_tokens": 664874057.0, + "step": 25693 + }, + { + "epoch": 2.821656050955414, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5944786071777344, + "learning_rate": 1e-06, + "loss": 0.8099, + "mean_token_accuracy": 0.7370429039001465, + "num_tokens": 664897031.0, + "step": 25694 + }, + { + "epoch": 2.821765868658028, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.433802366256714, + "learning_rate": 1e-06, + "loss": 0.9496, + "mean_token_accuracy": 0.7034042477607727, + "num_tokens": 664922382.0, + "step": 25695 + }, + { + "epoch": 2.8218756863606416, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.407761573791504, + "learning_rate": 1e-06, + "loss": 0.8531, + "mean_token_accuracy": 0.7318012714385986, + "num_tokens": 664946332.0, + "step": 25696 + }, + { + "epoch": 2.821985504063255, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4236745834350586, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7324097752571106, + "num_tokens": 664970614.0, + "step": 25697 + }, + { + "epoch": 2.8220953217658686, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3443233966827393, + "learning_rate": 1e-06, + "loss": 0.8638, + "mean_token_accuracy": 0.7273188829421997, + "num_tokens": 664997642.0, + "step": 25698 + }, + { + "epoch": 2.8222051394684824, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.16317081451416, + "learning_rate": 1e-06, + "loss": 0.815, + "mean_token_accuracy": 0.744499683380127, + "num_tokens": 665025423.0, + "step": 25699 + }, + { + "epoch": 2.8223149571710957, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.269862174987793, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7269947528839111, + "num_tokens": 665053357.0, + "step": 25700 + }, + { + "epoch": 2.82242477487371, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.47741436958313, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7369042634963989, + "num_tokens": 665078191.0, + "step": 25701 + }, + { + "epoch": 2.822534592576323, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6386473178863525, + "learning_rate": 1e-06, + "loss": 0.7478, + "mean_token_accuracy": 0.7570176124572754, + "num_tokens": 665100123.0, + "step": 25702 + }, + { + "epoch": 2.822644410278937, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1839823722839355, + "learning_rate": 1e-06, + "loss": 0.9815, + "mean_token_accuracy": 0.6964695453643799, + "num_tokens": 665130091.0, + "step": 25703 + }, + { + "epoch": 2.8227542279815507, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.9212348461151123, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7388701438903809, + "num_tokens": 665148247.0, + "step": 25704 + }, + { + "epoch": 2.822864045684164, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.087125062942505, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7122732996940613, + "num_tokens": 665178671.0, + "step": 25705 + }, + { + "epoch": 2.822973863386778, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.138899087905884, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.734981894493103, + "num_tokens": 665206942.0, + "step": 25706 + }, + { + "epoch": 2.8230836810893916, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.772825002670288, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.729251503944397, + "num_tokens": 665228016.0, + "step": 25707 + }, + { + "epoch": 2.8231934987920053, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4429893493652344, + "learning_rate": 1e-06, + "loss": 0.8947, + "mean_token_accuracy": 0.7268444299697876, + "num_tokens": 665252564.0, + "step": 25708 + }, + { + "epoch": 2.823303316494619, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3403358459472656, + "learning_rate": 1e-06, + "loss": 0.9307, + "mean_token_accuracy": 0.7155895233154297, + "num_tokens": 665279318.0, + "step": 25709 + }, + { + "epoch": 2.8234131341972324, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2645599842071533, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7411919832229614, + "num_tokens": 665305386.0, + "step": 25710 + }, + { + "epoch": 2.823522951899846, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.618802547454834, + "learning_rate": 1e-06, + "loss": 0.6934, + "mean_token_accuracy": 0.7730967998504639, + "num_tokens": 665327383.0, + "step": 25711 + }, + { + "epoch": 2.82363276960246, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1604466438293457, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7167403697967529, + "num_tokens": 665356505.0, + "step": 25712 + }, + { + "epoch": 2.8237425873050737, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4258992671966553, + "learning_rate": 1e-06, + "loss": 0.886, + "mean_token_accuracy": 0.7177081108093262, + "num_tokens": 665382325.0, + "step": 25713 + }, + { + "epoch": 2.8238524050076874, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2375428676605225, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7449396848678589, + "num_tokens": 665408629.0, + "step": 25714 + }, + { + "epoch": 2.8239622227103007, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2822272777557373, + "learning_rate": 1e-06, + "loss": 0.8835, + "mean_token_accuracy": 0.7246851921081543, + "num_tokens": 665436323.0, + "step": 25715 + }, + { + "epoch": 2.8240720404129145, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.225680351257324, + "learning_rate": 1e-06, + "loss": 0.743, + "mean_token_accuracy": 0.7727437019348145, + "num_tokens": 665463777.0, + "step": 25716 + }, + { + "epoch": 2.8241818581155282, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6123902797698975, + "learning_rate": 1e-06, + "loss": 0.8556, + "mean_token_accuracy": 0.7364094853401184, + "num_tokens": 665485836.0, + "step": 25717 + }, + { + "epoch": 2.824291675818142, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4334616661071777, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7247381210327148, + "num_tokens": 665510241.0, + "step": 25718 + }, + { + "epoch": 2.8244014935207558, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2472667694091797, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7125567197799683, + "num_tokens": 665539111.0, + "step": 25719 + }, + { + "epoch": 2.824511311223369, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.131183385848999, + "learning_rate": 1e-06, + "loss": 0.9631, + "mean_token_accuracy": 0.7116220593452454, + "num_tokens": 665572159.0, + "step": 25720 + }, + { + "epoch": 2.824621128925983, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2555527687072754, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7197571992874146, + "num_tokens": 665600417.0, + "step": 25721 + }, + { + "epoch": 2.8247309466285966, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5183398723602295, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7251661419868469, + "num_tokens": 665625111.0, + "step": 25722 + }, + { + "epoch": 2.8248407643312103, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7328906059265137, + "learning_rate": 1e-06, + "loss": 0.8341, + "mean_token_accuracy": 0.735637903213501, + "num_tokens": 665645082.0, + "step": 25723 + }, + { + "epoch": 2.824950582033824, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1403825283050537, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7183732986450195, + "num_tokens": 665675922.0, + "step": 25724 + }, + { + "epoch": 2.8250603997364374, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3945930004119873, + "learning_rate": 1e-06, + "loss": 0.786, + "mean_token_accuracy": 0.7454937696456909, + "num_tokens": 665701672.0, + "step": 25725 + }, + { + "epoch": 2.825170217439051, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.403933525085449, + "learning_rate": 1e-06, + "loss": 0.8101, + "mean_token_accuracy": 0.7412451505661011, + "num_tokens": 665726620.0, + "step": 25726 + }, + { + "epoch": 2.825280035141665, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3335371017456055, + "learning_rate": 1e-06, + "loss": 0.9093, + "mean_token_accuracy": 0.7124682664871216, + "num_tokens": 665752886.0, + "step": 25727 + }, + { + "epoch": 2.8253898528442782, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.45228910446167, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7271016836166382, + "num_tokens": 665776049.0, + "step": 25728 + }, + { + "epoch": 2.825499670546892, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.392446517944336, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7352675199508667, + "num_tokens": 665801166.0, + "step": 25729 + }, + { + "epoch": 2.8256094882495058, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2103774547576904, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7101907730102539, + "num_tokens": 665829909.0, + "step": 25730 + }, + { + "epoch": 2.8257193059521195, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.682143211364746, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7392371892929077, + "num_tokens": 665851708.0, + "step": 25731 + }, + { + "epoch": 2.8258291236547333, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.429483652114868, + "learning_rate": 1e-06, + "loss": 0.8636, + "mean_token_accuracy": 0.7261059284210205, + "num_tokens": 665875189.0, + "step": 25732 + }, + { + "epoch": 2.8259389413573466, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4632694721221924, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7253667116165161, + "num_tokens": 665897904.0, + "step": 25733 + }, + { + "epoch": 2.8260487590599603, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4829277992248535, + "learning_rate": 1e-06, + "loss": 0.7545, + "mean_token_accuracy": 0.7521661520004272, + "num_tokens": 665921091.0, + "step": 25734 + }, + { + "epoch": 2.826158576762574, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1923813819885254, + "learning_rate": 1e-06, + "loss": 0.8149, + "mean_token_accuracy": 0.7498005628585815, + "num_tokens": 665948538.0, + "step": 25735 + }, + { + "epoch": 2.826268394465188, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3964552879333496, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7206519842147827, + "num_tokens": 665974966.0, + "step": 25736 + }, + { + "epoch": 2.8263782121678016, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.766126871109009, + "learning_rate": 1e-06, + "loss": 0.8247, + "mean_token_accuracy": 0.7391878962516785, + "num_tokens": 665994760.0, + "step": 25737 + }, + { + "epoch": 2.826488029870415, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.291264533996582, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7281922101974487, + "num_tokens": 666021963.0, + "step": 25738 + }, + { + "epoch": 2.8265978475730287, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1390597820281982, + "learning_rate": 1e-06, + "loss": 0.8199, + "mean_token_accuracy": 0.7341819405555725, + "num_tokens": 666050709.0, + "step": 25739 + }, + { + "epoch": 2.8267076652756424, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1165876388549805, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7186157703399658, + "num_tokens": 666080470.0, + "step": 25740 + }, + { + "epoch": 2.826817482978256, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3294122219085693, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.7235303521156311, + "num_tokens": 666105655.0, + "step": 25741 + }, + { + "epoch": 2.82692730068087, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.523977756500244, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7346223592758179, + "num_tokens": 666129224.0, + "step": 25742 + }, + { + "epoch": 2.8270371183834833, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3501975536346436, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7330414652824402, + "num_tokens": 666153247.0, + "step": 25743 + }, + { + "epoch": 2.827146936086097, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6369755268096924, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7313483953475952, + "num_tokens": 666175056.0, + "step": 25744 + }, + { + "epoch": 2.8272567537887108, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1552367210388184, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7238917946815491, + "num_tokens": 666207477.0, + "step": 25745 + }, + { + "epoch": 2.8273665714913245, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.43789005279541, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7246954441070557, + "num_tokens": 666233368.0, + "step": 25746 + }, + { + "epoch": 2.8274763891939383, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4593276977539062, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.730915367603302, + "num_tokens": 666258159.0, + "step": 25747 + }, + { + "epoch": 2.8275862068965516, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.446668863296509, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7269501686096191, + "num_tokens": 666284219.0, + "step": 25748 + }, + { + "epoch": 2.8276960245991654, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.665778875350952, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7390097379684448, + "num_tokens": 666305589.0, + "step": 25749 + }, + { + "epoch": 2.827805842301779, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.484861373901367, + "learning_rate": 1e-06, + "loss": 0.8978, + "mean_token_accuracy": 0.7164674401283264, + "num_tokens": 666330272.0, + "step": 25750 + }, + { + "epoch": 2.827915660004393, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3603546619415283, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7349874973297119, + "num_tokens": 666356032.0, + "step": 25751 + }, + { + "epoch": 2.8280254777070066, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 3.081822633743286, + "learning_rate": 1e-06, + "loss": 0.7546, + "mean_token_accuracy": 0.7553932666778564, + "num_tokens": 666372562.0, + "step": 25752 + }, + { + "epoch": 2.82813529540962, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6835715770721436, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7367430329322815, + "num_tokens": 666395419.0, + "step": 25753 + }, + { + "epoch": 2.8282451131122337, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.279416561126709, + "learning_rate": 1e-06, + "loss": 0.7765, + "mean_token_accuracy": 0.7645553350448608, + "num_tokens": 666420506.0, + "step": 25754 + }, + { + "epoch": 2.8283549308148475, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.188631296157837, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.7105991244316101, + "num_tokens": 666450363.0, + "step": 25755 + }, + { + "epoch": 2.8284647485174608, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5469205379486084, + "learning_rate": 1e-06, + "loss": 0.8076, + "mean_token_accuracy": 0.7414140701293945, + "num_tokens": 666472731.0, + "step": 25756 + }, + { + "epoch": 2.8285745662200745, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4065146446228027, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7237998843193054, + "num_tokens": 666498792.0, + "step": 25757 + }, + { + "epoch": 2.8286843839226883, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.297034740447998, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7181636095046997, + "num_tokens": 666526752.0, + "step": 25758 + }, + { + "epoch": 2.828794201625302, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5741264820098877, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7610801458358765, + "num_tokens": 666548945.0, + "step": 25759 + }, + { + "epoch": 2.828904019327916, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6570916175842285, + "learning_rate": 1e-06, + "loss": 0.8991, + "mean_token_accuracy": 0.7220863699913025, + "num_tokens": 666569835.0, + "step": 25760 + }, + { + "epoch": 2.829013837030529, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.49392032623291, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7341029047966003, + "num_tokens": 666593958.0, + "step": 25761 + }, + { + "epoch": 2.829123654733143, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4698712825775146, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.725666880607605, + "num_tokens": 666618422.0, + "step": 25762 + }, + { + "epoch": 2.8292334724357566, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6831674575805664, + "learning_rate": 1e-06, + "loss": 0.7633, + "mean_token_accuracy": 0.7543648481369019, + "num_tokens": 666639019.0, + "step": 25763 + }, + { + "epoch": 2.8293432901383704, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.49468994140625, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.7323765158653259, + "num_tokens": 666662132.0, + "step": 25764 + }, + { + "epoch": 2.829453107840984, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.787916421890259, + "learning_rate": 1e-06, + "loss": 0.8242, + "mean_token_accuracy": 0.7445300221443176, + "num_tokens": 666680745.0, + "step": 25765 + }, + { + "epoch": 2.8295629255435975, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2676315307617188, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7469421625137329, + "num_tokens": 666708013.0, + "step": 25766 + }, + { + "epoch": 2.829672743246211, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.220144748687744, + "learning_rate": 1e-06, + "loss": 0.9787, + "mean_token_accuracy": 0.6952046155929565, + "num_tokens": 666738869.0, + "step": 25767 + }, + { + "epoch": 2.829782560948825, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3156373500823975, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.735463559627533, + "num_tokens": 666764646.0, + "step": 25768 + }, + { + "epoch": 2.8298923786514387, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2979848384857178, + "learning_rate": 1e-06, + "loss": 0.9031, + "mean_token_accuracy": 0.7149820923805237, + "num_tokens": 666792890.0, + "step": 25769 + }, + { + "epoch": 2.8300021963540525, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3693552017211914, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.7422479391098022, + "num_tokens": 666817196.0, + "step": 25770 + }, + { + "epoch": 2.830112014056666, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.678734540939331, + "learning_rate": 1e-06, + "loss": 0.7107, + "mean_token_accuracy": 0.7714095711708069, + "num_tokens": 666837965.0, + "step": 25771 + }, + { + "epoch": 2.8302218317592795, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.462907075881958, + "learning_rate": 1e-06, + "loss": 0.9384, + "mean_token_accuracy": 0.725191056728363, + "num_tokens": 666861412.0, + "step": 25772 + }, + { + "epoch": 2.8303316494618933, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.656377077102661, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7254632711410522, + "num_tokens": 666883947.0, + "step": 25773 + }, + { + "epoch": 2.830441467164507, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.718391180038452, + "learning_rate": 1e-06, + "loss": 0.7953, + "mean_token_accuracy": 0.7521106004714966, + "num_tokens": 666904341.0, + "step": 25774 + }, + { + "epoch": 2.830551284867121, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4021942615509033, + "learning_rate": 1e-06, + "loss": 0.8279, + "mean_token_accuracy": 0.7385237812995911, + "num_tokens": 666926188.0, + "step": 25775 + }, + { + "epoch": 2.830661102569734, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4781646728515625, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7332260608673096, + "num_tokens": 666950234.0, + "step": 25776 + }, + { + "epoch": 2.830770920272348, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3239991664886475, + "learning_rate": 1e-06, + "loss": 0.8643, + "mean_token_accuracy": 0.7246047854423523, + "num_tokens": 666976412.0, + "step": 25777 + }, + { + "epoch": 2.8308807379749616, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2078492641448975, + "learning_rate": 1e-06, + "loss": 0.9366, + "mean_token_accuracy": 0.7056534886360168, + "num_tokens": 667005457.0, + "step": 25778 + }, + { + "epoch": 2.830990555677575, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.607489824295044, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7387918829917908, + "num_tokens": 667028380.0, + "step": 25779 + }, + { + "epoch": 2.831100373380189, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.039222478866577, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7123279571533203, + "num_tokens": 667059785.0, + "step": 25780 + }, + { + "epoch": 2.8312101910828025, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3428804874420166, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7429249286651611, + "num_tokens": 667083512.0, + "step": 25781 + }, + { + "epoch": 2.8313200087854162, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.643151044845581, + "learning_rate": 1e-06, + "loss": 0.7045, + "mean_token_accuracy": 0.7634243965148926, + "num_tokens": 667104526.0, + "step": 25782 + }, + { + "epoch": 2.83142982648803, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.728447914123535, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.720497727394104, + "num_tokens": 667125859.0, + "step": 25783 + }, + { + "epoch": 2.8315396441906433, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3757760524749756, + "learning_rate": 1e-06, + "loss": 0.8353, + "mean_token_accuracy": 0.7349470853805542, + "num_tokens": 667151027.0, + "step": 25784 + }, + { + "epoch": 2.831649461893257, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 1.9066983461380005, + "learning_rate": 1e-06, + "loss": 0.8128, + "mean_token_accuracy": 0.7487002611160278, + "num_tokens": 667186120.0, + "step": 25785 + }, + { + "epoch": 2.831759279595871, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6791138648986816, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7252533435821533, + "num_tokens": 667209178.0, + "step": 25786 + }, + { + "epoch": 2.8318690972984846, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.421610116958618, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7436748743057251, + "num_tokens": 667233824.0, + "step": 25787 + }, + { + "epoch": 2.8319789150010983, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.690439462661743, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7451775074005127, + "num_tokens": 667255284.0, + "step": 25788 + }, + { + "epoch": 2.8320887327037116, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2440950870513916, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7135177254676819, + "num_tokens": 667284434.0, + "step": 25789 + }, + { + "epoch": 2.8321985504063254, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4940719604492188, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7291767001152039, + "num_tokens": 667307963.0, + "step": 25790 + }, + { + "epoch": 2.832308368108939, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7174487113952637, + "learning_rate": 1e-06, + "loss": 0.8241, + "mean_token_accuracy": 0.7349956035614014, + "num_tokens": 667330400.0, + "step": 25791 + }, + { + "epoch": 2.832418185811553, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4630465507507324, + "learning_rate": 1e-06, + "loss": 0.8832, + "mean_token_accuracy": 0.7169221639633179, + "num_tokens": 667353480.0, + "step": 25792 + }, + { + "epoch": 2.8325280035141667, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4511194229125977, + "learning_rate": 1e-06, + "loss": 0.8108, + "mean_token_accuracy": 0.7430073022842407, + "num_tokens": 667378113.0, + "step": 25793 + }, + { + "epoch": 2.83263782121678, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2792723178863525, + "learning_rate": 1e-06, + "loss": 0.7579, + "mean_token_accuracy": 0.7536293268203735, + "num_tokens": 667405448.0, + "step": 25794 + }, + { + "epoch": 2.8327476389193937, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2899656295776367, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7467336058616638, + "num_tokens": 667431979.0, + "step": 25795 + }, + { + "epoch": 2.8328574566220075, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5434858798980713, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.7469609975814819, + "num_tokens": 667454702.0, + "step": 25796 + }, + { + "epoch": 2.8329672743246213, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.732088327407837, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7273968458175659, + "num_tokens": 667476501.0, + "step": 25797 + }, + { + "epoch": 2.833077092027235, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.401618480682373, + "learning_rate": 1e-06, + "loss": 0.8577, + "mean_token_accuracy": 0.7291262745857239, + "num_tokens": 667502412.0, + "step": 25798 + }, + { + "epoch": 2.8331869097298483, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.767397403717041, + "learning_rate": 1e-06, + "loss": 0.7484, + "mean_token_accuracy": 0.7563192248344421, + "num_tokens": 667521563.0, + "step": 25799 + }, + { + "epoch": 2.833296727432462, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3409762382507324, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7306427955627441, + "num_tokens": 667548691.0, + "step": 25800 + }, + { + "epoch": 2.833406545135076, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2818124294281006, + "learning_rate": 1e-06, + "loss": 0.9006, + "mean_token_accuracy": 0.7210049033164978, + "num_tokens": 667575700.0, + "step": 25801 + }, + { + "epoch": 2.8335163628376896, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4536285400390625, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7337533235549927, + "num_tokens": 667599917.0, + "step": 25802 + }, + { + "epoch": 2.8336261805403034, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2066547870635986, + "learning_rate": 1e-06, + "loss": 0.9154, + "mean_token_accuracy": 0.7146610021591187, + "num_tokens": 667630484.0, + "step": 25803 + }, + { + "epoch": 2.8337359982429167, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0454609394073486, + "learning_rate": 1e-06, + "loss": 0.9425, + "mean_token_accuracy": 0.7079792022705078, + "num_tokens": 667662177.0, + "step": 25804 + }, + { + "epoch": 2.8338458159455304, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.371068000793457, + "learning_rate": 1e-06, + "loss": 0.9236, + "mean_token_accuracy": 0.7171767950057983, + "num_tokens": 667687948.0, + "step": 25805 + }, + { + "epoch": 2.833955633648144, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.478858709335327, + "learning_rate": 1e-06, + "loss": 0.8515, + "mean_token_accuracy": 0.7311111688613892, + "num_tokens": 667711620.0, + "step": 25806 + }, + { + "epoch": 2.8340654513507575, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5384867191314697, + "learning_rate": 1e-06, + "loss": 0.7932, + "mean_token_accuracy": 0.7559205293655396, + "num_tokens": 667733656.0, + "step": 25807 + }, + { + "epoch": 2.8341752690533712, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.046272039413452, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.7236136198043823, + "num_tokens": 667766584.0, + "step": 25808 + }, + { + "epoch": 2.834285086755985, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3081793785095215, + "learning_rate": 1e-06, + "loss": 0.9009, + "mean_token_accuracy": 0.7156002521514893, + "num_tokens": 667795665.0, + "step": 25809 + }, + { + "epoch": 2.8343949044585988, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5188982486724854, + "learning_rate": 1e-06, + "loss": 0.8343, + "mean_token_accuracy": 0.7348052859306335, + "num_tokens": 667820599.0, + "step": 25810 + }, + { + "epoch": 2.8345047221612125, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.2043216228485107, + "learning_rate": 1e-06, + "loss": 0.825, + "mean_token_accuracy": 0.7389421463012695, + "num_tokens": 667849042.0, + "step": 25811 + }, + { + "epoch": 2.834614539863826, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.1869428157806396, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7515235543251038, + "num_tokens": 667877786.0, + "step": 25812 + }, + { + "epoch": 2.8347243575664396, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.7063517570495605, + "learning_rate": 1e-06, + "loss": 0.802, + "mean_token_accuracy": 0.750601053237915, + "num_tokens": 667900474.0, + "step": 25813 + }, + { + "epoch": 2.8348341752690533, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.104829788208008, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7392619848251343, + "num_tokens": 667930778.0, + "step": 25814 + }, + { + "epoch": 2.834943992971667, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.335598945617676, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7532973885536194, + "num_tokens": 667956985.0, + "step": 25815 + }, + { + "epoch": 2.835053810674281, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.6812827587127686, + "learning_rate": 1e-06, + "loss": 0.7491, + "mean_token_accuracy": 0.7607255578041077, + "num_tokens": 667976498.0, + "step": 25816 + }, + { + "epoch": 2.835163628376894, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.570493698120117, + "learning_rate": 1e-06, + "loss": 0.9572, + "mean_token_accuracy": 0.7032525539398193, + "num_tokens": 667999562.0, + "step": 25817 + }, + { + "epoch": 2.835273446079508, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.445338726043701, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.7283377051353455, + "num_tokens": 668024673.0, + "step": 25818 + }, + { + "epoch": 2.8353832637821217, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.710221767425537, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7494789361953735, + "num_tokens": 668044984.0, + "step": 25819 + }, + { + "epoch": 2.8354930814847354, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.877530813217163, + "learning_rate": 1e-06, + "loss": 0.7807, + "mean_token_accuracy": 0.7476015090942383, + "num_tokens": 668062676.0, + "step": 25820 + }, + { + "epoch": 2.835602899187349, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.0467231273651123, + "learning_rate": 1e-06, + "loss": 0.9068, + "mean_token_accuracy": 0.7149908542633057, + "num_tokens": 668095964.0, + "step": 25821 + }, + { + "epoch": 2.8357127168899625, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.400279998779297, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7290782332420349, + "num_tokens": 668120610.0, + "step": 25822 + }, + { + "epoch": 2.8358225345925763, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.150219202041626, + "learning_rate": 1e-06, + "loss": 0.931, + "mean_token_accuracy": 0.7080954313278198, + "num_tokens": 668154666.0, + "step": 25823 + }, + { + "epoch": 2.83593235229519, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.5782222747802734, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7178711891174316, + "num_tokens": 668179455.0, + "step": 25824 + }, + { + "epoch": 2.836042169997804, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3785502910614014, + "learning_rate": 1e-06, + "loss": 0.7828, + "mean_token_accuracy": 0.7466650009155273, + "num_tokens": 668204743.0, + "step": 25825 + }, + { + "epoch": 2.8361519877004175, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4188878536224365, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7320889234542847, + "num_tokens": 668229429.0, + "step": 25826 + }, + { + "epoch": 2.836261805403031, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.486196756362915, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7474822998046875, + "num_tokens": 668252642.0, + "step": 25827 + }, + { + "epoch": 2.8363716231056446, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4734082221984863, + "learning_rate": 1e-06, + "loss": 0.8308, + "mean_token_accuracy": 0.733727216720581, + "num_tokens": 668274927.0, + "step": 25828 + }, + { + "epoch": 2.8364814408082584, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.502124786376953, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7303768396377563, + "num_tokens": 668300529.0, + "step": 25829 + }, + { + "epoch": 2.8365912585108717, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.491255044937134, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7461711168289185, + "num_tokens": 668323105.0, + "step": 25830 + }, + { + "epoch": 2.836701076213486, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.786646604537964, + "learning_rate": 1e-06, + "loss": 0.822, + "mean_token_accuracy": 0.7349385023117065, + "num_tokens": 668343240.0, + "step": 25831 + }, + { + "epoch": 2.836810893916099, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.519801378250122, + "learning_rate": 1e-06, + "loss": 0.9088, + "mean_token_accuracy": 0.7237488627433777, + "num_tokens": 668366235.0, + "step": 25832 + }, + { + "epoch": 2.836920711618713, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.4501445293426514, + "learning_rate": 1e-06, + "loss": 0.9156, + "mean_token_accuracy": 0.7125344276428223, + "num_tokens": 668390790.0, + "step": 25833 + }, + { + "epoch": 2.8370305293213267, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.3671622276306152, + "learning_rate": 1e-06, + "loss": 0.862, + "mean_token_accuracy": 0.7300007939338684, + "num_tokens": 668417367.0, + "step": 25834 + }, + { + "epoch": 2.83714034702394, + "ewc_loss": 1.633167266845703e-05, + "grad_norm": 2.914787530899048, + "learning_rate": 1e-06, + "loss": 0.7885, + "mean_token_accuracy": 0.7423981428146362, + "num_tokens": 668436434.0, + "step": 25835 + }, + { + "epoch": 2.837250164726554, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3465042114257812, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7464959621429443, + "num_tokens": 668462158.0, + "step": 25836 + }, + { + "epoch": 2.8373599824291675, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5102198123931885, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7262454628944397, + "num_tokens": 668484016.0, + "step": 25837 + }, + { + "epoch": 2.8374698001317813, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4327943325042725, + "learning_rate": 1e-06, + "loss": 0.8632, + "mean_token_accuracy": 0.7306520938873291, + "num_tokens": 668509697.0, + "step": 25838 + }, + { + "epoch": 2.837579617834395, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4938650131225586, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7320021986961365, + "num_tokens": 668534839.0, + "step": 25839 + }, + { + "epoch": 2.8376894355370084, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5407626628875732, + "learning_rate": 1e-06, + "loss": 0.8516, + "mean_token_accuracy": 0.7296421527862549, + "num_tokens": 668558902.0, + "step": 25840 + }, + { + "epoch": 2.837799253239622, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2730815410614014, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7173030376434326, + "num_tokens": 668589644.0, + "step": 25841 + }, + { + "epoch": 2.837909070942236, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.00734543800354, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7198233604431152, + "num_tokens": 668623296.0, + "step": 25842 + }, + { + "epoch": 2.8380188886448496, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.113722324371338, + "learning_rate": 1e-06, + "loss": 0.9111, + "mean_token_accuracy": 0.714002251625061, + "num_tokens": 668652015.0, + "step": 25843 + }, + { + "epoch": 2.8381287063474634, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2416627407073975, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7282744646072388, + "num_tokens": 668681762.0, + "step": 25844 + }, + { + "epoch": 2.8382385240500767, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2527740001678467, + "learning_rate": 1e-06, + "loss": 0.753, + "mean_token_accuracy": 0.7539185285568237, + "num_tokens": 668706177.0, + "step": 25845 + }, + { + "epoch": 2.8383483417526905, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8127706050872803, + "learning_rate": 1e-06, + "loss": 0.7287, + "mean_token_accuracy": 0.7623550295829773, + "num_tokens": 668724696.0, + "step": 25846 + }, + { + "epoch": 2.838458159455304, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.415497064590454, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7435410618782043, + "num_tokens": 668748726.0, + "step": 25847 + }, + { + "epoch": 2.838567977157918, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5797085762023926, + "learning_rate": 1e-06, + "loss": 0.7919, + "mean_token_accuracy": 0.7496451139450073, + "num_tokens": 668771087.0, + "step": 25848 + }, + { + "epoch": 2.8386777948605317, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4673869609832764, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7219462990760803, + "num_tokens": 668796119.0, + "step": 25849 + }, + { + "epoch": 2.838787612563145, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.380535364151001, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7319707274436951, + "num_tokens": 668821536.0, + "step": 25850 + }, + { + "epoch": 2.838897430265759, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3380494117736816, + "learning_rate": 1e-06, + "loss": 0.8738, + "mean_token_accuracy": 0.7319365739822388, + "num_tokens": 668849118.0, + "step": 25851 + }, + { + "epoch": 2.8390072479683726, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2993738651275635, + "learning_rate": 1e-06, + "loss": 0.915, + "mean_token_accuracy": 0.7097381353378296, + "num_tokens": 668878656.0, + "step": 25852 + }, + { + "epoch": 2.8391170656709863, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6955337524414062, + "learning_rate": 1e-06, + "loss": 0.7212, + "mean_token_accuracy": 0.7670167684555054, + "num_tokens": 668897345.0, + "step": 25853 + }, + { + "epoch": 2.8392268833736, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.500303030014038, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7279208898544312, + "num_tokens": 668921684.0, + "step": 25854 + }, + { + "epoch": 2.8393367010762134, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.269495964050293, + "learning_rate": 1e-06, + "loss": 0.8439, + "mean_token_accuracy": 0.7312499284744263, + "num_tokens": 668948081.0, + "step": 25855 + }, + { + "epoch": 2.839446518778827, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.112136125564575, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.7034897804260254, + "num_tokens": 668978585.0, + "step": 25856 + }, + { + "epoch": 2.839556336481441, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1237881183624268, + "learning_rate": 1e-06, + "loss": 0.8898, + "mean_token_accuracy": 0.7123138904571533, + "num_tokens": 669009970.0, + "step": 25857 + }, + { + "epoch": 2.839666154184054, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3820695877075195, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7235733270645142, + "num_tokens": 669036112.0, + "step": 25858 + }, + { + "epoch": 2.839775971886668, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3358402252197266, + "learning_rate": 1e-06, + "loss": 0.8246, + "mean_token_accuracy": 0.7397562265396118, + "num_tokens": 669062067.0, + "step": 25859 + }, + { + "epoch": 2.8398857895892817, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.297884225845337, + "learning_rate": 1e-06, + "loss": 0.8234, + "mean_token_accuracy": 0.747613251209259, + "num_tokens": 669089372.0, + "step": 25860 + }, + { + "epoch": 2.8399956072918955, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.284808874130249, + "learning_rate": 1e-06, + "loss": 0.8497, + "mean_token_accuracy": 0.7372642755508423, + "num_tokens": 669115387.0, + "step": 25861 + }, + { + "epoch": 2.8401054249945092, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4893763065338135, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7225924730300903, + "num_tokens": 669139887.0, + "step": 25862 + }, + { + "epoch": 2.8402152426971226, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2626709938049316, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.708903431892395, + "num_tokens": 669169332.0, + "step": 25863 + }, + { + "epoch": 2.8403250603997363, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6294469833374023, + "learning_rate": 1e-06, + "loss": 0.9005, + "mean_token_accuracy": 0.7264924645423889, + "num_tokens": 669193611.0, + "step": 25864 + }, + { + "epoch": 2.84043487810235, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.438431739807129, + "learning_rate": 1e-06, + "loss": 0.7928, + "mean_token_accuracy": 0.7504235506057739, + "num_tokens": 669217802.0, + "step": 25865 + }, + { + "epoch": 2.840544695804964, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4965014457702637, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7494422197341919, + "num_tokens": 669241755.0, + "step": 25866 + }, + { + "epoch": 2.8406545135075776, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7555594444274902, + "learning_rate": 1e-06, + "loss": 0.9375, + "mean_token_accuracy": 0.7089474201202393, + "num_tokens": 669263832.0, + "step": 25867 + }, + { + "epoch": 2.840764331210191, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.488938093185425, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7418210506439209, + "num_tokens": 669286704.0, + "step": 25868 + }, + { + "epoch": 2.8408741489128047, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.313185691833496, + "learning_rate": 1e-06, + "loss": 0.8285, + "mean_token_accuracy": 0.7328962087631226, + "num_tokens": 669313161.0, + "step": 25869 + }, + { + "epoch": 2.8409839666154184, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7468621730804443, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7199051976203918, + "num_tokens": 669334300.0, + "step": 25870 + }, + { + "epoch": 2.841093784318032, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.233344793319702, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7326748967170715, + "num_tokens": 669362793.0, + "step": 25871 + }, + { + "epoch": 2.841203602020646, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.224395751953125, + "learning_rate": 1e-06, + "loss": 0.9168, + "mean_token_accuracy": 0.7293976545333862, + "num_tokens": 669391541.0, + "step": 25872 + }, + { + "epoch": 2.8413134197232592, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1703853607177734, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7370690107345581, + "num_tokens": 669418882.0, + "step": 25873 + }, + { + "epoch": 2.841423237425873, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6349236965179443, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7204833030700684, + "num_tokens": 669441562.0, + "step": 25874 + }, + { + "epoch": 2.8415330551284868, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.297290086746216, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7059494256973267, + "num_tokens": 669468471.0, + "step": 25875 + }, + { + "epoch": 2.8416428728311005, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4653403759002686, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.716568112373352, + "num_tokens": 669493183.0, + "step": 25876 + }, + { + "epoch": 2.8417526905337143, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.588050365447998, + "learning_rate": 1e-06, + "loss": 0.8711, + "mean_token_accuracy": 0.7287176847457886, + "num_tokens": 669516414.0, + "step": 25877 + }, + { + "epoch": 2.8418625082363276, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.315629243850708, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7139855623245239, + "num_tokens": 669542733.0, + "step": 25878 + }, + { + "epoch": 2.8419723259389413, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5074727535247803, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7440603971481323, + "num_tokens": 669566008.0, + "step": 25879 + }, + { + "epoch": 2.842082143641555, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5221307277679443, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7325029373168945, + "num_tokens": 669593634.0, + "step": 25880 + }, + { + "epoch": 2.8421919613441684, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.258199453353882, + "learning_rate": 1e-06, + "loss": 0.8154, + "mean_token_accuracy": 0.7372576594352722, + "num_tokens": 669621603.0, + "step": 25881 + }, + { + "epoch": 2.8423017790467826, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.460190534591675, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7394951581954956, + "num_tokens": 669644770.0, + "step": 25882 + }, + { + "epoch": 2.842411596749396, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.426865577697754, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7346451878547668, + "num_tokens": 669669975.0, + "step": 25883 + }, + { + "epoch": 2.8425214144520097, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4928500652313232, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7073543667793274, + "num_tokens": 669693620.0, + "step": 25884 + }, + { + "epoch": 2.8426312321546234, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.250683307647705, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7259478569030762, + "num_tokens": 669720800.0, + "step": 25885 + }, + { + "epoch": 2.8427410498572367, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.578474283218384, + "learning_rate": 1e-06, + "loss": 0.7896, + "mean_token_accuracy": 0.7523010969161987, + "num_tokens": 669743147.0, + "step": 25886 + }, + { + "epoch": 2.8428508675598505, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2460193634033203, + "learning_rate": 1e-06, + "loss": 0.9239, + "mean_token_accuracy": 0.7236265540122986, + "num_tokens": 669773579.0, + "step": 25887 + }, + { + "epoch": 2.8429606852624643, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.304053544998169, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7228224277496338, + "num_tokens": 669801002.0, + "step": 25888 + }, + { + "epoch": 2.843070502965078, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4331014156341553, + "learning_rate": 1e-06, + "loss": 0.7754, + "mean_token_accuracy": 0.7523444294929504, + "num_tokens": 669823367.0, + "step": 25889 + }, + { + "epoch": 2.8431803206676918, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1105079650878906, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7308351993560791, + "num_tokens": 669853913.0, + "step": 25890 + }, + { + "epoch": 2.843290138370305, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.147202968597412, + "learning_rate": 1e-06, + "loss": 0.9254, + "mean_token_accuracy": 0.717910885810852, + "num_tokens": 669885033.0, + "step": 25891 + }, + { + "epoch": 2.843399956072919, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.204240560531616, + "learning_rate": 1e-06, + "loss": 0.9017, + "mean_token_accuracy": 0.7241154909133911, + "num_tokens": 669915585.0, + "step": 25892 + }, + { + "epoch": 2.8435097737755326, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.392735242843628, + "learning_rate": 1e-06, + "loss": 0.8719, + "mean_token_accuracy": 0.7244687676429749, + "num_tokens": 669940756.0, + "step": 25893 + }, + { + "epoch": 2.8436195914781464, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6994707584381104, + "learning_rate": 1e-06, + "loss": 0.8386, + "mean_token_accuracy": 0.742516040802002, + "num_tokens": 669961158.0, + "step": 25894 + }, + { + "epoch": 2.84372940918076, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.456258535385132, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7351168990135193, + "num_tokens": 669984561.0, + "step": 25895 + }, + { + "epoch": 2.8438392268833734, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.464736223220825, + "learning_rate": 1e-06, + "loss": 0.8551, + "mean_token_accuracy": 0.7332897186279297, + "num_tokens": 670008493.0, + "step": 25896 + }, + { + "epoch": 2.843949044585987, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.363304376602173, + "learning_rate": 1e-06, + "loss": 0.9061, + "mean_token_accuracy": 0.7165392637252808, + "num_tokens": 670033800.0, + "step": 25897 + }, + { + "epoch": 2.844058862288601, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2922232151031494, + "learning_rate": 1e-06, + "loss": 0.9033, + "mean_token_accuracy": 0.7283008694648743, + "num_tokens": 670060141.0, + "step": 25898 + }, + { + "epoch": 2.8441686799912147, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3078088760375977, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7412703633308411, + "num_tokens": 670087456.0, + "step": 25899 + }, + { + "epoch": 2.8442784976938285, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.09985089302063, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7327580451965332, + "num_tokens": 670118073.0, + "step": 25900 + }, + { + "epoch": 2.8443883153964418, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4496188163757324, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.734139084815979, + "num_tokens": 670142476.0, + "step": 25901 + }, + { + "epoch": 2.8444981330990555, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.289386510848999, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7329621911048889, + "num_tokens": 670168641.0, + "step": 25902 + }, + { + "epoch": 2.8446079508016693, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1961495876312256, + "learning_rate": 1e-06, + "loss": 0.917, + "mean_token_accuracy": 0.7234249114990234, + "num_tokens": 670201852.0, + "step": 25903 + }, + { + "epoch": 2.844717768504283, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6064388751983643, + "learning_rate": 1e-06, + "loss": 0.7864, + "mean_token_accuracy": 0.7477684020996094, + "num_tokens": 670224372.0, + "step": 25904 + }, + { + "epoch": 2.844827586206897, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.325611114501953, + "learning_rate": 1e-06, + "loss": 0.8981, + "mean_token_accuracy": 0.7305537462234497, + "num_tokens": 670252213.0, + "step": 25905 + }, + { + "epoch": 2.84493740390951, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.442310094833374, + "learning_rate": 1e-06, + "loss": 0.919, + "mean_token_accuracy": 0.7163544297218323, + "num_tokens": 670280173.0, + "step": 25906 + }, + { + "epoch": 2.845047221612124, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1764638423919678, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7489857077598572, + "num_tokens": 670310822.0, + "step": 25907 + }, + { + "epoch": 2.8451570393147376, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.100891351699829, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7370482683181763, + "num_tokens": 670338498.0, + "step": 25908 + }, + { + "epoch": 2.845266857017351, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5919201374053955, + "learning_rate": 1e-06, + "loss": 0.7584, + "mean_token_accuracy": 0.7600493431091309, + "num_tokens": 670360003.0, + "step": 25909 + }, + { + "epoch": 2.8453766747199647, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.15545654296875, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.722746729850769, + "num_tokens": 670392819.0, + "step": 25910 + }, + { + "epoch": 2.8454864924225785, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2751986980438232, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7218208909034729, + "num_tokens": 670420464.0, + "step": 25911 + }, + { + "epoch": 2.845596310125192, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1522436141967773, + "learning_rate": 1e-06, + "loss": 0.9463, + "mean_token_accuracy": 0.7085726857185364, + "num_tokens": 670451696.0, + "step": 25912 + }, + { + "epoch": 2.845706127827806, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.41631817817688, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7258321642875671, + "num_tokens": 670480806.0, + "step": 25913 + }, + { + "epoch": 2.8458159455304193, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.745032787322998, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7445510625839233, + "num_tokens": 670502025.0, + "step": 25914 + }, + { + "epoch": 2.845925763233033, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1569273471832275, + "learning_rate": 1e-06, + "loss": 0.8855, + "mean_token_accuracy": 0.7291542291641235, + "num_tokens": 670533258.0, + "step": 25915 + }, + { + "epoch": 2.846035580935647, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3658223152160645, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7328767776489258, + "num_tokens": 670560018.0, + "step": 25916 + }, + { + "epoch": 2.8461453986382605, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2194879055023193, + "learning_rate": 1e-06, + "loss": 0.974, + "mean_token_accuracy": 0.6997731924057007, + "num_tokens": 670590028.0, + "step": 25917 + }, + { + "epoch": 2.8462552163408743, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3737010955810547, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7481374144554138, + "num_tokens": 670615101.0, + "step": 25918 + }, + { + "epoch": 2.8463650340434876, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1221976280212402, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7474145293235779, + "num_tokens": 670646292.0, + "step": 25919 + }, + { + "epoch": 2.8464748517461014, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5415565967559814, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7353190779685974, + "num_tokens": 670667824.0, + "step": 25920 + }, + { + "epoch": 2.846584669448715, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.646834135055542, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7294824123382568, + "num_tokens": 670691045.0, + "step": 25921 + }, + { + "epoch": 2.846694487151329, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1238298416137695, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7261573672294617, + "num_tokens": 670722992.0, + "step": 25922 + }, + { + "epoch": 2.8468043048539426, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1594748497009277, + "learning_rate": 1e-06, + "loss": 0.8562, + "mean_token_accuracy": 0.7270988821983337, + "num_tokens": 670751400.0, + "step": 25923 + }, + { + "epoch": 2.846914122556556, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1869657039642334, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7104113698005676, + "num_tokens": 670780341.0, + "step": 25924 + }, + { + "epoch": 2.8470239402591697, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1370763778686523, + "learning_rate": 1e-06, + "loss": 0.8996, + "mean_token_accuracy": 0.7250794768333435, + "num_tokens": 670809210.0, + "step": 25925 + }, + { + "epoch": 2.8471337579617835, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2360920906066895, + "learning_rate": 1e-06, + "loss": 0.8894, + "mean_token_accuracy": 0.7236010432243347, + "num_tokens": 670837346.0, + "step": 25926 + }, + { + "epoch": 2.8472435756643972, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.0779736042022705, + "learning_rate": 1e-06, + "loss": 0.7941, + "mean_token_accuracy": 0.7442989945411682, + "num_tokens": 670866039.0, + "step": 25927 + }, + { + "epoch": 2.847353393367011, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.505147695541382, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7309370040893555, + "num_tokens": 670889048.0, + "step": 25928 + }, + { + "epoch": 2.8474632110696243, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2197952270507812, + "learning_rate": 1e-06, + "loss": 0.9324, + "mean_token_accuracy": 0.7056124210357666, + "num_tokens": 670918102.0, + "step": 25929 + }, + { + "epoch": 2.847573028772238, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.345449924468994, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7320585250854492, + "num_tokens": 670944479.0, + "step": 25930 + }, + { + "epoch": 2.847682846474852, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.330185651779175, + "learning_rate": 1e-06, + "loss": 0.8325, + "mean_token_accuracy": 0.7340130805969238, + "num_tokens": 670972970.0, + "step": 25931 + }, + { + "epoch": 2.8477926641774656, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5959784984588623, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7208176255226135, + "num_tokens": 670997133.0, + "step": 25932 + }, + { + "epoch": 2.8479024818800793, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4254865646362305, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7264312505722046, + "num_tokens": 671023043.0, + "step": 25933 + }, + { + "epoch": 2.8480122995826926, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.618757724761963, + "learning_rate": 1e-06, + "loss": 0.8909, + "mean_token_accuracy": 0.7267517447471619, + "num_tokens": 671046145.0, + "step": 25934 + }, + { + "epoch": 2.8481221172853064, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2500338554382324, + "learning_rate": 1e-06, + "loss": 0.9024, + "mean_token_accuracy": 0.7091684341430664, + "num_tokens": 671073990.0, + "step": 25935 + }, + { + "epoch": 2.84823193498792, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4746198654174805, + "learning_rate": 1e-06, + "loss": 0.7786, + "mean_token_accuracy": 0.7465088963508606, + "num_tokens": 671096700.0, + "step": 25936 + }, + { + "epoch": 2.8483417526905335, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3305180072784424, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7231307029724121, + "num_tokens": 671121925.0, + "step": 25937 + }, + { + "epoch": 2.8484515703931472, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.207785129547119, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7329764366149902, + "num_tokens": 671149683.0, + "step": 25938 + }, + { + "epoch": 2.848561388095761, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.417551040649414, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7301674485206604, + "num_tokens": 671174455.0, + "step": 25939 + }, + { + "epoch": 2.8486712057983747, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.469676971435547, + "learning_rate": 1e-06, + "loss": 0.7613, + "mean_token_accuracy": 0.7545938491821289, + "num_tokens": 671197544.0, + "step": 25940 + }, + { + "epoch": 2.8487810235009885, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.211777687072754, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7133901119232178, + "num_tokens": 671227099.0, + "step": 25941 + }, + { + "epoch": 2.848890841203602, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7687244415283203, + "learning_rate": 1e-06, + "loss": 0.7724, + "mean_token_accuracy": 0.7524951696395874, + "num_tokens": 671247278.0, + "step": 25942 + }, + { + "epoch": 2.8490006589062156, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7210068702697754, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7269527912139893, + "num_tokens": 671270021.0, + "step": 25943 + }, + { + "epoch": 2.8491104766088293, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3250157833099365, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7461952567100525, + "num_tokens": 671297070.0, + "step": 25944 + }, + { + "epoch": 2.849220294311443, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1619324684143066, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7193031311035156, + "num_tokens": 671327388.0, + "step": 25945 + }, + { + "epoch": 2.849330112014057, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.220869779586792, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7262625694274902, + "num_tokens": 671357515.0, + "step": 25946 + }, + { + "epoch": 2.84943992971667, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.294487237930298, + "learning_rate": 1e-06, + "loss": 0.7859, + "mean_token_accuracy": 0.7444744110107422, + "num_tokens": 671383917.0, + "step": 25947 + }, + { + "epoch": 2.849549747419284, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.599114179611206, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7382766008377075, + "num_tokens": 671405249.0, + "step": 25948 + }, + { + "epoch": 2.8496595651218977, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3641841411590576, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7185498476028442, + "num_tokens": 671434360.0, + "step": 25949 + }, + { + "epoch": 2.8497693828245114, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.202634811401367, + "learning_rate": 1e-06, + "loss": 0.8584, + "mean_token_accuracy": 0.7283124327659607, + "num_tokens": 671463945.0, + "step": 25950 + }, + { + "epoch": 2.849879200527125, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3583734035491943, + "learning_rate": 1e-06, + "loss": 0.7972, + "mean_token_accuracy": 0.7478584051132202, + "num_tokens": 671488614.0, + "step": 25951 + }, + { + "epoch": 2.8499890182297385, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4031927585601807, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7281649112701416, + "num_tokens": 671516910.0, + "step": 25952 + }, + { + "epoch": 2.8500988359323522, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.390915632247925, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7374941110610962, + "num_tokens": 671542002.0, + "step": 25953 + }, + { + "epoch": 2.850208653634966, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8280839920043945, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7287886142730713, + "num_tokens": 671564639.0, + "step": 25954 + }, + { + "epoch": 2.8503184713375798, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.395193576812744, + "learning_rate": 1e-06, + "loss": 0.8848, + "mean_token_accuracy": 0.7262045741081238, + "num_tokens": 671591875.0, + "step": 25955 + }, + { + "epoch": 2.8504282890401935, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3829493522644043, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7172849178314209, + "num_tokens": 671618230.0, + "step": 25956 + }, + { + "epoch": 2.850538106742807, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1297435760498047, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7330472469329834, + "num_tokens": 671650632.0, + "step": 25957 + }, + { + "epoch": 2.8506479244454206, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.494663953781128, + "learning_rate": 1e-06, + "loss": 0.8802, + "mean_token_accuracy": 0.7219850420951843, + "num_tokens": 671674132.0, + "step": 25958 + }, + { + "epoch": 2.8507577421480343, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.345510482788086, + "learning_rate": 1e-06, + "loss": 0.9143, + "mean_token_accuracy": 0.7177176475524902, + "num_tokens": 671701303.0, + "step": 25959 + }, + { + "epoch": 2.8508675598506477, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.193361282348633, + "learning_rate": 1e-06, + "loss": 0.8908, + "mean_token_accuracy": 0.7161900401115417, + "num_tokens": 671731503.0, + "step": 25960 + }, + { + "epoch": 2.850977377553262, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.299691915512085, + "learning_rate": 1e-06, + "loss": 0.7638, + "mean_token_accuracy": 0.7563226222991943, + "num_tokens": 671758936.0, + "step": 25961 + }, + { + "epoch": 2.851087195255875, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1776628494262695, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7195369601249695, + "num_tokens": 671789371.0, + "step": 25962 + }, + { + "epoch": 2.851197012958489, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.375037670135498, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7408897876739502, + "num_tokens": 671815168.0, + "step": 25963 + }, + { + "epoch": 2.8513068306611027, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.587764263153076, + "learning_rate": 1e-06, + "loss": 0.7422, + "mean_token_accuracy": 0.7596002817153931, + "num_tokens": 671835725.0, + "step": 25964 + }, + { + "epoch": 2.851416648363716, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1793713569641113, + "learning_rate": 1e-06, + "loss": 0.9257, + "mean_token_accuracy": 0.7134789228439331, + "num_tokens": 671864361.0, + "step": 25965 + }, + { + "epoch": 2.8515264660663298, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3023269176483154, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7181743383407593, + "num_tokens": 671893204.0, + "step": 25966 + }, + { + "epoch": 2.8516362837689435, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4973206520080566, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7313538789749146, + "num_tokens": 671917250.0, + "step": 25967 + }, + { + "epoch": 2.8517461014715573, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1578562259674072, + "learning_rate": 1e-06, + "loss": 0.8972, + "mean_token_accuracy": 0.71593177318573, + "num_tokens": 671948757.0, + "step": 25968 + }, + { + "epoch": 2.851855919174171, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.591918706893921, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7338963150978088, + "num_tokens": 671971315.0, + "step": 25969 + }, + { + "epoch": 2.8519657368767843, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.338836908340454, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7058731317520142, + "num_tokens": 672001667.0, + "step": 25970 + }, + { + "epoch": 2.852075554579398, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.661038637161255, + "learning_rate": 1e-06, + "loss": 0.7948, + "mean_token_accuracy": 0.7537142038345337, + "num_tokens": 672021785.0, + "step": 25971 + }, + { + "epoch": 2.852185372282012, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.286827325820923, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7342673540115356, + "num_tokens": 672047617.0, + "step": 25972 + }, + { + "epoch": 2.8522951899846256, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.659047842025757, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7260012030601501, + "num_tokens": 672068906.0, + "step": 25973 + }, + { + "epoch": 2.8524050076872394, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.204362630844116, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7232105731964111, + "num_tokens": 672096120.0, + "step": 25974 + }, + { + "epoch": 2.8525148253898527, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.340087890625, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7419827580451965, + "num_tokens": 672123519.0, + "step": 25975 + }, + { + "epoch": 2.8526246430924664, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4087331295013428, + "learning_rate": 1e-06, + "loss": 0.8492, + "mean_token_accuracy": 0.7326170206069946, + "num_tokens": 672148463.0, + "step": 25976 + }, + { + "epoch": 2.85273446079508, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.332364082336426, + "learning_rate": 1e-06, + "loss": 0.759, + "mean_token_accuracy": 0.7641610503196716, + "num_tokens": 672175014.0, + "step": 25977 + }, + { + "epoch": 2.852844278497694, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6189568042755127, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7356196045875549, + "num_tokens": 672196646.0, + "step": 25978 + }, + { + "epoch": 2.8529540962003077, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.392360210418701, + "learning_rate": 1e-06, + "loss": 0.8529, + "mean_token_accuracy": 0.7314891219139099, + "num_tokens": 672223961.0, + "step": 25979 + }, + { + "epoch": 2.853063913902921, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.208561658859253, + "learning_rate": 1e-06, + "loss": 0.8428, + "mean_token_accuracy": 0.7342881560325623, + "num_tokens": 672252712.0, + "step": 25980 + }, + { + "epoch": 2.853173731605535, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4111759662628174, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7312036752700806, + "num_tokens": 672276394.0, + "step": 25981 + }, + { + "epoch": 2.8532835493081485, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3716485500335693, + "learning_rate": 1e-06, + "loss": 0.7715, + "mean_token_accuracy": 0.7528377175331116, + "num_tokens": 672300503.0, + "step": 25982 + }, + { + "epoch": 2.8533933670107623, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.623936891555786, + "learning_rate": 1e-06, + "loss": 0.8988, + "mean_token_accuracy": 0.7362326383590698, + "num_tokens": 672322330.0, + "step": 25983 + }, + { + "epoch": 2.853503184713376, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4962828159332275, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.7334654331207275, + "num_tokens": 672345273.0, + "step": 25984 + }, + { + "epoch": 2.8536130024159894, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.324212074279785, + "learning_rate": 1e-06, + "loss": 0.895, + "mean_token_accuracy": 0.721997082233429, + "num_tokens": 672372616.0, + "step": 25985 + }, + { + "epoch": 2.853722820118603, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4167726039886475, + "learning_rate": 1e-06, + "loss": 0.6885, + "mean_token_accuracy": 0.7786245346069336, + "num_tokens": 672394953.0, + "step": 25986 + }, + { + "epoch": 2.853832637821217, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2151429653167725, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7223806977272034, + "num_tokens": 672423046.0, + "step": 25987 + }, + { + "epoch": 2.85394245552383, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7313826084136963, + "learning_rate": 1e-06, + "loss": 0.7862, + "mean_token_accuracy": 0.7454584836959839, + "num_tokens": 672443008.0, + "step": 25988 + }, + { + "epoch": 2.854052273226444, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.47280216217041, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7324793934822083, + "num_tokens": 672467994.0, + "step": 25989 + }, + { + "epoch": 2.8541620909290577, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2127017974853516, + "learning_rate": 1e-06, + "loss": 0.9397, + "mean_token_accuracy": 0.7080034017562866, + "num_tokens": 672497637.0, + "step": 25990 + }, + { + "epoch": 2.8542719086316715, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.279910087585449, + "learning_rate": 1e-06, + "loss": 0.8749, + "mean_token_accuracy": 0.7256239652633667, + "num_tokens": 672522773.0, + "step": 25991 + }, + { + "epoch": 2.854381726334285, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1534266471862793, + "learning_rate": 1e-06, + "loss": 0.8781, + "mean_token_accuracy": 0.7310376763343811, + "num_tokens": 672550357.0, + "step": 25992 + }, + { + "epoch": 2.8544915440368985, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5174777507781982, + "learning_rate": 1e-06, + "loss": 0.7572, + "mean_token_accuracy": 0.7582914233207703, + "num_tokens": 672572533.0, + "step": 25993 + }, + { + "epoch": 2.8546013617395123, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2523088455200195, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7317736148834229, + "num_tokens": 672599089.0, + "step": 25994 + }, + { + "epoch": 2.854711179442126, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3960180282592773, + "learning_rate": 1e-06, + "loss": 0.8172, + "mean_token_accuracy": 0.7421693801879883, + "num_tokens": 672626113.0, + "step": 25995 + }, + { + "epoch": 2.85482099714474, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4492533206939697, + "learning_rate": 1e-06, + "loss": 0.8817, + "mean_token_accuracy": 0.7258652448654175, + "num_tokens": 672651555.0, + "step": 25996 + }, + { + "epoch": 2.8549308148473536, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6144402027130127, + "learning_rate": 1e-06, + "loss": 0.8794, + "mean_token_accuracy": 0.7241137027740479, + "num_tokens": 672674340.0, + "step": 25997 + }, + { + "epoch": 2.855040632549967, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.151637315750122, + "learning_rate": 1e-06, + "loss": 0.7854, + "mean_token_accuracy": 0.7521381378173828, + "num_tokens": 672704680.0, + "step": 25998 + }, + { + "epoch": 2.8551504502525806, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.40423583984375, + "learning_rate": 1e-06, + "loss": 0.9098, + "mean_token_accuracy": 0.7114777565002441, + "num_tokens": 672731569.0, + "step": 25999 + }, + { + "epoch": 2.8552602679551944, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.0581719875335693, + "learning_rate": 1e-06, + "loss": 0.9409, + "mean_token_accuracy": 0.7081050872802734, + "num_tokens": 672763196.0, + "step": 26000 + }, + { + "epoch": 2.855370085657808, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.579587697982788, + "learning_rate": 1e-06, + "loss": 0.7367, + "mean_token_accuracy": 0.7585698366165161, + "num_tokens": 672784965.0, + "step": 26001 + }, + { + "epoch": 2.855479903360422, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1891002655029297, + "learning_rate": 1e-06, + "loss": 0.9488, + "mean_token_accuracy": 0.7027370929718018, + "num_tokens": 672816672.0, + "step": 26002 + }, + { + "epoch": 2.855589721063035, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4169697761535645, + "learning_rate": 1e-06, + "loss": 0.7907, + "mean_token_accuracy": 0.7529218196868896, + "num_tokens": 672840746.0, + "step": 26003 + }, + { + "epoch": 2.855699538765649, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.333881378173828, + "learning_rate": 1e-06, + "loss": 0.8143, + "mean_token_accuracy": 0.7470870614051819, + "num_tokens": 672866472.0, + "step": 26004 + }, + { + "epoch": 2.8558093564682627, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.296509265899658, + "learning_rate": 1e-06, + "loss": 0.8635, + "mean_token_accuracy": 0.7267894744873047, + "num_tokens": 672894541.0, + "step": 26005 + }, + { + "epoch": 2.8559191741708765, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4304685592651367, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7381643652915955, + "num_tokens": 672919064.0, + "step": 26006 + }, + { + "epoch": 2.8560289918734902, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.320904493331909, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.727571964263916, + "num_tokens": 672946396.0, + "step": 26007 + }, + { + "epoch": 2.8561388095761036, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.172273874282837, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7273242473602295, + "num_tokens": 672975469.0, + "step": 26008 + }, + { + "epoch": 2.8562486272787173, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1024229526519775, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.748657763004303, + "num_tokens": 673005393.0, + "step": 26009 + }, + { + "epoch": 2.856358444981331, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8032734394073486, + "learning_rate": 1e-06, + "loss": 0.7584, + "mean_token_accuracy": 0.7556970715522766, + "num_tokens": 673023441.0, + "step": 26010 + }, + { + "epoch": 2.8564682626839444, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3489372730255127, + "learning_rate": 1e-06, + "loss": 0.7464, + "mean_token_accuracy": 0.7584311962127686, + "num_tokens": 673049411.0, + "step": 26011 + }, + { + "epoch": 2.8565780803865586, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.559744119644165, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7351602911949158, + "num_tokens": 673071869.0, + "step": 26012 + }, + { + "epoch": 2.856687898089172, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3322479724884033, + "learning_rate": 1e-06, + "loss": 0.8974, + "mean_token_accuracy": 0.7121111154556274, + "num_tokens": 673098439.0, + "step": 26013 + }, + { + "epoch": 2.8567977157917857, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6064906120300293, + "learning_rate": 1e-06, + "loss": 0.8629, + "mean_token_accuracy": 0.736884355545044, + "num_tokens": 673120897.0, + "step": 26014 + }, + { + "epoch": 2.8569075334943994, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.652811050415039, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7364574074745178, + "num_tokens": 673141537.0, + "step": 26015 + }, + { + "epoch": 2.8570173511970127, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.340837001800537, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7309907674789429, + "num_tokens": 673169449.0, + "step": 26016 + }, + { + "epoch": 2.8571271688996265, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.172572135925293, + "learning_rate": 1e-06, + "loss": 0.8358, + "mean_token_accuracy": 0.7361570596694946, + "num_tokens": 673197705.0, + "step": 26017 + }, + { + "epoch": 2.8572369866022402, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1789896488189697, + "learning_rate": 1e-06, + "loss": 0.9181, + "mean_token_accuracy": 0.7146375179290771, + "num_tokens": 673227243.0, + "step": 26018 + }, + { + "epoch": 2.857346804304854, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4963948726654053, + "learning_rate": 1e-06, + "loss": 0.8594, + "mean_token_accuracy": 0.7241126894950867, + "num_tokens": 673252030.0, + "step": 26019 + }, + { + "epoch": 2.8574566220074678, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.208989143371582, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.738354504108429, + "num_tokens": 673280208.0, + "step": 26020 + }, + { + "epoch": 2.857566439710081, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.051280975341797, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7171338200569153, + "num_tokens": 673311114.0, + "step": 26021 + }, + { + "epoch": 2.857676257412695, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2624192237854004, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7359842658042908, + "num_tokens": 673336523.0, + "step": 26022 + }, + { + "epoch": 2.8577860751153086, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3164560794830322, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7203723192214966, + "num_tokens": 673363109.0, + "step": 26023 + }, + { + "epoch": 2.8578958928179223, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3242547512054443, + "learning_rate": 1e-06, + "loss": 0.82, + "mean_token_accuracy": 0.7497744560241699, + "num_tokens": 673388866.0, + "step": 26024 + }, + { + "epoch": 2.858005710520536, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.530151128768921, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7334384918212891, + "num_tokens": 673413250.0, + "step": 26025 + }, + { + "epoch": 2.8581155282231494, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7010080814361572, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7405452728271484, + "num_tokens": 673433658.0, + "step": 26026 + }, + { + "epoch": 2.858225345925763, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5495975017547607, + "learning_rate": 1e-06, + "loss": 0.7773, + "mean_token_accuracy": 0.7470948696136475, + "num_tokens": 673453861.0, + "step": 26027 + }, + { + "epoch": 2.858335163628377, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1801562309265137, + "learning_rate": 1e-06, + "loss": 0.9931, + "mean_token_accuracy": 0.6913647651672363, + "num_tokens": 673484844.0, + "step": 26028 + }, + { + "epoch": 2.8584449813309907, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4642722606658936, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7486762404441833, + "num_tokens": 673510058.0, + "step": 26029 + }, + { + "epoch": 2.8585547990336044, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2661781311035156, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7332952618598938, + "num_tokens": 673537680.0, + "step": 26030 + }, + { + "epoch": 2.8586646167362177, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4884934425354004, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7347971200942993, + "num_tokens": 673562118.0, + "step": 26031 + }, + { + "epoch": 2.8587744344388315, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.349935293197632, + "learning_rate": 1e-06, + "loss": 0.7652, + "mean_token_accuracy": 0.7497856020927429, + "num_tokens": 673587339.0, + "step": 26032 + }, + { + "epoch": 2.8588842521414453, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.448086738586426, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7328354120254517, + "num_tokens": 673612062.0, + "step": 26033 + }, + { + "epoch": 2.858994069844059, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1840975284576416, + "learning_rate": 1e-06, + "loss": 0.8304, + "mean_token_accuracy": 0.7361347675323486, + "num_tokens": 673639397.0, + "step": 26034 + }, + { + "epoch": 2.8591038875466728, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.0796306133270264, + "learning_rate": 1e-06, + "loss": 0.8215, + "mean_token_accuracy": 0.7383828163146973, + "num_tokens": 673672041.0, + "step": 26035 + }, + { + "epoch": 2.859213705249286, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6050961017608643, + "learning_rate": 1e-06, + "loss": 0.9316, + "mean_token_accuracy": 0.7229452729225159, + "num_tokens": 673696084.0, + "step": 26036 + }, + { + "epoch": 2.8593235229519, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1589112281799316, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7224330902099609, + "num_tokens": 673725454.0, + "step": 26037 + }, + { + "epoch": 2.8594333406545136, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.614145040512085, + "learning_rate": 1e-06, + "loss": 0.8214, + "mean_token_accuracy": 0.7393851280212402, + "num_tokens": 673746551.0, + "step": 26038 + }, + { + "epoch": 2.859543158357127, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.357377767562866, + "learning_rate": 1e-06, + "loss": 0.814, + "mean_token_accuracy": 0.7443089485168457, + "num_tokens": 673772122.0, + "step": 26039 + }, + { + "epoch": 2.8596529760597407, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.34214186668396, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7309293746948242, + "num_tokens": 673797623.0, + "step": 26040 + }, + { + "epoch": 2.8597627937623544, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1179444789886475, + "learning_rate": 1e-06, + "loss": 0.8778, + "mean_token_accuracy": 0.7225661277770996, + "num_tokens": 673828902.0, + "step": 26041 + }, + { + "epoch": 2.859872611464968, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1540966033935547, + "learning_rate": 1e-06, + "loss": 0.8604, + "mean_token_accuracy": 0.7299025058746338, + "num_tokens": 673859008.0, + "step": 26042 + }, + { + "epoch": 2.859982429167582, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4626963138580322, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.746173620223999, + "num_tokens": 673881917.0, + "step": 26043 + }, + { + "epoch": 2.8600922468701953, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.149078607559204, + "learning_rate": 1e-06, + "loss": 0.8441, + "mean_token_accuracy": 0.7349846363067627, + "num_tokens": 673912750.0, + "step": 26044 + }, + { + "epoch": 2.860202064572809, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5233640670776367, + "learning_rate": 1e-06, + "loss": 0.8731, + "mean_token_accuracy": 0.7302519083023071, + "num_tokens": 673934835.0, + "step": 26045 + }, + { + "epoch": 2.8603118822754228, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.226116418838501, + "learning_rate": 1e-06, + "loss": 0.8652, + "mean_token_accuracy": 0.7287328243255615, + "num_tokens": 673963165.0, + "step": 26046 + }, + { + "epoch": 2.8604216999780365, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.535266399383545, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7436622381210327, + "num_tokens": 673984440.0, + "step": 26047 + }, + { + "epoch": 2.8605315176806503, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2926108837127686, + "learning_rate": 1e-06, + "loss": 0.8008, + "mean_token_accuracy": 0.7459267377853394, + "num_tokens": 674009921.0, + "step": 26048 + }, + { + "epoch": 2.8606413353832636, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.517242908477783, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7227912545204163, + "num_tokens": 674034168.0, + "step": 26049 + }, + { + "epoch": 2.8607511530858774, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7148685455322266, + "learning_rate": 1e-06, + "loss": 0.7806, + "mean_token_accuracy": 0.7496098279953003, + "num_tokens": 674054595.0, + "step": 26050 + }, + { + "epoch": 2.860860970788491, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.8503196239471436, + "learning_rate": 1e-06, + "loss": 0.8571, + "mean_token_accuracy": 0.7342098951339722, + "num_tokens": 674073090.0, + "step": 26051 + }, + { + "epoch": 2.860970788491105, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1652657985687256, + "learning_rate": 1e-06, + "loss": 0.8734, + "mean_token_accuracy": 0.7271531820297241, + "num_tokens": 674102260.0, + "step": 26052 + }, + { + "epoch": 2.8610806061937186, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.6145358085632324, + "learning_rate": 1e-06, + "loss": 0.8726, + "mean_token_accuracy": 0.7222813963890076, + "num_tokens": 674123703.0, + "step": 26053 + }, + { + "epoch": 2.861190423896332, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3038692474365234, + "learning_rate": 1e-06, + "loss": 0.914, + "mean_token_accuracy": 0.7154274582862854, + "num_tokens": 674151101.0, + "step": 26054 + }, + { + "epoch": 2.8613002415989457, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.456648588180542, + "learning_rate": 1e-06, + "loss": 0.8753, + "mean_token_accuracy": 0.7202277183532715, + "num_tokens": 674176981.0, + "step": 26055 + }, + { + "epoch": 2.8614100593015594, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3394932746887207, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7157527804374695, + "num_tokens": 674202164.0, + "step": 26056 + }, + { + "epoch": 2.861519877004173, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4424221515655518, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7414058446884155, + "num_tokens": 674227273.0, + "step": 26057 + }, + { + "epoch": 2.861629694706787, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.389831304550171, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7419034242630005, + "num_tokens": 674251889.0, + "step": 26058 + }, + { + "epoch": 2.8617395124094003, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1773030757904053, + "learning_rate": 1e-06, + "loss": 0.9337, + "mean_token_accuracy": 0.7074025273323059, + "num_tokens": 674282868.0, + "step": 26059 + }, + { + "epoch": 2.861849330112014, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3163352012634277, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7266305088996887, + "num_tokens": 674310603.0, + "step": 26060 + }, + { + "epoch": 2.861959147814628, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5048394203186035, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7339800596237183, + "num_tokens": 674333678.0, + "step": 26061 + }, + { + "epoch": 2.862068965517241, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.282050132751465, + "learning_rate": 1e-06, + "loss": 0.7936, + "mean_token_accuracy": 0.7530426979064941, + "num_tokens": 674361458.0, + "step": 26062 + }, + { + "epoch": 2.8621787832198553, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5514721870422363, + "learning_rate": 1e-06, + "loss": 0.8009, + "mean_token_accuracy": 0.7466686964035034, + "num_tokens": 674384351.0, + "step": 26063 + }, + { + "epoch": 2.8622886009224686, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.222571849822998, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7097896933555603, + "num_tokens": 674411895.0, + "step": 26064 + }, + { + "epoch": 2.8623984186250824, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2375497817993164, + "learning_rate": 1e-06, + "loss": 0.9283, + "mean_token_accuracy": 0.705855131149292, + "num_tokens": 674441370.0, + "step": 26065 + }, + { + "epoch": 2.862508236327696, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.061802864074707, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.711307168006897, + "num_tokens": 674473057.0, + "step": 26066 + }, + { + "epoch": 2.8626180540303094, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.377291440963745, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7366572618484497, + "num_tokens": 674500205.0, + "step": 26067 + }, + { + "epoch": 2.862727871732923, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.5491552352905273, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7336286902427673, + "num_tokens": 674522500.0, + "step": 26068 + }, + { + "epoch": 2.862837689435537, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.1862282752990723, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7147397398948669, + "num_tokens": 674553971.0, + "step": 26069 + }, + { + "epoch": 2.8629475071381507, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.128293514251709, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7157610654830933, + "num_tokens": 674583947.0, + "step": 26070 + }, + { + "epoch": 2.8630573248407645, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.4469852447509766, + "learning_rate": 1e-06, + "loss": 0.897, + "mean_token_accuracy": 0.7236485481262207, + "num_tokens": 674609286.0, + "step": 26071 + }, + { + "epoch": 2.863167142543378, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.918151378631592, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7343615293502808, + "num_tokens": 674627776.0, + "step": 26072 + }, + { + "epoch": 2.8632769602459915, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.353238344192505, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7359473705291748, + "num_tokens": 674652947.0, + "step": 26073 + }, + { + "epoch": 2.8633867779486053, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2721188068389893, + "learning_rate": 1e-06, + "loss": 0.8912, + "mean_token_accuracy": 0.7224897742271423, + "num_tokens": 674680767.0, + "step": 26074 + }, + { + "epoch": 2.863496595651219, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.309995412826538, + "learning_rate": 1e-06, + "loss": 0.8994, + "mean_token_accuracy": 0.7194651961326599, + "num_tokens": 674706645.0, + "step": 26075 + }, + { + "epoch": 2.863606413353833, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3376541137695312, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7274287939071655, + "num_tokens": 674735371.0, + "step": 26076 + }, + { + "epoch": 2.863716231056446, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3979344367980957, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.7217516303062439, + "num_tokens": 674761652.0, + "step": 26077 + }, + { + "epoch": 2.86382604875906, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2204430103302, + "learning_rate": 1e-06, + "loss": 0.8388, + "mean_token_accuracy": 0.7406075596809387, + "num_tokens": 674790968.0, + "step": 26078 + }, + { + "epoch": 2.8639358664616736, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.2076926231384277, + "learning_rate": 1e-06, + "loss": 0.9978, + "mean_token_accuracy": 0.7024924755096436, + "num_tokens": 674820329.0, + "step": 26079 + }, + { + "epoch": 2.8640456841642874, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.401979684829712, + "learning_rate": 1e-06, + "loss": 0.8344, + "mean_token_accuracy": 0.7404302954673767, + "num_tokens": 674843544.0, + "step": 26080 + }, + { + "epoch": 2.864155501866901, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.597817897796631, + "learning_rate": 1e-06, + "loss": 0.7711, + "mean_token_accuracy": 0.755402684211731, + "num_tokens": 674866950.0, + "step": 26081 + }, + { + "epoch": 2.8642653195695145, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.402682065963745, + "learning_rate": 1e-06, + "loss": 0.9431, + "mean_token_accuracy": 0.7026422023773193, + "num_tokens": 674892458.0, + "step": 26082 + }, + { + "epoch": 2.8643751372721282, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.7261295318603516, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.7357257008552551, + "num_tokens": 674912625.0, + "step": 26083 + }, + { + "epoch": 2.864484954974742, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3247339725494385, + "learning_rate": 1e-06, + "loss": 0.8513, + "mean_token_accuracy": 0.7254196405410767, + "num_tokens": 674937917.0, + "step": 26084 + }, + { + "epoch": 2.8645947726773557, + "ewc_loss": 1.6450881958007812e-05, + "grad_norm": 2.3340256214141846, + "learning_rate": 1e-06, + "loss": 0.7756, + "mean_token_accuracy": 0.7508649826049805, + "num_tokens": 674963788.0, + "step": 26085 + }, + { + "epoch": 2.8647045903799695, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.631169080734253, + "learning_rate": 1e-06, + "loss": 0.8123, + "mean_token_accuracy": 0.7455025911331177, + "num_tokens": 674985862.0, + "step": 26086 + }, + { + "epoch": 2.864814408082583, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6299078464508057, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.733349084854126, + "num_tokens": 675006808.0, + "step": 26087 + }, + { + "epoch": 2.8649242257851966, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3387742042541504, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7330667972564697, + "num_tokens": 675033623.0, + "step": 26088 + }, + { + "epoch": 2.8650340434878103, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5768370628356934, + "learning_rate": 1e-06, + "loss": 0.7689, + "mean_token_accuracy": 0.7473347187042236, + "num_tokens": 675053083.0, + "step": 26089 + }, + { + "epoch": 2.8651438611904236, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.199099540710449, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7236042022705078, + "num_tokens": 675082485.0, + "step": 26090 + }, + { + "epoch": 2.8652536788930374, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7631094455718994, + "learning_rate": 1e-06, + "loss": 0.9395, + "mean_token_accuracy": 0.7140985727310181, + "num_tokens": 675105349.0, + "step": 26091 + }, + { + "epoch": 2.865363496595651, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.443558931350708, + "learning_rate": 1e-06, + "loss": 0.86, + "mean_token_accuracy": 0.7316761016845703, + "num_tokens": 675129468.0, + "step": 26092 + }, + { + "epoch": 2.865473314298265, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.201626777648926, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.731677770614624, + "num_tokens": 675158711.0, + "step": 26093 + }, + { + "epoch": 2.8655831320008787, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2493739128112793, + "learning_rate": 1e-06, + "loss": 0.7759, + "mean_token_accuracy": 0.7545892000198364, + "num_tokens": 675186157.0, + "step": 26094 + }, + { + "epoch": 2.865692949703492, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.194551944732666, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7275139093399048, + "num_tokens": 675213161.0, + "step": 26095 + }, + { + "epoch": 2.8658027674061057, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2528812885284424, + "learning_rate": 1e-06, + "loss": 0.8856, + "mean_token_accuracy": 0.7251052260398865, + "num_tokens": 675243061.0, + "step": 26096 + }, + { + "epoch": 2.8659125851087195, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4942266941070557, + "learning_rate": 1e-06, + "loss": 0.8189, + "mean_token_accuracy": 0.7442301511764526, + "num_tokens": 675267805.0, + "step": 26097 + }, + { + "epoch": 2.8660224028113332, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6756625175476074, + "learning_rate": 1e-06, + "loss": 0.7751, + "mean_token_accuracy": 0.7510223388671875, + "num_tokens": 675288863.0, + "step": 26098 + }, + { + "epoch": 2.866132220513947, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.333782434463501, + "learning_rate": 1e-06, + "loss": 0.7358, + "mean_token_accuracy": 0.7625699043273926, + "num_tokens": 675312823.0, + "step": 26099 + }, + { + "epoch": 2.8662420382165603, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2519795894622803, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7276324033737183, + "num_tokens": 675343825.0, + "step": 26100 + }, + { + "epoch": 2.866351855919174, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7299935817718506, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7238699793815613, + "num_tokens": 675364877.0, + "step": 26101 + }, + { + "epoch": 2.866461673621788, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3357865810394287, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7320895195007324, + "num_tokens": 675390248.0, + "step": 26102 + }, + { + "epoch": 2.8665714913244016, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1385738849639893, + "learning_rate": 1e-06, + "loss": 0.8602, + "mean_token_accuracy": 0.7281200885772705, + "num_tokens": 675418776.0, + "step": 26103 + }, + { + "epoch": 2.8666813090270153, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0029666423797607, + "learning_rate": 1e-06, + "loss": 0.8448, + "mean_token_accuracy": 0.7471018433570862, + "num_tokens": 675452449.0, + "step": 26104 + }, + { + "epoch": 2.8667911267296287, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.600846529006958, + "learning_rate": 1e-06, + "loss": 0.846, + "mean_token_accuracy": 0.7296820878982544, + "num_tokens": 675472964.0, + "step": 26105 + }, + { + "epoch": 2.8669009444322424, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3906402587890625, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7296441793441772, + "num_tokens": 675499433.0, + "step": 26106 + }, + { + "epoch": 2.867010762134856, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1260735988616943, + "learning_rate": 1e-06, + "loss": 0.9439, + "mean_token_accuracy": 0.709058940410614, + "num_tokens": 675531158.0, + "step": 26107 + }, + { + "epoch": 2.86712057983747, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.666963815689087, + "learning_rate": 1e-06, + "loss": 0.8112, + "mean_token_accuracy": 0.7462769746780396, + "num_tokens": 675552905.0, + "step": 26108 + }, + { + "epoch": 2.8672303975400837, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4348978996276855, + "learning_rate": 1e-06, + "loss": 0.8598, + "mean_token_accuracy": 0.7359977960586548, + "num_tokens": 675577409.0, + "step": 26109 + }, + { + "epoch": 2.867340215242697, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.477821111679077, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7475019693374634, + "num_tokens": 675600582.0, + "step": 26110 + }, + { + "epoch": 2.8674500329453108, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4719204902648926, + "learning_rate": 1e-06, + "loss": 0.8053, + "mean_token_accuracy": 0.75004643201828, + "num_tokens": 675623921.0, + "step": 26111 + }, + { + "epoch": 2.8675598506479245, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6604628562927246, + "learning_rate": 1e-06, + "loss": 0.7411, + "mean_token_accuracy": 0.7599611282348633, + "num_tokens": 675644555.0, + "step": 26112 + }, + { + "epoch": 2.8676696683505383, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6004230976104736, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.735087513923645, + "num_tokens": 675667029.0, + "step": 26113 + }, + { + "epoch": 2.867779486053152, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.033705949783325, + "learning_rate": 1e-06, + "loss": 0.9544, + "mean_token_accuracy": 0.7092326879501343, + "num_tokens": 675702160.0, + "step": 26114 + }, + { + "epoch": 2.8678893037557653, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4656410217285156, + "learning_rate": 1e-06, + "loss": 0.7734, + "mean_token_accuracy": 0.7517272233963013, + "num_tokens": 675723715.0, + "step": 26115 + }, + { + "epoch": 2.867999121458379, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.425455331802368, + "learning_rate": 1e-06, + "loss": 0.8527, + "mean_token_accuracy": 0.7255794405937195, + "num_tokens": 675749148.0, + "step": 26116 + }, + { + "epoch": 2.868108939160993, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.312460422515869, + "learning_rate": 1e-06, + "loss": 0.8876, + "mean_token_accuracy": 0.7199045419692993, + "num_tokens": 675776758.0, + "step": 26117 + }, + { + "epoch": 2.868218756863606, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4808266162872314, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7257705926895142, + "num_tokens": 675800674.0, + "step": 26118 + }, + { + "epoch": 2.86832857456622, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2498116493225098, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7109146118164062, + "num_tokens": 675827345.0, + "step": 26119 + }, + { + "epoch": 2.8684383922688337, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 1.9530874490737915, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7264517545700073, + "num_tokens": 675861466.0, + "step": 26120 + }, + { + "epoch": 2.8685482099714474, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.520310163497925, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7384575009346008, + "num_tokens": 675884135.0, + "step": 26121 + }, + { + "epoch": 2.868658027674061, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4391443729400635, + "learning_rate": 1e-06, + "loss": 0.8283, + "mean_token_accuracy": 0.7333461046218872, + "num_tokens": 675908730.0, + "step": 26122 + }, + { + "epoch": 2.8687678453766745, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4462530612945557, + "learning_rate": 1e-06, + "loss": 0.766, + "mean_token_accuracy": 0.7535851001739502, + "num_tokens": 675932288.0, + "step": 26123 + }, + { + "epoch": 2.8688776630792883, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.582214832305908, + "learning_rate": 1e-06, + "loss": 0.7957, + "mean_token_accuracy": 0.7459930777549744, + "num_tokens": 675954517.0, + "step": 26124 + }, + { + "epoch": 2.868987480781902, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.264683485031128, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7372344732284546, + "num_tokens": 675980211.0, + "step": 26125 + }, + { + "epoch": 2.869097298484516, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.331256151199341, + "learning_rate": 1e-06, + "loss": 0.7866, + "mean_token_accuracy": 0.751284122467041, + "num_tokens": 676005009.0, + "step": 26126 + }, + { + "epoch": 2.8692071161871295, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3827481269836426, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7319124937057495, + "num_tokens": 676030630.0, + "step": 26127 + }, + { + "epoch": 2.869316933889743, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3283560276031494, + "learning_rate": 1e-06, + "loss": 0.831, + "mean_token_accuracy": 0.7432514429092407, + "num_tokens": 676055394.0, + "step": 26128 + }, + { + "epoch": 2.8694267515923566, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.286468505859375, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7439894080162048, + "num_tokens": 676081325.0, + "step": 26129 + }, + { + "epoch": 2.8695365692949704, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3136701583862305, + "learning_rate": 1e-06, + "loss": 0.8906, + "mean_token_accuracy": 0.7143609523773193, + "num_tokens": 676107605.0, + "step": 26130 + }, + { + "epoch": 2.869646386997584, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5265722274780273, + "learning_rate": 1e-06, + "loss": 0.8486, + "mean_token_accuracy": 0.7357417345046997, + "num_tokens": 676130392.0, + "step": 26131 + }, + { + "epoch": 2.869756204700198, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5048270225524902, + "learning_rate": 1e-06, + "loss": 0.8536, + "mean_token_accuracy": 0.7273274660110474, + "num_tokens": 676155618.0, + "step": 26132 + }, + { + "epoch": 2.869866022402811, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.474637508392334, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7379485964775085, + "num_tokens": 676180193.0, + "step": 26133 + }, + { + "epoch": 2.869975840105425, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3571524620056152, + "learning_rate": 1e-06, + "loss": 0.9506, + "mean_token_accuracy": 0.702436089515686, + "num_tokens": 676207552.0, + "step": 26134 + }, + { + "epoch": 2.8700856578080387, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5292017459869385, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7307261228561401, + "num_tokens": 676231269.0, + "step": 26135 + }, + { + "epoch": 2.8701954755106525, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7090060710906982, + "learning_rate": 1e-06, + "loss": 0.7934, + "mean_token_accuracy": 0.7536734342575073, + "num_tokens": 676251030.0, + "step": 26136 + }, + { + "epoch": 2.870305293213266, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3697848320007324, + "learning_rate": 1e-06, + "loss": 0.7911, + "mean_token_accuracy": 0.7518392205238342, + "num_tokens": 676277159.0, + "step": 26137 + }, + { + "epoch": 2.8704151109158795, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4912142753601074, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.7333356142044067, + "num_tokens": 676302300.0, + "step": 26138 + }, + { + "epoch": 2.8705249286184933, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.590942144393921, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7461759448051453, + "num_tokens": 676322622.0, + "step": 26139 + }, + { + "epoch": 2.870634746321107, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.390256881713867, + "learning_rate": 1e-06, + "loss": 0.8824, + "mean_token_accuracy": 0.7191892862319946, + "num_tokens": 676347935.0, + "step": 26140 + }, + { + "epoch": 2.8707445640237204, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1260695457458496, + "learning_rate": 1e-06, + "loss": 0.9287, + "mean_token_accuracy": 0.7112855911254883, + "num_tokens": 676379047.0, + "step": 26141 + }, + { + "epoch": 2.8708543817263346, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.194549322128296, + "learning_rate": 1e-06, + "loss": 0.9799, + "mean_token_accuracy": 0.6929819583892822, + "num_tokens": 676408422.0, + "step": 26142 + }, + { + "epoch": 2.870964199428948, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.375847339630127, + "learning_rate": 1e-06, + "loss": 0.8589, + "mean_token_accuracy": 0.7275972962379456, + "num_tokens": 676435121.0, + "step": 26143 + }, + { + "epoch": 2.8710740171315616, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4043400287628174, + "learning_rate": 1e-06, + "loss": 0.8963, + "mean_token_accuracy": 0.7207750082015991, + "num_tokens": 676460750.0, + "step": 26144 + }, + { + "epoch": 2.8711838348341754, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4265475273132324, + "learning_rate": 1e-06, + "loss": 0.7947, + "mean_token_accuracy": 0.7468768358230591, + "num_tokens": 676482900.0, + "step": 26145 + }, + { + "epoch": 2.8712936525367887, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3276681900024414, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.73130202293396, + "num_tokens": 676510402.0, + "step": 26146 + }, + { + "epoch": 2.8714034702394025, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1255083084106445, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7385588884353638, + "num_tokens": 676541557.0, + "step": 26147 + }, + { + "epoch": 2.871513287942016, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.269411563873291, + "learning_rate": 1e-06, + "loss": 0.7699, + "mean_token_accuracy": 0.7507458925247192, + "num_tokens": 676567699.0, + "step": 26148 + }, + { + "epoch": 2.87162310564463, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5233750343322754, + "learning_rate": 1e-06, + "loss": 0.8054, + "mean_token_accuracy": 0.7379148006439209, + "num_tokens": 676590430.0, + "step": 26149 + }, + { + "epoch": 2.8717329233472437, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.399970293045044, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7273769974708557, + "num_tokens": 676616324.0, + "step": 26150 + }, + { + "epoch": 2.871842741049857, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.206007719039917, + "learning_rate": 1e-06, + "loss": 0.8469, + "mean_token_accuracy": 0.7355201244354248, + "num_tokens": 676645348.0, + "step": 26151 + }, + { + "epoch": 2.871952558752471, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7063281536102295, + "learning_rate": 1e-06, + "loss": 0.7993, + "mean_token_accuracy": 0.749527096748352, + "num_tokens": 676664263.0, + "step": 26152 + }, + { + "epoch": 2.8720623764550846, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.232564926147461, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7378489971160889, + "num_tokens": 676689852.0, + "step": 26153 + }, + { + "epoch": 2.8721721941576983, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4203555583953857, + "learning_rate": 1e-06, + "loss": 0.9197, + "mean_token_accuracy": 0.7121187448501587, + "num_tokens": 676716524.0, + "step": 26154 + }, + { + "epoch": 2.872282011860312, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2813944816589355, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7436953186988831, + "num_tokens": 676741984.0, + "step": 26155 + }, + { + "epoch": 2.8723918295629254, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.129793882369995, + "learning_rate": 1e-06, + "loss": 0.9295, + "mean_token_accuracy": 0.7100414037704468, + "num_tokens": 676774168.0, + "step": 26156 + }, + { + "epoch": 2.872501647265539, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0935957431793213, + "learning_rate": 1e-06, + "loss": 0.9284, + "mean_token_accuracy": 0.7177038192749023, + "num_tokens": 676806321.0, + "step": 26157 + }, + { + "epoch": 2.872611464968153, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4252586364746094, + "learning_rate": 1e-06, + "loss": 0.8006, + "mean_token_accuracy": 0.7450274229049683, + "num_tokens": 676830146.0, + "step": 26158 + }, + { + "epoch": 2.8727212826707667, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.252829074859619, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7447572946548462, + "num_tokens": 676856918.0, + "step": 26159 + }, + { + "epoch": 2.8728311003733804, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.722322940826416, + "learning_rate": 1e-06, + "loss": 0.7977, + "mean_token_accuracy": 0.7413542866706848, + "num_tokens": 676877334.0, + "step": 26160 + }, + { + "epoch": 2.8729409180759937, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.312932014465332, + "learning_rate": 1e-06, + "loss": 0.7975, + "mean_token_accuracy": 0.7365607023239136, + "num_tokens": 676902466.0, + "step": 26161 + }, + { + "epoch": 2.8730507357786075, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.528803825378418, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7362829446792603, + "num_tokens": 676923813.0, + "step": 26162 + }, + { + "epoch": 2.8731605534812212, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.467703342437744, + "learning_rate": 1e-06, + "loss": 0.8514, + "mean_token_accuracy": 0.735507071018219, + "num_tokens": 676946898.0, + "step": 26163 + }, + { + "epoch": 2.873270371183835, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.513429641723633, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7260516881942749, + "num_tokens": 676972284.0, + "step": 26164 + }, + { + "epoch": 2.8733801888864487, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.477976083755493, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7303262948989868, + "num_tokens": 676997213.0, + "step": 26165 + }, + { + "epoch": 2.873490006589062, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.533411979675293, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7258397340774536, + "num_tokens": 677020832.0, + "step": 26166 + }, + { + "epoch": 2.873599824291676, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.512455463409424, + "learning_rate": 1e-06, + "loss": 0.8934, + "mean_token_accuracy": 0.7205491065979004, + "num_tokens": 677044474.0, + "step": 26167 + }, + { + "epoch": 2.8737096419942896, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.226435422897339, + "learning_rate": 1e-06, + "loss": 0.7944, + "mean_token_accuracy": 0.7399155497550964, + "num_tokens": 677071226.0, + "step": 26168 + }, + { + "epoch": 2.873819459696903, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.457120895385742, + "learning_rate": 1e-06, + "loss": 0.7498, + "mean_token_accuracy": 0.754189133644104, + "num_tokens": 677094413.0, + "step": 26169 + }, + { + "epoch": 2.8739292773995166, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 1.9756340980529785, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7188197374343872, + "num_tokens": 677128378.0, + "step": 26170 + }, + { + "epoch": 2.8740390951021304, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5818071365356445, + "learning_rate": 1e-06, + "loss": 0.8792, + "mean_token_accuracy": 0.7245355248451233, + "num_tokens": 677151911.0, + "step": 26171 + }, + { + "epoch": 2.874148912804744, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1677117347717285, + "learning_rate": 1e-06, + "loss": 0.8409, + "mean_token_accuracy": 0.7423924207687378, + "num_tokens": 677180932.0, + "step": 26172 + }, + { + "epoch": 2.874258730507358, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3832945823669434, + "learning_rate": 1e-06, + "loss": 0.7618, + "mean_token_accuracy": 0.7518140077590942, + "num_tokens": 677204864.0, + "step": 26173 + }, + { + "epoch": 2.8743685482099712, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.365861415863037, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7270617485046387, + "num_tokens": 677229322.0, + "step": 26174 + }, + { + "epoch": 2.874478365912585, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.59340500831604, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7327001690864563, + "num_tokens": 677251165.0, + "step": 26175 + }, + { + "epoch": 2.8745881836151987, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.407811403274536, + "learning_rate": 1e-06, + "loss": 0.8798, + "mean_token_accuracy": 0.7316730618476868, + "num_tokens": 677276506.0, + "step": 26176 + }, + { + "epoch": 2.8746980013178125, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3976309299468994, + "learning_rate": 1e-06, + "loss": 0.8888, + "mean_token_accuracy": 0.7173742651939392, + "num_tokens": 677303683.0, + "step": 26177 + }, + { + "epoch": 2.8748078190204263, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.259009599685669, + "learning_rate": 1e-06, + "loss": 0.8018, + "mean_token_accuracy": 0.7485013008117676, + "num_tokens": 677329909.0, + "step": 26178 + }, + { + "epoch": 2.8749176367230396, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.549398422241211, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.72978675365448, + "num_tokens": 677352625.0, + "step": 26179 + }, + { + "epoch": 2.8750274544256533, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3542768955230713, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7345113754272461, + "num_tokens": 677377575.0, + "step": 26180 + }, + { + "epoch": 2.875137272128267, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.036320209503174, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7193915843963623, + "num_tokens": 677413143.0, + "step": 26181 + }, + { + "epoch": 2.875247089830881, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5012381076812744, + "learning_rate": 1e-06, + "loss": 0.8741, + "mean_token_accuracy": 0.7266467809677124, + "num_tokens": 677437420.0, + "step": 26182 + }, + { + "epoch": 2.8753569075334946, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4481890201568604, + "learning_rate": 1e-06, + "loss": 0.9498, + "mean_token_accuracy": 0.707470178604126, + "num_tokens": 677460817.0, + "step": 26183 + }, + { + "epoch": 2.875466725236108, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2298824787139893, + "learning_rate": 1e-06, + "loss": 0.9741, + "mean_token_accuracy": 0.7022586464881897, + "num_tokens": 677490990.0, + "step": 26184 + }, + { + "epoch": 2.8755765429387217, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3149006366729736, + "learning_rate": 1e-06, + "loss": 0.9638, + "mean_token_accuracy": 0.6997618079185486, + "num_tokens": 677519083.0, + "step": 26185 + }, + { + "epoch": 2.8756863606413354, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.686601161956787, + "learning_rate": 1e-06, + "loss": 0.9004, + "mean_token_accuracy": 0.7246006727218628, + "num_tokens": 677541676.0, + "step": 26186 + }, + { + "epoch": 2.875796178343949, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1906487941741943, + "learning_rate": 1e-06, + "loss": 0.9336, + "mean_token_accuracy": 0.708808958530426, + "num_tokens": 677574898.0, + "step": 26187 + }, + { + "epoch": 2.875905996046563, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.531609296798706, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7252997159957886, + "num_tokens": 677597511.0, + "step": 26188 + }, + { + "epoch": 2.8760158137491763, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4418272972106934, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.736663281917572, + "num_tokens": 677621779.0, + "step": 26189 + }, + { + "epoch": 2.87612563145179, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.515948534011841, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7214285135269165, + "num_tokens": 677645561.0, + "step": 26190 + }, + { + "epoch": 2.8762354491544038, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.33524227142334, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7393060922622681, + "num_tokens": 677669700.0, + "step": 26191 + }, + { + "epoch": 2.876345266857017, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2472667694091797, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7152413725852966, + "num_tokens": 677698827.0, + "step": 26192 + }, + { + "epoch": 2.8764550845596313, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1178243160247803, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7267632484436035, + "num_tokens": 677729257.0, + "step": 26193 + }, + { + "epoch": 2.8765649022622446, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3026885986328125, + "learning_rate": 1e-06, + "loss": 0.8921, + "mean_token_accuracy": 0.7187185287475586, + "num_tokens": 677755695.0, + "step": 26194 + }, + { + "epoch": 2.8766747199648584, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.310479164123535, + "learning_rate": 1e-06, + "loss": 0.8779, + "mean_token_accuracy": 0.7233573198318481, + "num_tokens": 677783088.0, + "step": 26195 + }, + { + "epoch": 2.876784537667472, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1893398761749268, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7311679124832153, + "num_tokens": 677814594.0, + "step": 26196 + }, + { + "epoch": 2.8768943553700854, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4929988384246826, + "learning_rate": 1e-06, + "loss": 0.7627, + "mean_token_accuracy": 0.753296971321106, + "num_tokens": 677836354.0, + "step": 26197 + }, + { + "epoch": 2.877004173072699, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2073495388031006, + "learning_rate": 1e-06, + "loss": 0.8007, + "mean_token_accuracy": 0.7445499897003174, + "num_tokens": 677865061.0, + "step": 26198 + }, + { + "epoch": 2.877113990775313, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.379314422607422, + "learning_rate": 1e-06, + "loss": 0.9598, + "mean_token_accuracy": 0.704538106918335, + "num_tokens": 677893485.0, + "step": 26199 + }, + { + "epoch": 2.8772238084779267, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4415838718414307, + "learning_rate": 1e-06, + "loss": 0.8496, + "mean_token_accuracy": 0.7241069078445435, + "num_tokens": 677917057.0, + "step": 26200 + }, + { + "epoch": 2.8773336261805404, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1145706176757812, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7418683767318726, + "num_tokens": 677946716.0, + "step": 26201 + }, + { + "epoch": 2.8774434438831538, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3817787170410156, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7127685546875, + "num_tokens": 677974654.0, + "step": 26202 + }, + { + "epoch": 2.8775532615857675, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.382392406463623, + "learning_rate": 1e-06, + "loss": 0.8935, + "mean_token_accuracy": 0.7205721139907837, + "num_tokens": 678000009.0, + "step": 26203 + }, + { + "epoch": 2.8776630792883813, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.17208194732666, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.736811637878418, + "num_tokens": 678027105.0, + "step": 26204 + }, + { + "epoch": 2.877772896990995, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0949695110321045, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7329181432723999, + "num_tokens": 678057760.0, + "step": 26205 + }, + { + "epoch": 2.877882714693609, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.401308059692383, + "learning_rate": 1e-06, + "loss": 0.8493, + "mean_token_accuracy": 0.7350612878799438, + "num_tokens": 678084386.0, + "step": 26206 + }, + { + "epoch": 2.877992532396222, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2922558784484863, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.724949300289154, + "num_tokens": 678110250.0, + "step": 26207 + }, + { + "epoch": 2.878102350098836, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2993338108062744, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7106974720954895, + "num_tokens": 678138519.0, + "step": 26208 + }, + { + "epoch": 2.8782121678014496, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5837278366088867, + "learning_rate": 1e-06, + "loss": 0.8601, + "mean_token_accuracy": 0.7251471281051636, + "num_tokens": 678160459.0, + "step": 26209 + }, + { + "epoch": 2.8783219855040634, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5412697792053223, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7389978766441345, + "num_tokens": 678183192.0, + "step": 26210 + }, + { + "epoch": 2.878431803206677, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2079882621765137, + "learning_rate": 1e-06, + "loss": 0.8338, + "mean_token_accuracy": 0.7335543036460876, + "num_tokens": 678212127.0, + "step": 26211 + }, + { + "epoch": 2.8785416209092904, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2444419860839844, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7363367080688477, + "num_tokens": 678240686.0, + "step": 26212 + }, + { + "epoch": 2.878651438611904, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1999616622924805, + "learning_rate": 1e-06, + "loss": 0.7714, + "mean_token_accuracy": 0.758634090423584, + "num_tokens": 678270121.0, + "step": 26213 + }, + { + "epoch": 2.878761256314518, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1412782669067383, + "learning_rate": 1e-06, + "loss": 0.9789, + "mean_token_accuracy": 0.7030709385871887, + "num_tokens": 678300259.0, + "step": 26214 + }, + { + "epoch": 2.8788710740171317, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.240243673324585, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7239261865615845, + "num_tokens": 678325595.0, + "step": 26215 + }, + { + "epoch": 2.8789808917197455, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4131109714508057, + "learning_rate": 1e-06, + "loss": 0.8693, + "mean_token_accuracy": 0.7260171175003052, + "num_tokens": 678351866.0, + "step": 26216 + }, + { + "epoch": 2.879090709422359, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.307141065597534, + "learning_rate": 1e-06, + "loss": 0.8679, + "mean_token_accuracy": 0.7205076217651367, + "num_tokens": 678380016.0, + "step": 26217 + }, + { + "epoch": 2.8792005271249725, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.385075330734253, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.743687629699707, + "num_tokens": 678406271.0, + "step": 26218 + }, + { + "epoch": 2.8793103448275863, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.264286756515503, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7386282682418823, + "num_tokens": 678433377.0, + "step": 26219 + }, + { + "epoch": 2.8794201625301996, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4344794750213623, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7382926940917969, + "num_tokens": 678457691.0, + "step": 26220 + }, + { + "epoch": 2.8795299802328134, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.210390567779541, + "learning_rate": 1e-06, + "loss": 0.8952, + "mean_token_accuracy": 0.7182099223136902, + "num_tokens": 678486304.0, + "step": 26221 + }, + { + "epoch": 2.879639797935427, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3183538913726807, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.7199993133544922, + "num_tokens": 678512215.0, + "step": 26222 + }, + { + "epoch": 2.879749615638041, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2429745197296143, + "learning_rate": 1e-06, + "loss": 0.8739, + "mean_token_accuracy": 0.7336090803146362, + "num_tokens": 678537965.0, + "step": 26223 + }, + { + "epoch": 2.8798594333406546, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1361119747161865, + "learning_rate": 1e-06, + "loss": 0.8456, + "mean_token_accuracy": 0.7368760108947754, + "num_tokens": 678567305.0, + "step": 26224 + }, + { + "epoch": 2.879969251043268, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6769275665283203, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7386454343795776, + "num_tokens": 678588741.0, + "step": 26225 + }, + { + "epoch": 2.8800790687458817, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3798577785491943, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7234959006309509, + "num_tokens": 678614744.0, + "step": 26226 + }, + { + "epoch": 2.8801888864484955, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.8522486686706543, + "learning_rate": 1e-06, + "loss": 0.729, + "mean_token_accuracy": 0.7515732049942017, + "num_tokens": 678633366.0, + "step": 26227 + }, + { + "epoch": 2.8802987041511092, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6253890991210938, + "learning_rate": 1e-06, + "loss": 0.8043, + "mean_token_accuracy": 0.7458503246307373, + "num_tokens": 678655254.0, + "step": 26228 + }, + { + "epoch": 2.880408521853723, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4428505897521973, + "learning_rate": 1e-06, + "loss": 0.7687, + "mean_token_accuracy": 0.7550694942474365, + "num_tokens": 678678121.0, + "step": 26229 + }, + { + "epoch": 2.8805183395563363, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0993223190307617, + "learning_rate": 1e-06, + "loss": 0.857, + "mean_token_accuracy": 0.7279915809631348, + "num_tokens": 678708706.0, + "step": 26230 + }, + { + "epoch": 2.88062815725895, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2598190307617188, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7166074514389038, + "num_tokens": 678736911.0, + "step": 26231 + }, + { + "epoch": 2.880737974961564, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3821136951446533, + "learning_rate": 1e-06, + "loss": 0.8466, + "mean_token_accuracy": 0.730161726474762, + "num_tokens": 678760187.0, + "step": 26232 + }, + { + "epoch": 2.8808477926641776, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4838061332702637, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7347859740257263, + "num_tokens": 678783427.0, + "step": 26233 + }, + { + "epoch": 2.8809576103667913, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5224411487579346, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7376022934913635, + "num_tokens": 678806767.0, + "step": 26234 + }, + { + "epoch": 2.8810674280694046, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5857930183410645, + "learning_rate": 1e-06, + "loss": 0.8606, + "mean_token_accuracy": 0.7285293340682983, + "num_tokens": 678831097.0, + "step": 26235 + }, + { + "epoch": 2.8811772457720184, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.228362798690796, + "learning_rate": 1e-06, + "loss": 0.9163, + "mean_token_accuracy": 0.7189884781837463, + "num_tokens": 678859565.0, + "step": 26236 + }, + { + "epoch": 2.881287063474632, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2737345695495605, + "learning_rate": 1e-06, + "loss": 0.8451, + "mean_token_accuracy": 0.7329871654510498, + "num_tokens": 678887636.0, + "step": 26237 + }, + { + "epoch": 2.881396881177246, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.315866231918335, + "learning_rate": 1e-06, + "loss": 0.8462, + "mean_token_accuracy": 0.7301771640777588, + "num_tokens": 678915873.0, + "step": 26238 + }, + { + "epoch": 2.8815066988798597, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3050289154052734, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7263344526290894, + "num_tokens": 678941675.0, + "step": 26239 + }, + { + "epoch": 2.881616516582473, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.490435838699341, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7391241192817688, + "num_tokens": 678965197.0, + "step": 26240 + }, + { + "epoch": 2.8817263342850867, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4558374881744385, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7099631428718567, + "num_tokens": 678990892.0, + "step": 26241 + }, + { + "epoch": 2.8818361519877005, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4987270832061768, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7456894516944885, + "num_tokens": 679012837.0, + "step": 26242 + }, + { + "epoch": 2.881945969690314, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1671879291534424, + "learning_rate": 1e-06, + "loss": 0.9697, + "mean_token_accuracy": 0.7045526504516602, + "num_tokens": 679045032.0, + "step": 26243 + }, + { + "epoch": 2.882055787392928, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.809286594390869, + "learning_rate": 1e-06, + "loss": 0.8102, + "mean_token_accuracy": 0.7386756539344788, + "num_tokens": 679064203.0, + "step": 26244 + }, + { + "epoch": 2.8821656050955413, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2680482864379883, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7285283803939819, + "num_tokens": 679091818.0, + "step": 26245 + }, + { + "epoch": 2.882275422798155, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4445886611938477, + "learning_rate": 1e-06, + "loss": 0.7368, + "mean_token_accuracy": 0.7617326378822327, + "num_tokens": 679117480.0, + "step": 26246 + }, + { + "epoch": 2.882385240500769, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2090346813201904, + "learning_rate": 1e-06, + "loss": 0.8776, + "mean_token_accuracy": 0.7221705913543701, + "num_tokens": 679147118.0, + "step": 26247 + }, + { + "epoch": 2.882495058203382, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4572646617889404, + "learning_rate": 1e-06, + "loss": 0.8955, + "mean_token_accuracy": 0.7214494943618774, + "num_tokens": 679171613.0, + "step": 26248 + }, + { + "epoch": 2.882604875905996, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5423102378845215, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.7313113808631897, + "num_tokens": 679200715.0, + "step": 26249 + }, + { + "epoch": 2.8827146936086097, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1527721881866455, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7396916151046753, + "num_tokens": 679229449.0, + "step": 26250 + }, + { + "epoch": 2.8828245113112234, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3867971897125244, + "learning_rate": 1e-06, + "loss": 0.9494, + "mean_token_accuracy": 0.7076486349105835, + "num_tokens": 679255059.0, + "step": 26251 + }, + { + "epoch": 2.882934329013837, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.729780912399292, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7382454872131348, + "num_tokens": 679274261.0, + "step": 26252 + }, + { + "epoch": 2.8830441467164505, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.062818765640259, + "learning_rate": 1e-06, + "loss": 0.8228, + "mean_token_accuracy": 0.7562159299850464, + "num_tokens": 679302410.0, + "step": 26253 + }, + { + "epoch": 2.8831539644190642, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.304351329803467, + "learning_rate": 1e-06, + "loss": 0.7631, + "mean_token_accuracy": 0.7551270723342896, + "num_tokens": 679327577.0, + "step": 26254 + }, + { + "epoch": 2.883263782121678, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3537678718566895, + "learning_rate": 1e-06, + "loss": 0.7829, + "mean_token_accuracy": 0.7452211380004883, + "num_tokens": 679353635.0, + "step": 26255 + }, + { + "epoch": 2.8833735998242918, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3437767028808594, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7255752086639404, + "num_tokens": 679382008.0, + "step": 26256 + }, + { + "epoch": 2.8834834175269055, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.270524501800537, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7205860018730164, + "num_tokens": 679410828.0, + "step": 26257 + }, + { + "epoch": 2.883593235229519, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0052268505096436, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7275736331939697, + "num_tokens": 679443533.0, + "step": 26258 + }, + { + "epoch": 2.8837030529321326, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.8196775913238525, + "learning_rate": 1e-06, + "loss": 0.7802, + "mean_token_accuracy": 0.7467730045318604, + "num_tokens": 679463114.0, + "step": 26259 + }, + { + "epoch": 2.8838128706347463, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.267831563949585, + "learning_rate": 1e-06, + "loss": 0.7376, + "mean_token_accuracy": 0.7566803097724915, + "num_tokens": 679490246.0, + "step": 26260 + }, + { + "epoch": 2.88392268833736, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1996665000915527, + "learning_rate": 1e-06, + "loss": 0.8399, + "mean_token_accuracy": 0.7367233633995056, + "num_tokens": 679517195.0, + "step": 26261 + }, + { + "epoch": 2.884032506039974, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.6080212593078613, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7419697046279907, + "num_tokens": 679540984.0, + "step": 26262 + }, + { + "epoch": 2.884142323742587, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.600728750228882, + "learning_rate": 1e-06, + "loss": 0.8616, + "mean_token_accuracy": 0.722359836101532, + "num_tokens": 679564476.0, + "step": 26263 + }, + { + "epoch": 2.884252141445201, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.255084276199341, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7318530082702637, + "num_tokens": 679591224.0, + "step": 26264 + }, + { + "epoch": 2.8843619591478147, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.252866744995117, + "learning_rate": 1e-06, + "loss": 0.7537, + "mean_token_accuracy": 0.7547461986541748, + "num_tokens": 679618867.0, + "step": 26265 + }, + { + "epoch": 2.8844717768504284, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.232107400894165, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7188670635223389, + "num_tokens": 679647701.0, + "step": 26266 + }, + { + "epoch": 2.884581594553042, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.840035915374756, + "learning_rate": 1e-06, + "loss": 0.755, + "mean_token_accuracy": 0.7518677711486816, + "num_tokens": 679666694.0, + "step": 26267 + }, + { + "epoch": 2.8846914122556555, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3800642490386963, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7450282573699951, + "num_tokens": 679690306.0, + "step": 26268 + }, + { + "epoch": 2.8848012299582693, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7956058979034424, + "learning_rate": 1e-06, + "loss": 0.8509, + "mean_token_accuracy": 0.7270331382751465, + "num_tokens": 679710856.0, + "step": 26269 + }, + { + "epoch": 2.884911047660883, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1624319553375244, + "learning_rate": 1e-06, + "loss": 0.8907, + "mean_token_accuracy": 0.716395378112793, + "num_tokens": 679740890.0, + "step": 26270 + }, + { + "epoch": 2.8850208653634963, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.101116418838501, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.7322040796279907, + "num_tokens": 679772541.0, + "step": 26271 + }, + { + "epoch": 2.88513068306611, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.689638614654541, + "learning_rate": 1e-06, + "loss": 0.7855, + "mean_token_accuracy": 0.7421709299087524, + "num_tokens": 679792942.0, + "step": 26272 + }, + { + "epoch": 2.885240500768724, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2463252544403076, + "learning_rate": 1e-06, + "loss": 0.9131, + "mean_token_accuracy": 0.7190015316009521, + "num_tokens": 679821926.0, + "step": 26273 + }, + { + "epoch": 2.8853503184713376, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3298118114471436, + "learning_rate": 1e-06, + "loss": 0.8047, + "mean_token_accuracy": 0.743493378162384, + "num_tokens": 679846469.0, + "step": 26274 + }, + { + "epoch": 2.8854601361739514, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.110889434814453, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7172276973724365, + "num_tokens": 679876950.0, + "step": 26275 + }, + { + "epoch": 2.8855699538765647, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.140772819519043, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7144507169723511, + "num_tokens": 679907480.0, + "step": 26276 + }, + { + "epoch": 2.8856797715791784, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.412214994430542, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7393866777420044, + "num_tokens": 679932416.0, + "step": 26277 + }, + { + "epoch": 2.885789589281792, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3701705932617188, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.7296446561813354, + "num_tokens": 679956498.0, + "step": 26278 + }, + { + "epoch": 2.885899406984406, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.20920467376709, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7203464508056641, + "num_tokens": 679983183.0, + "step": 26279 + }, + { + "epoch": 2.8860092246870197, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.120218276977539, + "learning_rate": 1e-06, + "loss": 0.8405, + "mean_token_accuracy": 0.7562923431396484, + "num_tokens": 680013070.0, + "step": 26280 + }, + { + "epoch": 2.886119042389633, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.479126453399658, + "learning_rate": 1e-06, + "loss": 0.9466, + "mean_token_accuracy": 0.71834397315979, + "num_tokens": 680038849.0, + "step": 26281 + }, + { + "epoch": 2.8862288600922468, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3170909881591797, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7314309477806091, + "num_tokens": 680064328.0, + "step": 26282 + }, + { + "epoch": 2.8863386777948605, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.397632122039795, + "learning_rate": 1e-06, + "loss": 0.942, + "mean_token_accuracy": 0.7108994722366333, + "num_tokens": 680091408.0, + "step": 26283 + }, + { + "epoch": 2.8864484954974743, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.080986976623535, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7385557293891907, + "num_tokens": 680123767.0, + "step": 26284 + }, + { + "epoch": 2.886558313200088, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.8659918308258057, + "learning_rate": 1e-06, + "loss": 0.8653, + "mean_token_accuracy": 0.7263916730880737, + "num_tokens": 680143812.0, + "step": 26285 + }, + { + "epoch": 2.8866681309027014, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3380062580108643, + "learning_rate": 1e-06, + "loss": 0.8098, + "mean_token_accuracy": 0.7557175159454346, + "num_tokens": 680169260.0, + "step": 26286 + }, + { + "epoch": 2.886777948605315, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.382185459136963, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7455301284790039, + "num_tokens": 680193558.0, + "step": 26287 + }, + { + "epoch": 2.886887766307929, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.325094699859619, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7257996201515198, + "num_tokens": 680220597.0, + "step": 26288 + }, + { + "epoch": 2.8869975840105426, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7468273639678955, + "learning_rate": 1e-06, + "loss": 0.8082, + "mean_token_accuracy": 0.751083493232727, + "num_tokens": 680240737.0, + "step": 26289 + }, + { + "epoch": 2.8871074017131564, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2968666553497314, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7422700524330139, + "num_tokens": 680266014.0, + "step": 26290 + }, + { + "epoch": 2.8872172194157697, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4281864166259766, + "learning_rate": 1e-06, + "loss": 0.8295, + "mean_token_accuracy": 0.7375892400741577, + "num_tokens": 680290115.0, + "step": 26291 + }, + { + "epoch": 2.8873270371183835, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2947728633880615, + "learning_rate": 1e-06, + "loss": 0.8569, + "mean_token_accuracy": 0.7271474003791809, + "num_tokens": 680318250.0, + "step": 26292 + }, + { + "epoch": 2.887436854820997, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3927903175354004, + "learning_rate": 1e-06, + "loss": 0.8567, + "mean_token_accuracy": 0.7298725843429565, + "num_tokens": 680342706.0, + "step": 26293 + }, + { + "epoch": 2.887546672523611, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.235783338546753, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.720257580280304, + "num_tokens": 680373243.0, + "step": 26294 + }, + { + "epoch": 2.8876564902262247, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0328874588012695, + "learning_rate": 1e-06, + "loss": 0.8823, + "mean_token_accuracy": 0.7274749875068665, + "num_tokens": 680404986.0, + "step": 26295 + }, + { + "epoch": 2.887766307928838, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.363516092300415, + "learning_rate": 1e-06, + "loss": 0.8564, + "mean_token_accuracy": 0.7271955609321594, + "num_tokens": 680430597.0, + "step": 26296 + }, + { + "epoch": 2.887876125631452, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2295982837677, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.7462068796157837, + "num_tokens": 680458734.0, + "step": 26297 + }, + { + "epoch": 2.8879859433340656, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.224487781524658, + "learning_rate": 1e-06, + "loss": 0.9612, + "mean_token_accuracy": 0.7053467035293579, + "num_tokens": 680490034.0, + "step": 26298 + }, + { + "epoch": 2.888095761036679, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.326909303665161, + "learning_rate": 1e-06, + "loss": 0.7848, + "mean_token_accuracy": 0.7451959848403931, + "num_tokens": 680515547.0, + "step": 26299 + }, + { + "epoch": 2.8882055787392926, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.461484432220459, + "learning_rate": 1e-06, + "loss": 0.8383, + "mean_token_accuracy": 0.73728346824646, + "num_tokens": 680538587.0, + "step": 26300 + }, + { + "epoch": 2.8883153964419064, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3295445442199707, + "learning_rate": 1e-06, + "loss": 0.8764, + "mean_token_accuracy": 0.7263540625572205, + "num_tokens": 680565636.0, + "step": 26301 + }, + { + "epoch": 2.88842521414452, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3834142684936523, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7391315698623657, + "num_tokens": 680589704.0, + "step": 26302 + }, + { + "epoch": 2.888535031847134, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5407040119171143, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7238250970840454, + "num_tokens": 680611431.0, + "step": 26303 + }, + { + "epoch": 2.888644849549747, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.516354560852051, + "learning_rate": 1e-06, + "loss": 0.8852, + "mean_token_accuracy": 0.722601592540741, + "num_tokens": 680634011.0, + "step": 26304 + }, + { + "epoch": 2.888754667252361, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.387651205062866, + "learning_rate": 1e-06, + "loss": 0.7829, + "mean_token_accuracy": 0.74951171875, + "num_tokens": 680657458.0, + "step": 26305 + }, + { + "epoch": 2.8888644849549747, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1811110973358154, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.7588946223258972, + "num_tokens": 680683301.0, + "step": 26306 + }, + { + "epoch": 2.8889743026575885, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.397172689437866, + "learning_rate": 1e-06, + "loss": 0.9196, + "mean_token_accuracy": 0.7157506942749023, + "num_tokens": 680709705.0, + "step": 26307 + }, + { + "epoch": 2.8890841203602022, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.356193780899048, + "learning_rate": 1e-06, + "loss": 0.8818, + "mean_token_accuracy": 0.723732590675354, + "num_tokens": 680736867.0, + "step": 26308 + }, + { + "epoch": 2.8891939380628155, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4706382751464844, + "learning_rate": 1e-06, + "loss": 0.8633, + "mean_token_accuracy": 0.7318462133407593, + "num_tokens": 680761365.0, + "step": 26309 + }, + { + "epoch": 2.8893037557654293, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4377658367156982, + "learning_rate": 1e-06, + "loss": 0.8983, + "mean_token_accuracy": 0.7153880596160889, + "num_tokens": 680785840.0, + "step": 26310 + }, + { + "epoch": 2.889413573468043, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4663994312286377, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7304303050041199, + "num_tokens": 680811875.0, + "step": 26311 + }, + { + "epoch": 2.889523391170657, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.457430124282837, + "learning_rate": 1e-06, + "loss": 0.7403, + "mean_token_accuracy": 0.7661110758781433, + "num_tokens": 680834686.0, + "step": 26312 + }, + { + "epoch": 2.8896332088732706, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.191474676132202, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7244645357131958, + "num_tokens": 680863684.0, + "step": 26313 + }, + { + "epoch": 2.889743026575884, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4714245796203613, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.736301064491272, + "num_tokens": 680886106.0, + "step": 26314 + }, + { + "epoch": 2.8898528442784976, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2316689491271973, + "learning_rate": 1e-06, + "loss": 0.866, + "mean_token_accuracy": 0.7232959270477295, + "num_tokens": 680914405.0, + "step": 26315 + }, + { + "epoch": 2.8899626619811114, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1929006576538086, + "learning_rate": 1e-06, + "loss": 0.8666, + "mean_token_accuracy": 0.7293162941932678, + "num_tokens": 680943760.0, + "step": 26316 + }, + { + "epoch": 2.890072479683725, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2800214290618896, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7418391108512878, + "num_tokens": 680971377.0, + "step": 26317 + }, + { + "epoch": 2.890182297386339, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0898256301879883, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7266595959663391, + "num_tokens": 681002931.0, + "step": 26318 + }, + { + "epoch": 2.8902921150889522, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2703137397766113, + "learning_rate": 1e-06, + "loss": 0.8786, + "mean_token_accuracy": 0.7278395891189575, + "num_tokens": 681030516.0, + "step": 26319 + }, + { + "epoch": 2.890401932791566, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.18143367767334, + "learning_rate": 1e-06, + "loss": 0.9221, + "mean_token_accuracy": 0.7062829732894897, + "num_tokens": 681059702.0, + "step": 26320 + }, + { + "epoch": 2.8905117504941797, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.2089741230010986, + "learning_rate": 1e-06, + "loss": 1.0039, + "mean_token_accuracy": 0.6875835061073303, + "num_tokens": 681092970.0, + "step": 26321 + }, + { + "epoch": 2.890621568196793, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.5427327156066895, + "learning_rate": 1e-06, + "loss": 0.8165, + "mean_token_accuracy": 0.7490760087966919, + "num_tokens": 681116302.0, + "step": 26322 + }, + { + "epoch": 2.8907313858994073, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4125616550445557, + "learning_rate": 1e-06, + "loss": 0.8661, + "mean_token_accuracy": 0.7237191200256348, + "num_tokens": 681140960.0, + "step": 26323 + }, + { + "epoch": 2.8908412036020206, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4726319313049316, + "learning_rate": 1e-06, + "loss": 0.8628, + "mean_token_accuracy": 0.7230852842330933, + "num_tokens": 681166078.0, + "step": 26324 + }, + { + "epoch": 2.8909510213046343, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.473614454269409, + "learning_rate": 1e-06, + "loss": 0.9648, + "mean_token_accuracy": 0.7118116021156311, + "num_tokens": 681194378.0, + "step": 26325 + }, + { + "epoch": 2.891060839007248, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.3176095485687256, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7288841009140015, + "num_tokens": 681220387.0, + "step": 26326 + }, + { + "epoch": 2.8911706567098614, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.380906343460083, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7242113947868347, + "num_tokens": 681246376.0, + "step": 26327 + }, + { + "epoch": 2.891280474412475, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.1991565227508545, + "learning_rate": 1e-06, + "loss": 0.9335, + "mean_token_accuracy": 0.7156867980957031, + "num_tokens": 681277320.0, + "step": 26328 + }, + { + "epoch": 2.891390292115089, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4604198932647705, + "learning_rate": 1e-06, + "loss": 0.8049, + "mean_token_accuracy": 0.7466680407524109, + "num_tokens": 681301160.0, + "step": 26329 + }, + { + "epoch": 2.8915001098177027, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.588876962661743, + "learning_rate": 1e-06, + "loss": 0.8159, + "mean_token_accuracy": 0.7405654191970825, + "num_tokens": 681322226.0, + "step": 26330 + }, + { + "epoch": 2.8916099275203164, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.487534999847412, + "learning_rate": 1e-06, + "loss": 0.8051, + "mean_token_accuracy": 0.7444554567337036, + "num_tokens": 681345895.0, + "step": 26331 + }, + { + "epoch": 2.8917197452229297, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.55267071723938, + "learning_rate": 1e-06, + "loss": 0.8119, + "mean_token_accuracy": 0.7430859804153442, + "num_tokens": 681369058.0, + "step": 26332 + }, + { + "epoch": 2.8918295629255435, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4430606365203857, + "learning_rate": 1e-06, + "loss": 0.9089, + "mean_token_accuracy": 0.7226998805999756, + "num_tokens": 681393988.0, + "step": 26333 + }, + { + "epoch": 2.8919393806281573, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4326813220977783, + "learning_rate": 1e-06, + "loss": 0.7987, + "mean_token_accuracy": 0.7445476055145264, + "num_tokens": 681419271.0, + "step": 26334 + }, + { + "epoch": 2.892049198330771, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4451632499694824, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7162055373191833, + "num_tokens": 681445181.0, + "step": 26335 + }, + { + "epoch": 2.8921590160333848, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.0566985607147217, + "learning_rate": 1e-06, + "loss": 0.9157, + "mean_token_accuracy": 0.7161334753036499, + "num_tokens": 681478406.0, + "step": 26336 + }, + { + "epoch": 2.892268833735998, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4483890533447266, + "learning_rate": 1e-06, + "loss": 0.8851, + "mean_token_accuracy": 0.728985071182251, + "num_tokens": 681502338.0, + "step": 26337 + }, + { + "epoch": 2.892378651438612, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.4745242595672607, + "learning_rate": 1e-06, + "loss": 0.9245, + "mean_token_accuracy": 0.7124297618865967, + "num_tokens": 681531407.0, + "step": 26338 + }, + { + "epoch": 2.8924884691412256, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.126476287841797, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7130142450332642, + "num_tokens": 681563462.0, + "step": 26339 + }, + { + "epoch": 2.8925982868438394, + "ewc_loss": 1.6570091247558594e-05, + "grad_norm": 2.7018353939056396, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.75104820728302, + "num_tokens": 681584214.0, + "step": 26340 + }, + { + "epoch": 2.892708104546453, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.2137937545776367, + "learning_rate": 1e-06, + "loss": 0.8174, + "mean_token_accuracy": 0.7429821491241455, + "num_tokens": 681613791.0, + "step": 26341 + }, + { + "epoch": 2.8928179222490664, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.211418628692627, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7195309400558472, + "num_tokens": 681646841.0, + "step": 26342 + }, + { + "epoch": 2.89292773995168, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.3087105751037598, + "learning_rate": 1e-06, + "loss": 0.9342, + "mean_token_accuracy": 0.7157776355743408, + "num_tokens": 681672239.0, + "step": 26343 + }, + { + "epoch": 2.893037557654294, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.7403581142425537, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7273023128509521, + "num_tokens": 681692934.0, + "step": 26344 + }, + { + "epoch": 2.8931473753569077, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.4112374782562256, + "learning_rate": 1e-06, + "loss": 0.7683, + "mean_token_accuracy": 0.7521350383758545, + "num_tokens": 681716216.0, + "step": 26345 + }, + { + "epoch": 2.8932571930595214, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.217501401901245, + "learning_rate": 1e-06, + "loss": 0.8651, + "mean_token_accuracy": 0.7257087230682373, + "num_tokens": 681745947.0, + "step": 26346 + }, + { + "epoch": 2.8933670107621348, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.7330312728881836, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7462825775146484, + "num_tokens": 681764836.0, + "step": 26347 + }, + { + "epoch": 2.8934768284647485, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.4378809928894043, + "learning_rate": 1e-06, + "loss": 0.9069, + "mean_token_accuracy": 0.7216790914535522, + "num_tokens": 681790826.0, + "step": 26348 + }, + { + "epoch": 2.8935866461673623, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.241410732269287, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.7386050224304199, + "num_tokens": 681818482.0, + "step": 26349 + }, + { + "epoch": 2.8936964638699756, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.5135498046875, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7150965929031372, + "num_tokens": 681843663.0, + "step": 26350 + }, + { + "epoch": 2.8938062815725893, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.316868543624878, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7305251359939575, + "num_tokens": 681870011.0, + "step": 26351 + }, + { + "epoch": 2.893916099275203, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.4015390872955322, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7462913990020752, + "num_tokens": 681894347.0, + "step": 26352 + }, + { + "epoch": 2.894025916977817, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.181718587875366, + "learning_rate": 1e-06, + "loss": 0.7963, + "mean_token_accuracy": 0.7438775300979614, + "num_tokens": 681921147.0, + "step": 26353 + }, + { + "epoch": 2.8941357346804306, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.34686279296875, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7425495386123657, + "num_tokens": 681946221.0, + "step": 26354 + }, + { + "epoch": 2.894245552383044, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.2080161571502686, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7018862366676331, + "num_tokens": 681977249.0, + "step": 26355 + }, + { + "epoch": 2.8943553700856577, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.259298801422119, + "learning_rate": 1e-06, + "loss": 0.8676, + "mean_token_accuracy": 0.7306361198425293, + "num_tokens": 682004527.0, + "step": 26356 + }, + { + "epoch": 2.8944651877882714, + "ewc_loss": 1.6689300537109375e-05, + "grad_norm": 2.4274098873138428, + "learning_rate": 1e-06, + "loss": 0.7539, + "mean_token_accuracy": 0.756767213344574, + "num_tokens": 682028877.0, + "step": 26357 + }, + { + "epoch": 2.894575005490885, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.287259578704834, + "learning_rate": 1e-06, + "loss": 0.7903, + "mean_token_accuracy": 0.7436143159866333, + "num_tokens": 682055359.0, + "step": 26358 + }, + { + "epoch": 2.894684823193499, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6909472942352295, + "learning_rate": 1e-06, + "loss": 0.8122, + "mean_token_accuracy": 0.7393739223480225, + "num_tokens": 682077733.0, + "step": 26359 + }, + { + "epoch": 2.8947946408961123, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6674530506134033, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.724410891532898, + "num_tokens": 682099772.0, + "step": 26360 + }, + { + "epoch": 2.894904458598726, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.321244478225708, + "learning_rate": 1e-06, + "loss": 0.8292, + "mean_token_accuracy": 0.7393116354942322, + "num_tokens": 682127088.0, + "step": 26361 + }, + { + "epoch": 2.89501427630134, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6098639965057373, + "learning_rate": 1e-06, + "loss": 0.8336, + "mean_token_accuracy": 0.7370345592498779, + "num_tokens": 682149027.0, + "step": 26362 + }, + { + "epoch": 2.8951240940039535, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.562758684158325, + "learning_rate": 1e-06, + "loss": 0.9021, + "mean_token_accuracy": 0.7238530516624451, + "num_tokens": 682172011.0, + "step": 26363 + }, + { + "epoch": 2.8952339117065673, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.336888313293457, + "learning_rate": 1e-06, + "loss": 0.9126, + "mean_token_accuracy": 0.7180368900299072, + "num_tokens": 682199704.0, + "step": 26364 + }, + { + "epoch": 2.8953437294091806, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.64679217338562, + "learning_rate": 1e-06, + "loss": 0.7036, + "mean_token_accuracy": 0.7716154456138611, + "num_tokens": 682219918.0, + "step": 26365 + }, + { + "epoch": 2.8954535471117944, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.725515842437744, + "learning_rate": 1e-06, + "loss": 0.7869, + "mean_token_accuracy": 0.7449929714202881, + "num_tokens": 682241287.0, + "step": 26366 + }, + { + "epoch": 2.895563364814408, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.366070032119751, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7280453443527222, + "num_tokens": 682266925.0, + "step": 26367 + }, + { + "epoch": 2.895673182517022, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.792914390563965, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7442805767059326, + "num_tokens": 682287503.0, + "step": 26368 + }, + { + "epoch": 2.8957830002196356, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.362578868865967, + "learning_rate": 1e-06, + "loss": 0.7679, + "mean_token_accuracy": 0.7573134303092957, + "num_tokens": 682311142.0, + "step": 26369 + }, + { + "epoch": 2.895892817922249, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0033576488494873, + "learning_rate": 1e-06, + "loss": 0.8948, + "mean_token_accuracy": 0.716924786567688, + "num_tokens": 682346342.0, + "step": 26370 + }, + { + "epoch": 2.8960026356248627, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4780893325805664, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.7282420992851257, + "num_tokens": 682370628.0, + "step": 26371 + }, + { + "epoch": 2.8961124533274765, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6279168128967285, + "learning_rate": 1e-06, + "loss": 0.8508, + "mean_token_accuracy": 0.7337068915367126, + "num_tokens": 682393547.0, + "step": 26372 + }, + { + "epoch": 2.89622227103009, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2510931491851807, + "learning_rate": 1e-06, + "loss": 0.8517, + "mean_token_accuracy": 0.7415919303894043, + "num_tokens": 682421741.0, + "step": 26373 + }, + { + "epoch": 2.896332088732704, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.275195360183716, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7267506122589111, + "num_tokens": 682449528.0, + "step": 26374 + }, + { + "epoch": 2.8964419064353173, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.482973337173462, + "learning_rate": 1e-06, + "loss": 0.9023, + "mean_token_accuracy": 0.7188577055931091, + "num_tokens": 682473028.0, + "step": 26375 + }, + { + "epoch": 2.896551724137931, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4002513885498047, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.7231186032295227, + "num_tokens": 682500082.0, + "step": 26376 + }, + { + "epoch": 2.896661541840545, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2574775218963623, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.721294641494751, + "num_tokens": 682530425.0, + "step": 26377 + }, + { + "epoch": 2.896771359543158, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5866341590881348, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7230581641197205, + "num_tokens": 682554091.0, + "step": 26378 + }, + { + "epoch": 2.896881177245772, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.366227626800537, + "learning_rate": 1e-06, + "loss": 0.8649, + "mean_token_accuracy": 0.7258068323135376, + "num_tokens": 682579492.0, + "step": 26379 + }, + { + "epoch": 2.8969909949483856, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4917564392089844, + "learning_rate": 1e-06, + "loss": 0.7892, + "mean_token_accuracy": 0.7539520263671875, + "num_tokens": 682603660.0, + "step": 26380 + }, + { + "epoch": 2.8971008126509994, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.364943265914917, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7204846739768982, + "num_tokens": 682631859.0, + "step": 26381 + }, + { + "epoch": 2.897210630353613, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4721181392669678, + "learning_rate": 1e-06, + "loss": 0.8141, + "mean_token_accuracy": 0.7450554966926575, + "num_tokens": 682656306.0, + "step": 26382 + }, + { + "epoch": 2.8973204480562265, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.500119924545288, + "learning_rate": 1e-06, + "loss": 0.8275, + "mean_token_accuracy": 0.7349562644958496, + "num_tokens": 682679307.0, + "step": 26383 + }, + { + "epoch": 2.89743026575884, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.327676773071289, + "learning_rate": 1e-06, + "loss": 0.8277, + "mean_token_accuracy": 0.7382153272628784, + "num_tokens": 682704005.0, + "step": 26384 + }, + { + "epoch": 2.897540083461454, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2913951873779297, + "learning_rate": 1e-06, + "loss": 0.8031, + "mean_token_accuracy": 0.7445274591445923, + "num_tokens": 682730369.0, + "step": 26385 + }, + { + "epoch": 2.8976499011640677, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1070406436920166, + "learning_rate": 1e-06, + "loss": 0.9461, + "mean_token_accuracy": 0.7193770408630371, + "num_tokens": 682761701.0, + "step": 26386 + }, + { + "epoch": 2.8977597188666815, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.238832950592041, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7392745018005371, + "num_tokens": 682787398.0, + "step": 26387 + }, + { + "epoch": 2.897869536569295, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4064910411834717, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7232620716094971, + "num_tokens": 682813490.0, + "step": 26388 + }, + { + "epoch": 2.8979793542719086, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.409791946411133, + "learning_rate": 1e-06, + "loss": 0.8305, + "mean_token_accuracy": 0.7422994375228882, + "num_tokens": 682839551.0, + "step": 26389 + }, + { + "epoch": 2.8980891719745223, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4018445014953613, + "learning_rate": 1e-06, + "loss": 0.877, + "mean_token_accuracy": 0.7267682552337646, + "num_tokens": 682864892.0, + "step": 26390 + }, + { + "epoch": 2.898198989677136, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.359733819961548, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7136601805686951, + "num_tokens": 682892896.0, + "step": 26391 + }, + { + "epoch": 2.89830880737975, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.626194715499878, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7289965152740479, + "num_tokens": 682915798.0, + "step": 26392 + }, + { + "epoch": 2.898418625082363, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.335810661315918, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7308434844017029, + "num_tokens": 682940538.0, + "step": 26393 + }, + { + "epoch": 2.898528442784977, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.228938102722168, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.707011342048645, + "num_tokens": 682972109.0, + "step": 26394 + }, + { + "epoch": 2.8986382604875907, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2737581729888916, + "learning_rate": 1e-06, + "loss": 0.8625, + "mean_token_accuracy": 0.7236001491546631, + "num_tokens": 682998786.0, + "step": 26395 + }, + { + "epoch": 2.8987480781902044, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4353010654449463, + "learning_rate": 1e-06, + "loss": 0.9185, + "mean_token_accuracy": 0.7098038196563721, + "num_tokens": 683025291.0, + "step": 26396 + }, + { + "epoch": 2.898857895892818, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1454596519470215, + "learning_rate": 1e-06, + "loss": 0.8155, + "mean_token_accuracy": 0.7348916530609131, + "num_tokens": 683056370.0, + "step": 26397 + }, + { + "epoch": 2.8989677135954315, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.324827194213867, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7308991551399231, + "num_tokens": 683084465.0, + "step": 26398 + }, + { + "epoch": 2.8990775312980452, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.609649658203125, + "learning_rate": 1e-06, + "loss": 0.8461, + "mean_token_accuracy": 0.7313129901885986, + "num_tokens": 683108337.0, + "step": 26399 + }, + { + "epoch": 2.899187349000659, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.417112112045288, + "learning_rate": 1e-06, + "loss": 0.8671, + "mean_token_accuracy": 0.7264388799667358, + "num_tokens": 683134858.0, + "step": 26400 + }, + { + "epoch": 2.8992971667032723, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5523030757904053, + "learning_rate": 1e-06, + "loss": 0.8752, + "mean_token_accuracy": 0.7304490804672241, + "num_tokens": 683158821.0, + "step": 26401 + }, + { + "epoch": 2.899406984405886, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.320859432220459, + "learning_rate": 1e-06, + "loss": 0.8712, + "mean_token_accuracy": 0.7220685482025146, + "num_tokens": 683185864.0, + "step": 26402 + }, + { + "epoch": 2.8995168021085, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3841378688812256, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.725785493850708, + "num_tokens": 683211009.0, + "step": 26403 + }, + { + "epoch": 2.8996266198111136, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1575210094451904, + "learning_rate": 1e-06, + "loss": 0.7971, + "mean_token_accuracy": 0.7440639138221741, + "num_tokens": 683240253.0, + "step": 26404 + }, + { + "epoch": 2.8997364375137273, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5598292350769043, + "learning_rate": 1e-06, + "loss": 0.885, + "mean_token_accuracy": 0.7210787534713745, + "num_tokens": 683264149.0, + "step": 26405 + }, + { + "epoch": 2.8998462552163407, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3859808444976807, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7207118272781372, + "num_tokens": 683292588.0, + "step": 26406 + }, + { + "epoch": 2.8999560729189544, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5175461769104004, + "learning_rate": 1e-06, + "loss": 0.7853, + "mean_token_accuracy": 0.7449504733085632, + "num_tokens": 683314964.0, + "step": 26407 + }, + { + "epoch": 2.900065890621568, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.218064308166504, + "learning_rate": 1e-06, + "loss": 0.7075, + "mean_token_accuracy": 0.7698713541030884, + "num_tokens": 683340361.0, + "step": 26408 + }, + { + "epoch": 2.900175708324182, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6072640419006348, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7360214591026306, + "num_tokens": 683362288.0, + "step": 26409 + }, + { + "epoch": 2.9002855260267957, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.332977771759033, + "learning_rate": 1e-06, + "loss": 0.9041, + "mean_token_accuracy": 0.7291675209999084, + "num_tokens": 683387158.0, + "step": 26410 + }, + { + "epoch": 2.900395343729409, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7048556804656982, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7239913940429688, + "num_tokens": 683408163.0, + "step": 26411 + }, + { + "epoch": 2.9005051614320227, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4196958541870117, + "learning_rate": 1e-06, + "loss": 0.869, + "mean_token_accuracy": 0.7255533933639526, + "num_tokens": 683434116.0, + "step": 26412 + }, + { + "epoch": 2.9006149791346365, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.211697578430176, + "learning_rate": 1e-06, + "loss": 0.8641, + "mean_token_accuracy": 0.7298778891563416, + "num_tokens": 683463104.0, + "step": 26413 + }, + { + "epoch": 2.9007247968372503, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3120501041412354, + "learning_rate": 1e-06, + "loss": 0.8673, + "mean_token_accuracy": 0.7253693342208862, + "num_tokens": 683488989.0, + "step": 26414 + }, + { + "epoch": 2.900834614539864, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.121992826461792, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7240287661552429, + "num_tokens": 683520293.0, + "step": 26415 + }, + { + "epoch": 2.9009444322424773, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7209012508392334, + "learning_rate": 1e-06, + "loss": 0.7344, + "mean_token_accuracy": 0.7555863857269287, + "num_tokens": 683542115.0, + "step": 26416 + }, + { + "epoch": 2.901054249945091, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.20979642868042, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.732749342918396, + "num_tokens": 683572409.0, + "step": 26417 + }, + { + "epoch": 2.901164067647705, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.451542377471924, + "learning_rate": 1e-06, + "loss": 0.8619, + "mean_token_accuracy": 0.734555721282959, + "num_tokens": 683597498.0, + "step": 26418 + }, + { + "epoch": 2.9012738853503186, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0087921619415283, + "learning_rate": 1e-06, + "loss": 0.8578, + "mean_token_accuracy": 0.7441248893737793, + "num_tokens": 683632653.0, + "step": 26419 + }, + { + "epoch": 2.9013837030529324, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.185436248779297, + "learning_rate": 1e-06, + "loss": 0.8446, + "mean_token_accuracy": 0.7315765619277954, + "num_tokens": 683662090.0, + "step": 26420 + }, + { + "epoch": 2.9014935207555457, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.620689868927002, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7352961897850037, + "num_tokens": 683684165.0, + "step": 26421 + }, + { + "epoch": 2.9016033384581594, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.417954683303833, + "learning_rate": 1e-06, + "loss": 0.9112, + "mean_token_accuracy": 0.71728515625, + "num_tokens": 683710140.0, + "step": 26422 + }, + { + "epoch": 2.901713156160773, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3561887741088867, + "learning_rate": 1e-06, + "loss": 0.93, + "mean_token_accuracy": 0.7135907411575317, + "num_tokens": 683738357.0, + "step": 26423 + }, + { + "epoch": 2.9018229738633865, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1765663623809814, + "learning_rate": 1e-06, + "loss": 0.7902, + "mean_token_accuracy": 0.7544560432434082, + "num_tokens": 683765150.0, + "step": 26424 + }, + { + "epoch": 2.9019327915660007, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4149210453033447, + "learning_rate": 1e-06, + "loss": 0.9448, + "mean_token_accuracy": 0.7094954252243042, + "num_tokens": 683791127.0, + "step": 26425 + }, + { + "epoch": 2.902042609268614, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4219377040863037, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7376356720924377, + "num_tokens": 683814724.0, + "step": 26426 + }, + { + "epoch": 2.9021524269712278, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.719440460205078, + "learning_rate": 1e-06, + "loss": 0.75, + "mean_token_accuracy": 0.7652210593223572, + "num_tokens": 683834075.0, + "step": 26427 + }, + { + "epoch": 2.9022622446738415, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0648882389068604, + "learning_rate": 1e-06, + "loss": 0.7954, + "mean_token_accuracy": 0.7443641424179077, + "num_tokens": 683865794.0, + "step": 26428 + }, + { + "epoch": 2.902372062376455, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4223031997680664, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7362763285636902, + "num_tokens": 683891188.0, + "step": 26429 + }, + { + "epoch": 2.9024818800790686, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0631539821624756, + "learning_rate": 1e-06, + "loss": 0.9949, + "mean_token_accuracy": 0.6975685358047485, + "num_tokens": 683926150.0, + "step": 26430 + }, + { + "epoch": 2.9025916977816824, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7267866134643555, + "learning_rate": 1e-06, + "loss": 0.7383, + "mean_token_accuracy": 0.7651057243347168, + "num_tokens": 683946763.0, + "step": 26431 + }, + { + "epoch": 2.902701515484296, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3516106605529785, + "learning_rate": 1e-06, + "loss": 0.8512, + "mean_token_accuracy": 0.7364044189453125, + "num_tokens": 683972539.0, + "step": 26432 + }, + { + "epoch": 2.90281133318691, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.546736001968384, + "learning_rate": 1e-06, + "loss": 0.7909, + "mean_token_accuracy": 0.7454966306686401, + "num_tokens": 683995041.0, + "step": 26433 + }, + { + "epoch": 2.902921150889523, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0452141761779785, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7100124359130859, + "num_tokens": 684028733.0, + "step": 26434 + }, + { + "epoch": 2.903030968592137, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2944343090057373, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7260609865188599, + "num_tokens": 684056832.0, + "step": 26435 + }, + { + "epoch": 2.9031407862947507, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1645264625549316, + "learning_rate": 1e-06, + "loss": 0.7425, + "mean_token_accuracy": 0.7585592269897461, + "num_tokens": 684085015.0, + "step": 26436 + }, + { + "epoch": 2.9032506039973645, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.557743549346924, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7488875985145569, + "num_tokens": 684109006.0, + "step": 26437 + }, + { + "epoch": 2.903360421699978, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5462989807128906, + "learning_rate": 1e-06, + "loss": 0.7636, + "mean_token_accuracy": 0.7485953569412231, + "num_tokens": 684132450.0, + "step": 26438 + }, + { + "epoch": 2.9034702394025915, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0494415760040283, + "learning_rate": 1e-06, + "loss": 0.8663, + "mean_token_accuracy": 0.7307834625244141, + "num_tokens": 684164024.0, + "step": 26439 + }, + { + "epoch": 2.9035800571052053, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1640970706939697, + "learning_rate": 1e-06, + "loss": 0.9247, + "mean_token_accuracy": 0.7202727794647217, + "num_tokens": 684193418.0, + "step": 26440 + }, + { + "epoch": 2.903689874807819, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.539966106414795, + "learning_rate": 1e-06, + "loss": 0.8861, + "mean_token_accuracy": 0.7255886793136597, + "num_tokens": 684219420.0, + "step": 26441 + }, + { + "epoch": 2.903799692510433, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.384758949279785, + "learning_rate": 1e-06, + "loss": 0.6563, + "mean_token_accuracy": 0.7840114831924438, + "num_tokens": 684241687.0, + "step": 26442 + }, + { + "epoch": 2.9039095102130466, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2635750770568848, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7372546792030334, + "num_tokens": 684270672.0, + "step": 26443 + }, + { + "epoch": 2.90401932791566, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4800546169281006, + "learning_rate": 1e-06, + "loss": 0.8964, + "mean_token_accuracy": 0.723186731338501, + "num_tokens": 684296671.0, + "step": 26444 + }, + { + "epoch": 2.9041291456182736, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.380291223526001, + "learning_rate": 1e-06, + "loss": 0.8655, + "mean_token_accuracy": 0.7261029481887817, + "num_tokens": 684322906.0, + "step": 26445 + }, + { + "epoch": 2.9042389633208874, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.218017578125, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7212224006652832, + "num_tokens": 684352381.0, + "step": 26446 + }, + { + "epoch": 2.904348781023501, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4828970432281494, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7368278503417969, + "num_tokens": 684376296.0, + "step": 26447 + }, + { + "epoch": 2.904458598726115, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4602649211883545, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7335577011108398, + "num_tokens": 684400919.0, + "step": 26448 + }, + { + "epoch": 2.904568416428728, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3799142837524414, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7231786847114563, + "num_tokens": 684428314.0, + "step": 26449 + }, + { + "epoch": 2.904678234131342, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.509438991546631, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7168611288070679, + "num_tokens": 684452540.0, + "step": 26450 + }, + { + "epoch": 2.9047880518339557, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4693586826324463, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7220432758331299, + "num_tokens": 684477807.0, + "step": 26451 + }, + { + "epoch": 2.904897869536569, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.8384554386138916, + "learning_rate": 1e-06, + "loss": 0.7753, + "mean_token_accuracy": 0.7580233812332153, + "num_tokens": 684496994.0, + "step": 26452 + }, + { + "epoch": 2.905007687239183, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.082181215286255, + "learning_rate": 1e-06, + "loss": 0.8919, + "mean_token_accuracy": 0.7181525230407715, + "num_tokens": 684529066.0, + "step": 26453 + }, + { + "epoch": 2.9051175049417965, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.50877046585083, + "learning_rate": 1e-06, + "loss": 0.7677, + "mean_token_accuracy": 0.7504363059997559, + "num_tokens": 684551104.0, + "step": 26454 + }, + { + "epoch": 2.9052273226444103, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.35064959526062, + "learning_rate": 1e-06, + "loss": 0.8464, + "mean_token_accuracy": 0.7328853607177734, + "num_tokens": 684577229.0, + "step": 26455 + }, + { + "epoch": 2.905337140347024, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2155845165252686, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7364379167556763, + "num_tokens": 684605572.0, + "step": 26456 + }, + { + "epoch": 2.9054469580496374, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1420843601226807, + "learning_rate": 1e-06, + "loss": 0.8533, + "mean_token_accuracy": 0.7358787059783936, + "num_tokens": 684636023.0, + "step": 26457 + }, + { + "epoch": 2.905556775752251, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4211525917053223, + "learning_rate": 1e-06, + "loss": 0.911, + "mean_token_accuracy": 0.7230349779129028, + "num_tokens": 684663762.0, + "step": 26458 + }, + { + "epoch": 2.905666593454865, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5797181129455566, + "learning_rate": 1e-06, + "loss": 0.7882, + "mean_token_accuracy": 0.747194766998291, + "num_tokens": 684685399.0, + "step": 26459 + }, + { + "epoch": 2.9057764111574786, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.530979871749878, + "learning_rate": 1e-06, + "loss": 0.7396, + "mean_token_accuracy": 0.7566127181053162, + "num_tokens": 684706935.0, + "step": 26460 + }, + { + "epoch": 2.9058862288600924, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2461302280426025, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.736677348613739, + "num_tokens": 684735676.0, + "step": 26461 + }, + { + "epoch": 2.9059960465627057, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.147367477416992, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7270631790161133, + "num_tokens": 684767002.0, + "step": 26462 + }, + { + "epoch": 2.9061058642653195, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.329821825027466, + "learning_rate": 1e-06, + "loss": 0.8878, + "mean_token_accuracy": 0.7190658450126648, + "num_tokens": 684794623.0, + "step": 26463 + }, + { + "epoch": 2.9062156819679332, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.39208722114563, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7309427857398987, + "num_tokens": 684820573.0, + "step": 26464 + }, + { + "epoch": 2.906325499670547, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1567723751068115, + "learning_rate": 1e-06, + "loss": 0.8465, + "mean_token_accuracy": 0.7288529276847839, + "num_tokens": 684848905.0, + "step": 26465 + }, + { + "epoch": 2.9064353173731607, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.472365140914917, + "learning_rate": 1e-06, + "loss": 0.7434, + "mean_token_accuracy": 0.7592384219169617, + "num_tokens": 684872597.0, + "step": 26466 + }, + { + "epoch": 2.906545135075774, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.439650058746338, + "learning_rate": 1e-06, + "loss": 0.9229, + "mean_token_accuracy": 0.7051725387573242, + "num_tokens": 684899741.0, + "step": 26467 + }, + { + "epoch": 2.906654952778388, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5352296829223633, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.745097279548645, + "num_tokens": 684922939.0, + "step": 26468 + }, + { + "epoch": 2.9067647704810016, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.328010082244873, + "learning_rate": 1e-06, + "loss": 0.8459, + "mean_token_accuracy": 0.73018479347229, + "num_tokens": 684948105.0, + "step": 26469 + }, + { + "epoch": 2.9068745881836153, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0890026092529297, + "learning_rate": 1e-06, + "loss": 0.918, + "mean_token_accuracy": 0.7177333831787109, + "num_tokens": 684979234.0, + "step": 26470 + }, + { + "epoch": 2.906984405886229, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5926311016082764, + "learning_rate": 1e-06, + "loss": 0.8093, + "mean_token_accuracy": 0.744534432888031, + "num_tokens": 685001535.0, + "step": 26471 + }, + { + "epoch": 2.9070942235888424, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5437800884246826, + "learning_rate": 1e-06, + "loss": 0.8645, + "mean_token_accuracy": 0.7335619926452637, + "num_tokens": 685024754.0, + "step": 26472 + }, + { + "epoch": 2.907204041291456, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.221266746520996, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.734720766544342, + "num_tokens": 685054084.0, + "step": 26473 + }, + { + "epoch": 2.90731385899407, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.169072389602661, + "learning_rate": 1e-06, + "loss": 0.9052, + "mean_token_accuracy": 0.7212240099906921, + "num_tokens": 685084931.0, + "step": 26474 + }, + { + "epoch": 2.9074236766966837, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3039824962615967, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7216497659683228, + "num_tokens": 685112701.0, + "step": 26475 + }, + { + "epoch": 2.9075334943992974, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.318394899368286, + "learning_rate": 1e-06, + "loss": 0.8769, + "mean_token_accuracy": 0.727320671081543, + "num_tokens": 685139446.0, + "step": 26476 + }, + { + "epoch": 2.9076433121019107, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.76265549659729, + "learning_rate": 1e-06, + "loss": 0.7942, + "mean_token_accuracy": 0.7519198060035706, + "num_tokens": 685160733.0, + "step": 26477 + }, + { + "epoch": 2.9077531298045245, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.162872076034546, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7239187955856323, + "num_tokens": 685189134.0, + "step": 26478 + }, + { + "epoch": 2.9078629475071383, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.578662157058716, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7561638355255127, + "num_tokens": 685211228.0, + "step": 26479 + }, + { + "epoch": 2.9079727652097516, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.417956590652466, + "learning_rate": 1e-06, + "loss": 0.8667, + "mean_token_accuracy": 0.7274112701416016, + "num_tokens": 685236995.0, + "step": 26480 + }, + { + "epoch": 2.9080825829123653, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4257137775421143, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7253557443618774, + "num_tokens": 685262502.0, + "step": 26481 + }, + { + "epoch": 2.908192400614979, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5111074447631836, + "learning_rate": 1e-06, + "loss": 0.8073, + "mean_token_accuracy": 0.7415900230407715, + "num_tokens": 685283923.0, + "step": 26482 + }, + { + "epoch": 2.908302218317593, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2201600074768066, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7314929962158203, + "num_tokens": 685311142.0, + "step": 26483 + }, + { + "epoch": 2.9084120360202066, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4861271381378174, + "learning_rate": 1e-06, + "loss": 0.7804, + "mean_token_accuracy": 0.7449180483818054, + "num_tokens": 685334840.0, + "step": 26484 + }, + { + "epoch": 2.90852185372282, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3446733951568604, + "learning_rate": 1e-06, + "loss": 0.8684, + "mean_token_accuracy": 0.7293376326560974, + "num_tokens": 685362203.0, + "step": 26485 + }, + { + "epoch": 2.9086316714254337, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4891512393951416, + "learning_rate": 1e-06, + "loss": 0.8487, + "mean_token_accuracy": 0.7262587547302246, + "num_tokens": 685385678.0, + "step": 26486 + }, + { + "epoch": 2.9087414891280474, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1606924533843994, + "learning_rate": 1e-06, + "loss": 0.9344, + "mean_token_accuracy": 0.7167654633522034, + "num_tokens": 685416591.0, + "step": 26487 + }, + { + "epoch": 2.908851306830661, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.532102346420288, + "learning_rate": 1e-06, + "loss": 0.9134, + "mean_token_accuracy": 0.7306113839149475, + "num_tokens": 685439488.0, + "step": 26488 + }, + { + "epoch": 2.908961124533275, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.224414348602295, + "learning_rate": 1e-06, + "loss": 0.8503, + "mean_token_accuracy": 0.7207140922546387, + "num_tokens": 685467009.0, + "step": 26489 + }, + { + "epoch": 2.9090709422358882, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.258423089981079, + "learning_rate": 1e-06, + "loss": 0.8081, + "mean_token_accuracy": 0.7377783060073853, + "num_tokens": 685492831.0, + "step": 26490 + }, + { + "epoch": 2.909180759938502, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3624801635742188, + "learning_rate": 1e-06, + "loss": 0.9191, + "mean_token_accuracy": 0.722888171672821, + "num_tokens": 685520269.0, + "step": 26491 + }, + { + "epoch": 2.9092905776411158, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.706040143966675, + "learning_rate": 1e-06, + "loss": 0.8479, + "mean_token_accuracy": 0.742249608039856, + "num_tokens": 685539928.0, + "step": 26492 + }, + { + "epoch": 2.9094003953437295, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6712255477905273, + "learning_rate": 1e-06, + "loss": 0.9238, + "mean_token_accuracy": 0.7144923210144043, + "num_tokens": 685564126.0, + "step": 26493 + }, + { + "epoch": 2.9095102130463433, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2709460258483887, + "learning_rate": 1e-06, + "loss": 0.8447, + "mean_token_accuracy": 0.732286810874939, + "num_tokens": 685590910.0, + "step": 26494 + }, + { + "epoch": 2.9096200307489566, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3617849349975586, + "learning_rate": 1e-06, + "loss": 0.8869, + "mean_token_accuracy": 0.7217262983322144, + "num_tokens": 685620084.0, + "step": 26495 + }, + { + "epoch": 2.9097298484515703, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.240659713745117, + "learning_rate": 1e-06, + "loss": 0.7839, + "mean_token_accuracy": 0.7486791014671326, + "num_tokens": 685647562.0, + "step": 26496 + }, + { + "epoch": 2.909839666154184, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1817824840545654, + "learning_rate": 1e-06, + "loss": 0.8788, + "mean_token_accuracy": 0.72418212890625, + "num_tokens": 685675349.0, + "step": 26497 + }, + { + "epoch": 2.909949483856798, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4410505294799805, + "learning_rate": 1e-06, + "loss": 0.7955, + "mean_token_accuracy": 0.7466563582420349, + "num_tokens": 685698449.0, + "step": 26498 + }, + { + "epoch": 2.9100593015594116, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.283282995223999, + "learning_rate": 1e-06, + "loss": 0.9323, + "mean_token_accuracy": 0.7055106163024902, + "num_tokens": 685727170.0, + "step": 26499 + }, + { + "epoch": 2.910169119262025, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1435041427612305, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7176965475082397, + "num_tokens": 685759830.0, + "step": 26500 + }, + { + "epoch": 2.9102789369646387, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.134381055831909, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7335811853408813, + "num_tokens": 685790878.0, + "step": 26501 + }, + { + "epoch": 2.9103887546672524, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.108074903488159, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7378919124603271, + "num_tokens": 685822032.0, + "step": 26502 + }, + { + "epoch": 2.9104985723698658, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.479931592941284, + "learning_rate": 1e-06, + "loss": 0.8883, + "mean_token_accuracy": 0.7253866791725159, + "num_tokens": 685849320.0, + "step": 26503 + }, + { + "epoch": 2.91060839007248, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.446850061416626, + "learning_rate": 1e-06, + "loss": 0.8875, + "mean_token_accuracy": 0.718161404132843, + "num_tokens": 685873771.0, + "step": 26504 + }, + { + "epoch": 2.9107182077750933, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1730637550354004, + "learning_rate": 1e-06, + "loss": 0.9294, + "mean_token_accuracy": 0.7138770818710327, + "num_tokens": 685904227.0, + "step": 26505 + }, + { + "epoch": 2.910828025477707, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.312842607498169, + "learning_rate": 1e-06, + "loss": 0.8208, + "mean_token_accuracy": 0.7378106713294983, + "num_tokens": 685931888.0, + "step": 26506 + }, + { + "epoch": 2.910937843180321, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.466946601867676, + "learning_rate": 1e-06, + "loss": 0.835, + "mean_token_accuracy": 0.7373892664909363, + "num_tokens": 685955824.0, + "step": 26507 + }, + { + "epoch": 2.911047660882934, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3671510219573975, + "learning_rate": 1e-06, + "loss": 0.8725, + "mean_token_accuracy": 0.7273569703102112, + "num_tokens": 685980405.0, + "step": 26508 + }, + { + "epoch": 2.911157478585548, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.731907844543457, + "learning_rate": 1e-06, + "loss": 0.8181, + "mean_token_accuracy": 0.7385092377662659, + "num_tokens": 686000283.0, + "step": 26509 + }, + { + "epoch": 2.9112672962881616, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.416440963745117, + "learning_rate": 1e-06, + "loss": 0.7992, + "mean_token_accuracy": 0.7449907660484314, + "num_tokens": 686024486.0, + "step": 26510 + }, + { + "epoch": 2.9113771139907754, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.318776845932007, + "learning_rate": 1e-06, + "loss": 0.7447, + "mean_token_accuracy": 0.7575069665908813, + "num_tokens": 686049863.0, + "step": 26511 + }, + { + "epoch": 2.911486931693389, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3804991245269775, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7391730546951294, + "num_tokens": 686074858.0, + "step": 26512 + }, + { + "epoch": 2.9115967493960024, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4355313777923584, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7268872261047363, + "num_tokens": 686098909.0, + "step": 26513 + }, + { + "epoch": 2.911706567098616, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3949921131134033, + "learning_rate": 1e-06, + "loss": 0.8686, + "mean_token_accuracy": 0.7233296632766724, + "num_tokens": 686122999.0, + "step": 26514 + }, + { + "epoch": 2.91181638480123, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6201279163360596, + "learning_rate": 1e-06, + "loss": 0.8257, + "mean_token_accuracy": 0.7410430908203125, + "num_tokens": 686146006.0, + "step": 26515 + }, + { + "epoch": 2.9119262025038437, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0625147819519043, + "learning_rate": 1e-06, + "loss": 0.9016, + "mean_token_accuracy": 0.7204136848449707, + "num_tokens": 686179381.0, + "step": 26516 + }, + { + "epoch": 2.9120360202064575, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.275684118270874, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7265757918357849, + "num_tokens": 686207753.0, + "step": 26517 + }, + { + "epoch": 2.912145837909071, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4249191284179688, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7221069931983948, + "num_tokens": 686233578.0, + "step": 26518 + }, + { + "epoch": 2.9122556556116845, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4173314571380615, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7420855164527893, + "num_tokens": 686258649.0, + "step": 26519 + }, + { + "epoch": 2.9123654733142983, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.312976837158203, + "learning_rate": 1e-06, + "loss": 0.7046, + "mean_token_accuracy": 0.7748651504516602, + "num_tokens": 686282824.0, + "step": 26520 + }, + { + "epoch": 2.912475291016912, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 1.9983198642730713, + "learning_rate": 1e-06, + "loss": 0.938, + "mean_token_accuracy": 0.7218714952468872, + "num_tokens": 686315799.0, + "step": 26521 + }, + { + "epoch": 2.912585108719526, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.492645502090454, + "learning_rate": 1e-06, + "loss": 0.795, + "mean_token_accuracy": 0.7453427314758301, + "num_tokens": 686339348.0, + "step": 26522 + }, + { + "epoch": 2.912694926422139, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.714695453643799, + "learning_rate": 1e-06, + "loss": 0.7584, + "mean_token_accuracy": 0.7569378018379211, + "num_tokens": 686360029.0, + "step": 26523 + }, + { + "epoch": 2.912804744124753, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3671157360076904, + "learning_rate": 1e-06, + "loss": 0.9237, + "mean_token_accuracy": 0.7109020948410034, + "num_tokens": 686387739.0, + "step": 26524 + }, + { + "epoch": 2.9129145618273666, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.528383255004883, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7368379831314087, + "num_tokens": 686410635.0, + "step": 26525 + }, + { + "epoch": 2.9130243795299804, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5440919399261475, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7437212467193604, + "num_tokens": 686436426.0, + "step": 26526 + }, + { + "epoch": 2.913134197232594, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.076148271560669, + "learning_rate": 1e-06, + "loss": 0.884, + "mean_token_accuracy": 0.7181103229522705, + "num_tokens": 686466815.0, + "step": 26527 + }, + { + "epoch": 2.9132440149352075, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3719124794006348, + "learning_rate": 1e-06, + "loss": 0.9075, + "mean_token_accuracy": 0.7174451947212219, + "num_tokens": 686491711.0, + "step": 26528 + }, + { + "epoch": 2.913353832637821, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.427600383758545, + "learning_rate": 1e-06, + "loss": 0.8877, + "mean_token_accuracy": 0.7179253697395325, + "num_tokens": 686515784.0, + "step": 26529 + }, + { + "epoch": 2.913463650340435, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6472034454345703, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7254780530929565, + "num_tokens": 686538227.0, + "step": 26530 + }, + { + "epoch": 2.9135734680430483, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2657785415649414, + "learning_rate": 1e-06, + "loss": 0.9317, + "mean_token_accuracy": 0.7096431255340576, + "num_tokens": 686568077.0, + "step": 26531 + }, + { + "epoch": 2.913683285745662, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.22221040725708, + "learning_rate": 1e-06, + "loss": 0.8744, + "mean_token_accuracy": 0.730546772480011, + "num_tokens": 686597018.0, + "step": 26532 + }, + { + "epoch": 2.913793103448276, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6945550441741943, + "learning_rate": 1e-06, + "loss": 0.817, + "mean_token_accuracy": 0.7376396059989929, + "num_tokens": 686618219.0, + "step": 26533 + }, + { + "epoch": 2.9139029211508896, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4046552181243896, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.731399416923523, + "num_tokens": 686642533.0, + "step": 26534 + }, + { + "epoch": 2.9140127388535033, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5103096961975098, + "learning_rate": 1e-06, + "loss": 0.9289, + "mean_token_accuracy": 0.7151103019714355, + "num_tokens": 686665114.0, + "step": 26535 + }, + { + "epoch": 2.9141225565561166, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.282357692718506, + "learning_rate": 1e-06, + "loss": 0.927, + "mean_token_accuracy": 0.7088980078697205, + "num_tokens": 686695160.0, + "step": 26536 + }, + { + "epoch": 2.9142323742587304, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.608146905899048, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7169697284698486, + "num_tokens": 686717975.0, + "step": 26537 + }, + { + "epoch": 2.914342191961344, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.378051996231079, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7323540449142456, + "num_tokens": 686744794.0, + "step": 26538 + }, + { + "epoch": 2.914452009663958, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.365858554840088, + "learning_rate": 1e-06, + "loss": 0.9146, + "mean_token_accuracy": 0.7118949890136719, + "num_tokens": 686770704.0, + "step": 26539 + }, + { + "epoch": 2.9145618273665717, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.084599256515503, + "learning_rate": 1e-06, + "loss": 0.8194, + "mean_token_accuracy": 0.7395597696304321, + "num_tokens": 686800460.0, + "step": 26540 + }, + { + "epoch": 2.914671645069185, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0154943466186523, + "learning_rate": 1e-06, + "loss": 0.996, + "mean_token_accuracy": 0.690686821937561, + "num_tokens": 686835550.0, + "step": 26541 + }, + { + "epoch": 2.9147814627717987, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3081345558166504, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.736384928226471, + "num_tokens": 686860874.0, + "step": 26542 + }, + { + "epoch": 2.9148912804744125, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.275526762008667, + "learning_rate": 1e-06, + "loss": 0.8902, + "mean_token_accuracy": 0.720845103263855, + "num_tokens": 686889331.0, + "step": 26543 + }, + { + "epoch": 2.9150010981770262, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.240651845932007, + "learning_rate": 1e-06, + "loss": 0.8697, + "mean_token_accuracy": 0.7247065305709839, + "num_tokens": 686919202.0, + "step": 26544 + }, + { + "epoch": 2.91511091587964, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5071940422058105, + "learning_rate": 1e-06, + "loss": 0.8528, + "mean_token_accuracy": 0.7247843742370605, + "num_tokens": 686942904.0, + "step": 26545 + }, + { + "epoch": 2.9152207335822533, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1710751056671143, + "learning_rate": 1e-06, + "loss": 0.8634, + "mean_token_accuracy": 0.7318621873855591, + "num_tokens": 686973126.0, + "step": 26546 + }, + { + "epoch": 2.915330551284867, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.471144437789917, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7370721101760864, + "num_tokens": 686998311.0, + "step": 26547 + }, + { + "epoch": 2.915440368987481, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2805795669555664, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7252597808837891, + "num_tokens": 687025114.0, + "step": 26548 + }, + { + "epoch": 2.9155501866900946, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.325247049331665, + "learning_rate": 1e-06, + "loss": 0.6912, + "mean_token_accuracy": 0.762901782989502, + "num_tokens": 687050089.0, + "step": 26549 + }, + { + "epoch": 2.9156600043927083, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.191850423812866, + "learning_rate": 1e-06, + "loss": 0.9014, + "mean_token_accuracy": 0.7161291837692261, + "num_tokens": 687079672.0, + "step": 26550 + }, + { + "epoch": 2.9157698220953217, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.424633502960205, + "learning_rate": 1e-06, + "loss": 0.828, + "mean_token_accuracy": 0.7397170066833496, + "num_tokens": 687104830.0, + "step": 26551 + }, + { + "epoch": 2.9158796397979354, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.8270483016967773, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7529867887496948, + "num_tokens": 687124592.0, + "step": 26552 + }, + { + "epoch": 2.915989457500549, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.409926414489746, + "learning_rate": 1e-06, + "loss": 0.7923, + "mean_token_accuracy": 0.7498705983161926, + "num_tokens": 687149660.0, + "step": 26553 + }, + { + "epoch": 2.9160992752031625, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5516891479492188, + "learning_rate": 1e-06, + "loss": 0.9119, + "mean_token_accuracy": 0.7131832242012024, + "num_tokens": 687173940.0, + "step": 26554 + }, + { + "epoch": 2.9162090929057767, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1460025310516357, + "learning_rate": 1e-06, + "loss": 0.8705, + "mean_token_accuracy": 0.7231392860412598, + "num_tokens": 687206727.0, + "step": 26555 + }, + { + "epoch": 2.91631891060839, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.581148624420166, + "learning_rate": 1e-06, + "loss": 0.8218, + "mean_token_accuracy": 0.7420815229415894, + "num_tokens": 687227811.0, + "step": 26556 + }, + { + "epoch": 2.9164287283110037, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1948394775390625, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7310563921928406, + "num_tokens": 687256730.0, + "step": 26557 + }, + { + "epoch": 2.9165385460136175, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2831544876098633, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7482722997665405, + "num_tokens": 687281628.0, + "step": 26558 + }, + { + "epoch": 2.916648363716231, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1153087615966797, + "learning_rate": 1e-06, + "loss": 0.8449, + "mean_token_accuracy": 0.7375562191009521, + "num_tokens": 687310682.0, + "step": 26559 + }, + { + "epoch": 2.9167581814188446, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1562867164611816, + "learning_rate": 1e-06, + "loss": 0.9105, + "mean_token_accuracy": 0.7135396003723145, + "num_tokens": 687340045.0, + "step": 26560 + }, + { + "epoch": 2.9168679991214583, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.295681953430176, + "learning_rate": 1e-06, + "loss": 0.8342, + "mean_token_accuracy": 0.7318958640098572, + "num_tokens": 687365898.0, + "step": 26561 + }, + { + "epoch": 2.916977816824072, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.250030517578125, + "learning_rate": 1e-06, + "loss": 0.797, + "mean_token_accuracy": 0.7559788227081299, + "num_tokens": 687390425.0, + "step": 26562 + }, + { + "epoch": 2.917087634526686, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.678950548171997, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.738960862159729, + "num_tokens": 687412248.0, + "step": 26563 + }, + { + "epoch": 2.917197452229299, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.399346113204956, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7291063666343689, + "num_tokens": 687438042.0, + "step": 26564 + }, + { + "epoch": 2.917307269931913, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3300888538360596, + "learning_rate": 1e-06, + "loss": 0.8853, + "mean_token_accuracy": 0.7175999283790588, + "num_tokens": 687464699.0, + "step": 26565 + }, + { + "epoch": 2.9174170876345267, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.276815414428711, + "learning_rate": 1e-06, + "loss": 0.8253, + "mean_token_accuracy": 0.7394784092903137, + "num_tokens": 687490154.0, + "step": 26566 + }, + { + "epoch": 2.9175269053371404, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.547898054122925, + "learning_rate": 1e-06, + "loss": 0.739, + "mean_token_accuracy": 0.7535943388938904, + "num_tokens": 687510151.0, + "step": 26567 + }, + { + "epoch": 2.917636723039754, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4171512126922607, + "learning_rate": 1e-06, + "loss": 0.8982, + "mean_token_accuracy": 0.7240257859230042, + "num_tokens": 687534651.0, + "step": 26568 + }, + { + "epoch": 2.9177465407423675, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.511967658996582, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7184425592422485, + "num_tokens": 687557368.0, + "step": 26569 + }, + { + "epoch": 2.9178563584449813, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.400273323059082, + "learning_rate": 1e-06, + "loss": 0.8326, + "mean_token_accuracy": 0.7375146746635437, + "num_tokens": 687581983.0, + "step": 26570 + }, + { + "epoch": 2.917966176147595, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.216829299926758, + "learning_rate": 1e-06, + "loss": 0.8814, + "mean_token_accuracy": 0.7183226346969604, + "num_tokens": 687610019.0, + "step": 26571 + }, + { + "epoch": 2.9180759938502088, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0724129676818848, + "learning_rate": 1e-06, + "loss": 0.9244, + "mean_token_accuracy": 0.7081032991409302, + "num_tokens": 687644128.0, + "step": 26572 + }, + { + "epoch": 2.9181858115528225, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.71182918548584, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.714397132396698, + "num_tokens": 687664735.0, + "step": 26573 + }, + { + "epoch": 2.918295629255436, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3787424564361572, + "learning_rate": 1e-06, + "loss": 0.7218, + "mean_token_accuracy": 0.762801468372345, + "num_tokens": 687687794.0, + "step": 26574 + }, + { + "epoch": 2.9184054469580496, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4791641235351562, + "learning_rate": 1e-06, + "loss": 0.88, + "mean_token_accuracy": 0.7224346399307251, + "num_tokens": 687711311.0, + "step": 26575 + }, + { + "epoch": 2.9185152646606634, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6373536586761475, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7408961653709412, + "num_tokens": 687733156.0, + "step": 26576 + }, + { + "epoch": 2.918625082363277, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5813138484954834, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7393711805343628, + "num_tokens": 687755265.0, + "step": 26577 + }, + { + "epoch": 2.918734900065891, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3013830184936523, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7380312085151672, + "num_tokens": 687780821.0, + "step": 26578 + }, + { + "epoch": 2.918844717768504, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.544198751449585, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7507638931274414, + "num_tokens": 687803688.0, + "step": 26579 + }, + { + "epoch": 2.918954535471118, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.360133171081543, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7417252063751221, + "num_tokens": 687829995.0, + "step": 26580 + }, + { + "epoch": 2.9190643531737317, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.211923122406006, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7256685495376587, + "num_tokens": 687859682.0, + "step": 26581 + }, + { + "epoch": 2.919174170876345, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0459718704223633, + "learning_rate": 1e-06, + "loss": 0.9278, + "mean_token_accuracy": 0.7135729789733887, + "num_tokens": 687893976.0, + "step": 26582 + }, + { + "epoch": 2.9192839885789588, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.209407091140747, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7505201101303101, + "num_tokens": 687922005.0, + "step": 26583 + }, + { + "epoch": 2.9193938062815725, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5362377166748047, + "learning_rate": 1e-06, + "loss": 0.8394, + "mean_token_accuracy": 0.7353090643882751, + "num_tokens": 687946333.0, + "step": 26584 + }, + { + "epoch": 2.9195036239841863, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.61122465133667, + "learning_rate": 1e-06, + "loss": 0.7618, + "mean_token_accuracy": 0.7588125467300415, + "num_tokens": 687966792.0, + "step": 26585 + }, + { + "epoch": 2.9196134416868, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.304497718811035, + "learning_rate": 1e-06, + "loss": 0.8611, + "mean_token_accuracy": 0.7349388599395752, + "num_tokens": 687993284.0, + "step": 26586 + }, + { + "epoch": 2.9197232593894134, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.579197883605957, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7370500564575195, + "num_tokens": 688016354.0, + "step": 26587 + }, + { + "epoch": 2.919833077092027, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.285356044769287, + "learning_rate": 1e-06, + "loss": 0.7981, + "mean_token_accuracy": 0.7480485439300537, + "num_tokens": 688042962.0, + "step": 26588 + }, + { + "epoch": 2.919942894794641, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6123132705688477, + "learning_rate": 1e-06, + "loss": 0.8926, + "mean_token_accuracy": 0.7244691848754883, + "num_tokens": 688065254.0, + "step": 26589 + }, + { + "epoch": 2.9200527124972546, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2042737007141113, + "learning_rate": 1e-06, + "loss": 0.9537, + "mean_token_accuracy": 0.7019002437591553, + "num_tokens": 688094167.0, + "step": 26590 + }, + { + "epoch": 2.9201625301998684, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.30965256690979, + "learning_rate": 1e-06, + "loss": 0.8434, + "mean_token_accuracy": 0.7365658283233643, + "num_tokens": 688121059.0, + "step": 26591 + }, + { + "epoch": 2.9202723479024817, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3413493633270264, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.729374885559082, + "num_tokens": 688148228.0, + "step": 26592 + }, + { + "epoch": 2.9203821656050954, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7330596446990967, + "learning_rate": 1e-06, + "loss": 0.8024, + "mean_token_accuracy": 0.7415981292724609, + "num_tokens": 688168188.0, + "step": 26593 + }, + { + "epoch": 2.920491983307709, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3465466499328613, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7289914488792419, + "num_tokens": 688193376.0, + "step": 26594 + }, + { + "epoch": 2.920601801010323, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1632299423217773, + "learning_rate": 1e-06, + "loss": 0.8833, + "mean_token_accuracy": 0.723542332649231, + "num_tokens": 688223449.0, + "step": 26595 + }, + { + "epoch": 2.9207116187129367, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3607728481292725, + "learning_rate": 1e-06, + "loss": 0.8368, + "mean_token_accuracy": 0.7365289926528931, + "num_tokens": 688248470.0, + "step": 26596 + }, + { + "epoch": 2.92082143641555, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2369673252105713, + "learning_rate": 1e-06, + "loss": 0.7455, + "mean_token_accuracy": 0.7572976350784302, + "num_tokens": 688275671.0, + "step": 26597 + }, + { + "epoch": 2.920931254118164, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.286890983581543, + "learning_rate": 1e-06, + "loss": 0.8365, + "mean_token_accuracy": 0.7325528264045715, + "num_tokens": 688305394.0, + "step": 26598 + }, + { + "epoch": 2.9210410718207775, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6096041202545166, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7299087643623352, + "num_tokens": 688327261.0, + "step": 26599 + }, + { + "epoch": 2.9211508895233913, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.269899845123291, + "learning_rate": 1e-06, + "loss": 0.8004, + "mean_token_accuracy": 0.7511530518531799, + "num_tokens": 688354440.0, + "step": 26600 + }, + { + "epoch": 2.921260707226005, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.646299123764038, + "learning_rate": 1e-06, + "loss": 0.8328, + "mean_token_accuracy": 0.7341074347496033, + "num_tokens": 688376412.0, + "step": 26601 + }, + { + "epoch": 2.9213705249286184, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.201359748840332, + "learning_rate": 1e-06, + "loss": 0.9567, + "mean_token_accuracy": 0.7051005363464355, + "num_tokens": 688408083.0, + "step": 26602 + }, + { + "epoch": 2.921480342631232, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.470283269882202, + "learning_rate": 1e-06, + "loss": 0.7574, + "mean_token_accuracy": 0.7543582916259766, + "num_tokens": 688430832.0, + "step": 26603 + }, + { + "epoch": 2.921590160333846, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4220640659332275, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7305582761764526, + "num_tokens": 688455208.0, + "step": 26604 + }, + { + "epoch": 2.921699978036459, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3514230251312256, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7391999363899231, + "num_tokens": 688480867.0, + "step": 26605 + }, + { + "epoch": 2.9218097957390734, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3522748947143555, + "learning_rate": 1e-06, + "loss": 0.9042, + "mean_token_accuracy": 0.7187883853912354, + "num_tokens": 688508797.0, + "step": 26606 + }, + { + "epoch": 2.9219196134416867, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2491912841796875, + "learning_rate": 1e-06, + "loss": 0.9474, + "mean_token_accuracy": 0.7081607580184937, + "num_tokens": 688539193.0, + "step": 26607 + }, + { + "epoch": 2.9220294311443005, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1319782733917236, + "learning_rate": 1e-06, + "loss": 0.8873, + "mean_token_accuracy": 0.7268247008323669, + "num_tokens": 688568712.0, + "step": 26608 + }, + { + "epoch": 2.9221392488469142, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.568845272064209, + "learning_rate": 1e-06, + "loss": 0.8539, + "mean_token_accuracy": 0.7271671295166016, + "num_tokens": 688592866.0, + "step": 26609 + }, + { + "epoch": 2.9222490665495275, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5070362091064453, + "learning_rate": 1e-06, + "loss": 0.8282, + "mean_token_accuracy": 0.7400049567222595, + "num_tokens": 688617035.0, + "step": 26610 + }, + { + "epoch": 2.9223588842521413, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.294513463973999, + "learning_rate": 1e-06, + "loss": 0.8152, + "mean_token_accuracy": 0.7421168684959412, + "num_tokens": 688642534.0, + "step": 26611 + }, + { + "epoch": 2.922468701954755, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.470659017562866, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.730826735496521, + "num_tokens": 688666128.0, + "step": 26612 + }, + { + "epoch": 2.922578519657369, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5299999713897705, + "learning_rate": 1e-06, + "loss": 0.7339, + "mean_token_accuracy": 0.7617426514625549, + "num_tokens": 688688505.0, + "step": 26613 + }, + { + "epoch": 2.9226883373599826, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0948679447174072, + "learning_rate": 1e-06, + "loss": 0.8773, + "mean_token_accuracy": 0.7180636525154114, + "num_tokens": 688719595.0, + "step": 26614 + }, + { + "epoch": 2.922798155062596, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.633598566055298, + "learning_rate": 1e-06, + "loss": 0.9378, + "mean_token_accuracy": 0.7154983878135681, + "num_tokens": 688741262.0, + "step": 26615 + }, + { + "epoch": 2.9229079727652096, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5210132598876953, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.7344312071800232, + "num_tokens": 688764920.0, + "step": 26616 + }, + { + "epoch": 2.9230177904678234, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3176498413085938, + "learning_rate": 1e-06, + "loss": 0.965, + "mean_token_accuracy": 0.7080708146095276, + "num_tokens": 688796330.0, + "step": 26617 + }, + { + "epoch": 2.923127608170437, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5313334465026855, + "learning_rate": 1e-06, + "loss": 0.7646, + "mean_token_accuracy": 0.7563958168029785, + "num_tokens": 688817560.0, + "step": 26618 + }, + { + "epoch": 2.923237425873051, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4362785816192627, + "learning_rate": 1e-06, + "loss": 0.7945, + "mean_token_accuracy": 0.740673303604126, + "num_tokens": 688840841.0, + "step": 26619 + }, + { + "epoch": 2.9233472435756642, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.576378345489502, + "learning_rate": 1e-06, + "loss": 0.8369, + "mean_token_accuracy": 0.7351668477058411, + "num_tokens": 688863396.0, + "step": 26620 + }, + { + "epoch": 2.923457061278278, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6539669036865234, + "learning_rate": 1e-06, + "loss": 0.79, + "mean_token_accuracy": 0.7490423917770386, + "num_tokens": 688884373.0, + "step": 26621 + }, + { + "epoch": 2.9235668789808917, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5571205615997314, + "learning_rate": 1e-06, + "loss": 0.7554, + "mean_token_accuracy": 0.7617158889770508, + "num_tokens": 688906250.0, + "step": 26622 + }, + { + "epoch": 2.9236766966835055, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5761725902557373, + "learning_rate": 1e-06, + "loss": 0.8136, + "mean_token_accuracy": 0.7397463321685791, + "num_tokens": 688927619.0, + "step": 26623 + }, + { + "epoch": 2.9237865143861193, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.183651924133301, + "learning_rate": 1e-06, + "loss": 0.8041, + "mean_token_accuracy": 0.7530397176742554, + "num_tokens": 688955302.0, + "step": 26624 + }, + { + "epoch": 2.9238963320887326, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.221360683441162, + "learning_rate": 1e-06, + "loss": 0.9261, + "mean_token_accuracy": 0.7170149683952332, + "num_tokens": 688985863.0, + "step": 26625 + }, + { + "epoch": 2.9240061497913463, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3278756141662598, + "learning_rate": 1e-06, + "loss": 0.7868, + "mean_token_accuracy": 0.7474279403686523, + "num_tokens": 689013871.0, + "step": 26626 + }, + { + "epoch": 2.92411596749396, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6048026084899902, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.732162356376648, + "num_tokens": 689036331.0, + "step": 26627 + }, + { + "epoch": 2.924225785196574, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3419861793518066, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7300205230712891, + "num_tokens": 689064330.0, + "step": 26628 + }, + { + "epoch": 2.9243356028991876, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.492640256881714, + "learning_rate": 1e-06, + "loss": 0.8879, + "mean_token_accuracy": 0.71990966796875, + "num_tokens": 689089193.0, + "step": 26629 + }, + { + "epoch": 2.924445420601801, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2357211112976074, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7271588444709778, + "num_tokens": 689117563.0, + "step": 26630 + }, + { + "epoch": 2.9245552383044147, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.472085475921631, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7386154532432556, + "num_tokens": 689141597.0, + "step": 26631 + }, + { + "epoch": 2.9246650560070284, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3128902912139893, + "learning_rate": 1e-06, + "loss": 0.7803, + "mean_token_accuracy": 0.7498270273208618, + "num_tokens": 689167430.0, + "step": 26632 + }, + { + "epoch": 2.9247748737096417, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.102715253829956, + "learning_rate": 1e-06, + "loss": 0.8144, + "mean_token_accuracy": 0.7423743605613708, + "num_tokens": 689199668.0, + "step": 26633 + }, + { + "epoch": 2.924884691412256, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.20206356048584, + "learning_rate": 1e-06, + "loss": 0.8473, + "mean_token_accuracy": 0.7307021617889404, + "num_tokens": 689232059.0, + "step": 26634 + }, + { + "epoch": 2.9249945091148692, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3273000717163086, + "learning_rate": 1e-06, + "loss": 0.9094, + "mean_token_accuracy": 0.7168903350830078, + "num_tokens": 689259070.0, + "step": 26635 + }, + { + "epoch": 2.925104326817483, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.310612440109253, + "learning_rate": 1e-06, + "loss": 0.9495, + "mean_token_accuracy": 0.7053900957107544, + "num_tokens": 689284076.0, + "step": 26636 + }, + { + "epoch": 2.9252141445200968, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.541489601135254, + "learning_rate": 1e-06, + "loss": 0.9001, + "mean_token_accuracy": 0.7218779921531677, + "num_tokens": 689308431.0, + "step": 26637 + }, + { + "epoch": 2.92532396222271, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.038882255554199, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7319775223731995, + "num_tokens": 689340239.0, + "step": 26638 + }, + { + "epoch": 2.925433779925324, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.220407247543335, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.733750581741333, + "num_tokens": 689369272.0, + "step": 26639 + }, + { + "epoch": 2.9255435976279376, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2207789421081543, + "learning_rate": 1e-06, + "loss": 0.9501, + "mean_token_accuracy": 0.7024635076522827, + "num_tokens": 689400323.0, + "step": 26640 + }, + { + "epoch": 2.9256534153305513, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7008354663848877, + "learning_rate": 1e-06, + "loss": 0.8188, + "mean_token_accuracy": 0.7367371320724487, + "num_tokens": 689422422.0, + "step": 26641 + }, + { + "epoch": 2.925763233033165, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.252605676651001, + "learning_rate": 1e-06, + "loss": 0.923, + "mean_token_accuracy": 0.7088061571121216, + "num_tokens": 689451044.0, + "step": 26642 + }, + { + "epoch": 2.9258730507357784, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5259389877319336, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7456129789352417, + "num_tokens": 689473795.0, + "step": 26643 + }, + { + "epoch": 2.925982868438392, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2557034492492676, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7232509255409241, + "num_tokens": 689503380.0, + "step": 26644 + }, + { + "epoch": 2.926092686141006, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5088582038879395, + "learning_rate": 1e-06, + "loss": 0.8614, + "mean_token_accuracy": 0.72669517993927, + "num_tokens": 689526467.0, + "step": 26645 + }, + { + "epoch": 2.9262025038436197, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2647650241851807, + "learning_rate": 1e-06, + "loss": 0.8993, + "mean_token_accuracy": 0.7164023518562317, + "num_tokens": 689555347.0, + "step": 26646 + }, + { + "epoch": 2.9263123215462334, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 1.9914246797561646, + "learning_rate": 1e-06, + "loss": 0.9829, + "mean_token_accuracy": 0.6977565884590149, + "num_tokens": 689592022.0, + "step": 26647 + }, + { + "epoch": 2.9264221392488468, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5608716011047363, + "learning_rate": 1e-06, + "loss": 0.8322, + "mean_token_accuracy": 0.7335478067398071, + "num_tokens": 689612055.0, + "step": 26648 + }, + { + "epoch": 2.9265319569514605, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4022767543792725, + "learning_rate": 1e-06, + "loss": 0.9076, + "mean_token_accuracy": 0.7138984799385071, + "num_tokens": 689638227.0, + "step": 26649 + }, + { + "epoch": 2.9266417746540743, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.188960313796997, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7292590141296387, + "num_tokens": 689669711.0, + "step": 26650 + }, + { + "epoch": 2.926751592356688, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.242645502090454, + "learning_rate": 1e-06, + "loss": 0.8915, + "mean_token_accuracy": 0.7172368764877319, + "num_tokens": 689699272.0, + "step": 26651 + }, + { + "epoch": 2.926861410059302, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.225731372833252, + "learning_rate": 1e-06, + "loss": 0.9091, + "mean_token_accuracy": 0.7146936655044556, + "num_tokens": 689729717.0, + "step": 26652 + }, + { + "epoch": 2.926971227761915, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2610561847686768, + "learning_rate": 1e-06, + "loss": 0.8495, + "mean_token_accuracy": 0.7263974547386169, + "num_tokens": 689755552.0, + "step": 26653 + }, + { + "epoch": 2.927081045464529, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.514942169189453, + "learning_rate": 1e-06, + "loss": 0.9626, + "mean_token_accuracy": 0.7023589611053467, + "num_tokens": 689780275.0, + "step": 26654 + }, + { + "epoch": 2.9271908631671426, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2788503170013428, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7271692752838135, + "num_tokens": 689806057.0, + "step": 26655 + }, + { + "epoch": 2.9273006808697564, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.485093116760254, + "learning_rate": 1e-06, + "loss": 0.8468, + "mean_token_accuracy": 0.7362335324287415, + "num_tokens": 689831959.0, + "step": 26656 + }, + { + "epoch": 2.92741049857237, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.087543487548828, + "learning_rate": 1e-06, + "loss": 0.829, + "mean_token_accuracy": 0.7341673374176025, + "num_tokens": 689863640.0, + "step": 26657 + }, + { + "epoch": 2.9275203162749834, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3473799228668213, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7213395833969116, + "num_tokens": 689892249.0, + "step": 26658 + }, + { + "epoch": 2.927630133977597, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.401405096054077, + "learning_rate": 1e-06, + "loss": 0.8397, + "mean_token_accuracy": 0.7343437671661377, + "num_tokens": 689916482.0, + "step": 26659 + }, + { + "epoch": 2.927739951680211, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.36032772064209, + "learning_rate": 1e-06, + "loss": 0.816, + "mean_token_accuracy": 0.7417966723442078, + "num_tokens": 689941566.0, + "step": 26660 + }, + { + "epoch": 2.9278497693828243, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2425453662872314, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7349634766578674, + "num_tokens": 689968884.0, + "step": 26661 + }, + { + "epoch": 2.927959587085438, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2730205059051514, + "learning_rate": 1e-06, + "loss": 0.8995, + "mean_token_accuracy": 0.7197315096855164, + "num_tokens": 689997689.0, + "step": 26662 + }, + { + "epoch": 2.9280694047880518, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.494236707687378, + "learning_rate": 1e-06, + "loss": 0.8131, + "mean_token_accuracy": 0.7442610859870911, + "num_tokens": 690020953.0, + "step": 26663 + }, + { + "epoch": 2.9281792224906655, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2854979038238525, + "learning_rate": 1e-06, + "loss": 0.8946, + "mean_token_accuracy": 0.7216498851776123, + "num_tokens": 690049019.0, + "step": 26664 + }, + { + "epoch": 2.9282890401932793, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.367957592010498, + "learning_rate": 1e-06, + "loss": 0.8089, + "mean_token_accuracy": 0.7386070489883423, + "num_tokens": 690075642.0, + "step": 26665 + }, + { + "epoch": 2.9283988578958926, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.122554063796997, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7272663116455078, + "num_tokens": 690107687.0, + "step": 26666 + }, + { + "epoch": 2.9285086755985064, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.921694755554199, + "learning_rate": 1e-06, + "loss": 0.8683, + "mean_token_accuracy": 0.7287721633911133, + "num_tokens": 690128154.0, + "step": 26667 + }, + { + "epoch": 2.92861849330112, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6243538856506348, + "learning_rate": 1e-06, + "loss": 0.7588, + "mean_token_accuracy": 0.7602927684783936, + "num_tokens": 690149424.0, + "step": 26668 + }, + { + "epoch": 2.928728311003734, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.405705213546753, + "learning_rate": 1e-06, + "loss": 0.8362, + "mean_token_accuracy": 0.7403978109359741, + "num_tokens": 690174474.0, + "step": 26669 + }, + { + "epoch": 2.9288381287063476, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5645859241485596, + "learning_rate": 1e-06, + "loss": 0.8417, + "mean_token_accuracy": 0.737118661403656, + "num_tokens": 690198756.0, + "step": 26670 + }, + { + "epoch": 2.928947946408961, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6517834663391113, + "learning_rate": 1e-06, + "loss": 0.7208, + "mean_token_accuracy": 0.7662957906723022, + "num_tokens": 690219581.0, + "step": 26671 + }, + { + "epoch": 2.9290577641115747, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1490519046783447, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7190759181976318, + "num_tokens": 690249638.0, + "step": 26672 + }, + { + "epoch": 2.9291675818141885, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.36165189743042, + "learning_rate": 1e-06, + "loss": 0.9364, + "mean_token_accuracy": 0.7153939008712769, + "num_tokens": 690275987.0, + "step": 26673 + }, + { + "epoch": 2.929277399516802, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7140445709228516, + "learning_rate": 1e-06, + "loss": 0.765, + "mean_token_accuracy": 0.7552317380905151, + "num_tokens": 690295474.0, + "step": 26674 + }, + { + "epoch": 2.929387217219416, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7311694622039795, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7402694225311279, + "num_tokens": 690315944.0, + "step": 26675 + }, + { + "epoch": 2.9294970349220293, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.433651924133301, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7353081703186035, + "num_tokens": 690340715.0, + "step": 26676 + }, + { + "epoch": 2.929606852624643, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.544855833053589, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7365361452102661, + "num_tokens": 690361929.0, + "step": 26677 + }, + { + "epoch": 2.929716670327257, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1585776805877686, + "learning_rate": 1e-06, + "loss": 0.8969, + "mean_token_accuracy": 0.7180749773979187, + "num_tokens": 690391351.0, + "step": 26678 + }, + { + "epoch": 2.9298264880298706, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1568942070007324, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7238864898681641, + "num_tokens": 690421413.0, + "step": 26679 + }, + { + "epoch": 2.9299363057324843, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.505530595779419, + "learning_rate": 1e-06, + "loss": 0.9541, + "mean_token_accuracy": 0.7075362205505371, + "num_tokens": 690446312.0, + "step": 26680 + }, + { + "epoch": 2.9300461234350976, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.545159339904785, + "learning_rate": 1e-06, + "loss": 0.9479, + "mean_token_accuracy": 0.7051987051963806, + "num_tokens": 690471830.0, + "step": 26681 + }, + { + "epoch": 2.9301559411377114, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3708548545837402, + "learning_rate": 1e-06, + "loss": 0.9248, + "mean_token_accuracy": 0.7127257585525513, + "num_tokens": 690501404.0, + "step": 26682 + }, + { + "epoch": 2.930265758840325, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.452000856399536, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.7273515462875366, + "num_tokens": 690524742.0, + "step": 26683 + }, + { + "epoch": 2.9303755765429385, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.562081813812256, + "learning_rate": 1e-06, + "loss": 0.7253, + "mean_token_accuracy": 0.7722100019454956, + "num_tokens": 690546333.0, + "step": 26684 + }, + { + "epoch": 2.9304853942455527, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.494981050491333, + "learning_rate": 1e-06, + "loss": 0.8357, + "mean_token_accuracy": 0.7450333833694458, + "num_tokens": 690569697.0, + "step": 26685 + }, + { + "epoch": 2.930595211948166, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.388587713241577, + "learning_rate": 1e-06, + "loss": 0.6988, + "mean_token_accuracy": 0.7739367485046387, + "num_tokens": 690593988.0, + "step": 26686 + }, + { + "epoch": 2.9307050296507797, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.685136556625366, + "learning_rate": 1e-06, + "loss": 0.8933, + "mean_token_accuracy": 0.7235888838768005, + "num_tokens": 690618230.0, + "step": 26687 + }, + { + "epoch": 2.9308148473533935, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7135252952575684, + "learning_rate": 1e-06, + "loss": 0.8072, + "mean_token_accuracy": 0.7405754327774048, + "num_tokens": 690639264.0, + "step": 26688 + }, + { + "epoch": 2.930924665056007, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5527398586273193, + "learning_rate": 1e-06, + "loss": 0.7716, + "mean_token_accuracy": 0.7540758848190308, + "num_tokens": 690661980.0, + "step": 26689 + }, + { + "epoch": 2.9310344827586206, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.560253143310547, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7398228645324707, + "num_tokens": 690685725.0, + "step": 26690 + }, + { + "epoch": 2.9311443004612343, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5144600868225098, + "learning_rate": 1e-06, + "loss": 0.8698, + "mean_token_accuracy": 0.7289792895317078, + "num_tokens": 690710122.0, + "step": 26691 + }, + { + "epoch": 2.931254118163848, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5503923892974854, + "learning_rate": 1e-06, + "loss": 0.8703, + "mean_token_accuracy": 0.7332425117492676, + "num_tokens": 690737360.0, + "step": 26692 + }, + { + "epoch": 2.931363935866462, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6474053859710693, + "learning_rate": 1e-06, + "loss": 0.8558, + "mean_token_accuracy": 0.727519690990448, + "num_tokens": 690758151.0, + "step": 26693 + }, + { + "epoch": 2.931473753569075, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2743325233459473, + "learning_rate": 1e-06, + "loss": 0.7211, + "mean_token_accuracy": 0.763529896736145, + "num_tokens": 690782626.0, + "step": 26694 + }, + { + "epoch": 2.931583571271689, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.645751714706421, + "learning_rate": 1e-06, + "loss": 0.7851, + "mean_token_accuracy": 0.7473129034042358, + "num_tokens": 690804268.0, + "step": 26695 + }, + { + "epoch": 2.9316933889743026, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.126622200012207, + "learning_rate": 1e-06, + "loss": 0.8844, + "mean_token_accuracy": 0.7260386943817139, + "num_tokens": 690833758.0, + "step": 26696 + }, + { + "epoch": 2.9318032066769164, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.137392044067383, + "learning_rate": 1e-06, + "loss": 0.9356, + "mean_token_accuracy": 0.706343412399292, + "num_tokens": 690867639.0, + "step": 26697 + }, + { + "epoch": 2.93191302437953, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 1.9977083206176758, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7163859605789185, + "num_tokens": 690902197.0, + "step": 26698 + }, + { + "epoch": 2.9320228420821435, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.276991844177246, + "learning_rate": 1e-06, + "loss": 0.92, + "mean_token_accuracy": 0.7132838368415833, + "num_tokens": 690932496.0, + "step": 26699 + }, + { + "epoch": 2.9321326597847572, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6949820518493652, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7389541864395142, + "num_tokens": 690956064.0, + "step": 26700 + }, + { + "epoch": 2.932242477487371, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.357306718826294, + "learning_rate": 1e-06, + "loss": 0.8829, + "mean_token_accuracy": 0.7224758267402649, + "num_tokens": 690980200.0, + "step": 26701 + }, + { + "epoch": 2.9323522951899847, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4060049057006836, + "learning_rate": 1e-06, + "loss": 0.8622, + "mean_token_accuracy": 0.7366108894348145, + "num_tokens": 691003199.0, + "step": 26702 + }, + { + "epoch": 2.9324621128925985, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5839850902557373, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.7346419095993042, + "num_tokens": 691025235.0, + "step": 26703 + }, + { + "epoch": 2.932571930595212, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3369293212890625, + "learning_rate": 1e-06, + "loss": 0.7491, + "mean_token_accuracy": 0.758675217628479, + "num_tokens": 691049426.0, + "step": 26704 + }, + { + "epoch": 2.9326817482978256, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1517727375030518, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7297848463058472, + "num_tokens": 691080489.0, + "step": 26705 + }, + { + "epoch": 2.9327915660004393, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2625796794891357, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7231617569923401, + "num_tokens": 691109836.0, + "step": 26706 + }, + { + "epoch": 2.932901383703053, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.545461654663086, + "learning_rate": 1e-06, + "loss": 0.8547, + "mean_token_accuracy": 0.734485387802124, + "num_tokens": 691133883.0, + "step": 26707 + }, + { + "epoch": 2.933011201405667, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3446834087371826, + "learning_rate": 1e-06, + "loss": 0.8395, + "mean_token_accuracy": 0.7317784428596497, + "num_tokens": 691159481.0, + "step": 26708 + }, + { + "epoch": 2.93312101910828, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 3.0524237155914307, + "learning_rate": 1e-06, + "loss": 0.8055, + "mean_token_accuracy": 0.7390334010124207, + "num_tokens": 691178012.0, + "step": 26709 + }, + { + "epoch": 2.933230836810894, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6996312141418457, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7288567423820496, + "num_tokens": 691200489.0, + "step": 26710 + }, + { + "epoch": 2.9333406545135077, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.671372175216675, + "learning_rate": 1e-06, + "loss": 0.8058, + "mean_token_accuracy": 0.7442443370819092, + "num_tokens": 691220284.0, + "step": 26711 + }, + { + "epoch": 2.933450472216121, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2488906383514404, + "learning_rate": 1e-06, + "loss": 0.7789, + "mean_token_accuracy": 0.7588232755661011, + "num_tokens": 691247468.0, + "step": 26712 + }, + { + "epoch": 2.9335602899187347, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3714544773101807, + "learning_rate": 1e-06, + "loss": 0.8548, + "mean_token_accuracy": 0.7289892435073853, + "num_tokens": 691274301.0, + "step": 26713 + }, + { + "epoch": 2.9336701076213485, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7982072830200195, + "learning_rate": 1e-06, + "loss": 0.7774, + "mean_token_accuracy": 0.7460631132125854, + "num_tokens": 691294714.0, + "step": 26714 + }, + { + "epoch": 2.9337799253239623, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2921946048736572, + "learning_rate": 1e-06, + "loss": 0.898, + "mean_token_accuracy": 0.7144149541854858, + "num_tokens": 691320042.0, + "step": 26715 + }, + { + "epoch": 2.933889743026576, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 3.035780191421509, + "learning_rate": 1e-06, + "loss": 0.7568, + "mean_token_accuracy": 0.7560294270515442, + "num_tokens": 691336829.0, + "step": 26716 + }, + { + "epoch": 2.9339995607291893, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.082163095474243, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7497007250785828, + "num_tokens": 691366283.0, + "step": 26717 + }, + { + "epoch": 2.934109378431803, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5105693340301514, + "learning_rate": 1e-06, + "loss": 0.8183, + "mean_token_accuracy": 0.7444947361946106, + "num_tokens": 691390445.0, + "step": 26718 + }, + { + "epoch": 2.934219196134417, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.332608461380005, + "learning_rate": 1e-06, + "loss": 0.8077, + "mean_token_accuracy": 0.740402340888977, + "num_tokens": 691416398.0, + "step": 26719 + }, + { + "epoch": 2.9343290138370306, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 1.9780982732772827, + "learning_rate": 1e-06, + "loss": 0.9311, + "mean_token_accuracy": 0.7118732333183289, + "num_tokens": 691451529.0, + "step": 26720 + }, + { + "epoch": 2.9344388315396444, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.458991765975952, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.740509033203125, + "num_tokens": 691476310.0, + "step": 26721 + }, + { + "epoch": 2.9345486492422577, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.640632152557373, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7440664768218994, + "num_tokens": 691499514.0, + "step": 26722 + }, + { + "epoch": 2.9346584669448714, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.65878963470459, + "learning_rate": 1e-06, + "loss": 0.843, + "mean_token_accuracy": 0.7295247316360474, + "num_tokens": 691521563.0, + "step": 26723 + }, + { + "epoch": 2.934768284647485, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3916265964508057, + "learning_rate": 1e-06, + "loss": 0.875, + "mean_token_accuracy": 0.7268691062927246, + "num_tokens": 691545425.0, + "step": 26724 + }, + { + "epoch": 2.934878102350099, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3233883380889893, + "learning_rate": 1e-06, + "loss": 0.9282, + "mean_token_accuracy": 0.7068014144897461, + "num_tokens": 691571772.0, + "step": 26725 + }, + { + "epoch": 2.9349879200527127, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4446895122528076, + "learning_rate": 1e-06, + "loss": 0.901, + "mean_token_accuracy": 0.7168352603912354, + "num_tokens": 691597828.0, + "step": 26726 + }, + { + "epoch": 2.935097737755326, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.255735397338867, + "learning_rate": 1e-06, + "loss": 0.9524, + "mean_token_accuracy": 0.7046127319335938, + "num_tokens": 691627954.0, + "step": 26727 + }, + { + "epoch": 2.9352075554579398, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.351828098297119, + "learning_rate": 1e-06, + "loss": 0.9418, + "mean_token_accuracy": 0.7039715647697449, + "num_tokens": 691655678.0, + "step": 26728 + }, + { + "epoch": 2.9353173731605535, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4486703872680664, + "learning_rate": 1e-06, + "loss": 0.9095, + "mean_token_accuracy": 0.7257954478263855, + "num_tokens": 691681956.0, + "step": 26729 + }, + { + "epoch": 2.9354271908631673, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.443441152572632, + "learning_rate": 1e-06, + "loss": 0.864, + "mean_token_accuracy": 0.7335227727890015, + "num_tokens": 691708417.0, + "step": 26730 + }, + { + "epoch": 2.935537008565781, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6060259342193604, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7299547791481018, + "num_tokens": 691728792.0, + "step": 26731 + }, + { + "epoch": 2.9356468262683943, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.50437068939209, + "learning_rate": 1e-06, + "loss": 0.7771, + "mean_token_accuracy": 0.7449908256530762, + "num_tokens": 691751611.0, + "step": 26732 + }, + { + "epoch": 2.935756643971008, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6245622634887695, + "learning_rate": 1e-06, + "loss": 0.8805, + "mean_token_accuracy": 0.7285714149475098, + "num_tokens": 691773539.0, + "step": 26733 + }, + { + "epoch": 2.935866461673622, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.488619327545166, + "learning_rate": 1e-06, + "loss": 0.8117, + "mean_token_accuracy": 0.7351385354995728, + "num_tokens": 691796543.0, + "step": 26734 + }, + { + "epoch": 2.935976279376235, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.998321294784546, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7382413148880005, + "num_tokens": 691824616.0, + "step": 26735 + }, + { + "epoch": 2.9360860970788494, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.520212173461914, + "learning_rate": 1e-06, + "loss": 0.9051, + "mean_token_accuracy": 0.7192427515983582, + "num_tokens": 691847901.0, + "step": 26736 + }, + { + "epoch": 2.9361959147814627, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1331100463867188, + "learning_rate": 1e-06, + "loss": 0.8255, + "mean_token_accuracy": 0.7426737546920776, + "num_tokens": 691875459.0, + "step": 26737 + }, + { + "epoch": 2.9363057324840764, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2578370571136475, + "learning_rate": 1e-06, + "loss": 0.8892, + "mean_token_accuracy": 0.7249769568443298, + "num_tokens": 691905817.0, + "step": 26738 + }, + { + "epoch": 2.93641555018669, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2794876098632812, + "learning_rate": 1e-06, + "loss": 0.8506, + "mean_token_accuracy": 0.7401465177536011, + "num_tokens": 691932060.0, + "step": 26739 + }, + { + "epoch": 2.9365253678893035, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1961748600006104, + "learning_rate": 1e-06, + "loss": 0.9115, + "mean_token_accuracy": 0.713677167892456, + "num_tokens": 691959423.0, + "step": 26740 + }, + { + "epoch": 2.9366351855919173, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1568052768707275, + "learning_rate": 1e-06, + "loss": 0.8716, + "mean_token_accuracy": 0.735001802444458, + "num_tokens": 691989589.0, + "step": 26741 + }, + { + "epoch": 2.936745003294531, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2311742305755615, + "learning_rate": 1e-06, + "loss": 0.8272, + "mean_token_accuracy": 0.73381108045578, + "num_tokens": 692018265.0, + "step": 26742 + }, + { + "epoch": 2.936854820997145, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3026299476623535, + "learning_rate": 1e-06, + "loss": 0.8729, + "mean_token_accuracy": 0.731277346611023, + "num_tokens": 692044060.0, + "step": 26743 + }, + { + "epoch": 2.9369646386997585, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.497274398803711, + "learning_rate": 1e-06, + "loss": 0.8316, + "mean_token_accuracy": 0.7316128015518188, + "num_tokens": 692071309.0, + "step": 26744 + }, + { + "epoch": 2.937074456402372, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.366860866546631, + "learning_rate": 1e-06, + "loss": 0.7997, + "mean_token_accuracy": 0.7483575940132141, + "num_tokens": 692095403.0, + "step": 26745 + }, + { + "epoch": 2.9371842741049856, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4040932655334473, + "learning_rate": 1e-06, + "loss": 0.882, + "mean_token_accuracy": 0.7214826345443726, + "num_tokens": 692120477.0, + "step": 26746 + }, + { + "epoch": 2.9372940918075994, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2603814601898193, + "learning_rate": 1e-06, + "loss": 0.8507, + "mean_token_accuracy": 0.7230238914489746, + "num_tokens": 692147261.0, + "step": 26747 + }, + { + "epoch": 2.937403909510213, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.530348777770996, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.716172456741333, + "num_tokens": 692170175.0, + "step": 26748 + }, + { + "epoch": 2.937513727212827, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 1.9454443454742432, + "learning_rate": 1e-06, + "loss": 0.8809, + "mean_token_accuracy": 0.7180105447769165, + "num_tokens": 692207159.0, + "step": 26749 + }, + { + "epoch": 2.93762354491544, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6492562294006348, + "learning_rate": 1e-06, + "loss": 0.8059, + "mean_token_accuracy": 0.7486307621002197, + "num_tokens": 692228771.0, + "step": 26750 + }, + { + "epoch": 2.937733362618054, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4878153800964355, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.727603018283844, + "num_tokens": 692252062.0, + "step": 26751 + }, + { + "epoch": 2.9378431803206677, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4885761737823486, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.740784764289856, + "num_tokens": 692276146.0, + "step": 26752 + }, + { + "epoch": 2.9379529980232815, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3218955993652344, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7380335330963135, + "num_tokens": 692302206.0, + "step": 26753 + }, + { + "epoch": 2.9380628157258952, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.497894525527954, + "learning_rate": 1e-06, + "loss": 0.801, + "mean_token_accuracy": 0.7368888854980469, + "num_tokens": 692325652.0, + "step": 26754 + }, + { + "epoch": 2.9381726334285085, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1769001483917236, + "learning_rate": 1e-06, + "loss": 0.8882, + "mean_token_accuracy": 0.7202863097190857, + "num_tokens": 692355280.0, + "step": 26755 + }, + { + "epoch": 2.9382824511311223, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3817250728607178, + "learning_rate": 1e-06, + "loss": 0.8815, + "mean_token_accuracy": 0.7287233471870422, + "num_tokens": 692379783.0, + "step": 26756 + }, + { + "epoch": 2.938392268833736, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.53334379196167, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7357891798019409, + "num_tokens": 692402919.0, + "step": 26757 + }, + { + "epoch": 2.93850208653635, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.363377571105957, + "learning_rate": 1e-06, + "loss": 0.8236, + "mean_token_accuracy": 0.739835262298584, + "num_tokens": 692429074.0, + "step": 26758 + }, + { + "epoch": 2.9386119042389636, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6130447387695312, + "learning_rate": 1e-06, + "loss": 0.8563, + "mean_token_accuracy": 0.7313631772994995, + "num_tokens": 692452534.0, + "step": 26759 + }, + { + "epoch": 2.938721721941577, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4560654163360596, + "learning_rate": 1e-06, + "loss": 0.8259, + "mean_token_accuracy": 0.736970841884613, + "num_tokens": 692477274.0, + "step": 26760 + }, + { + "epoch": 2.9388315396441906, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5624897480010986, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7412979602813721, + "num_tokens": 692500243.0, + "step": 26761 + }, + { + "epoch": 2.9389413573468044, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.517547845840454, + "learning_rate": 1e-06, + "loss": 0.8407, + "mean_token_accuracy": 0.7292574644088745, + "num_tokens": 692523281.0, + "step": 26762 + }, + { + "epoch": 2.9390511750494177, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.234384298324585, + "learning_rate": 1e-06, + "loss": 0.8243, + "mean_token_accuracy": 0.7384015917778015, + "num_tokens": 692550813.0, + "step": 26763 + }, + { + "epoch": 2.9391609927520315, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.767317295074463, + "learning_rate": 1e-06, + "loss": 0.8411, + "mean_token_accuracy": 0.7435815334320068, + "num_tokens": 692571744.0, + "step": 26764 + }, + { + "epoch": 2.939270810454645, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5531136989593506, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7412243485450745, + "num_tokens": 692594231.0, + "step": 26765 + }, + { + "epoch": 2.939380628157259, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.669128894805908, + "learning_rate": 1e-06, + "loss": 0.8069, + "mean_token_accuracy": 0.7420494556427002, + "num_tokens": 692614966.0, + "step": 26766 + }, + { + "epoch": 2.9394904458598727, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4990296363830566, + "learning_rate": 1e-06, + "loss": 0.8335, + "mean_token_accuracy": 0.735030472278595, + "num_tokens": 692638923.0, + "step": 26767 + }, + { + "epoch": 2.939600263562486, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3883819580078125, + "learning_rate": 1e-06, + "loss": 0.9026, + "mean_token_accuracy": 0.7188777923583984, + "num_tokens": 692664177.0, + "step": 26768 + }, + { + "epoch": 2.9397100812651, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.51259446144104, + "learning_rate": 1e-06, + "loss": 0.8212, + "mean_token_accuracy": 0.7424201965332031, + "num_tokens": 692686427.0, + "step": 26769 + }, + { + "epoch": 2.9398198989677136, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2622039318084717, + "learning_rate": 1e-06, + "loss": 0.8736, + "mean_token_accuracy": 0.7255936861038208, + "num_tokens": 692714060.0, + "step": 26770 + }, + { + "epoch": 2.9399297166703273, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5439958572387695, + "learning_rate": 1e-06, + "loss": 0.6817, + "mean_token_accuracy": 0.7737728953361511, + "num_tokens": 692735925.0, + "step": 26771 + }, + { + "epoch": 2.940039534372941, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.779402494430542, + "learning_rate": 1e-06, + "loss": 0.8656, + "mean_token_accuracy": 0.7325857281684875, + "num_tokens": 692755877.0, + "step": 26772 + }, + { + "epoch": 2.9401493520755544, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.548618793487549, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7140816450119019, + "num_tokens": 692779742.0, + "step": 26773 + }, + { + "epoch": 2.940259169778168, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3029744625091553, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7218577861785889, + "num_tokens": 692807891.0, + "step": 26774 + }, + { + "epoch": 2.940368987480782, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7528605461120605, + "learning_rate": 1e-06, + "loss": 0.8267, + "mean_token_accuracy": 0.7359371185302734, + "num_tokens": 692827979.0, + "step": 26775 + }, + { + "epoch": 2.9404788051833957, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.303302764892578, + "learning_rate": 1e-06, + "loss": 0.9322, + "mean_token_accuracy": 0.7047939896583557, + "num_tokens": 692857900.0, + "step": 26776 + }, + { + "epoch": 2.9405886228860094, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2139761447906494, + "learning_rate": 1e-06, + "loss": 0.9285, + "mean_token_accuracy": 0.7150574922561646, + "num_tokens": 692887200.0, + "step": 26777 + }, + { + "epoch": 2.9406984405886227, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1753385066986084, + "learning_rate": 1e-06, + "loss": 0.8408, + "mean_token_accuracy": 0.7376391291618347, + "num_tokens": 692916475.0, + "step": 26778 + }, + { + "epoch": 2.9408082582912365, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.401390314102173, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7248737812042236, + "num_tokens": 692941391.0, + "step": 26779 + }, + { + "epoch": 2.9409180759938502, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3660833835601807, + "learning_rate": 1e-06, + "loss": 0.8034, + "mean_token_accuracy": 0.7380890846252441, + "num_tokens": 692965948.0, + "step": 26780 + }, + { + "epoch": 2.941027893696464, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.534994602203369, + "learning_rate": 1e-06, + "loss": 0.8323, + "mean_token_accuracy": 0.7335647940635681, + "num_tokens": 692988724.0, + "step": 26781 + }, + { + "epoch": 2.9411377113990778, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.681462049484253, + "learning_rate": 1e-06, + "loss": 0.8458, + "mean_token_accuracy": 0.7337608337402344, + "num_tokens": 693010367.0, + "step": 26782 + }, + { + "epoch": 2.941247529101691, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.301254987716675, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7384074926376343, + "num_tokens": 693036425.0, + "step": 26783 + }, + { + "epoch": 2.941357346804305, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.977041244506836, + "learning_rate": 1e-06, + "loss": 0.7805, + "mean_token_accuracy": 0.7423579692840576, + "num_tokens": 693055801.0, + "step": 26784 + }, + { + "epoch": 2.9414671645069186, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.7353174686431885, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7303472757339478, + "num_tokens": 693076977.0, + "step": 26785 + }, + { + "epoch": 2.9415769822095323, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2698841094970703, + "learning_rate": 1e-06, + "loss": 0.9135, + "mean_token_accuracy": 0.7130946516990662, + "num_tokens": 693105603.0, + "step": 26786 + }, + { + "epoch": 2.941686799912146, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4281160831451416, + "learning_rate": 1e-06, + "loss": 0.8494, + "mean_token_accuracy": 0.73735111951828, + "num_tokens": 693132176.0, + "step": 26787 + }, + { + "epoch": 2.9417966176147594, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3111188411712646, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.743373453617096, + "num_tokens": 693158898.0, + "step": 26788 + }, + { + "epoch": 2.941906435317373, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.498420000076294, + "learning_rate": 1e-06, + "loss": 0.819, + "mean_token_accuracy": 0.7447615265846252, + "num_tokens": 693181428.0, + "step": 26789 + }, + { + "epoch": 2.942016253019987, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2819294929504395, + "learning_rate": 1e-06, + "loss": 0.8742, + "mean_token_accuracy": 0.7301820516586304, + "num_tokens": 693209373.0, + "step": 26790 + }, + { + "epoch": 2.9421260707226002, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5086755752563477, + "learning_rate": 1e-06, + "loss": 0.8137, + "mean_token_accuracy": 0.7474260330200195, + "num_tokens": 693233344.0, + "step": 26791 + }, + { + "epoch": 2.942235888425214, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3916029930114746, + "learning_rate": 1e-06, + "loss": 0.8213, + "mean_token_accuracy": 0.7365180253982544, + "num_tokens": 693256521.0, + "step": 26792 + }, + { + "epoch": 2.9423457061278278, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6601064205169678, + "learning_rate": 1e-06, + "loss": 0.842, + "mean_token_accuracy": 0.7358788251876831, + "num_tokens": 693282044.0, + "step": 26793 + }, + { + "epoch": 2.9424555238304415, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2996504306793213, + "learning_rate": 1e-06, + "loss": 0.8401, + "mean_token_accuracy": 0.7300597429275513, + "num_tokens": 693306197.0, + "step": 26794 + }, + { + "epoch": 2.9425653415330553, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2688100337982178, + "learning_rate": 1e-06, + "loss": 0.8986, + "mean_token_accuracy": 0.7179772257804871, + "num_tokens": 693333976.0, + "step": 26795 + }, + { + "epoch": 2.9426751592356686, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5377445220947266, + "learning_rate": 1e-06, + "loss": 0.8135, + "mean_token_accuracy": 0.7386796474456787, + "num_tokens": 693355923.0, + "step": 26796 + }, + { + "epoch": 2.9427849769382823, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.228926181793213, + "learning_rate": 1e-06, + "loss": 0.8816, + "mean_token_accuracy": 0.7205647230148315, + "num_tokens": 693383526.0, + "step": 26797 + }, + { + "epoch": 2.942894794640896, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3548171520233154, + "learning_rate": 1e-06, + "loss": 0.8962, + "mean_token_accuracy": 0.724075973033905, + "num_tokens": 693408827.0, + "step": 26798 + }, + { + "epoch": 2.94300461234351, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.41927170753479, + "learning_rate": 1e-06, + "loss": 0.8463, + "mean_token_accuracy": 0.7372487783432007, + "num_tokens": 693434950.0, + "step": 26799 + }, + { + "epoch": 2.9431144300461236, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3033864498138428, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7362259030342102, + "num_tokens": 693461097.0, + "step": 26800 + }, + { + "epoch": 2.943224247748737, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4574854373931885, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7239514589309692, + "num_tokens": 693486772.0, + "step": 26801 + }, + { + "epoch": 2.9433340654513507, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.168436050415039, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7261142730712891, + "num_tokens": 693514765.0, + "step": 26802 + }, + { + "epoch": 2.9434438831539644, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.495943784713745, + "learning_rate": 1e-06, + "loss": 0.9202, + "mean_token_accuracy": 0.7215649485588074, + "num_tokens": 693538919.0, + "step": 26803 + }, + { + "epoch": 2.943553700856578, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.55843448638916, + "learning_rate": 1e-06, + "loss": 0.754, + "mean_token_accuracy": 0.7547762393951416, + "num_tokens": 693560016.0, + "step": 26804 + }, + { + "epoch": 2.943663518559192, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1939444541931152, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7238093614578247, + "num_tokens": 693590642.0, + "step": 26805 + }, + { + "epoch": 2.9437733362618053, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.559983491897583, + "learning_rate": 1e-06, + "loss": 0.8502, + "mean_token_accuracy": 0.7261024117469788, + "num_tokens": 693614369.0, + "step": 26806 + }, + { + "epoch": 2.943883153964419, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2037532329559326, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7281385660171509, + "num_tokens": 693643798.0, + "step": 26807 + }, + { + "epoch": 2.9439929716670328, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.0679264068603516, + "learning_rate": 1e-06, + "loss": 0.9305, + "mean_token_accuracy": 0.709274411201477, + "num_tokens": 693676593.0, + "step": 26808 + }, + { + "epoch": 2.9441027893696465, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6417431831359863, + "learning_rate": 1e-06, + "loss": 0.8129, + "mean_token_accuracy": 0.7441985607147217, + "num_tokens": 693698578.0, + "step": 26809 + }, + { + "epoch": 2.9442126070722603, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1398353576660156, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7332697510719299, + "num_tokens": 693728855.0, + "step": 26810 + }, + { + "epoch": 2.9443224247748736, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.22745943069458, + "learning_rate": 1e-06, + "loss": 0.9588, + "mean_token_accuracy": 0.7031494975090027, + "num_tokens": 693757497.0, + "step": 26811 + }, + { + "epoch": 2.9444322424774874, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 3.488685369491577, + "learning_rate": 1e-06, + "loss": 0.804, + "mean_token_accuracy": 0.7427927255630493, + "num_tokens": 693780563.0, + "step": 26812 + }, + { + "epoch": 2.944542060180101, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.212655544281006, + "learning_rate": 1e-06, + "loss": 0.8854, + "mean_token_accuracy": 0.7183345556259155, + "num_tokens": 693807523.0, + "step": 26813 + }, + { + "epoch": 2.9446518778827144, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.46856951713562, + "learning_rate": 1e-06, + "loss": 0.8767, + "mean_token_accuracy": 0.7248064279556274, + "num_tokens": 693831195.0, + "step": 26814 + }, + { + "epoch": 2.9447616955853286, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.204263687133789, + "learning_rate": 1e-06, + "loss": 0.9698, + "mean_token_accuracy": 0.7101518511772156, + "num_tokens": 693860382.0, + "step": 26815 + }, + { + "epoch": 2.944871513287942, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3348891735076904, + "learning_rate": 1e-06, + "loss": 0.8967, + "mean_token_accuracy": 0.7226419448852539, + "num_tokens": 693887384.0, + "step": 26816 + }, + { + "epoch": 2.9449813309905557, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.478017568588257, + "learning_rate": 1e-06, + "loss": 0.8044, + "mean_token_accuracy": 0.741256594657898, + "num_tokens": 693910703.0, + "step": 26817 + }, + { + "epoch": 2.9450911486931695, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4022765159606934, + "learning_rate": 1e-06, + "loss": 0.9212, + "mean_token_accuracy": 0.7132875919342041, + "num_tokens": 693935462.0, + "step": 26818 + }, + { + "epoch": 2.9452009663957828, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.52712345123291, + "learning_rate": 1e-06, + "loss": 0.7766, + "mean_token_accuracy": 0.749117374420166, + "num_tokens": 693956186.0, + "step": 26819 + }, + { + "epoch": 2.9453107840983965, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.177967071533203, + "learning_rate": 1e-06, + "loss": 0.8587, + "mean_token_accuracy": 0.7286412715911865, + "num_tokens": 693985511.0, + "step": 26820 + }, + { + "epoch": 2.9454206018010103, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.422886371612549, + "learning_rate": 1e-06, + "loss": 0.7466, + "mean_token_accuracy": 0.7534387111663818, + "num_tokens": 694008428.0, + "step": 26821 + }, + { + "epoch": 2.945530419503624, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2670955657958984, + "learning_rate": 1e-06, + "loss": 0.8674, + "mean_token_accuracy": 0.7329061627388, + "num_tokens": 694035928.0, + "step": 26822 + }, + { + "epoch": 2.945640237206238, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3718056678771973, + "learning_rate": 1e-06, + "loss": 0.9071, + "mean_token_accuracy": 0.7116458415985107, + "num_tokens": 694061531.0, + "step": 26823 + }, + { + "epoch": 2.945750054908851, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.461956262588501, + "learning_rate": 1e-06, + "loss": 0.863, + "mean_token_accuracy": 0.7287039756774902, + "num_tokens": 694084998.0, + "step": 26824 + }, + { + "epoch": 2.945859872611465, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.56071138381958, + "learning_rate": 1e-06, + "loss": 0.8118, + "mean_token_accuracy": 0.7400645017623901, + "num_tokens": 694106285.0, + "step": 26825 + }, + { + "epoch": 2.9459696903140786, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2373359203338623, + "learning_rate": 1e-06, + "loss": 0.8745, + "mean_token_accuracy": 0.7236668467521667, + "num_tokens": 694135134.0, + "step": 26826 + }, + { + "epoch": 2.9460795080166924, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4953548908233643, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.7426457405090332, + "num_tokens": 694160220.0, + "step": 26827 + }, + { + "epoch": 2.946189325719306, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.6344125270843506, + "learning_rate": 1e-06, + "loss": 0.8708, + "mean_token_accuracy": 0.7266572117805481, + "num_tokens": 694181966.0, + "step": 26828 + }, + { + "epoch": 2.9462991434219195, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.39630389213562, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7315282225608826, + "num_tokens": 694207810.0, + "step": 26829 + }, + { + "epoch": 2.946408961124533, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2156407833099365, + "learning_rate": 1e-06, + "loss": 0.8672, + "mean_token_accuracy": 0.7277460098266602, + "num_tokens": 694238163.0, + "step": 26830 + }, + { + "epoch": 2.946518778827147, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.592397689819336, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7521791458129883, + "num_tokens": 694260657.0, + "step": 26831 + }, + { + "epoch": 2.9466285965297607, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1169028282165527, + "learning_rate": 1e-06, + "loss": 0.9277, + "mean_token_accuracy": 0.7068263292312622, + "num_tokens": 694290003.0, + "step": 26832 + }, + { + "epoch": 2.9467384142323745, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.3605563640594482, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7317849397659302, + "num_tokens": 694315162.0, + "step": 26833 + }, + { + "epoch": 2.946848231934988, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2719979286193848, + "learning_rate": 1e-06, + "loss": 0.8525, + "mean_token_accuracy": 0.730002760887146, + "num_tokens": 694342414.0, + "step": 26834 + }, + { + "epoch": 2.9469580496376016, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.382431745529175, + "learning_rate": 1e-06, + "loss": 0.7693, + "mean_token_accuracy": 0.7502850294113159, + "num_tokens": 694367724.0, + "step": 26835 + }, + { + "epoch": 2.9470678673402153, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.656233787536621, + "learning_rate": 1e-06, + "loss": 0.771, + "mean_token_accuracy": 0.756316602230072, + "num_tokens": 694388463.0, + "step": 26836 + }, + { + "epoch": 2.947177685042829, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.338078260421753, + "learning_rate": 1e-06, + "loss": 0.8027, + "mean_token_accuracy": 0.7439327836036682, + "num_tokens": 694415520.0, + "step": 26837 + }, + { + "epoch": 2.947287502745443, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5284557342529297, + "learning_rate": 1e-06, + "loss": 0.8868, + "mean_token_accuracy": 0.7207682132720947, + "num_tokens": 694438833.0, + "step": 26838 + }, + { + "epoch": 2.947397320448056, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.217311382293701, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.7481900453567505, + "num_tokens": 694466746.0, + "step": 26839 + }, + { + "epoch": 2.94750713815067, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5956814289093018, + "learning_rate": 1e-06, + "loss": 0.7406, + "mean_token_accuracy": 0.7563265562057495, + "num_tokens": 694488620.0, + "step": 26840 + }, + { + "epoch": 2.9476169558532836, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.486635208129883, + "learning_rate": 1e-06, + "loss": 0.8618, + "mean_token_accuracy": 0.7254754304885864, + "num_tokens": 694513493.0, + "step": 26841 + }, + { + "epoch": 2.947726773555897, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.607428550720215, + "learning_rate": 1e-06, + "loss": 0.8291, + "mean_token_accuracy": 0.7410331964492798, + "num_tokens": 694535953.0, + "step": 26842 + }, + { + "epoch": 2.9478365912585107, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.265390396118164, + "learning_rate": 1e-06, + "loss": 0.812, + "mean_token_accuracy": 0.7406255006790161, + "num_tokens": 694563545.0, + "step": 26843 + }, + { + "epoch": 2.9479464089611245, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.24530029296875, + "learning_rate": 1e-06, + "loss": 0.8351, + "mean_token_accuracy": 0.7377892136573792, + "num_tokens": 694592406.0, + "step": 26844 + }, + { + "epoch": 2.9480562266637382, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.5579612255096436, + "learning_rate": 1e-06, + "loss": 0.8929, + "mean_token_accuracy": 0.7195819616317749, + "num_tokens": 694615286.0, + "step": 26845 + }, + { + "epoch": 2.948166044366352, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.2995307445526123, + "learning_rate": 1e-06, + "loss": 0.9413, + "mean_token_accuracy": 0.708477258682251, + "num_tokens": 694643591.0, + "step": 26846 + }, + { + "epoch": 2.9482758620689653, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.716765880584717, + "learning_rate": 1e-06, + "loss": 0.7825, + "mean_token_accuracy": 0.7427825927734375, + "num_tokens": 694664116.0, + "step": 26847 + }, + { + "epoch": 2.948385679771579, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4334473609924316, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7328532934188843, + "num_tokens": 694688798.0, + "step": 26848 + }, + { + "epoch": 2.948495497474193, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4938607215881348, + "learning_rate": 1e-06, + "loss": 0.7323, + "mean_token_accuracy": 0.757195234298706, + "num_tokens": 694710847.0, + "step": 26849 + }, + { + "epoch": 2.9486053151768066, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.781059503555298, + "learning_rate": 1e-06, + "loss": 0.7994, + "mean_token_accuracy": 0.7422113418579102, + "num_tokens": 694729400.0, + "step": 26850 + }, + { + "epoch": 2.9487151328794203, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.1263370513916016, + "learning_rate": 1e-06, + "loss": 0.9315, + "mean_token_accuracy": 0.7081491947174072, + "num_tokens": 694760140.0, + "step": 26851 + }, + { + "epoch": 2.9488249505820336, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.628509759902954, + "learning_rate": 1e-06, + "loss": 0.9011, + "mean_token_accuracy": 0.7226799130439758, + "num_tokens": 694782721.0, + "step": 26852 + }, + { + "epoch": 2.9489347682846474, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.343040704727173, + "learning_rate": 1e-06, + "loss": 0.8791, + "mean_token_accuracy": 0.7339352369308472, + "num_tokens": 694807509.0, + "step": 26853 + }, + { + "epoch": 2.949044585987261, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.4627182483673096, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7324690818786621, + "num_tokens": 694833524.0, + "step": 26854 + }, + { + "epoch": 2.949154403689875, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.475529193878174, + "learning_rate": 1e-06, + "loss": 0.7695, + "mean_token_accuracy": 0.7608098387718201, + "num_tokens": 694855875.0, + "step": 26855 + }, + { + "epoch": 2.9492642213924887, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.655219078063965, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.730861485004425, + "num_tokens": 694877608.0, + "step": 26856 + }, + { + "epoch": 2.949374039095102, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.342423439025879, + "learning_rate": 1e-06, + "loss": 0.848, + "mean_token_accuracy": 0.7331843376159668, + "num_tokens": 694901942.0, + "step": 26857 + }, + { + "epoch": 2.9494838567977157, + "ewc_loss": 1.6808509826660156e-05, + "grad_norm": 2.489992618560791, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.7430189251899719, + "num_tokens": 694925693.0, + "step": 26858 + }, + { + "epoch": 2.9495936745003295, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2213966846466064, + "learning_rate": 1e-06, + "loss": 0.8499, + "mean_token_accuracy": 0.7302329540252686, + "num_tokens": 694957187.0, + "step": 26859 + }, + { + "epoch": 2.9497034922029433, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4608442783355713, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7305088639259338, + "num_tokens": 694982495.0, + "step": 26860 + }, + { + "epoch": 2.949813309905557, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.442404270172119, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7398784160614014, + "num_tokens": 695006353.0, + "step": 26861 + }, + { + "epoch": 2.9499231276081703, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.388892650604248, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.742997407913208, + "num_tokens": 695030589.0, + "step": 26862 + }, + { + "epoch": 2.950032945310784, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.478454351425171, + "learning_rate": 1e-06, + "loss": 0.9198, + "mean_token_accuracy": 0.7230600714683533, + "num_tokens": 695055152.0, + "step": 26863 + }, + { + "epoch": 2.950142763013398, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5307235717773438, + "learning_rate": 1e-06, + "loss": 0.8454, + "mean_token_accuracy": 0.7397516965866089, + "num_tokens": 695077887.0, + "step": 26864 + }, + { + "epoch": 2.950252580716011, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1476922035217285, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7176963090896606, + "num_tokens": 695108433.0, + "step": 26865 + }, + { + "epoch": 2.9503623984186254, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4161460399627686, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7245723009109497, + "num_tokens": 695134430.0, + "step": 26866 + }, + { + "epoch": 2.9504722161212387, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5798916816711426, + "learning_rate": 1e-06, + "loss": 0.8669, + "mean_token_accuracy": 0.7330501079559326, + "num_tokens": 695157626.0, + "step": 26867 + }, + { + "epoch": 2.9505820338238524, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.276526689529419, + "learning_rate": 1e-06, + "loss": 0.9029, + "mean_token_accuracy": 0.7175877094268799, + "num_tokens": 695185852.0, + "step": 26868 + }, + { + "epoch": 2.950691851526466, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.333108425140381, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7241990566253662, + "num_tokens": 695213099.0, + "step": 26869 + }, + { + "epoch": 2.9508016692290795, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5026988983154297, + "learning_rate": 1e-06, + "loss": 0.9162, + "mean_token_accuracy": 0.7152483463287354, + "num_tokens": 695239514.0, + "step": 26870 + }, + { + "epoch": 2.9509114869316933, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1767473220825195, + "learning_rate": 1e-06, + "loss": 0.8621, + "mean_token_accuracy": 0.7237244844436646, + "num_tokens": 695267943.0, + "step": 26871 + }, + { + "epoch": 2.951021304634307, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.683368444442749, + "learning_rate": 1e-06, + "loss": 0.7247, + "mean_token_accuracy": 0.7660455703735352, + "num_tokens": 695287393.0, + "step": 26872 + }, + { + "epoch": 2.9511311223369208, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.455972194671631, + "learning_rate": 1e-06, + "loss": 0.8538, + "mean_token_accuracy": 0.7321261763572693, + "num_tokens": 695310829.0, + "step": 26873 + }, + { + "epoch": 2.9512409400395345, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3788540363311768, + "learning_rate": 1e-06, + "loss": 0.8148, + "mean_token_accuracy": 0.7454842329025269, + "num_tokens": 695335851.0, + "step": 26874 + }, + { + "epoch": 2.951350757742148, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4079418182373047, + "learning_rate": 1e-06, + "loss": 0.8432, + "mean_token_accuracy": 0.7317233681678772, + "num_tokens": 695359724.0, + "step": 26875 + }, + { + "epoch": 2.9514605754447616, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.437187671661377, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7245519161224365, + "num_tokens": 695383927.0, + "step": 26876 + }, + { + "epoch": 2.9515703931473753, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.077052593231201, + "learning_rate": 1e-06, + "loss": 0.8576, + "mean_token_accuracy": 0.7331739068031311, + "num_tokens": 695415377.0, + "step": 26877 + }, + { + "epoch": 2.951680210849989, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4200875759124756, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7096804976463318, + "num_tokens": 695444063.0, + "step": 26878 + }, + { + "epoch": 2.951790028552603, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4647629261016846, + "learning_rate": 1e-06, + "loss": 0.8156, + "mean_token_accuracy": 0.7407183051109314, + "num_tokens": 695468197.0, + "step": 26879 + }, + { + "epoch": 2.951899846255216, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.168895721435547, + "learning_rate": 1e-06, + "loss": 0.9592, + "mean_token_accuracy": 0.7030191421508789, + "num_tokens": 695498017.0, + "step": 26880 + }, + { + "epoch": 2.95200966395783, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4360361099243164, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7055973410606384, + "num_tokens": 695523901.0, + "step": 26881 + }, + { + "epoch": 2.9521194816604437, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2934460639953613, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7317723035812378, + "num_tokens": 695552061.0, + "step": 26882 + }, + { + "epoch": 2.9522292993630574, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2249374389648438, + "learning_rate": 1e-06, + "loss": 0.963, + "mean_token_accuracy": 0.7016709446907043, + "num_tokens": 695582133.0, + "step": 26883 + }, + { + "epoch": 2.952339117065671, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2344179153442383, + "learning_rate": 1e-06, + "loss": 0.8524, + "mean_token_accuracy": 0.727107048034668, + "num_tokens": 695609439.0, + "step": 26884 + }, + { + "epoch": 2.9524489347682845, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5615639686584473, + "learning_rate": 1e-06, + "loss": 0.9092, + "mean_token_accuracy": 0.7284379601478577, + "num_tokens": 695634479.0, + "step": 26885 + }, + { + "epoch": 2.9525587524708983, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4203152656555176, + "learning_rate": 1e-06, + "loss": 0.8565, + "mean_token_accuracy": 0.7210280299186707, + "num_tokens": 695659018.0, + "step": 26886 + }, + { + "epoch": 2.952668570173512, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2191879749298096, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7294211387634277, + "num_tokens": 695685941.0, + "step": 26887 + }, + { + "epoch": 2.952778387876126, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7232370376586914, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7241075038909912, + "num_tokens": 695706360.0, + "step": 26888 + }, + { + "epoch": 2.9528882055787395, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.269341468811035, + "learning_rate": 1e-06, + "loss": 0.8232, + "mean_token_accuracy": 0.7422528266906738, + "num_tokens": 695736570.0, + "step": 26889 + }, + { + "epoch": 2.952998023281353, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5418245792388916, + "learning_rate": 1e-06, + "loss": 0.8209, + "mean_token_accuracy": 0.7382752895355225, + "num_tokens": 695758780.0, + "step": 26890 + }, + { + "epoch": 2.9531078409839666, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.619934558868408, + "learning_rate": 1e-06, + "loss": 0.8593, + "mean_token_accuracy": 0.7290880680084229, + "num_tokens": 695781814.0, + "step": 26891 + }, + { + "epoch": 2.9532176586865804, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.009627103805542, + "learning_rate": 1e-06, + "loss": 0.8217, + "mean_token_accuracy": 0.7373145818710327, + "num_tokens": 695816254.0, + "step": 26892 + }, + { + "epoch": 2.9533274763891937, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4048755168914795, + "learning_rate": 1e-06, + "loss": 0.8748, + "mean_token_accuracy": 0.7262604832649231, + "num_tokens": 695841352.0, + "step": 26893 + }, + { + "epoch": 2.9534372940918074, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.258570432662964, + "learning_rate": 1e-06, + "loss": 0.8775, + "mean_token_accuracy": 0.722074568271637, + "num_tokens": 695869778.0, + "step": 26894 + }, + { + "epoch": 2.953547111794421, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4321365356445312, + "learning_rate": 1e-06, + "loss": 0.8717, + "mean_token_accuracy": 0.7236514687538147, + "num_tokens": 695895169.0, + "step": 26895 + }, + { + "epoch": 2.953656929497035, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0101234912872314, + "learning_rate": 1e-06, + "loss": 0.8313, + "mean_token_accuracy": 0.7417578101158142, + "num_tokens": 695929771.0, + "step": 26896 + }, + { + "epoch": 2.9537667471996487, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.522313117980957, + "learning_rate": 1e-06, + "loss": 0.8192, + "mean_token_accuracy": 0.7390441298484802, + "num_tokens": 695952223.0, + "step": 26897 + }, + { + "epoch": 2.953876564902262, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0556960105895996, + "learning_rate": 1e-06, + "loss": 0.932, + "mean_token_accuracy": 0.7150388360023499, + "num_tokens": 695983239.0, + "step": 26898 + }, + { + "epoch": 2.953986382604876, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.423335313796997, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7307546138763428, + "num_tokens": 696007787.0, + "step": 26899 + }, + { + "epoch": 2.9540962003074895, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4835898876190186, + "learning_rate": 1e-06, + "loss": 0.8424, + "mean_token_accuracy": 0.7332545518875122, + "num_tokens": 696032528.0, + "step": 26900 + }, + { + "epoch": 2.9542060180101033, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.290040969848633, + "learning_rate": 1e-06, + "loss": 0.9122, + "mean_token_accuracy": 0.709456205368042, + "num_tokens": 696061706.0, + "step": 26901 + }, + { + "epoch": 2.954315835712717, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2424464225769043, + "learning_rate": 1e-06, + "loss": 0.8624, + "mean_token_accuracy": 0.7321849465370178, + "num_tokens": 696090386.0, + "step": 26902 + }, + { + "epoch": 2.9544256534153304, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.337066650390625, + "learning_rate": 1e-06, + "loss": 0.8038, + "mean_token_accuracy": 0.7414292097091675, + "num_tokens": 696114610.0, + "step": 26903 + }, + { + "epoch": 2.954535471117944, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.021681785583496, + "learning_rate": 1e-06, + "loss": 0.8376, + "mean_token_accuracy": 0.7394620776176453, + "num_tokens": 696147123.0, + "step": 26904 + }, + { + "epoch": 2.954645288820558, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.167036771774292, + "learning_rate": 1e-06, + "loss": 0.8299, + "mean_token_accuracy": 0.7340385913848877, + "num_tokens": 696176382.0, + "step": 26905 + }, + { + "epoch": 2.9547551065231716, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3964335918426514, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7265681624412537, + "num_tokens": 696207007.0, + "step": 26906 + }, + { + "epoch": 2.9548649242257854, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3742051124572754, + "learning_rate": 1e-06, + "loss": 0.8595, + "mean_token_accuracy": 0.7336466312408447, + "num_tokens": 696233648.0, + "step": 26907 + }, + { + "epoch": 2.9549747419283987, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.494676351547241, + "learning_rate": 1e-06, + "loss": 0.8481, + "mean_token_accuracy": 0.7318531274795532, + "num_tokens": 696261850.0, + "step": 26908 + }, + { + "epoch": 2.9550845596310125, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.66068172454834, + "learning_rate": 1e-06, + "loss": 0.8046, + "mean_token_accuracy": 0.7513240575790405, + "num_tokens": 696283083.0, + "step": 26909 + }, + { + "epoch": 2.955194377333626, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0409135818481445, + "learning_rate": 1e-06, + "loss": 0.9502, + "mean_token_accuracy": 0.6996122598648071, + "num_tokens": 696318250.0, + "step": 26910 + }, + { + "epoch": 2.95530419503624, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6111643314361572, + "learning_rate": 1e-06, + "loss": 0.8179, + "mean_token_accuracy": 0.7422484159469604, + "num_tokens": 696338696.0, + "step": 26911 + }, + { + "epoch": 2.9554140127388537, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7658421993255615, + "learning_rate": 1e-06, + "loss": 0.8429, + "mean_token_accuracy": 0.7441949248313904, + "num_tokens": 696359347.0, + "step": 26912 + }, + { + "epoch": 2.955523830441467, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6702051162719727, + "learning_rate": 1e-06, + "loss": 0.7707, + "mean_token_accuracy": 0.7503631114959717, + "num_tokens": 696379979.0, + "step": 26913 + }, + { + "epoch": 2.955633648144081, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.217830181121826, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7149019241333008, + "num_tokens": 696410744.0, + "step": 26914 + }, + { + "epoch": 2.9557434658466946, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1394455432891846, + "learning_rate": 1e-06, + "loss": 0.899, + "mean_token_accuracy": 0.7236579060554504, + "num_tokens": 696442078.0, + "step": 26915 + }, + { + "epoch": 2.955853283549308, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5749447345733643, + "learning_rate": 1e-06, + "loss": 0.8193, + "mean_token_accuracy": 0.7380969524383545, + "num_tokens": 696464422.0, + "step": 26916 + }, + { + "epoch": 2.955963101251922, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4465854167938232, + "learning_rate": 1e-06, + "loss": 0.8254, + "mean_token_accuracy": 0.7409430742263794, + "num_tokens": 696488774.0, + "step": 26917 + }, + { + "epoch": 2.9560729189545354, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2702910900115967, + "learning_rate": 1e-06, + "loss": 0.9265, + "mean_token_accuracy": 0.7141796350479126, + "num_tokens": 696516657.0, + "step": 26918 + }, + { + "epoch": 2.956182736657149, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7035279273986816, + "learning_rate": 1e-06, + "loss": 0.8605, + "mean_token_accuracy": 0.7321181297302246, + "num_tokens": 696537618.0, + "step": 26919 + }, + { + "epoch": 2.956292554359763, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.33595609664917, + "learning_rate": 1e-06, + "loss": 0.7876, + "mean_token_accuracy": 0.7519935369491577, + "num_tokens": 696562197.0, + "step": 26920 + }, + { + "epoch": 2.956402372062376, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3618075847625732, + "learning_rate": 1e-06, + "loss": 0.8813, + "mean_token_accuracy": 0.7233287692070007, + "num_tokens": 696588380.0, + "step": 26921 + }, + { + "epoch": 2.95651218976499, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.421008825302124, + "learning_rate": 1e-06, + "loss": 0.8927, + "mean_token_accuracy": 0.7278263568878174, + "num_tokens": 696612837.0, + "step": 26922 + }, + { + "epoch": 2.9566220074676037, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6004960536956787, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7348557710647583, + "num_tokens": 696633842.0, + "step": 26923 + }, + { + "epoch": 2.9567318251702175, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.270749807357788, + "learning_rate": 1e-06, + "loss": 0.8482, + "mean_token_accuracy": 0.730139970779419, + "num_tokens": 696660609.0, + "step": 26924 + }, + { + "epoch": 2.9568416428728312, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2772347927093506, + "learning_rate": 1e-06, + "loss": 0.9172, + "mean_token_accuracy": 0.7224862575531006, + "num_tokens": 696688257.0, + "step": 26925 + }, + { + "epoch": 2.9569514605754446, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4759490489959717, + "learning_rate": 1e-06, + "loss": 0.8537, + "mean_token_accuracy": 0.7461181282997131, + "num_tokens": 696712726.0, + "step": 26926 + }, + { + "epoch": 2.9570612782780583, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0407168865203857, + "learning_rate": 1e-06, + "loss": 0.8414, + "mean_token_accuracy": 0.7369449138641357, + "num_tokens": 696743551.0, + "step": 26927 + }, + { + "epoch": 2.957171095980672, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1167123317718506, + "learning_rate": 1e-06, + "loss": 0.7917, + "mean_token_accuracy": 0.7458015084266663, + "num_tokens": 696772162.0, + "step": 26928 + }, + { + "epoch": 2.957280913683286, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.341297149658203, + "learning_rate": 1e-06, + "loss": 0.8355, + "mean_token_accuracy": 0.7397180795669556, + "num_tokens": 696796890.0, + "step": 26929 + }, + { + "epoch": 2.9573907313858996, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5271713733673096, + "learning_rate": 1e-06, + "loss": 0.8718, + "mean_token_accuracy": 0.7288353443145752, + "num_tokens": 696821425.0, + "step": 26930 + }, + { + "epoch": 2.957500549088513, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3281407356262207, + "learning_rate": 1e-06, + "loss": 0.8784, + "mean_token_accuracy": 0.7300076484680176, + "num_tokens": 696849118.0, + "step": 26931 + }, + { + "epoch": 2.9576103667911267, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3356130123138428, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.718658447265625, + "num_tokens": 696875342.0, + "step": 26932 + }, + { + "epoch": 2.9577201844937404, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4163818359375, + "learning_rate": 1e-06, + "loss": 0.8613, + "mean_token_accuracy": 0.7284852266311646, + "num_tokens": 696900139.0, + "step": 26933 + }, + { + "epoch": 2.957830002196354, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4429705142974854, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7169933915138245, + "num_tokens": 696925544.0, + "step": 26934 + }, + { + "epoch": 2.957939819898968, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.168869972229004, + "learning_rate": 1e-06, + "loss": 0.8808, + "mean_token_accuracy": 0.729651153087616, + "num_tokens": 696957200.0, + "step": 26935 + }, + { + "epoch": 2.9580496376015812, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3188273906707764, + "learning_rate": 1e-06, + "loss": 0.8976, + "mean_token_accuracy": 0.7175996899604797, + "num_tokens": 696984092.0, + "step": 26936 + }, + { + "epoch": 2.958159455304195, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.518606185913086, + "learning_rate": 1e-06, + "loss": 0.8084, + "mean_token_accuracy": 0.7439783811569214, + "num_tokens": 697009435.0, + "step": 26937 + }, + { + "epoch": 2.9582692730068088, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.120840072631836, + "learning_rate": 1e-06, + "loss": 0.8648, + "mean_token_accuracy": 0.7384920716285706, + "num_tokens": 697040127.0, + "step": 26938 + }, + { + "epoch": 2.9583790907094225, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1406283378601074, + "learning_rate": 1e-06, + "loss": 0.8825, + "mean_token_accuracy": 0.7282792329788208, + "num_tokens": 697068990.0, + "step": 26939 + }, + { + "epoch": 2.9584889084120363, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4198808670043945, + "learning_rate": 1e-06, + "loss": 0.861, + "mean_token_accuracy": 0.7243326902389526, + "num_tokens": 697093253.0, + "step": 26940 + }, + { + "epoch": 2.9585987261146496, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.036329507827759, + "learning_rate": 1e-06, + "loss": 0.9087, + "mean_token_accuracy": 0.7168013453483582, + "num_tokens": 697127762.0, + "step": 26941 + }, + { + "epoch": 2.9587085438172633, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4001829624176025, + "learning_rate": 1e-06, + "loss": 0.8793, + "mean_token_accuracy": 0.7177054286003113, + "num_tokens": 697152355.0, + "step": 26942 + }, + { + "epoch": 2.958818361519877, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3402891159057617, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7520028352737427, + "num_tokens": 697177570.0, + "step": 26943 + }, + { + "epoch": 2.9589281792224904, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.252901792526245, + "learning_rate": 1e-06, + "loss": 0.7287, + "mean_token_accuracy": 0.7632997035980225, + "num_tokens": 697203614.0, + "step": 26944 + }, + { + "epoch": 2.959037996925104, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4934511184692383, + "learning_rate": 1e-06, + "loss": 0.8045, + "mean_token_accuracy": 0.7474057674407959, + "num_tokens": 697227800.0, + "step": 26945 + }, + { + "epoch": 2.959147814627718, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.814284324645996, + "learning_rate": 1e-06, + "loss": 0.8687, + "mean_token_accuracy": 0.732343852519989, + "num_tokens": 697249865.0, + "step": 26946 + }, + { + "epoch": 2.9592576323303317, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.599031686782837, + "learning_rate": 1e-06, + "loss": 0.8349, + "mean_token_accuracy": 0.7325245141983032, + "num_tokens": 697272909.0, + "step": 26947 + }, + { + "epoch": 2.9593674500329454, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1914522647857666, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7480810880661011, + "num_tokens": 697300674.0, + "step": 26948 + }, + { + "epoch": 2.9594772677355587, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3391408920288086, + "learning_rate": 1e-06, + "loss": 0.7503, + "mean_token_accuracy": 0.7640613317489624, + "num_tokens": 697327265.0, + "step": 26949 + }, + { + "epoch": 2.9595870854381725, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3068127632141113, + "learning_rate": 1e-06, + "loss": 0.7583, + "mean_token_accuracy": 0.7549300789833069, + "num_tokens": 697352568.0, + "step": 26950 + }, + { + "epoch": 2.9596969031407863, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0467042922973633, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7246520519256592, + "num_tokens": 697383626.0, + "step": 26951 + }, + { + "epoch": 2.9598067208434, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.589398145675659, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7307015657424927, + "num_tokens": 697406230.0, + "step": 26952 + }, + { + "epoch": 2.9599165385460138, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5322914123535156, + "learning_rate": 1e-06, + "loss": 0.8301, + "mean_token_accuracy": 0.7417454719543457, + "num_tokens": 697428583.0, + "step": 26953 + }, + { + "epoch": 2.960026356248627, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6221933364868164, + "learning_rate": 1e-06, + "loss": 0.8707, + "mean_token_accuracy": 0.7314473390579224, + "num_tokens": 697450490.0, + "step": 26954 + }, + { + "epoch": 2.960136173951241, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5205466747283936, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.718598484992981, + "num_tokens": 697474413.0, + "step": 26955 + }, + { + "epoch": 2.9602459916538546, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.369419574737549, + "learning_rate": 1e-06, + "loss": 0.9109, + "mean_token_accuracy": 0.7125803232192993, + "num_tokens": 697500328.0, + "step": 26956 + }, + { + "epoch": 2.9603558093564684, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.093048334121704, + "learning_rate": 1e-06, + "loss": 0.8226, + "mean_token_accuracy": 0.7456047534942627, + "num_tokens": 697531539.0, + "step": 26957 + }, + { + "epoch": 2.960465627059082, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.288409471511841, + "learning_rate": 1e-06, + "loss": 0.9301, + "mean_token_accuracy": 0.7127457857131958, + "num_tokens": 697557611.0, + "step": 26958 + }, + { + "epoch": 2.9605754447616954, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.331737518310547, + "learning_rate": 1e-06, + "loss": 0.8637, + "mean_token_accuracy": 0.7266567945480347, + "num_tokens": 697584171.0, + "step": 26959 + }, + { + "epoch": 2.960685262464309, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.323713779449463, + "learning_rate": 1e-06, + "loss": 0.8737, + "mean_token_accuracy": 0.7326741814613342, + "num_tokens": 697610562.0, + "step": 26960 + }, + { + "epoch": 2.960795080166923, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.610966682434082, + "learning_rate": 1e-06, + "loss": 0.8436, + "mean_token_accuracy": 0.7391495704650879, + "num_tokens": 697632479.0, + "step": 26961 + }, + { + "epoch": 2.9609048978695367, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.27034068107605, + "learning_rate": 1e-06, + "loss": 0.8746, + "mean_token_accuracy": 0.7249352335929871, + "num_tokens": 697659337.0, + "step": 26962 + }, + { + "epoch": 2.9610147155721505, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.076934814453125, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7332316637039185, + "num_tokens": 697692917.0, + "step": 26963 + }, + { + "epoch": 2.9611245332747638, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.133854866027832, + "learning_rate": 1e-06, + "loss": 0.9129, + "mean_token_accuracy": 0.7201039791107178, + "num_tokens": 697722169.0, + "step": 26964 + }, + { + "epoch": 2.9612343509773775, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2855963706970215, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7294812798500061, + "num_tokens": 697750091.0, + "step": 26965 + }, + { + "epoch": 2.9613441686799913, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.451359510421753, + "learning_rate": 1e-06, + "loss": 0.9055, + "mean_token_accuracy": 0.720658540725708, + "num_tokens": 697776446.0, + "step": 26966 + }, + { + "epoch": 2.961453986382605, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.718311071395874, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.7280370593070984, + "num_tokens": 697796895.0, + "step": 26967 + }, + { + "epoch": 2.961563804085219, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2068653106689453, + "learning_rate": 1e-06, + "loss": 0.9629, + "mean_token_accuracy": 0.707787036895752, + "num_tokens": 697827015.0, + "step": 26968 + }, + { + "epoch": 2.961673621787832, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.672593832015991, + "learning_rate": 1e-06, + "loss": 0.796, + "mean_token_accuracy": 0.7618470191955566, + "num_tokens": 697849556.0, + "step": 26969 + }, + { + "epoch": 2.961783439490446, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.399141311645508, + "learning_rate": 1e-06, + "loss": 0.7908, + "mean_token_accuracy": 0.7505124807357788, + "num_tokens": 697873031.0, + "step": 26970 + }, + { + "epoch": 2.9618932571930596, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.277885675430298, + "learning_rate": 1e-06, + "loss": 0.8615, + "mean_token_accuracy": 0.7246791124343872, + "num_tokens": 697902202.0, + "step": 26971 + }, + { + "epoch": 2.962003074895673, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.24442195892334, + "learning_rate": 1e-06, + "loss": 0.8352, + "mean_token_accuracy": 0.7399744391441345, + "num_tokens": 697929995.0, + "step": 26972 + }, + { + "epoch": 2.9621128925982867, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.39152193069458, + "learning_rate": 1e-06, + "loss": 0.8078, + "mean_token_accuracy": 0.7399486303329468, + "num_tokens": 697954809.0, + "step": 26973 + }, + { + "epoch": 2.9622227103009005, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3607304096221924, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7056117653846741, + "num_tokens": 697982219.0, + "step": 26974 + }, + { + "epoch": 2.962332528003514, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2657759189605713, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7218124866485596, + "num_tokens": 698010780.0, + "step": 26975 + }, + { + "epoch": 2.962442345706128, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2555112838745117, + "learning_rate": 1e-06, + "loss": 0.8318, + "mean_token_accuracy": 0.7409254312515259, + "num_tokens": 698036413.0, + "step": 26976 + }, + { + "epoch": 2.9625521634087413, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.290480136871338, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7445294260978699, + "num_tokens": 698062900.0, + "step": 26977 + }, + { + "epoch": 2.962661981111355, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5297765731811523, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7230740785598755, + "num_tokens": 698087121.0, + "step": 26978 + }, + { + "epoch": 2.962771798813969, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.36897349357605, + "learning_rate": 1e-06, + "loss": 0.8975, + "mean_token_accuracy": 0.712504506111145, + "num_tokens": 698113106.0, + "step": 26979 + }, + { + "epoch": 2.9628816165165826, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3461852073669434, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7441446781158447, + "num_tokens": 698136953.0, + "step": 26980 + }, + { + "epoch": 2.9629914342191963, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5905847549438477, + "learning_rate": 1e-06, + "loss": 0.908, + "mean_token_accuracy": 0.7098324298858643, + "num_tokens": 698160688.0, + "step": 26981 + }, + { + "epoch": 2.9631012519218096, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3846843242645264, + "learning_rate": 1e-06, + "loss": 0.8373, + "mean_token_accuracy": 0.7402467727661133, + "num_tokens": 698186106.0, + "step": 26982 + }, + { + "epoch": 2.9632110696244234, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3180551528930664, + "learning_rate": 1e-06, + "loss": 0.8418, + "mean_token_accuracy": 0.7331240773200989, + "num_tokens": 698213086.0, + "step": 26983 + }, + { + "epoch": 2.963320887327037, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.088895082473755, + "learning_rate": 1e-06, + "loss": 0.8158, + "mean_token_accuracy": 0.7466757297515869, + "num_tokens": 698243181.0, + "step": 26984 + }, + { + "epoch": 2.963430705029651, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4861035346984863, + "learning_rate": 1e-06, + "loss": 0.8178, + "mean_token_accuracy": 0.7413027286529541, + "num_tokens": 698267595.0, + "step": 26985 + }, + { + "epoch": 2.9635405227322646, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5218238830566406, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.7483959794044495, + "num_tokens": 698291108.0, + "step": 26986 + }, + { + "epoch": 2.963650340434878, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1651768684387207, + "learning_rate": 1e-06, + "loss": 0.8334, + "mean_token_accuracy": 0.7343388199806213, + "num_tokens": 698321436.0, + "step": 26987 + }, + { + "epoch": 2.9637601581374917, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.391779661178589, + "learning_rate": 1e-06, + "loss": 0.8706, + "mean_token_accuracy": 0.7352007627487183, + "num_tokens": 698348173.0, + "step": 26988 + }, + { + "epoch": 2.9638699758401055, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.089495897293091, + "learning_rate": 1e-06, + "loss": 0.9624, + "mean_token_accuracy": 0.7047619223594666, + "num_tokens": 698382618.0, + "step": 26989 + }, + { + "epoch": 2.9639797935427192, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 1.9699243307113647, + "learning_rate": 1e-06, + "loss": 0.9054, + "mean_token_accuracy": 0.7217203378677368, + "num_tokens": 698419430.0, + "step": 26990 + }, + { + "epoch": 2.964089611245333, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6137561798095703, + "learning_rate": 1e-06, + "loss": 0.7974, + "mean_token_accuracy": 0.748515248298645, + "num_tokens": 698440289.0, + "step": 26991 + }, + { + "epoch": 2.9641994289479463, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4242751598358154, + "learning_rate": 1e-06, + "loss": 0.8484, + "mean_token_accuracy": 0.7309686541557312, + "num_tokens": 698468329.0, + "step": 26992 + }, + { + "epoch": 2.96430924665056, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.272942543029785, + "learning_rate": 1e-06, + "loss": 0.8265, + "mean_token_accuracy": 0.7347793579101562, + "num_tokens": 698493506.0, + "step": 26993 + }, + { + "epoch": 2.964419064353174, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.734812021255493, + "learning_rate": 1e-06, + "loss": 0.78, + "mean_token_accuracy": 0.7511531114578247, + "num_tokens": 698512949.0, + "step": 26994 + }, + { + "epoch": 2.964528882055787, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.328296422958374, + "learning_rate": 1e-06, + "loss": 0.9121, + "mean_token_accuracy": 0.7217038869857788, + "num_tokens": 698540485.0, + "step": 26995 + }, + { + "epoch": 2.9646386997584013, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.272594928741455, + "learning_rate": 1e-06, + "loss": 0.9433, + "mean_token_accuracy": 0.7070915699005127, + "num_tokens": 698568590.0, + "step": 26996 + }, + { + "epoch": 2.9647485174610146, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.383985996246338, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7311279773712158, + "num_tokens": 698595857.0, + "step": 26997 + }, + { + "epoch": 2.9648583351636284, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0453829765319824, + "learning_rate": 1e-06, + "loss": 0.9291, + "mean_token_accuracy": 0.7104036808013916, + "num_tokens": 698630604.0, + "step": 26998 + }, + { + "epoch": 2.964968152866242, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4811253547668457, + "learning_rate": 1e-06, + "loss": 0.8261, + "mean_token_accuracy": 0.7476527094841003, + "num_tokens": 698655461.0, + "step": 26999 + }, + { + "epoch": 2.9650779705688555, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.0740694999694824, + "learning_rate": 1e-06, + "loss": 0.928, + "mean_token_accuracy": 0.7101910710334778, + "num_tokens": 698690583.0, + "step": 27000 + }, + { + "epoch": 2.9651877882714692, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.229884386062622, + "learning_rate": 1e-06, + "loss": 0.9306, + "mean_token_accuracy": 0.7151855230331421, + "num_tokens": 698717539.0, + "step": 27001 + }, + { + "epoch": 2.965297605974083, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5426108837127686, + "learning_rate": 1e-06, + "loss": 0.8014, + "mean_token_accuracy": 0.7474466562271118, + "num_tokens": 698740933.0, + "step": 27002 + }, + { + "epoch": 2.9654074236766967, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.416865110397339, + "learning_rate": 1e-06, + "loss": 0.8379, + "mean_token_accuracy": 0.7343142032623291, + "num_tokens": 698767324.0, + "step": 27003 + }, + { + "epoch": 2.9655172413793105, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2833008766174316, + "learning_rate": 1e-06, + "loss": 0.8709, + "mean_token_accuracy": 0.7252926230430603, + "num_tokens": 698795914.0, + "step": 27004 + }, + { + "epoch": 2.965627059081924, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.500661849975586, + "learning_rate": 1e-06, + "loss": 0.8804, + "mean_token_accuracy": 0.7271684408187866, + "num_tokens": 698821613.0, + "step": 27005 + }, + { + "epoch": 2.9657368767845376, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6461288928985596, + "learning_rate": 1e-06, + "loss": 0.8743, + "mean_token_accuracy": 0.7270790934562683, + "num_tokens": 698842556.0, + "step": 27006 + }, + { + "epoch": 2.9658466944871513, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3144516944885254, + "learning_rate": 1e-06, + "loss": 0.8911, + "mean_token_accuracy": 0.730167031288147, + "num_tokens": 698871271.0, + "step": 27007 + }, + { + "epoch": 2.965956512189765, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3316245079040527, + "learning_rate": 1e-06, + "loss": 0.9201, + "mean_token_accuracy": 0.7130194902420044, + "num_tokens": 698897929.0, + "step": 27008 + }, + { + "epoch": 2.966066329892379, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4755349159240723, + "learning_rate": 1e-06, + "loss": 0.906, + "mean_token_accuracy": 0.7149890661239624, + "num_tokens": 698922956.0, + "step": 27009 + }, + { + "epoch": 2.966176147594992, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1480042934417725, + "learning_rate": 1e-06, + "loss": 0.805, + "mean_token_accuracy": 0.7423220276832581, + "num_tokens": 698952848.0, + "step": 27010 + }, + { + "epoch": 2.966285965297606, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4012608528137207, + "learning_rate": 1e-06, + "loss": 0.8359, + "mean_token_accuracy": 0.736810564994812, + "num_tokens": 698977153.0, + "step": 27011 + }, + { + "epoch": 2.9663957830002197, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.898864269256592, + "learning_rate": 1e-06, + "loss": 0.7277, + "mean_token_accuracy": 0.7663002014160156, + "num_tokens": 698996246.0, + "step": 27012 + }, + { + "epoch": 2.9665056007028334, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.585198163986206, + "learning_rate": 1e-06, + "loss": 0.8375, + "mean_token_accuracy": 0.7369269132614136, + "num_tokens": 699018202.0, + "step": 27013 + }, + { + "epoch": 2.966615418405447, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6731300354003906, + "learning_rate": 1e-06, + "loss": 0.826, + "mean_token_accuracy": 0.7325908541679382, + "num_tokens": 699039269.0, + "step": 27014 + }, + { + "epoch": 2.9667252361080605, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.220069646835327, + "learning_rate": 1e-06, + "loss": 0.8772, + "mean_token_accuracy": 0.7217787504196167, + "num_tokens": 699067947.0, + "step": 27015 + }, + { + "epoch": 2.9668350538106742, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.341395616531372, + "learning_rate": 1e-06, + "loss": 0.8617, + "mean_token_accuracy": 0.7268064022064209, + "num_tokens": 699093164.0, + "step": 27016 + }, + { + "epoch": 2.966944871513288, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4367189407348633, + "learning_rate": 1e-06, + "loss": 0.765, + "mean_token_accuracy": 0.7508256435394287, + "num_tokens": 699116466.0, + "step": 27017 + }, + { + "epoch": 2.9670546892159018, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5291059017181396, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7597103118896484, + "num_tokens": 699138646.0, + "step": 27018 + }, + { + "epoch": 2.9671645069185155, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.408085823059082, + "learning_rate": 1e-06, + "loss": 0.8162, + "mean_token_accuracy": 0.7348225116729736, + "num_tokens": 699163571.0, + "step": 27019 + }, + { + "epoch": 2.967274324621129, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1657917499542236, + "learning_rate": 1e-06, + "loss": 0.9303, + "mean_token_accuracy": 0.7182433009147644, + "num_tokens": 699194222.0, + "step": 27020 + }, + { + "epoch": 2.9673841423237426, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.076409339904785, + "learning_rate": 1e-06, + "loss": 0.8535, + "mean_token_accuracy": 0.7274723052978516, + "num_tokens": 699225044.0, + "step": 27021 + }, + { + "epoch": 2.9674939600263563, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.204714775085449, + "learning_rate": 1e-06, + "loss": 0.9078, + "mean_token_accuracy": 0.7131905555725098, + "num_tokens": 699254793.0, + "step": 27022 + }, + { + "epoch": 2.9676037777289697, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5390584468841553, + "learning_rate": 1e-06, + "loss": 0.8489, + "mean_token_accuracy": 0.7315698862075806, + "num_tokens": 699276903.0, + "step": 27023 + }, + { + "epoch": 2.9677135954315834, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.149277687072754, + "learning_rate": 1e-06, + "loss": 0.8087, + "mean_token_accuracy": 0.7433719635009766, + "num_tokens": 699306065.0, + "step": 27024 + }, + { + "epoch": 2.967823413134197, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7514588832855225, + "learning_rate": 1e-06, + "loss": 0.9046, + "mean_token_accuracy": 0.7268123626708984, + "num_tokens": 699329481.0, + "step": 27025 + }, + { + "epoch": 2.967933230836811, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1148715019226074, + "learning_rate": 1e-06, + "loss": 0.9174, + "mean_token_accuracy": 0.712742030620575, + "num_tokens": 699361293.0, + "step": 27026 + }, + { + "epoch": 2.9680430485394247, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.370148181915283, + "learning_rate": 1e-06, + "loss": 0.81, + "mean_token_accuracy": 0.7419456243515015, + "num_tokens": 699387539.0, + "step": 27027 + }, + { + "epoch": 2.968152866242038, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7232344150543213, + "learning_rate": 1e-06, + "loss": 0.839, + "mean_token_accuracy": 0.7440975904464722, + "num_tokens": 699407609.0, + "step": 27028 + }, + { + "epoch": 2.9682626839446518, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2360756397247314, + "learning_rate": 1e-06, + "loss": 0.8874, + "mean_token_accuracy": 0.7229700684547424, + "num_tokens": 699435534.0, + "step": 27029 + }, + { + "epoch": 2.9683725016472655, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.434760332107544, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7291349172592163, + "num_tokens": 699460880.0, + "step": 27030 + }, + { + "epoch": 2.9684823193498793, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.6039469242095947, + "learning_rate": 1e-06, + "loss": 0.7521, + "mean_token_accuracy": 0.7564568519592285, + "num_tokens": 699481884.0, + "step": 27031 + }, + { + "epoch": 2.968592137052493, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4196507930755615, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7343617677688599, + "num_tokens": 699506015.0, + "step": 27032 + }, + { + "epoch": 2.9687019547551063, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.361971616744995, + "learning_rate": 1e-06, + "loss": 0.8887, + "mean_token_accuracy": 0.7224836349487305, + "num_tokens": 699532009.0, + "step": 27033 + }, + { + "epoch": 2.96881177245772, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.546071767807007, + "learning_rate": 1e-06, + "loss": 0.8586, + "mean_token_accuracy": 0.7350874543190002, + "num_tokens": 699553030.0, + "step": 27034 + }, + { + "epoch": 2.968921590160334, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.268573045730591, + "learning_rate": 1e-06, + "loss": 0.8442, + "mean_token_accuracy": 0.7407246828079224, + "num_tokens": 699579827.0, + "step": 27035 + }, + { + "epoch": 2.9690314078629476, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2723827362060547, + "learning_rate": 1e-06, + "loss": 0.8659, + "mean_token_accuracy": 0.7286579608917236, + "num_tokens": 699606097.0, + "step": 27036 + }, + { + "epoch": 2.9691412255655614, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.428314685821533, + "learning_rate": 1e-06, + "loss": 0.8025, + "mean_token_accuracy": 0.745887815952301, + "num_tokens": 699628139.0, + "step": 27037 + }, + { + "epoch": 2.9692510432681747, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.282217025756836, + "learning_rate": 1e-06, + "loss": 0.8715, + "mean_token_accuracy": 0.726559042930603, + "num_tokens": 699656894.0, + "step": 27038 + }, + { + "epoch": 2.9693608609707884, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5060629844665527, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.725142240524292, + "num_tokens": 699682476.0, + "step": 27039 + }, + { + "epoch": 2.969470678673402, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.400294065475464, + "learning_rate": 1e-06, + "loss": 0.8722, + "mean_token_accuracy": 0.7255313396453857, + "num_tokens": 699708794.0, + "step": 27040 + }, + { + "epoch": 2.969580496376016, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5638978481292725, + "learning_rate": 1e-06, + "loss": 0.7874, + "mean_token_accuracy": 0.7459431886672974, + "num_tokens": 699731205.0, + "step": 27041 + }, + { + "epoch": 2.9696903140786297, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.302234411239624, + "learning_rate": 1e-06, + "loss": 0.8474, + "mean_token_accuracy": 0.7394601702690125, + "num_tokens": 699757450.0, + "step": 27042 + }, + { + "epoch": 2.969800131781243, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5970566272735596, + "learning_rate": 1e-06, + "loss": 0.7733, + "mean_token_accuracy": 0.7482736110687256, + "num_tokens": 699778921.0, + "step": 27043 + }, + { + "epoch": 2.969909949483857, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.231353521347046, + "learning_rate": 1e-06, + "loss": 0.8403, + "mean_token_accuracy": 0.7303091883659363, + "num_tokens": 699808168.0, + "step": 27044 + }, + { + "epoch": 2.9700197671864705, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2486419677734375, + "learning_rate": 1e-06, + "loss": 0.8871, + "mean_token_accuracy": 0.7293479442596436, + "num_tokens": 699838825.0, + "step": 27045 + }, + { + "epoch": 2.970129584889084, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.259772300720215, + "learning_rate": 1e-06, + "loss": 0.8932, + "mean_token_accuracy": 0.7237293124198914, + "num_tokens": 699866348.0, + "step": 27046 + }, + { + "epoch": 2.970239402591698, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4709219932556152, + "learning_rate": 1e-06, + "loss": 0.7702, + "mean_token_accuracy": 0.7460028529167175, + "num_tokens": 699889755.0, + "step": 27047 + }, + { + "epoch": 2.9703492202943114, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.139674186706543, + "learning_rate": 1e-06, + "loss": 0.9454, + "mean_token_accuracy": 0.7103871703147888, + "num_tokens": 699922208.0, + "step": 27048 + }, + { + "epoch": 2.970459037996925, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.8304946422576904, + "learning_rate": 1e-06, + "loss": 0.8782, + "mean_token_accuracy": 0.7268247604370117, + "num_tokens": 699941048.0, + "step": 27049 + }, + { + "epoch": 2.970568855699539, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.170938491821289, + "learning_rate": 1e-06, + "loss": 0.8957, + "mean_token_accuracy": 0.7153085470199585, + "num_tokens": 699969241.0, + "step": 27050 + }, + { + "epoch": 2.970678673402152, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.329637289047241, + "learning_rate": 1e-06, + "loss": 0.8103, + "mean_token_accuracy": 0.73856520652771, + "num_tokens": 699993630.0, + "step": 27051 + }, + { + "epoch": 2.970788491104766, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.353027582168579, + "learning_rate": 1e-06, + "loss": 0.807, + "mean_token_accuracy": 0.7387964129447937, + "num_tokens": 700020713.0, + "step": 27052 + }, + { + "epoch": 2.9708983088073797, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5990335941314697, + "learning_rate": 1e-06, + "loss": 0.8426, + "mean_token_accuracy": 0.7385978698730469, + "num_tokens": 700043333.0, + "step": 27053 + }, + { + "epoch": 2.9710081265099935, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.302393913269043, + "learning_rate": 1e-06, + "loss": 0.9086, + "mean_token_accuracy": 0.7220820784568787, + "num_tokens": 700071824.0, + "step": 27054 + }, + { + "epoch": 2.971117944212607, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3160359859466553, + "learning_rate": 1e-06, + "loss": 0.7818, + "mean_token_accuracy": 0.7504739761352539, + "num_tokens": 700096285.0, + "step": 27055 + }, + { + "epoch": 2.9712277619152205, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.379897356033325, + "learning_rate": 1e-06, + "loss": 0.8662, + "mean_token_accuracy": 0.7261104583740234, + "num_tokens": 700122349.0, + "step": 27056 + }, + { + "epoch": 2.9713375796178343, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1224770545959473, + "learning_rate": 1e-06, + "loss": 0.8944, + "mean_token_accuracy": 0.7353560924530029, + "num_tokens": 700152605.0, + "step": 27057 + }, + { + "epoch": 2.971447397320448, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5385243892669678, + "learning_rate": 1e-06, + "loss": 0.7815, + "mean_token_accuracy": 0.7520622611045837, + "num_tokens": 700173611.0, + "step": 27058 + }, + { + "epoch": 2.971557215023062, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2253003120422363, + "learning_rate": 1e-06, + "loss": 0.873, + "mean_token_accuracy": 0.7237987518310547, + "num_tokens": 700205386.0, + "step": 27059 + }, + { + "epoch": 2.9716670327256756, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2116360664367676, + "learning_rate": 1e-06, + "loss": 0.8124, + "mean_token_accuracy": 0.7355270385742188, + "num_tokens": 700235046.0, + "step": 27060 + }, + { + "epoch": 2.971776850428289, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1711833477020264, + "learning_rate": 1e-06, + "loss": 0.8677, + "mean_token_accuracy": 0.7276128530502319, + "num_tokens": 700265594.0, + "step": 27061 + }, + { + "epoch": 2.9718866681309026, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.9951510429382324, + "learning_rate": 1e-06, + "loss": 0.7813, + "mean_token_accuracy": 0.747711181640625, + "num_tokens": 700284713.0, + "step": 27062 + }, + { + "epoch": 2.9719964858335164, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5026421546936035, + "learning_rate": 1e-06, + "loss": 0.8642, + "mean_token_accuracy": 0.7289519906044006, + "num_tokens": 700311304.0, + "step": 27063 + }, + { + "epoch": 2.97210630353613, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.562863826751709, + "learning_rate": 1e-06, + "loss": 0.8389, + "mean_token_accuracy": 0.7349941730499268, + "num_tokens": 700334528.0, + "step": 27064 + }, + { + "epoch": 2.972216121238744, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1547536849975586, + "learning_rate": 1e-06, + "loss": 0.8925, + "mean_token_accuracy": 0.7201001644134521, + "num_tokens": 700365145.0, + "step": 27065 + }, + { + "epoch": 2.972325938941357, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.154705762863159, + "learning_rate": 1e-06, + "loss": 0.8478, + "mean_token_accuracy": 0.7389589548110962, + "num_tokens": 700394174.0, + "step": 27066 + }, + { + "epoch": 2.972435756643971, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2434194087982178, + "learning_rate": 1e-06, + "loss": 0.8321, + "mean_token_accuracy": 0.737405002117157, + "num_tokens": 700422187.0, + "step": 27067 + }, + { + "epoch": 2.9725455743465847, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4071877002716064, + "learning_rate": 1e-06, + "loss": 0.9556, + "mean_token_accuracy": 0.7028764486312866, + "num_tokens": 700449559.0, + "step": 27068 + }, + { + "epoch": 2.9726553920491985, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.345176935195923, + "learning_rate": 1e-06, + "loss": 0.9048, + "mean_token_accuracy": 0.7207910418510437, + "num_tokens": 700479833.0, + "step": 27069 + }, + { + "epoch": 2.9727652097518122, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.469125986099243, + "learning_rate": 1e-06, + "loss": 0.7918, + "mean_token_accuracy": 0.7423882484436035, + "num_tokens": 700501218.0, + "step": 27070 + }, + { + "epoch": 2.9728750274544256, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.212914228439331, + "learning_rate": 1e-06, + "loss": 0.939, + "mean_token_accuracy": 0.7036679983139038, + "num_tokens": 700529618.0, + "step": 27071 + }, + { + "epoch": 2.9729848451570393, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.270699977874756, + "learning_rate": 1e-06, + "loss": 0.8293, + "mean_token_accuracy": 0.7342842817306519, + "num_tokens": 700555811.0, + "step": 27072 + }, + { + "epoch": 2.973094662859653, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.86861515045166, + "learning_rate": 1e-06, + "loss": 0.7192, + "mean_token_accuracy": 0.7689204216003418, + "num_tokens": 700574821.0, + "step": 27073 + }, + { + "epoch": 2.9732044805622664, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4709432125091553, + "learning_rate": 1e-06, + "loss": 0.8822, + "mean_token_accuracy": 0.7216483354568481, + "num_tokens": 700599791.0, + "step": 27074 + }, + { + "epoch": 2.97331429826488, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.309920310974121, + "learning_rate": 1e-06, + "loss": 0.8467, + "mean_token_accuracy": 0.7362889051437378, + "num_tokens": 700627048.0, + "step": 27075 + }, + { + "epoch": 2.973424115967494, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.170039415359497, + "learning_rate": 1e-06, + "loss": 0.9167, + "mean_token_accuracy": 0.7165571451187134, + "num_tokens": 700656950.0, + "step": 27076 + }, + { + "epoch": 2.9735339336701077, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.21345591545105, + "learning_rate": 1e-06, + "loss": 0.8891, + "mean_token_accuracy": 0.7160037755966187, + "num_tokens": 700686909.0, + "step": 27077 + }, + { + "epoch": 2.9736437513727214, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4604008197784424, + "learning_rate": 1e-06, + "loss": 0.7611, + "mean_token_accuracy": 0.752615749835968, + "num_tokens": 700711077.0, + "step": 27078 + }, + { + "epoch": 2.9737535690753347, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3485376834869385, + "learning_rate": 1e-06, + "loss": 0.8572, + "mean_token_accuracy": 0.7326066493988037, + "num_tokens": 700737323.0, + "step": 27079 + }, + { + "epoch": 2.9738633867779485, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.487924098968506, + "learning_rate": 1e-06, + "loss": 0.871, + "mean_token_accuracy": 0.7257087230682373, + "num_tokens": 700763459.0, + "step": 27080 + }, + { + "epoch": 2.9739732044805622, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3538918495178223, + "learning_rate": 1e-06, + "loss": 0.7378, + "mean_token_accuracy": 0.7626094818115234, + "num_tokens": 700790020.0, + "step": 27081 + }, + { + "epoch": 2.974083022183176, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.7047338485717773, + "learning_rate": 1e-06, + "loss": 0.7328, + "mean_token_accuracy": 0.7672703862190247, + "num_tokens": 700810259.0, + "step": 27082 + }, + { + "epoch": 2.9741928398857898, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2872369289398193, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.719602108001709, + "num_tokens": 700837021.0, + "step": 27083 + }, + { + "epoch": 2.974302657588403, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2144839763641357, + "learning_rate": 1e-06, + "loss": 0.8366, + "mean_token_accuracy": 0.7367512583732605, + "num_tokens": 700864837.0, + "step": 27084 + }, + { + "epoch": 2.974412475291017, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.21508526802063, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7326666116714478, + "num_tokens": 700893517.0, + "step": 27085 + }, + { + "epoch": 2.9745222929936306, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.412689447402954, + "learning_rate": 1e-06, + "loss": 0.9036, + "mean_token_accuracy": 0.7200285792350769, + "num_tokens": 700918292.0, + "step": 27086 + }, + { + "epoch": 2.9746321106962443, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.390815258026123, + "learning_rate": 1e-06, + "loss": 0.8107, + "mean_token_accuracy": 0.736680805683136, + "num_tokens": 700942703.0, + "step": 27087 + }, + { + "epoch": 2.974741928398858, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.5387840270996094, + "learning_rate": 1e-06, + "loss": 0.8177, + "mean_token_accuracy": 0.7464123964309692, + "num_tokens": 700966432.0, + "step": 27088 + }, + { + "epoch": 2.9748517461014714, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.226975679397583, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7307543754577637, + "num_tokens": 700994539.0, + "step": 27089 + }, + { + "epoch": 2.974961563804085, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.047753095626831, + "learning_rate": 1e-06, + "loss": 0.9082, + "mean_token_accuracy": 0.7210987210273743, + "num_tokens": 701028518.0, + "step": 27090 + }, + { + "epoch": 2.975071381506699, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.291995048522949, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7402336001396179, + "num_tokens": 701053782.0, + "step": 27091 + }, + { + "epoch": 2.9751811992093127, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.4460830688476562, + "learning_rate": 1e-06, + "loss": 0.7785, + "mean_token_accuracy": 0.7535655498504639, + "num_tokens": 701077793.0, + "step": 27092 + }, + { + "epoch": 2.9752910169119264, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.352688789367676, + "learning_rate": 1e-06, + "loss": 0.9917, + "mean_token_accuracy": 0.7083719968795776, + "num_tokens": 701104306.0, + "step": 27093 + }, + { + "epoch": 2.9754008346145397, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.298250436782837, + "learning_rate": 1e-06, + "loss": 0.803, + "mean_token_accuracy": 0.7408892512321472, + "num_tokens": 701129809.0, + "step": 27094 + }, + { + "epoch": 2.9755106523171535, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.2747323513031006, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7412828803062439, + "num_tokens": 701157533.0, + "step": 27095 + }, + { + "epoch": 2.9756204700197673, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.1550350189208984, + "learning_rate": 1e-06, + "loss": 0.8579, + "mean_token_accuracy": 0.7297576665878296, + "num_tokens": 701188408.0, + "step": 27096 + }, + { + "epoch": 2.9757302877223806, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.493408203125, + "learning_rate": 1e-06, + "loss": 0.9113, + "mean_token_accuracy": 0.7218443155288696, + "num_tokens": 701212371.0, + "step": 27097 + }, + { + "epoch": 2.9758401054249948, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.391829490661621, + "learning_rate": 1e-06, + "loss": 0.7752, + "mean_token_accuracy": 0.7506659030914307, + "num_tokens": 701235803.0, + "step": 27098 + }, + { + "epoch": 2.975949923127608, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.342741012573242, + "learning_rate": 1e-06, + "loss": 0.8789, + "mean_token_accuracy": 0.7224946022033691, + "num_tokens": 701263089.0, + "step": 27099 + }, + { + "epoch": 2.976059740830222, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.384127140045166, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7237374782562256, + "num_tokens": 701288344.0, + "step": 27100 + }, + { + "epoch": 2.9761695585328356, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.3185198307037354, + "learning_rate": 1e-06, + "loss": 0.8938, + "mean_token_accuracy": 0.7268617153167725, + "num_tokens": 701314693.0, + "step": 27101 + }, + { + "epoch": 2.976279376235449, + "ewc_loss": 1.6927719116210938e-05, + "grad_norm": 2.729172706604004, + "learning_rate": 1e-06, + "loss": 0.761, + "mean_token_accuracy": 0.7551302313804626, + "num_tokens": 701332861.0, + "step": 27102 + }, + { + "epoch": 2.9763891939380627, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2995660305023193, + "learning_rate": 1e-06, + "loss": 0.9241, + "mean_token_accuracy": 0.7103509306907654, + "num_tokens": 701362650.0, + "step": 27103 + }, + { + "epoch": 2.9764990116406764, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.438509225845337, + "learning_rate": 1e-06, + "loss": 0.8803, + "mean_token_accuracy": 0.7213287353515625, + "num_tokens": 701387886.0, + "step": 27104 + }, + { + "epoch": 2.97660882934329, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4084930419921875, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.729476809501648, + "num_tokens": 701411826.0, + "step": 27105 + }, + { + "epoch": 2.976718647045904, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2956387996673584, + "learning_rate": 1e-06, + "loss": 0.8151, + "mean_token_accuracy": 0.7406997680664062, + "num_tokens": 701437492.0, + "step": 27106 + }, + { + "epoch": 2.9768284647485173, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1714255809783936, + "learning_rate": 1e-06, + "loss": 0.8444, + "mean_token_accuracy": 0.7278899550437927, + "num_tokens": 701467832.0, + "step": 27107 + }, + { + "epoch": 2.976938282451131, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3315296173095703, + "learning_rate": 1e-06, + "loss": 0.8532, + "mean_token_accuracy": 0.7279360294342041, + "num_tokens": 701495198.0, + "step": 27108 + }, + { + "epoch": 2.9770481001537448, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.198227643966675, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7223556637763977, + "num_tokens": 701523997.0, + "step": 27109 + }, + { + "epoch": 2.9771579178563585, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5230250358581543, + "learning_rate": 1e-06, + "loss": 0.9149, + "mean_token_accuracy": 0.7123321294784546, + "num_tokens": 701549938.0, + "step": 27110 + }, + { + "epoch": 2.9772677355589723, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.310413360595703, + "learning_rate": 1e-06, + "loss": 0.8839, + "mean_token_accuracy": 0.7251671552658081, + "num_tokens": 701576929.0, + "step": 27111 + }, + { + "epoch": 2.9773775532615856, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 3.0116846561431885, + "learning_rate": 1e-06, + "loss": 0.8175, + "mean_token_accuracy": 0.7334067821502686, + "num_tokens": 701594938.0, + "step": 27112 + }, + { + "epoch": 2.9774873709641994, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3427467346191406, + "learning_rate": 1e-06, + "loss": 0.9034, + "mean_token_accuracy": 0.7159265279769897, + "num_tokens": 701620664.0, + "step": 27113 + }, + { + "epoch": 2.977597188666813, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6153719425201416, + "learning_rate": 1e-06, + "loss": 0.7704, + "mean_token_accuracy": 0.7509475946426392, + "num_tokens": 701642530.0, + "step": 27114 + }, + { + "epoch": 2.977707006369427, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2976253032684326, + "learning_rate": 1e-06, + "loss": 0.8845, + "mean_token_accuracy": 0.726346492767334, + "num_tokens": 701667475.0, + "step": 27115 + }, + { + "epoch": 2.9778168240720406, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.777912139892578, + "learning_rate": 1e-06, + "loss": 0.8922, + "mean_token_accuracy": 0.7253819704055786, + "num_tokens": 701689185.0, + "step": 27116 + }, + { + "epoch": 2.977926641774654, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2537107467651367, + "learning_rate": 1e-06, + "loss": 0.7499, + "mean_token_accuracy": 0.753064751625061, + "num_tokens": 701717697.0, + "step": 27117 + }, + { + "epoch": 2.9780364594772677, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2333180904388428, + "learning_rate": 1e-06, + "loss": 0.7855, + "mean_token_accuracy": 0.7464035749435425, + "num_tokens": 701743206.0, + "step": 27118 + }, + { + "epoch": 2.9781462771798815, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.402985095977783, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7381641268730164, + "num_tokens": 701766654.0, + "step": 27119 + }, + { + "epoch": 2.978256094882495, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5628020763397217, + "learning_rate": 1e-06, + "loss": 0.8755, + "mean_token_accuracy": 0.7243069410324097, + "num_tokens": 701789536.0, + "step": 27120 + }, + { + "epoch": 2.978365912585109, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4560725688934326, + "learning_rate": 1e-06, + "loss": 0.8111, + "mean_token_accuracy": 0.7350153923034668, + "num_tokens": 701812744.0, + "step": 27121 + }, + { + "epoch": 2.9784757302877223, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.416839599609375, + "learning_rate": 1e-06, + "loss": 0.7964, + "mean_token_accuracy": 0.7465623021125793, + "num_tokens": 701836365.0, + "step": 27122 + }, + { + "epoch": 2.978585547990336, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.7098121643066406, + "learning_rate": 1e-06, + "loss": 0.8207, + "mean_token_accuracy": 0.7380928993225098, + "num_tokens": 701857860.0, + "step": 27123 + }, + { + "epoch": 2.97869536569295, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.292311906814575, + "learning_rate": 1e-06, + "loss": 0.8916, + "mean_token_accuracy": 0.7170363664627075, + "num_tokens": 701886383.0, + "step": 27124 + }, + { + "epoch": 2.978805183395563, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.66207218170166, + "learning_rate": 1e-06, + "loss": 0.7442, + "mean_token_accuracy": 0.7709633111953735, + "num_tokens": 701907847.0, + "step": 27125 + }, + { + "epoch": 2.978915001098177, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3676469326019287, + "learning_rate": 1e-06, + "loss": 0.8296, + "mean_token_accuracy": 0.7345187664031982, + "num_tokens": 701936129.0, + "step": 27126 + }, + { + "epoch": 2.9790248188007906, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.540286064147949, + "learning_rate": 1e-06, + "loss": 0.8727, + "mean_token_accuracy": 0.7306830286979675, + "num_tokens": 701959243.0, + "step": 27127 + }, + { + "epoch": 2.9791346365034044, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2433910369873047, + "learning_rate": 1e-06, + "loss": 0.8431, + "mean_token_accuracy": 0.7360109090805054, + "num_tokens": 701985609.0, + "step": 27128 + }, + { + "epoch": 2.979244454206018, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.444277048110962, + "learning_rate": 1e-06, + "loss": 0.8865, + "mean_token_accuracy": 0.7230107188224792, + "num_tokens": 702010856.0, + "step": 27129 + }, + { + "epoch": 2.9793542719086314, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.399746894836426, + "learning_rate": 1e-06, + "loss": 0.9084, + "mean_token_accuracy": 0.7195956707000732, + "num_tokens": 702036322.0, + "step": 27130 + }, + { + "epoch": 2.979464089611245, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2189013957977295, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7460653185844421, + "num_tokens": 702064103.0, + "step": 27131 + }, + { + "epoch": 2.979573907313859, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3749637603759766, + "learning_rate": 1e-06, + "loss": 0.8337, + "mean_token_accuracy": 0.7364578247070312, + "num_tokens": 702089732.0, + "step": 27132 + }, + { + "epoch": 2.9796837250164727, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.509336471557617, + "learning_rate": 1e-06, + "loss": 0.7544, + "mean_token_accuracy": 0.7590219974517822, + "num_tokens": 702111620.0, + "step": 27133 + }, + { + "epoch": 2.9797935427190865, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.422382116317749, + "learning_rate": 1e-06, + "loss": 0.9355, + "mean_token_accuracy": 0.7056939005851746, + "num_tokens": 702136971.0, + "step": 27134 + }, + { + "epoch": 2.9799033604217, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1174943447113037, + "learning_rate": 1e-06, + "loss": 0.8544, + "mean_token_accuracy": 0.7279583811759949, + "num_tokens": 702166116.0, + "step": 27135 + }, + { + "epoch": 2.9800131781243135, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 3.2368054389953613, + "learning_rate": 1e-06, + "loss": 0.7129, + "mean_token_accuracy": 0.7628882527351379, + "num_tokens": 702180819.0, + "step": 27136 + }, + { + "epoch": 2.9801229958269273, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.690150022506714, + "learning_rate": 1e-06, + "loss": 0.8182, + "mean_token_accuracy": 0.7416384220123291, + "num_tokens": 702201448.0, + "step": 27137 + }, + { + "epoch": 2.980232813529541, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4533915519714355, + "learning_rate": 1e-06, + "loss": 0.8582, + "mean_token_accuracy": 0.7328882217407227, + "num_tokens": 702225696.0, + "step": 27138 + }, + { + "epoch": 2.980342631232155, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4254300594329834, + "learning_rate": 1e-06, + "loss": 0.8483, + "mean_token_accuracy": 0.7298556566238403, + "num_tokens": 702252138.0, + "step": 27139 + }, + { + "epoch": 2.980452448934768, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5624284744262695, + "learning_rate": 1e-06, + "loss": 0.8173, + "mean_token_accuracy": 0.742300271987915, + "num_tokens": 702275638.0, + "step": 27140 + }, + { + "epoch": 2.980562266637382, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.505143880844116, + "learning_rate": 1e-06, + "loss": 0.7988, + "mean_token_accuracy": 0.7419551014900208, + "num_tokens": 702298503.0, + "step": 27141 + }, + { + "epoch": 2.9806720843399956, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.599083423614502, + "learning_rate": 1e-06, + "loss": 0.7965, + "mean_token_accuracy": 0.7529683113098145, + "num_tokens": 702321521.0, + "step": 27142 + }, + { + "epoch": 2.9807819020426094, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.513585090637207, + "learning_rate": 1e-06, + "loss": 0.8691, + "mean_token_accuracy": 0.7220331430435181, + "num_tokens": 702343966.0, + "step": 27143 + }, + { + "epoch": 2.980891719745223, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6102652549743652, + "learning_rate": 1e-06, + "loss": 0.8864, + "mean_token_accuracy": 0.7221460342407227, + "num_tokens": 702366429.0, + "step": 27144 + }, + { + "epoch": 2.9810015374478365, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6246933937072754, + "learning_rate": 1e-06, + "loss": 0.8348, + "mean_token_accuracy": 0.7324431538581848, + "num_tokens": 702388325.0, + "step": 27145 + }, + { + "epoch": 2.9811113551504502, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.245657444000244, + "learning_rate": 1e-06, + "loss": 0.9357, + "mean_token_accuracy": 0.7177002429962158, + "num_tokens": 702417213.0, + "step": 27146 + }, + { + "epoch": 2.981221172853064, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4649453163146973, + "learning_rate": 1e-06, + "loss": 0.8826, + "mean_token_accuracy": 0.7221651077270508, + "num_tokens": 702441237.0, + "step": 27147 + }, + { + "epoch": 2.9813309905556777, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0760385990142822, + "learning_rate": 1e-06, + "loss": 0.9396, + "mean_token_accuracy": 0.7077308297157288, + "num_tokens": 702474699.0, + "step": 27148 + }, + { + "epoch": 2.9814408082582915, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.211477041244507, + "learning_rate": 1e-06, + "loss": 0.8658, + "mean_token_accuracy": 0.7280119061470032, + "num_tokens": 702504887.0, + "step": 27149 + }, + { + "epoch": 2.981550625960905, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4836971759796143, + "learning_rate": 1e-06, + "loss": 0.9144, + "mean_token_accuracy": 0.7138057947158813, + "num_tokens": 702528723.0, + "step": 27150 + }, + { + "epoch": 2.9816604436635186, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.473519802093506, + "learning_rate": 1e-06, + "loss": 0.7811, + "mean_token_accuracy": 0.7496225833892822, + "num_tokens": 702551150.0, + "step": 27151 + }, + { + "epoch": 2.9817702613661323, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4602153301239014, + "learning_rate": 1e-06, + "loss": 0.7931, + "mean_token_accuracy": 0.7540050148963928, + "num_tokens": 702574087.0, + "step": 27152 + }, + { + "epoch": 2.9818800790687456, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6007132530212402, + "learning_rate": 1e-06, + "loss": 0.8702, + "mean_token_accuracy": 0.7227389812469482, + "num_tokens": 702601130.0, + "step": 27153 + }, + { + "epoch": 2.9819898967713594, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1225898265838623, + "learning_rate": 1e-06, + "loss": 0.878, + "mean_token_accuracy": 0.7305344343185425, + "num_tokens": 702633442.0, + "step": 27154 + }, + { + "epoch": 2.982099714473973, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2631900310516357, + "learning_rate": 1e-06, + "loss": 0.7979, + "mean_token_accuracy": 0.7410954236984253, + "num_tokens": 702659621.0, + "step": 27155 + }, + { + "epoch": 2.982209532176587, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1092491149902344, + "learning_rate": 1e-06, + "loss": 0.8452, + "mean_token_accuracy": 0.735465943813324, + "num_tokens": 702687776.0, + "step": 27156 + }, + { + "epoch": 2.9823193498792007, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.473386526107788, + "learning_rate": 1e-06, + "loss": 0.8425, + "mean_token_accuracy": 0.7359127998352051, + "num_tokens": 702712274.0, + "step": 27157 + }, + { + "epoch": 2.982429167581814, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2207159996032715, + "learning_rate": 1e-06, + "loss": 0.8309, + "mean_token_accuracy": 0.7375599145889282, + "num_tokens": 702740462.0, + "step": 27158 + }, + { + "epoch": 2.9825389852844277, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.430013656616211, + "learning_rate": 1e-06, + "loss": 0.7929, + "mean_token_accuracy": 0.7397683262825012, + "num_tokens": 702764661.0, + "step": 27159 + }, + { + "epoch": 2.9826488029870415, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3923277854919434, + "learning_rate": 1e-06, + "loss": 0.8491, + "mean_token_accuracy": 0.7296184301376343, + "num_tokens": 702792044.0, + "step": 27160 + }, + { + "epoch": 2.9827586206896552, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.606900691986084, + "learning_rate": 1e-06, + "loss": 0.8223, + "mean_token_accuracy": 0.7327142953872681, + "num_tokens": 702813905.0, + "step": 27161 + }, + { + "epoch": 2.982868438392269, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6760613918304443, + "learning_rate": 1e-06, + "loss": 0.8374, + "mean_token_accuracy": 0.7363961935043335, + "num_tokens": 702834398.0, + "step": 27162 + }, + { + "epoch": 2.9829782560948823, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.454639196395874, + "learning_rate": 1e-06, + "loss": 0.8998, + "mean_token_accuracy": 0.7206202745437622, + "num_tokens": 702860805.0, + "step": 27163 + }, + { + "epoch": 2.983088073797496, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.443328380584717, + "learning_rate": 1e-06, + "loss": 0.8294, + "mean_token_accuracy": 0.7347186207771301, + "num_tokens": 702887110.0, + "step": 27164 + }, + { + "epoch": 2.98319789150011, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3531854152679443, + "learning_rate": 1e-06, + "loss": 0.8765, + "mean_token_accuracy": 0.7212809324264526, + "num_tokens": 702914694.0, + "step": 27165 + }, + { + "epoch": 2.9833077092027236, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3374531269073486, + "learning_rate": 1e-06, + "loss": 0.8904, + "mean_token_accuracy": 0.717149019241333, + "num_tokens": 702941666.0, + "step": 27166 + }, + { + "epoch": 2.9834175269053373, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.8184471130371094, + "learning_rate": 1e-06, + "loss": 0.7937, + "mean_token_accuracy": 0.7426595687866211, + "num_tokens": 702962865.0, + "step": 27167 + }, + { + "epoch": 2.9835273446079507, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.169501543045044, + "learning_rate": 1e-06, + "loss": 0.7743, + "mean_token_accuracy": 0.7435454726219177, + "num_tokens": 702991618.0, + "step": 27168 + }, + { + "epoch": 2.9836371623105644, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.159400463104248, + "learning_rate": 1e-06, + "loss": 0.9252, + "mean_token_accuracy": 0.712590217590332, + "num_tokens": 703023298.0, + "step": 27169 + }, + { + "epoch": 2.983746980013178, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1529695987701416, + "learning_rate": 1e-06, + "loss": 0.9019, + "mean_token_accuracy": 0.7234253883361816, + "num_tokens": 703053813.0, + "step": 27170 + }, + { + "epoch": 2.983856797715792, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.973392963409424, + "learning_rate": 1e-06, + "loss": 0.7666, + "mean_token_accuracy": 0.7521398067474365, + "num_tokens": 703072059.0, + "step": 27171 + }, + { + "epoch": 2.9839666154184057, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.476679563522339, + "learning_rate": 1e-06, + "loss": 0.8842, + "mean_token_accuracy": 0.7225433588027954, + "num_tokens": 703096897.0, + "step": 27172 + }, + { + "epoch": 2.984076433121019, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.7031490802764893, + "learning_rate": 1e-06, + "loss": 0.8385, + "mean_token_accuracy": 0.7276061773300171, + "num_tokens": 703117645.0, + "step": 27173 + }, + { + "epoch": 2.9841862508236328, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3623201847076416, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7604359984397888, + "num_tokens": 703139908.0, + "step": 27174 + }, + { + "epoch": 2.9842960685262465, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3209269046783447, + "learning_rate": 1e-06, + "loss": 0.7743, + "mean_token_accuracy": 0.7532157301902771, + "num_tokens": 703163662.0, + "step": 27175 + }, + { + "epoch": 2.98440588622886, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.366039752960205, + "learning_rate": 1e-06, + "loss": 0.8361, + "mean_token_accuracy": 0.7431498765945435, + "num_tokens": 703188222.0, + "step": 27176 + }, + { + "epoch": 2.984515703931474, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3393259048461914, + "learning_rate": 1e-06, + "loss": 0.859, + "mean_token_accuracy": 0.733364462852478, + "num_tokens": 703213448.0, + "step": 27177 + }, + { + "epoch": 2.9846255216340873, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.332413673400879, + "learning_rate": 1e-06, + "loss": 0.8268, + "mean_token_accuracy": 0.7378108501434326, + "num_tokens": 703239870.0, + "step": 27178 + }, + { + "epoch": 2.984735339336701, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.512399435043335, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7327702045440674, + "num_tokens": 703262075.0, + "step": 27179 + }, + { + "epoch": 2.984845157039315, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.491281747817993, + "learning_rate": 1e-06, + "loss": 0.8472, + "mean_token_accuracy": 0.7301552295684814, + "num_tokens": 703286704.0, + "step": 27180 + }, + { + "epoch": 2.984954974741928, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3059980869293213, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7116620540618896, + "num_tokens": 703315082.0, + "step": 27181 + }, + { + "epoch": 2.985064792444542, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3152124881744385, + "learning_rate": 1e-06, + "loss": 0.7895, + "mean_token_accuracy": 0.7436789870262146, + "num_tokens": 703342324.0, + "step": 27182 + }, + { + "epoch": 2.9851746101471557, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1514079570770264, + "learning_rate": 1e-06, + "loss": 0.8581, + "mean_token_accuracy": 0.7291260361671448, + "num_tokens": 703373898.0, + "step": 27183 + }, + { + "epoch": 2.9852844278497694, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.656663417816162, + "learning_rate": 1e-06, + "loss": 0.7962, + "mean_token_accuracy": 0.7454067468643188, + "num_tokens": 703394399.0, + "step": 27184 + }, + { + "epoch": 2.985394245552383, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5751147270202637, + "learning_rate": 1e-06, + "loss": 0.8924, + "mean_token_accuracy": 0.7249389886856079, + "num_tokens": 703418105.0, + "step": 27185 + }, + { + "epoch": 2.9855040632549965, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.310903310775757, + "learning_rate": 1e-06, + "loss": 0.8863, + "mean_token_accuracy": 0.7216889262199402, + "num_tokens": 703445569.0, + "step": 27186 + }, + { + "epoch": 2.9856138809576103, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.685884475708008, + "learning_rate": 1e-06, + "loss": 0.8849, + "mean_token_accuracy": 0.7256943583488464, + "num_tokens": 703466434.0, + "step": 27187 + }, + { + "epoch": 2.985723698660224, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.344050168991089, + "learning_rate": 1e-06, + "loss": 0.8227, + "mean_token_accuracy": 0.7352348566055298, + "num_tokens": 703493661.0, + "step": 27188 + }, + { + "epoch": 2.985833516362838, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6782686710357666, + "learning_rate": 1e-06, + "loss": 0.8597, + "mean_token_accuracy": 0.7379902005195618, + "num_tokens": 703515129.0, + "step": 27189 + }, + { + "epoch": 2.9859433340654515, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.9489119052886963, + "learning_rate": 1e-06, + "loss": 0.8035, + "mean_token_accuracy": 0.748324990272522, + "num_tokens": 703531781.0, + "step": 27190 + }, + { + "epoch": 2.986053151768065, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0887744426727295, + "learning_rate": 1e-06, + "loss": 0.9403, + "mean_token_accuracy": 0.7151617407798767, + "num_tokens": 703564640.0, + "step": 27191 + }, + { + "epoch": 2.9861629694706786, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.265491485595703, + "learning_rate": 1e-06, + "loss": 0.845, + "mean_token_accuracy": 0.7319395542144775, + "num_tokens": 703590692.0, + "step": 27192 + }, + { + "epoch": 2.9862727871732924, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.464904308319092, + "learning_rate": 1e-06, + "loss": 0.8423, + "mean_token_accuracy": 0.7294417023658752, + "num_tokens": 703614104.0, + "step": 27193 + }, + { + "epoch": 2.986382604875906, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2084052562713623, + "learning_rate": 1e-06, + "loss": 0.865, + "mean_token_accuracy": 0.7371859550476074, + "num_tokens": 703643107.0, + "step": 27194 + }, + { + "epoch": 2.98649242257852, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.127638816833496, + "learning_rate": 1e-06, + "loss": 0.9161, + "mean_token_accuracy": 0.7090592384338379, + "num_tokens": 703675675.0, + "step": 27195 + }, + { + "epoch": 2.986602240281133, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.166146993637085, + "learning_rate": 1e-06, + "loss": 0.8608, + "mean_token_accuracy": 0.7266502380371094, + "num_tokens": 703705056.0, + "step": 27196 + }, + { + "epoch": 2.986712057983747, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.386554002761841, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7585793733596802, + "num_tokens": 703728771.0, + "step": 27197 + }, + { + "epoch": 2.9868218756863607, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1954331398010254, + "learning_rate": 1e-06, + "loss": 0.8819, + "mean_token_accuracy": 0.7344751358032227, + "num_tokens": 703757949.0, + "step": 27198 + }, + { + "epoch": 2.9869316933889745, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.533334493637085, + "learning_rate": 1e-06, + "loss": 0.8074, + "mean_token_accuracy": 0.7390002012252808, + "num_tokens": 703781668.0, + "step": 27199 + }, + { + "epoch": 2.987041511091588, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.284046173095703, + "learning_rate": 1e-06, + "loss": 0.8195, + "mean_token_accuracy": 0.7413854598999023, + "num_tokens": 703810232.0, + "step": 27200 + }, + { + "epoch": 2.9871513287942015, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2464842796325684, + "learning_rate": 1e-06, + "loss": 0.8522, + "mean_token_accuracy": 0.7355127334594727, + "num_tokens": 703837505.0, + "step": 27201 + }, + { + "epoch": 2.9872611464968153, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.302487850189209, + "learning_rate": 1e-06, + "loss": 0.892, + "mean_token_accuracy": 0.7246978282928467, + "num_tokens": 703864601.0, + "step": 27202 + }, + { + "epoch": 2.987370964199429, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2973577976226807, + "learning_rate": 1e-06, + "loss": 0.9779, + "mean_token_accuracy": 0.6978105306625366, + "num_tokens": 703896134.0, + "step": 27203 + }, + { + "epoch": 2.9874807819020424, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6101834774017334, + "learning_rate": 1e-06, + "loss": 0.8161, + "mean_token_accuracy": 0.7446014285087585, + "num_tokens": 703916331.0, + "step": 27204 + }, + { + "epoch": 2.987590599604656, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2478044033050537, + "learning_rate": 1e-06, + "loss": 0.8699, + "mean_token_accuracy": 0.7269835472106934, + "num_tokens": 703944317.0, + "step": 27205 + }, + { + "epoch": 2.98770041730727, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.585217237472534, + "learning_rate": 1e-06, + "loss": 0.8393, + "mean_token_accuracy": 0.7363167405128479, + "num_tokens": 703965788.0, + "step": 27206 + }, + { + "epoch": 2.9878102350098836, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.310025930404663, + "learning_rate": 1e-06, + "loss": 0.844, + "mean_token_accuracy": 0.7329928874969482, + "num_tokens": 703991689.0, + "step": 27207 + }, + { + "epoch": 2.9879200527124974, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3217742443084717, + "learning_rate": 1e-06, + "loss": 0.8202, + "mean_token_accuracy": 0.7394295930862427, + "num_tokens": 704018020.0, + "step": 27208 + }, + { + "epoch": 2.9880298704151107, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.518496513366699, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.727867841720581, + "num_tokens": 704040366.0, + "step": 27209 + }, + { + "epoch": 2.9881396881177245, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3747434616088867, + "learning_rate": 1e-06, + "loss": 0.8732, + "mean_token_accuracy": 0.730368971824646, + "num_tokens": 704064833.0, + "step": 27210 + }, + { + "epoch": 2.988249505820338, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.637209415435791, + "learning_rate": 1e-06, + "loss": 0.8251, + "mean_token_accuracy": 0.7431625723838806, + "num_tokens": 704087321.0, + "step": 27211 + }, + { + "epoch": 2.988359323522952, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2903690338134766, + "learning_rate": 1e-06, + "loss": 0.9532, + "mean_token_accuracy": 0.7058656215667725, + "num_tokens": 704115895.0, + "step": 27212 + }, + { + "epoch": 2.9884691412255657, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0502257347106934, + "learning_rate": 1e-06, + "loss": 0.8644, + "mean_token_accuracy": 0.736981987953186, + "num_tokens": 704148074.0, + "step": 27213 + }, + { + "epoch": 2.988578958928179, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.35298490524292, + "learning_rate": 1e-06, + "loss": 0.8846, + "mean_token_accuracy": 0.7237272262573242, + "num_tokens": 704175384.0, + "step": 27214 + }, + { + "epoch": 2.988688776630793, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4064714908599854, + "learning_rate": 1e-06, + "loss": 0.8196, + "mean_token_accuracy": 0.7443677186965942, + "num_tokens": 704198495.0, + "step": 27215 + }, + { + "epoch": 2.9887985943334066, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1923863887786865, + "learning_rate": 1e-06, + "loss": 0.8939, + "mean_token_accuracy": 0.7165729999542236, + "num_tokens": 704226538.0, + "step": 27216 + }, + { + "epoch": 2.9889084120360203, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.674145221710205, + "learning_rate": 1e-06, + "loss": 0.7626, + "mean_token_accuracy": 0.7525331974029541, + "num_tokens": 704248083.0, + "step": 27217 + }, + { + "epoch": 2.989018229738634, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6427364349365234, + "learning_rate": 1e-06, + "loss": 0.7423, + "mean_token_accuracy": 0.755744457244873, + "num_tokens": 704267508.0, + "step": 27218 + }, + { + "epoch": 2.9891280474412474, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.183403968811035, + "learning_rate": 1e-06, + "loss": 0.9333, + "mean_token_accuracy": 0.7101011276245117, + "num_tokens": 704298993.0, + "step": 27219 + }, + { + "epoch": 2.989237865143861, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3749735355377197, + "learning_rate": 1e-06, + "loss": 0.8147, + "mean_token_accuracy": 0.7425618767738342, + "num_tokens": 704323108.0, + "step": 27220 + }, + { + "epoch": 2.989347682846475, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.7936861515045166, + "learning_rate": 1e-06, + "loss": 0.847, + "mean_token_accuracy": 0.7346232533454895, + "num_tokens": 704343373.0, + "step": 27221 + }, + { + "epoch": 2.9894575005490887, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5101139545440674, + "learning_rate": 1e-06, + "loss": 0.9153, + "mean_token_accuracy": 0.72121262550354, + "num_tokens": 704366538.0, + "step": 27222 + }, + { + "epoch": 2.9895673182517024, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.680866241455078, + "learning_rate": 1e-06, + "loss": 0.8664, + "mean_token_accuracy": 0.7269080877304077, + "num_tokens": 704390386.0, + "step": 27223 + }, + { + "epoch": 2.9896771359543157, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4027156829833984, + "learning_rate": 1e-06, + "loss": 0.9235, + "mean_token_accuracy": 0.7193626761436462, + "num_tokens": 704415032.0, + "step": 27224 + }, + { + "epoch": 2.9897869536569295, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.281114101409912, + "learning_rate": 1e-06, + "loss": 0.797, + "mean_token_accuracy": 0.7490347623825073, + "num_tokens": 704442716.0, + "step": 27225 + }, + { + "epoch": 2.9898967713595432, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.354240655899048, + "learning_rate": 1e-06, + "loss": 0.8092, + "mean_token_accuracy": 0.7437170743942261, + "num_tokens": 704467676.0, + "step": 27226 + }, + { + "epoch": 2.9900065890621566, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.581707000732422, + "learning_rate": 1e-06, + "loss": 0.7925, + "mean_token_accuracy": 0.7398968935012817, + "num_tokens": 704488473.0, + "step": 27227 + }, + { + "epoch": 2.9901164067647708, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 1.9933485984802246, + "learning_rate": 1e-06, + "loss": 0.8542, + "mean_token_accuracy": 0.7268818616867065, + "num_tokens": 704520933.0, + "step": 27228 + }, + { + "epoch": 2.990226224467384, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1651813983917236, + "learning_rate": 1e-06, + "loss": 0.8106, + "mean_token_accuracy": 0.739128589630127, + "num_tokens": 704550513.0, + "step": 27229 + }, + { + "epoch": 2.990336042169998, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3456485271453857, + "learning_rate": 1e-06, + "loss": 0.9444, + "mean_token_accuracy": 0.7134360074996948, + "num_tokens": 704577609.0, + "step": 27230 + }, + { + "epoch": 2.9904458598726116, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3400676250457764, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7343041896820068, + "num_tokens": 704601933.0, + "step": 27231 + }, + { + "epoch": 2.990555677575225, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5653395652770996, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7314504384994507, + "num_tokens": 704624988.0, + "step": 27232 + }, + { + "epoch": 2.9906654952778386, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2697246074676514, + "learning_rate": 1e-06, + "loss": 0.8333, + "mean_token_accuracy": 0.7328324317932129, + "num_tokens": 704653433.0, + "step": 27233 + }, + { + "epoch": 2.9907753129804524, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1405465602874756, + "learning_rate": 1e-06, + "loss": 0.8266, + "mean_token_accuracy": 0.7365168929100037, + "num_tokens": 704681291.0, + "step": 27234 + }, + { + "epoch": 2.990885130683066, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5458531379699707, + "learning_rate": 1e-06, + "loss": 0.7747, + "mean_token_accuracy": 0.7444279193878174, + "num_tokens": 704704250.0, + "step": 27235 + }, + { + "epoch": 2.99099494838568, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4144465923309326, + "learning_rate": 1e-06, + "loss": 0.8245, + "mean_token_accuracy": 0.750848114490509, + "num_tokens": 704729508.0, + "step": 27236 + }, + { + "epoch": 2.9911047660882932, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5099215507507324, + "learning_rate": 1e-06, + "loss": 0.8219, + "mean_token_accuracy": 0.7365017533302307, + "num_tokens": 704753815.0, + "step": 27237 + }, + { + "epoch": 2.991214583790907, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2464115619659424, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.7067755460739136, + "num_tokens": 704785221.0, + "step": 27238 + }, + { + "epoch": 2.9913244014935207, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1279354095458984, + "learning_rate": 1e-06, + "loss": 0.9175, + "mean_token_accuracy": 0.7253133058547974, + "num_tokens": 704815571.0, + "step": 27239 + }, + { + "epoch": 2.9914342191961345, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.348604679107666, + "learning_rate": 1e-06, + "loss": 0.858, + "mean_token_accuracy": 0.7288432121276855, + "num_tokens": 704841008.0, + "step": 27240 + }, + { + "epoch": 2.9915440368987483, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2865352630615234, + "learning_rate": 1e-06, + "loss": 0.8521, + "mean_token_accuracy": 0.7356878519058228, + "num_tokens": 704867545.0, + "step": 27241 + }, + { + "epoch": 2.9916538546013616, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6005868911743164, + "learning_rate": 1e-06, + "loss": 0.8347, + "mean_token_accuracy": 0.7333163619041443, + "num_tokens": 704891252.0, + "step": 27242 + }, + { + "epoch": 2.9917636723039753, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.604173183441162, + "learning_rate": 1e-06, + "loss": 0.8378, + "mean_token_accuracy": 0.7447358965873718, + "num_tokens": 704910961.0, + "step": 27243 + }, + { + "epoch": 2.991873490006589, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0503857135772705, + "learning_rate": 1e-06, + "loss": 0.9319, + "mean_token_accuracy": 0.7097817063331604, + "num_tokens": 704943947.0, + "step": 27244 + }, + { + "epoch": 2.991983307709203, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2041525840759277, + "learning_rate": 1e-06, + "loss": 0.8471, + "mean_token_accuracy": 0.7288919687271118, + "num_tokens": 704974192.0, + "step": 27245 + }, + { + "epoch": 2.9920931254118166, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5936803817749023, + "learning_rate": 1e-06, + "loss": 0.8224, + "mean_token_accuracy": 0.7446913123130798, + "num_tokens": 704995669.0, + "step": 27246 + }, + { + "epoch": 2.99220294311443, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.385636568069458, + "learning_rate": 1e-06, + "loss": 0.8011, + "mean_token_accuracy": 0.7389209270477295, + "num_tokens": 705020032.0, + "step": 27247 + }, + { + "epoch": 2.9923127608170437, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3153154850006104, + "learning_rate": 1e-06, + "loss": 0.933, + "mean_token_accuracy": 0.7140171527862549, + "num_tokens": 705048849.0, + "step": 27248 + }, + { + "epoch": 2.9924225785196574, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.291982412338257, + "learning_rate": 1e-06, + "loss": 0.8346, + "mean_token_accuracy": 0.7363831996917725, + "num_tokens": 705077489.0, + "step": 27249 + }, + { + "epoch": 2.992532396222271, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.148517608642578, + "learning_rate": 1e-06, + "loss": 0.8286, + "mean_token_accuracy": 0.7409987449645996, + "num_tokens": 705105991.0, + "step": 27250 + }, + { + "epoch": 2.992642213924885, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.725285053253174, + "learning_rate": 1e-06, + "loss": 0.8252, + "mean_token_accuracy": 0.7388023734092712, + "num_tokens": 705126974.0, + "step": 27251 + }, + { + "epoch": 2.9927520316274983, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3291659355163574, + "learning_rate": 1e-06, + "loss": 0.9345, + "mean_token_accuracy": 0.709017813205719, + "num_tokens": 705151444.0, + "step": 27252 + }, + { + "epoch": 2.992861849330112, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.639939546585083, + "learning_rate": 1e-06, + "loss": 0.811, + "mean_token_accuracy": 0.7460870742797852, + "num_tokens": 705172830.0, + "step": 27253 + }, + { + "epoch": 2.9929716670327258, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.247018337249756, + "learning_rate": 1e-06, + "loss": 0.8827, + "mean_token_accuracy": 0.7285240888595581, + "num_tokens": 705202731.0, + "step": 27254 + }, + { + "epoch": 2.993081484735339, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.694798231124878, + "learning_rate": 1e-06, + "loss": 0.8176, + "mean_token_accuracy": 0.7355378866195679, + "num_tokens": 705223406.0, + "step": 27255 + }, + { + "epoch": 2.993191302437953, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2260639667510986, + "learning_rate": 1e-06, + "loss": 0.8881, + "mean_token_accuracy": 0.7177776098251343, + "num_tokens": 705252641.0, + "step": 27256 + }, + { + "epoch": 2.9933011201405666, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0687291622161865, + "learning_rate": 1e-06, + "loss": 0.8762, + "mean_token_accuracy": 0.7251174449920654, + "num_tokens": 705283871.0, + "step": 27257 + }, + { + "epoch": 2.9934109378431804, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3646275997161865, + "learning_rate": 1e-06, + "loss": 0.8169, + "mean_token_accuracy": 0.737213134765625, + "num_tokens": 705308264.0, + "step": 27258 + }, + { + "epoch": 2.993520755545794, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.495734930038452, + "learning_rate": 1e-06, + "loss": 0.7787, + "mean_token_accuracy": 0.7467990517616272, + "num_tokens": 705331376.0, + "step": 27259 + }, + { + "epoch": 2.9936305732484074, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.334984302520752, + "learning_rate": 1e-06, + "loss": 0.9451, + "mean_token_accuracy": 0.702772319316864, + "num_tokens": 705359289.0, + "step": 27260 + }, + { + "epoch": 2.993740390951021, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.280329942703247, + "learning_rate": 1e-06, + "loss": 0.8801, + "mean_token_accuracy": 0.721330463886261, + "num_tokens": 705386707.0, + "step": 27261 + }, + { + "epoch": 2.993850208653635, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.090693712234497, + "learning_rate": 1e-06, + "loss": 0.7996, + "mean_token_accuracy": 0.739944577217102, + "num_tokens": 705418513.0, + "step": 27262 + }, + { + "epoch": 2.9939600263562487, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.090062141418457, + "learning_rate": 1e-06, + "loss": 0.7912, + "mean_token_accuracy": 0.7478365898132324, + "num_tokens": 705447497.0, + "step": 27263 + }, + { + "epoch": 2.9940698440588625, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.223701238632202, + "learning_rate": 1e-06, + "loss": 0.852, + "mean_token_accuracy": 0.7308976650238037, + "num_tokens": 705474844.0, + "step": 27264 + }, + { + "epoch": 2.9941796617614758, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.561389207839966, + "learning_rate": 1e-06, + "loss": 0.757, + "mean_token_accuracy": 0.7573740482330322, + "num_tokens": 705496732.0, + "step": 27265 + }, + { + "epoch": 2.9942894794640895, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4472200870513916, + "learning_rate": 1e-06, + "loss": 0.8096, + "mean_token_accuracy": 0.7381449937820435, + "num_tokens": 705519700.0, + "step": 27266 + }, + { + "epoch": 2.9943992971667033, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2533414363861084, + "learning_rate": 1e-06, + "loss": 0.8153, + "mean_token_accuracy": 0.742698609828949, + "num_tokens": 705546443.0, + "step": 27267 + }, + { + "epoch": 2.994509114869317, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.463948965072632, + "learning_rate": 1e-06, + "loss": 0.8413, + "mean_token_accuracy": 0.7404037714004517, + "num_tokens": 705569904.0, + "step": 27268 + }, + { + "epoch": 2.994618932571931, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5527517795562744, + "learning_rate": 1e-06, + "loss": 0.8391, + "mean_token_accuracy": 0.7395679354667664, + "num_tokens": 705593737.0, + "step": 27269 + }, + { + "epoch": 2.994728750274544, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.324342966079712, + "learning_rate": 1e-06, + "loss": 0.8307, + "mean_token_accuracy": 0.7358397245407104, + "num_tokens": 705620490.0, + "step": 27270 + }, + { + "epoch": 2.994838567977158, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6480319499969482, + "learning_rate": 1e-06, + "loss": 0.8086, + "mean_token_accuracy": 0.7469995021820068, + "num_tokens": 705641312.0, + "step": 27271 + }, + { + "epoch": 2.9949483856797716, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.348526954650879, + "learning_rate": 1e-06, + "loss": 0.8104, + "mean_token_accuracy": 0.7427264451980591, + "num_tokens": 705668248.0, + "step": 27272 + }, + { + "epoch": 2.9950582033823854, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.212796926498413, + "learning_rate": 1e-06, + "loss": 0.8557, + "mean_token_accuracy": 0.7222536206245422, + "num_tokens": 705696773.0, + "step": 27273 + }, + { + "epoch": 2.995168021084999, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2916572093963623, + "learning_rate": 1e-06, + "loss": 0.9038, + "mean_token_accuracy": 0.7199180126190186, + "num_tokens": 705728765.0, + "step": 27274 + }, + { + "epoch": 2.9952778387876124, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4150807857513428, + "learning_rate": 1e-06, + "loss": 0.9039, + "mean_token_accuracy": 0.7123445272445679, + "num_tokens": 705755618.0, + "step": 27275 + }, + { + "epoch": 2.995387656490226, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.301319122314453, + "learning_rate": 1e-06, + "loss": 0.8728, + "mean_token_accuracy": 0.7277653813362122, + "num_tokens": 705784998.0, + "step": 27276 + }, + { + "epoch": 2.99549747419284, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4907467365264893, + "learning_rate": 1e-06, + "loss": 0.7523, + "mean_token_accuracy": 0.7645708322525024, + "num_tokens": 705807142.0, + "step": 27277 + }, + { + "epoch": 2.9956072918954533, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.264967918395996, + "learning_rate": 1e-06, + "loss": 0.8511, + "mean_token_accuracy": 0.7274025678634644, + "num_tokens": 705834156.0, + "step": 27278 + }, + { + "epoch": 2.9957171095980675, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.377197027206421, + "learning_rate": 1e-06, + "loss": 0.8498, + "mean_token_accuracy": 0.7302054166793823, + "num_tokens": 705858564.0, + "step": 27279 + }, + { + "epoch": 2.995826927300681, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.268258810043335, + "learning_rate": 1e-06, + "loss": 0.8956, + "mean_token_accuracy": 0.7190762162208557, + "num_tokens": 705888824.0, + "step": 27280 + }, + { + "epoch": 2.9959367450032945, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.477916955947876, + "learning_rate": 1e-06, + "loss": 0.8777, + "mean_token_accuracy": 0.7332585453987122, + "num_tokens": 705912030.0, + "step": 27281 + }, + { + "epoch": 2.9960465627059083, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.394533634185791, + "learning_rate": 1e-06, + "loss": 0.8654, + "mean_token_accuracy": 0.7288134098052979, + "num_tokens": 705937277.0, + "step": 27282 + }, + { + "epoch": 2.9961563804085216, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.646573305130005, + "learning_rate": 1e-06, + "loss": 0.7788, + "mean_token_accuracy": 0.7466193437576294, + "num_tokens": 705959095.0, + "step": 27283 + }, + { + "epoch": 2.9962661981111354, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6737732887268066, + "learning_rate": 1e-06, + "loss": 0.7878, + "mean_token_accuracy": 0.7468715310096741, + "num_tokens": 705980298.0, + "step": 27284 + }, + { + "epoch": 2.996376015813749, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4120564460754395, + "learning_rate": 1e-06, + "loss": 0.7609, + "mean_token_accuracy": 0.7607412338256836, + "num_tokens": 706002573.0, + "step": 27285 + }, + { + "epoch": 2.996485833516363, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5806515216827393, + "learning_rate": 1e-06, + "loss": 0.8757, + "mean_token_accuracy": 0.7260504961013794, + "num_tokens": 706024834.0, + "step": 27286 + }, + { + "epoch": 2.9965956512189766, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.321047306060791, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7153232097625732, + "num_tokens": 706050593.0, + "step": 27287 + }, + { + "epoch": 2.99670546892159, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.291459083557129, + "learning_rate": 1e-06, + "loss": 0.9246, + "mean_token_accuracy": 0.7153040766716003, + "num_tokens": 706076878.0, + "step": 27288 + }, + { + "epoch": 2.9968152866242037, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5783681869506836, + "learning_rate": 1e-06, + "loss": 0.7434, + "mean_token_accuracy": 0.7597215175628662, + "num_tokens": 706098592.0, + "step": 27289 + }, + { + "epoch": 2.9969251043268175, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4268481731414795, + "learning_rate": 1e-06, + "loss": 0.7637, + "mean_token_accuracy": 0.7535932064056396, + "num_tokens": 706122101.0, + "step": 27290 + }, + { + "epoch": 2.9970349220294312, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2505691051483154, + "learning_rate": 1e-06, + "loss": 0.8387, + "mean_token_accuracy": 0.7328680157661438, + "num_tokens": 706150293.0, + "step": 27291 + }, + { + "epoch": 2.997144739732045, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.585198163986206, + "learning_rate": 1e-06, + "loss": 0.8015, + "mean_token_accuracy": 0.7570888996124268, + "num_tokens": 706172552.0, + "step": 27292 + }, + { + "epoch": 2.9972545574346583, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2001190185546875, + "learning_rate": 1e-06, + "loss": 0.8657, + "mean_token_accuracy": 0.7262966632843018, + "num_tokens": 706202391.0, + "step": 27293 + }, + { + "epoch": 2.997364375137272, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.480623245239258, + "learning_rate": 1e-06, + "loss": 0.8943, + "mean_token_accuracy": 0.7207897305488586, + "num_tokens": 706226644.0, + "step": 27294 + }, + { + "epoch": 2.997474192839886, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.735086679458618, + "learning_rate": 1e-06, + "loss": 0.7838, + "mean_token_accuracy": 0.7470910549163818, + "num_tokens": 706245912.0, + "step": 27295 + }, + { + "epoch": 2.9975840105424996, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.6804168224334717, + "learning_rate": 1e-06, + "loss": 0.8363, + "mean_token_accuracy": 0.7354326248168945, + "num_tokens": 706268033.0, + "step": 27296 + }, + { + "epoch": 2.9976938282451133, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.484604835510254, + "learning_rate": 1e-06, + "loss": 0.7438, + "mean_token_accuracy": 0.7570509910583496, + "num_tokens": 706290252.0, + "step": 27297 + }, + { + "epoch": 2.9978036459477266, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.1568336486816406, + "learning_rate": 1e-06, + "loss": 0.9477, + "mean_token_accuracy": 0.7084763050079346, + "num_tokens": 706320660.0, + "step": 27298 + }, + { + "epoch": 2.9979134636503404, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5460712909698486, + "learning_rate": 1e-06, + "loss": 0.8575, + "mean_token_accuracy": 0.7252973318099976, + "num_tokens": 706342544.0, + "step": 27299 + }, + { + "epoch": 2.998023281352954, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2655789852142334, + "learning_rate": 1e-06, + "loss": 0.8094, + "mean_token_accuracy": 0.7439537048339844, + "num_tokens": 706369294.0, + "step": 27300 + }, + { + "epoch": 2.998133099055568, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.511504888534546, + "learning_rate": 1e-06, + "loss": 0.7156, + "mean_token_accuracy": 0.770453155040741, + "num_tokens": 706392264.0, + "step": 27301 + }, + { + "epoch": 2.9982429167581817, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.134657382965088, + "learning_rate": 1e-06, + "loss": 0.8903, + "mean_token_accuracy": 0.7251728773117065, + "num_tokens": 706423052.0, + "step": 27302 + }, + { + "epoch": 2.998352734460795, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.295243740081787, + "learning_rate": 1e-06, + "loss": 0.7609, + "mean_token_accuracy": 0.7563621401786804, + "num_tokens": 706448674.0, + "step": 27303 + }, + { + "epoch": 2.9984625521634087, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.349416732788086, + "learning_rate": 1e-06, + "loss": 0.823, + "mean_token_accuracy": 0.7368612289428711, + "num_tokens": 706475897.0, + "step": 27304 + }, + { + "epoch": 2.9985723698660225, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 1.9713045358657837, + "learning_rate": 1e-06, + "loss": 0.8033, + "mean_token_accuracy": 0.7446698546409607, + "num_tokens": 706508259.0, + "step": 27305 + }, + { + "epoch": 2.998682187568636, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.439138412475586, + "learning_rate": 1e-06, + "loss": 0.8591, + "mean_token_accuracy": 0.7331465482711792, + "num_tokens": 706531279.0, + "step": 27306 + }, + { + "epoch": 2.9987920052712496, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3008906841278076, + "learning_rate": 1e-06, + "loss": 0.8889, + "mean_token_accuracy": 0.7205398082733154, + "num_tokens": 706557018.0, + "step": 27307 + }, + { + "epoch": 2.9989018229738633, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.231318235397339, + "learning_rate": 1e-06, + "loss": 0.9022, + "mean_token_accuracy": 0.7249277830123901, + "num_tokens": 706587542.0, + "step": 27308 + }, + { + "epoch": 2.999011640676477, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.8782827854156494, + "learning_rate": 1e-06, + "loss": 0.8534, + "mean_token_accuracy": 0.7351442575454712, + "num_tokens": 706606653.0, + "step": 27309 + }, + { + "epoch": 2.999121458379091, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3622145652770996, + "learning_rate": 1e-06, + "loss": 0.8249, + "mean_token_accuracy": 0.7453508377075195, + "num_tokens": 706631700.0, + "step": 27310 + }, + { + "epoch": 2.999231276081704, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.3751280307769775, + "learning_rate": 1e-06, + "loss": 0.8132, + "mean_token_accuracy": 0.7390028238296509, + "num_tokens": 706656969.0, + "step": 27311 + }, + { + "epoch": 2.999341093784318, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.5087764263153076, + "learning_rate": 1e-06, + "loss": 0.8415, + "mean_token_accuracy": 0.7307515144348145, + "num_tokens": 706681011.0, + "step": 27312 + }, + { + "epoch": 2.9994509114869317, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.2234480381011963, + "learning_rate": 1e-06, + "loss": 0.838, + "mean_token_accuracy": 0.7366110682487488, + "num_tokens": 706707342.0, + "step": 27313 + }, + { + "epoch": 2.9995607291895454, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.363271474838257, + "learning_rate": 1e-06, + "loss": 0.8163, + "mean_token_accuracy": 0.7384752035140991, + "num_tokens": 706731699.0, + "step": 27314 + }, + { + "epoch": 2.999670546892159, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.4032692909240723, + "learning_rate": 1e-06, + "loss": 0.8445, + "mean_token_accuracy": 0.7312161922454834, + "num_tokens": 706757635.0, + "step": 27315 + }, + { + "epoch": 2.9997803645947725, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.124324083328247, + "learning_rate": 1e-06, + "loss": 0.8021, + "mean_token_accuracy": 0.7466285228729248, + "num_tokens": 706787813.0, + "step": 27316 + }, + { + "epoch": 2.9998901822973862, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.151148796081543, + "learning_rate": 1e-06, + "loss": 0.9459, + "mean_token_accuracy": 0.7014310359954834, + "num_tokens": 706821902.0, + "step": 27317 + }, + { + "epoch": 3.0, + "ewc_loss": 1.71661376953125e-05, + "grad_norm": 2.0712335109710693, + "learning_rate": 1e-06, + "loss": 0.8164, + "mean_token_accuracy": 0.7377893924713135, + "num_tokens": 706854136.0, + "step": 27318 + }, + { + "epoch": 3.0, + "ewc_loss": 1.71661376953125e-05, + "step": 27318, + "total_flos": 3.1829322264618205e+19, + "train_loss": 0.923321149223798, + "train_runtime": 37069.0608, + "train_samples_per_second": 11.791, + "train_steps_per_second": 0.737 + } + ], + "logging_steps": 1, + "max_steps": 27318, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 13659, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 3.1829322264618205e+19, + "train_batch_size": 4, + "trial_name": null, + "trial_params": null +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..215e326 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b9c1a3a4169ffd62d65435ed3608b24bf44c6dba6ec25ba3bf4f54c1c8e2762 +size 13329